1 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
4 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
6 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
7 make_temp_ssa_name): move to tree-ssanames.h
8 * tree-ssa-alias.h: Move prototype.
9 * tree-ssa.h: Include tree-ssanames.h.
10 * tree-ssanames.c (FREE_SSANAMES): Move to here.
11 * tree-ssanames.h: New. Move items from tree-flow*.h
12 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
14 2013-09-12 Richard Biener <rguenther@suse.de>
16 PR tree-optimization/58404
17 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
18 propagate non-invariant addresses into dereferences wrapped
19 in component references.
21 2013-09-12 Richard Biener <rguenther@suse.de>
23 PR tree-optimization/58402
24 * passes.def: Move pass_late_warn_uninitialized later.
26 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
28 * tree-ssa.h: New. Move content from tree-flow.h and
30 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
31 Move prototypes belonging to tree-ssa.c.
32 * tree-flow-inline.h (redirect_edge_var_map_def,
33 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
35 * gimple.h: Adjust prototypes.
36 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move to...
37 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
38 * tree.h: Move prototype to tree-ssa.h.
39 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
40 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
41 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
42 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
43 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
44 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
45 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
46 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
47 graphite-blocking.c, graphite-clast-to-gimple.c, graphite-dependences.c,
48 graphite-interchange.c, graphite-optimize-isl.c, graphite-poly.c,
49 graphite-scop-detection.c, graphite-sese-to-poly.c, graphite.c,
50 ipa-cp.c, ipa-inline-analysis.c, ipa-inline-transform.c, ipa-inline.c,
51 ipa-prop.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa-utils.c,
52 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
53 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
54 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
55 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
56 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
57 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
58 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
59 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
60 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
61 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
62 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
63 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
64 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
65 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
66 value-prof.c, var-tracking.c,
67 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
69 2013-09-12 Richard Biener <rguenther@suse.de>
71 PR tree-optimization/58396
72 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
73 (build_rdg): Take a loop-nest parameter, fix memleaks.
74 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
76 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
78 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
81 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
83 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
84 The FMA instruction names should have a 'v' prefix.
86 2013-09-12 Richard Biener <rguenther@suse.de>
88 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
89 (dot_rdg): Use popen instead of system in optional code.
90 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
91 (already_processed_vertex_p): Adjust.
92 (has_anti_or_output_dependence, predecessor_has_mem_write,
93 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
94 rdg_flag_uses): Remove.
95 (rdg_flag_vertex): Simplify.
96 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
98 (build_rdg_partition_for_component): Process the first vertex
100 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
102 2013-09-12 Alan Modra <amodra@gmail.com>
104 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
106 2013-09-11 DJ Delorie <dj@redhat.com>
107 Nick Clifton <nickc@redhat.com>
109 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
110 (rl78_stack_based_mem): New.
111 * config/rl78/constraints.md (Iv08): New.
120 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
122 (movsi): Change from expand to insn-and-split.
123 (ashrsi3): Clobber AX.
127 * config/rl78/rl78.md (CC_REG): Fix.
128 (addsi3): Allow memory and immediate operands.
129 (addsi3_internal): Split into...
130 (addsi3_internal_virt): ...new, and ...
131 (addsi3_internal_real): ...new.
133 (subsi3_internal_virt): New.
134 (subsi3_internal_real): New.
135 (mulsi3): Add memory operand.
136 (mulsi3_rl78): Likewise.
137 (mulsi3_g13): Likewise.
138 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
139 (cbranchqi4_real): Add more constraint options.
140 (cbranchhi4_real): Expand pattern.
141 (cbranchhi4_real_signed): New.
142 (cbranchhi4_real_inverted): New.
143 (cbranchsi4_real_lt): New.
144 (cbranchsi4_real_ge): New.
145 (cbranchsi4_real_signed): New.
146 (cbranchsi4_real): New.
148 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
150 (lshrsi3_virt): Likewise.
151 (ashlsi3_virt): Likewise.
152 (cbranchqi4_virt_signed): New.
153 (cbranchhi4_virt_signed): New.
154 (cbranchsi4_virt): New.
155 * config/rl78/rl78.c: Whitespace fixes throughout.
156 (move_elim_pass): New.
157 (pass_data_rl78_move_elim): New.
158 (pass_rl78_move_elim): New.
159 (make_pass_rl78_move_elim): New.
160 (rl78_devirt_info): Run devirt earlier.
161 (rl78_move_elim_info): New.
162 (rl78_asm_file_start): Register it.
163 (rl78_split_movsi): New.
164 (rl78_as_legitimate_address): Allow virtual base registers when
166 (rl78_addr_space_convert): Remove spurious debug stuff.
167 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
168 (rl78_print_operand): More cases for not printing '#'.
169 (rl78_expand_compare): Remove most of the logic.
170 (content_memory): New.
171 (clear_content_memory): New.
172 (get_content_index): New.
173 (get_content_name): New.
174 (display_content_memory): New.
175 (update_content): New.
176 (record_content): New.
177 (already_contains): New.
178 (insn_ok_now): Re-recog insns with virtual registers.
179 (add_postponed_content_update): New.
180 (process_postponed_content_update): New.
181 (gen_and_emit_move): New.
182 (transcode_memory_rtx): Record new location content.
183 Use gen_and_emit_move.
184 (force_into_acc): New.
185 (move_to_acc): Use gen_and_emit_move.
186 (move_from_acc): Likewise.
187 (move_acc_to_reg): Likewise.
188 (move_to_x): Likewise.
189 (move_to_hl): Likewise.
190 (move_to_de): Likewise.
191 (rl78_alloc_physical_registers_op1): Record location content.
192 (has_constraint): New.
193 (rl78_alloc_physical_registers_op2): Record location content.
195 (rl78_alloc_physical_registers_ro1): Likewise.
196 (rl78_alloc_physical_registers_cmp): Likewise.
197 (rl78_alloc_physical_registers_umul): Likewise.
198 (rl78_alloc_address_registers_macax): New.
199 (rl78_alloc_physical_registers): Initialize and set location
200 content memory as needed.
201 (rl78_reorg): Make sure split2 is called.
202 (rl78_rtx_costs): New.
204 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
206 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
207 for (not (neg ...)) and (neg (not ...)) cases.
209 2013-09-11 Richard Biener <rguenther@suse.de>
212 * passes.def: Split critical edges before late uninit warning passes.
213 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
215 2013-09-11 Jakub Jelinek <jakub@redhat.com>
217 PR tree-optimization/58385
218 * fold-const.c (build_range_check): If both low and high are NULL,
219 use omit_one_operand_loc to preserve exp side-effects.
221 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
223 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
225 2013-09-11 Richard Biener <rguenther@suse.de>
227 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
228 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
229 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
230 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
231 stmts_from_loop, known_dependences_p, build_empty_rdg,
232 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
233 * tree-loop-distribution.c: ... here.
234 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
235 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
236 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
237 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
238 * tree-loop-distribution.c: ... here.
239 * tree-loop-distribution.c: Include gimple-pretty-print.h.
240 (struct partition_s): Add loops member.
241 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
242 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
243 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
245 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
246 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
247 Sergey Lega <sergey.s.lega@intel.com>
248 Anna Tikhonova <anna.tikhonova@intel.com>
249 Ilya Tocar <ilya.tocar@intel.com>
250 Andrey Turetskiy <andrey.turetskiy@intel.com>
251 Ilya Verbin <ilya.verbin@intel.com>
252 Kirill Yukhin <kirill.yukhin@intel.com>
253 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
255 * config/i386/constraints.md (k): New.
257 * config/i386/i386.c (const regclass_map): Add new mask registers.
258 (dbx_register_map): Ditto.
259 (dbx64_register_map): Ditto.
260 (svr4_dbx_register_map): Ditto.
261 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
263 (ix86_preferred_reload_class): Disable constants for mask registers.
264 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
265 (ix86_hard_regno_mode_ok): Support new mask registers.
266 (x86_order_regs_for_local_alloc): Ditto.
267 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
268 (FIXED_REGISTERS): Add new mask registers.
269 (CALL_USED_REGISTERS): Ditto.
270 (REG_ALLOC_ORDER): Ditto.
271 (VALID_MASK_REG_MODE): New.
272 (FIRST_MASK_REG): Ditto.
273 (LAST_MASK_REG): Ditto.
274 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
275 (MAYBE_MASK_CLASS_P): New.
276 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
277 (REG_CLASS_CONTENTS): Ditto.
279 (ANY_MASK_REG_P): Ditto.
280 (HI_REGISTER_NAMES): Add new mask registers.
281 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
282 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
284 (attribute "type"): Add mskmov, msklog.
285 (attribute "length_immediate"): Support them.
286 (attribute "memory"): Ditto.
287 (attribute "prefix_0f"): Ditto.
288 (*movhi_internal): Support new mask registers.
289 (*movqi_internal): Ditto.
290 (define_split): Split out clobber pattern is a logic
291 insn on mask registers.
292 (*k<logic><mode>): New.
293 (*andhi_1): Extend to support mask regs.
294 (*andqi_1): Extend to support mask regs.
296 (define_split): Split and-not to and and not if operands
298 (*<code><mode>_1): Separate HI mode to new pattern...
300 (*<code>qi_1): Extend to support mask regs.
305 (*one_cmpl<mode>2_1): Remove HImode and handle it...
306 (*one_cmplhi2_1): ...Here, now with mask registers support.
307 (*one_cmplqi2_1): Support new mask registers.
308 (HI/QImode arithmetics splitter): Don't split if mask registers
310 (HI/QImode not splitter): Ditto.
311 * config/i386/predicated.md (mask_reg_operand): New.
312 (general_reg_operand): Ditto.
314 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
316 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
317 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
319 2013-09-10 Jeff Law <law@redhat.com>
321 PR tree-optimization/58380
322 * tree-ssa-threadupdate.c (thread_block): Recognize another case
323 of threading through a buried loop header.
325 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
326 return value for single successor case.
328 2013-09-10 Jan Hubicka <jh@suse.cz>
330 * ipa-devirt.c (ipa_devirt): Enable with LTO.
332 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
335 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
336 support conditional execution.
337 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
339 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
341 * lra.c (lra): Clear lra_optional_reload_pseudos before every
343 * lra-constraints.c (curr_insn_transform): Switch on optional
344 reloads. Check destination too to check move insn.
345 (undo_optional_reloads): Add check that the original peudo did not
346 changed its allocation and the optional reload was inherited on
347 last inheritance pass. Break loop after deciding to keep optional
349 (lra_undo_inheritance): Add check that inherited pseudo still in
352 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
354 * config/aarch64/aarch64.md (generic_sched): New.
355 * config/aarch64/aarch64-generic.md (load): Make conditional
356 on generic_sched attribute.
359 2013-09-10 Jan Hubicka <jh@suse.cz>
361 * lto-cgraph.c: Include ipa-utils.h.
362 (compute_ltrans_boundary): Also add possible targets into the boundary.
364 2013-09-10 Jan Hubicka <jh@suse.cz>
366 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
367 VAR_DECL of vtable rather than full expression.
369 2013-09-10 Jan Hubicka <jh@suse.cz>
370 Paolo Carlini <paolo.carlini@oracle.com>
372 * cgraphunit.c (analyze_functions): Save input_location, set it
373 to UNKNOWN_LOCATION and restore it at the end.
375 2013-09-10 Martin Jambor <mjambor@suse.cz>
377 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
378 represented by a thunk.
380 2013-09-10 Jeff Law <law@redhat.com>
382 PR tree-optimization/58343
383 * tree-ssa-threadupdate.c (thread_block): Identify and disable
384 jump threading requests through loop headers buried in the middle
385 of a jump threading path.
387 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
388 in return value/type.
390 2013-09-10 Jakub Jelinek <jakub@redhat.com>
392 PR rtl-optimization/58365
393 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
394 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
397 2013-09-10 Richard Biener <rguenther@suse.de>
399 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
400 * tree-data-ref.c (create_rdg_vertices): Collect all data
401 references, signal failure to the caller, use data-ref API.
402 (build_rdg): Compute data references only once. Maintain lifetime
403 of data references and data dependences from within RDG.
404 (free_rdg): Free dependence relations.
405 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
406 inventing extra dependences.
407 (distribute_loop): Update for RDG API changes.
409 2013-09-10 Kai Tietz <ktietz@redhat.com>
411 * doc/invoke.texi (fms-extensions): Document changed
412 behavior for ms-abi targets.
413 * config/i386/i386.c (ix86_option_override_internal):
414 Set default value of option -fms-extension for ms-abi targets.
416 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
418 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
420 2013-09-10 Alan Modra <amodra@gmail.com>
423 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
425 2013-09-10 Alan Modra <amodra@gmail.com>
427 * config/rs6000/predicates.md (add_cint_operand): New.
428 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
429 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
430 using add_cint_operand.
431 (largetoc_high_plus_aix): Likewise.
433 2013-09-09 Jakub Jelinek <jakub@redhat.com>
435 PR tree-optimization/58364
436 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
437 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
438 the current range can't be an unconditional true or false.
440 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
442 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
444 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
446 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
448 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
451 * doc/invoke.texi (-Wdelete-incomplete): Document it.
453 2013-09-09 Ian Bolton <ian.bolton@arm.com>
455 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
456 NO_REGS for immediate that can't be moved directly into FP_REGS.
458 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
460 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
461 comparison with negated operand.
462 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
465 2013-09-09 Richard Biener <rguenther@suse.de>
468 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
469 of a subloop record all its block as affecting loop-closed SSA form.
471 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
473 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
474 of an rtx/bitpos pair.
475 (store_fixed_bit_field): Update accordingly.
477 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
479 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
481 * builtins.c (expand_errno_check): Likewise.
482 * dwarf2cfi.c (init_return_column_size): Likewise.
483 * except.c (sjlj_mark_call_sites): Likewise.
484 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
485 * lra-constraints.c (emit_inc): Likewise.
486 * ree.c (combine_set_extension): Likewise.
487 * regmove.c (fixup_match_2): Likewise.
488 * reload1.c (inc_for_reload): Likewise.
490 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
492 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
493 (force_to_mode, simplify_shift_const_1, simplify_comparison):
494 Use gen_int_mode with the mode of the associated simplify_* call.
495 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
496 * expmed.c (expand_shift_1): Likewise.
497 * function.c (instantiate_virtual_regs_in_insn): Likewise.
498 * loop-iv.c (iv_number_of_iterations): Likewise.
499 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
500 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
502 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
504 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
505 of the associated expand_* call.
506 (asan_emit_stack_protection): Likewise.
507 * builtins.c (round_trampoline_addr): Likewise.
508 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
509 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
510 (emit_store_flag): Likewise.
511 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
512 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
514 * function.c (instantiate_virtual_regs_in_insn): Likewise.
515 * ifcvt.c (noce_try_store_flag_constants): Likewise.
516 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
517 * modulo-sched.c (generate_prolog_epilog): Likewise.
518 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
519 (expand_ctz, expand_ffs, expand_unop): Likewise.
521 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
523 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
524 of the associated gen_rtx_* call.
525 * caller-save.c (init_caller_save): Likewise.
526 * combine.c (find_split_point, make_extraction): Likewise.
527 (make_compound_operation): Likewise.
528 * dwarf2out.c (mem_loc_descriptor): Likewise.
529 * explow.c (plus_constant, probe_stack_range): Likewise.
530 * expmed.c (expand_mult_const): Likewise.
531 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
532 * reload1.c (init_reload): Likewise.
533 * valtrack.c (cleanup_auto_inc_dec): Likewise.
534 * var-tracking.c (adjust_mems): Likewise.
535 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
536 rather than gen_rtx_fmt_ee.
538 2013-09-09 Jan Hubicka <jh@suse.cz>
541 * value-prof.c (gimple_ic): Copy also abnormal edges.
543 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
545 * asan.c (asan_shadow_cst): Use gen_int_mode.
547 2013-09-08 Jan Hubicka <jh@suse.cz>
549 * ipa-profile.c: Add toplevel comment.
550 (ipa_propagate_frequency_1): Be more conservative when profile is read.
551 (contains_hot_call_p): New function.
552 (ipa_propagate_frequency): Set frequencies based on counts when
554 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
555 profile; do not tamper with profile after inlining if it is read.
557 2013-09-08 Jan Hubicka <jh@suse.cz>
559 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
562 2013-09-08 Jan Hubicka <jh@suse.cz>
564 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
567 2013-09-08 Jeff Law <law@redhat.com>
570 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
573 2013-09-08 Andi Kleen <ak@linux.intel.com>
575 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
577 2013-09-08 Jan Hubicka <jh@suse.cz>
579 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
581 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
583 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
585 * lra.c (new_insn_reg): Take the containing insn as a parameter.
586 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
587 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
590 2013-09-08 Jan Hubicka <jh@suse.cz>
592 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
593 targets and devirtualize to BUILT_IN_UNREACHABLE.
594 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
595 * ipa.c (walk_polymorphic_call_targets): New function.
596 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
597 functions; use the new timevar.
598 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
599 was removed from the program.
600 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
601 not consider it in the walk when its vtable is dead.
602 (possible_polymorphic_call_targets_1): Pass anonymous flag to
604 (devirt_variable_node_removal_hook): New function.
605 (possible_polymorphic_call_targets): Also register
606 devirt_variable_node_removal_hook.
607 (ipa_devirt): Do not do non-speculative devirtualization.
608 (gate_ipa_devirt): One execute if devirtualizing speculatively.
610 2013-09-08 Jan Hubicka <jh@suse.cz>
612 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
613 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
614 varpool_remove_variable_insertion_hook): Declare.
615 * varpool.c (varpool_node_hook_list): New structure.
616 (first_varpool_node_removal_hook,
617 first_varpool_variable_insertion_hook): New variables.
618 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
619 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
620 varpool_remove_variable_insertion_hook,
621 varpool_call_variable_insertion_hooks): New functions.
622 (varpool_remove_node): Use it.
624 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
627 * diagnostic.c (diagnostic_build_prefix): When s.file is
628 "<built-in>" don't output line and column numbers.
630 2013-09-06 Jan Hubicka <jh@suse.cz>
632 * cgraphunit.c (expand_thunk): Get body before touching arguments.
633 * lto-streamer-out.c: Stream thunks, too.
634 * lto-streamer-in.c (input_function): Pop cfun here
635 (lto_read_body): Instead of here.
637 2013-09-06 Caroline Tice <cmtice@google.com>
639 * doc/install.texi: Add documentation for the --enable-vtable-verify
640 and the --disable-libvtv configure options.
642 2013-09-06 Jeff Law <law@redhat.com>
644 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
645 edge implied equivalences into successor phis.
647 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
649 * resource.c (mark_referenced_resources): Handle COND_EXEC.
651 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
653 * resource.c (mark_target_live_regs): Compute resources taking
654 into account if a call is predicated or not.
656 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
658 * toplev.c (output_stack_usage): Be prepared for suffixes created by
659 the compiler in the function names.
661 2013-09-06 Jan Hubicka <jh@suse.cz>
664 * ipa-inline.c (has_caller_p): New function.
665 (want_inline_function_to_all_callers_p): Use it.
666 (sum_callers, inline_to_all_callers): Break out from ...
667 (ipa_inline): ... here.
669 2013-09-06 Jan Hubicka <jh@suse.cz>
671 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
672 only when AVX is enabled.
674 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
676 * config/aarch64/aarch64.md
677 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
678 is fpsimd_<load/store>2.
679 (load_pair<mode>): Likewise.
680 (store_pair<mode>): Likewise.
682 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
684 * config/arm/types.md (type): Add "mrs" type.
685 * config/aarch64/aarch64.md
686 (aarch64_load_tp_hard): Make type "mrs".
688 (load_tp_hard): Make type "mrs".
689 * config/arm/cortex-a15.md: Update with new attributes.
690 * config/arm/cortex-a5.md: Update with new attributes.
691 * config/arm/cortex-a53.md: Update with new attributes.
692 * config/arm/cortex-a7.md: Update with new attributes.
693 * config/arm/cortex-a8.md: Update with new attributes.
694 * config/arm/cortex-a9.md: Update with new attributes.
695 * config/arm/cortex-m4.md: Update with new attributes.
696 * config/arm/cortex-r4.md: Update with new attributes.
697 * config/arm/fa526.md: Update with new attributes.
698 * config/arm/fa606te.md: Update with new attributes.
699 * config/arm/fa626te.md: Update with new attributes.
700 * config/arm/fa726te.md: Update with new attributes.
702 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
704 * config/aarch64/aarch64.md
705 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
706 (*movtf_aarch64): Likewise.
708 (thumb1_movdi_insn): Use "multiple" for type where more than one
709 instruction is used for a move.
710 (*arm32_movhf): Likewise.
711 (*thumb_movdf_insn): Likewise.
713 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
715 * config/arm/types.md (type): Rename fcpys to fmov.
717 (*arm_movsi_vfp): Rename type fcpys as fmov.
718 (*thumb2_movsi_vfp): Likewise
719 (*movhf_vfp_neon): Likewise
720 (*movhf_vfp): Likewise
721 (*movsf_vfp): Likewise
722 (*thumb2_movsf_vfp): Likewise
723 (*movsfcc_vfp): Likewise
724 (*thumb2_movsfcc_vfp): Likewise
725 * config/aarch64/aarch64-simd.md
726 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
727 * config/aarch64/aarch64.md
728 (*movsi_aarch64): Replace type mov_reg with fmovs.
729 (*movdi_aarch64): Likewise
730 (*movsf_aarch64): Likewise
731 (*movdf_aarch64): Likewise
733 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
734 * config/arm/iwmmxt.md
735 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
736 * config/arm/arm1020e.md: Update with new attributes.
737 * config/arm/cortex-a15-neon.md: Update with new attributes.
738 * config/arm/cortex-a5.md: Update with new attributes.
739 * config/arm/cortex-a53.md: Update with new attributes.
740 * config/arm/cortex-a7.md: Update with new attributes.
741 * config/arm/cortex-a8-neon.md: Update with new attributes.
742 * config/arm/cortex-a9.md: Update with new attributes.
743 * config/arm/cortex-m4-fpu.md: Update with new attributes.
744 * config/arm/cortex-r4f.md: Update with new attributes.
745 * config/arm/marvell-pj4.md: Update with new attributes.
746 * config/arm/vfp11.md: Update with new attributes.
748 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
750 * config/aarch64/aarch64.md
751 (*madd<mode>): Fix type attribute.
752 (*maddsi_uxtw): Likewise.
753 (*msub<mode>): Likewise.
754 (*msubsi_uxtw): Likewise.
755 (<su_optab>maddsidi4): Likewise.
756 (<su_optab>msubsidi4): Likewise.
758 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
760 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
761 * config/arm/arm.md (core_cycles): Remove fdiv.
763 (*sqrtsf2_vfp): Update for attribute changes.
764 (*sqrtdf2_vfp): Likewise.
765 * config/aarch64/aarch64.md:
766 (sqrt<mode>2): Update for attribute changes.
767 * config/arm/arm1020e.md: Update with new attributes.
768 * config/arm/cortex-a15-neon.md: Update with new attributes.
769 * config/arm/cortex-a5.md: Update with new attributes.
770 * config/arm/cortex-a53.md: Update with new attributes.
771 * config/arm/cortex-a7.md: Update with new attributes.
772 * config/arm/cortex-a8-neon.md: Update with new attributes.
773 * config/arm/cortex-a9.md: Update with new attributes.
774 * config/arm/cortex-m4-fpu.md: Update with new attributes.
775 * config/arm/cortex-r4f.md: Update with new attributes.
776 * config/arm/marvell-pj4.md: Update with new attributes.
777 * config/arm/vfp11.md: Update with new attributes.
779 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
781 * config/arm/types.md
782 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
783 * config/aarch64/aarch64.md
784 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
786 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
787 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
788 (float<GPI:mode><GPF:mode>2): Likewise.
790 (*truncsisf2_vfp): Update with new attributes.
791 (*truncsidf2_vfp): Likewise.
792 (fixuns_truncsfsi2): Likewise.
793 (fixuns_truncdfsi2): Likewise.
794 (*floatsisf2_vfp): Likewise.
795 (*floatsidf2_vfp): Likewise.
796 (floatunssisf2): Likewise.
797 (floatunssidf2): Likewise.
798 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
799 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
800 * config/arm/arm1020e.md: Update with new attributes.
801 * config/arm/cortex-a15-neon.md: Update with new attributes.
802 * config/arm/cortex-a5.md: Update with new attributes.
803 * config/arm/cortex-a53.md: Update with new attributes.
804 * config/arm/cortex-a7.md: Update with new attributes.
805 * config/arm/cortex-a8-neon.md: Update with new attributes.
806 * config/arm/cortex-a9.md: Update with new attributes.
807 * config/arm/cortex-m4-fpu.md: Update with new attributes.
808 * config/arm/cortex-r4f.md: Update with new attributes.
809 * config/arm/marvell-pj4.md: Update with new attributes.
810 * config/arm/vfp11.md: Update with new attributes.
812 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
814 * config/aarch64/arm_neon.h
815 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
816 (vqtbx<1,2,3,4><q>_s8): Likewise.
818 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
820 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
821 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
822 (add<mode>3): Add type attribute.
823 (add<mode>3): Likewise.
824 (usadd<mode>3): Likewise.
825 (ssadd<mode>3): Likewise.
826 (sub<mode>3): Likewise.
827 (sub<mode>3): Likewise.
828 (ussub<mode>3): Likewise.
829 (sssub<mode>3): Likewise.
830 (ssmulsa3): Likewise.
831 (usmulusa3): Likewise.
832 (arm_usatsihi): Likewise.
834 (*movdi_vfp): Add types for all instructions.
835 (*movdi_vfp_cortexa8): Likewise.
836 (*movhf_vfp_neon): Likewise.
837 (*movhf_vfp): Likewise.
838 (*movdf_vfp): Likewise.
839 (*thumb2_movdf_vfp): Likewise.
840 (*thumb2_movdfcc_vfp): Likewise.
841 * config/arm/arm.md: Add type attribute to all insn patterns.
842 (*thumb1_adddi3): Add type attribute.
843 (*arm_adddi3): Likewise.
844 (*adddi_sesidi_di): Likewise.
845 (*adddi_zesidi_di): Likewise.
846 (*thumb1_addsi3): Likewise.
847 (addsi3_compare0): Likewise.
848 (*addsi3_compare0_scratch): Likewise.
849 (*compare_negsi_si): Likewise.
850 (cmpsi2_addneg): Likewise.
851 (*addsi3_carryin_<optab>): Likewise.
852 (*addsi3_carryin_alt2_<optab>): Likewise.
853 (*addsi3_carryin_clobercc_<optab>): Likewise.
854 (*subsi3_carryin): Likewise.
855 (*subsi3_carryin_const): Likewise.
856 (*subsi3_carryin_compare): Likewise.
857 (*subsi3_carryin_compare_const): Likewise.
858 (*arm_subdi3): Likewise.
859 (*thumb_subdi3): Likewise.
860 (*subdi_di_zesidi): Likewise.
861 (*subdi_di_sesidi): Likewise.
862 (*subdi_zesidi_di): Likewise.
863 (*subdi_sesidi_di): Likewise.
864 (*subdi_zesidi_ze): Likewise.
865 (thumb1_subsi3_insn): Likewise.
866 (*arm_subsi3_insn): Likewise.
867 (*anddi3_insn): Likewise.
868 (*anddi_zesidi_di): Likewise.
869 (*anddi_sesdi_di): Likewise.
870 (*ne_zeroextracts): Likewise.
871 (*ne_zeroextracts): Likewise.
872 (*ite_ne_zeroextr): Likewise.
873 (*ite_ne_zeroextr): Likewise.
874 (*anddi_notdi_di): Likewise.
875 (*anddi_notzesidi): Likewise.
876 (*anddi_notsesidi): Likewise.
877 (andsi_notsi_si): Likewise.
878 (thumb1_bicsi3): Likewise.
879 (*iordi3_insn): Likewise.
880 (*iordi_zesidi_di): Likewise.
881 (*iordi_sesidi_di): Likewise.
882 (*thumb1_iorsi3_insn): Likewise.
883 (*xordi3_insn): Likewise.
884 (*xordi_zesidi_di): Likewise.
885 (*xordi_sesidi_di): Likewise.
886 (*arm_xorsi3): Likewise.
887 (*andsi_iorsi3_no): Likewise.
889 (*smax_m1): Likewise.
890 (*arm_smax_insn): Likewise.
892 (*arm_smin_insn): Likewise.
893 (*arm_umaxsi3): Likewise.
894 (*arm_uminsi3): Likewise.
895 (*minmax_arithsi): Likewise.
896 (*minmax_arithsi_): Likewise.
897 (*satsi_<SAT:code>): Likewise.
898 (arm_ashldi3_1bit): Likewise.
899 (arm_ashrdi3_1bit): Likewise.
900 (arm_lshrdi3_1bit): Likewise.
901 (*arm_negdi2): Likewise.
902 (*thumb1_negdi2): Likewise.
903 (*arm_negsi2): Likewise.
904 (*thumb1_negsi2): Likewise.
905 (*negdi_extendsid): Likewise.
906 (*negdi_zero_extend): Likewise.
907 (*arm_abssi2): Likewise.
908 (*thumb1_abssi2): Likewise.
909 (*arm_neg_abssi2): Likewise.
910 (*thumb1_neg_abss): Likewise.
911 (one_cmpldi2): Likewise.
912 (extend<mode>di2): Likewise.
913 (*compareqi_eq0): Likewise.
914 (*arm_extendhisi2addsi): Likewise.
915 (*arm_movdi): Likewise.
916 (*thumb1_movdi_insn): Likewise.
917 (*arm_movt): Likewise.
918 (*thumb1_movsi_insn): Likewise.
919 (pic_add_dot_plus_four): Likewise.
920 (pic_add_dot_plus_eight): Likewise.
921 (tls_load_dot_plus_eight): Likewise.
922 (*thumb1_movhi_insn): Likewise.
923 (*thumb1_movsf_insn): Likewise.
924 (*movdf_soft_insn): Likewise.
925 (*thumb_movdf_insn): Likewise.
926 (cbranchsi4_insn): Likewise.
927 (cbranchsi4_scratch): Likewise.
928 (*negated_cbranchsi4): Likewise.
929 (*tbit_cbranch): Likewise.
930 (*tlobits_cbranch): Likewise.
931 (*tstsi3_cbranch): Likewise.
932 (*cbranchne_decr1): Likewise.
933 (*addsi3_cbranch): Likewise.
934 (*addsi3_cbranch_scratch): Likewise.
935 (*arm_cmpdi_insn): Likewise.
936 (*arm_cmpdi_unsig): Likewise.
937 (*arm_cmpdi_zero): Likewise.
938 (*thumb_cmpdi_zero): Likewise.
939 (*deleted_compare): Likewise.
940 (*mov_scc): Likewise.
941 (*mov_negscc): Likewise.
942 (*mov_notscc): Likewise.
943 (*cstoresi_eq0_thumb1_insn): Likewise.
944 (cstoresi_nltu_thumb1): Likewise.
945 (cstoresi_ltu_thu): Likewise.
946 (thumb1_addsi3_addgeu): Likewise.
947 (*arm_jump): Likewise.
948 (*thumb_jump): Likewise.
949 (*check_arch2): Likewise.
950 (arm_casesi_internal): Likewise.
951 (thumb1_casesi_dispatch): Likewise.
952 (*arm_indirect_jump): Likewise.
953 (*thumb1_indirect_jump): Likewise.
955 (*and_scc): Likewise.
956 (*ior_scc): Likewise.
957 (*compare_scc): Likewise.
958 (*cond_move): Likewise.
959 (*cond_arith): Likewise.
960 (*cond_sub): Likewise.
961 (*cmp_ite0): Likewise.
962 (*cmp_ite1): Likewise.
963 (*cmp_and): Likewise.
964 (*cmp_ior): Likewise.
965 (*ior_scc_scc): Likewise.
966 (*ior_scc_scc_cmp): Likewise.
967 (*and_scc_scc): Likewise.
968 (*and_scc_scc_cmp): Likewise.
969 (*and_scc_scc_nod): Likewise.
971 (movcond_addsi): Likewise.
973 (*ifcompare_plus_move): Likewise.
974 (*if_plus_move): Likewise.
975 (*ifcompare_move_plus): Likewise.
976 (*if_move_plus): Likewise.
977 (*ifcompare_arith_arith): Likewise.
978 (*if_arith_arith): Likewise.
979 (*ifcompare_arith_move): Likewise.
980 (*if_arith_move): Likewise.
981 (*ifcompare_move_arith): Likewise.
982 (*if_move_arith): Likewise.
983 (*ifcompare_move_not): Likewise.
984 (*if_move_not): Likewise.
985 (*ifcompare_not_move): Likewise.
986 (*if_not_move): Likewise.
987 (*ifcompare_shift_move): Likewise.
988 (*if_shift_move): Likewise.
989 (*ifcompare_move_shift): Likewise.
990 (*if_move_shift): Likewise.
991 (*ifcompare_shift_shift): Likewise.
992 (*ifcompare_not_arith): Likewise.
993 (*ifcompare_arith_not): Likewise.
994 (*if_arith_not): Likewise.
995 (*ifcompare_neg_move): Likewise.
996 (*if_neg_move): Likewise.
997 (*ifcompare_move_neg): Likewise.
998 (*if_move_neg): Likewise.
999 (prologue_thumb1_interwork): Likewise.
1000 (*cond_move_not): Likewise.
1001 (*sign_extract_onebit): Likewise.
1002 (*not_signextract_onebit): Likewise.
1003 (stack_tie): Likewise.
1004 (align_4): Likewise.
1005 (align_8): Likewise.
1006 (consttable_end): Likewise.
1007 (consttable_1): Likewise.
1008 (consttable_2): Likewise.
1009 (consttable_4): Likewise.
1010 (consttable_8): Likewise.
1011 (consttable_16): Likewise.
1012 (*thumb1_tablejump): Likewise.
1013 (prefetch): Likewise.
1014 (force_register_use): Likewise.
1015 (thumb_eh_return): Likewise.
1016 (load_tp_hard): Likewise.
1017 (load_tp_soft): Likewise.
1018 (tlscall): Likewise.
1019 (*arm_movtas_ze): Likewise.
1020 (*arm_rev): Likewise.
1021 (*arm_revsh): Likewise.
1022 (*arm_rev16): Likewise.
1023 * config/arm/thumb2.md
1024 (*thumb2_smaxsi3): Likewise.
1025 (*thumb2_sminsi3): Likewise.
1026 (*thumb32_umaxsi3): Likewise.
1027 (*thumb2_uminsi3): Likewise.
1028 (*thumb2_negdi2): Likewise.
1029 (*thumb2_abssi2): Likewise.
1030 (*thumb2_neg_abss): Likewise.
1031 (*thumb2_movsi_insn): Likewise.
1032 (tls_load_dot_plus_four): Likewise.
1033 (*thumb2_movhi_insn): Likewise.
1034 (*thumb2_mov_scc): Likewise.
1035 (*thumb2_mov_negs): Likewise.
1036 (*thumb2_mov_negs): Likewise.
1037 (*thumb2_mov_nots): Likewise.
1038 (*thumb2_mov_nots): Likewise.
1039 (*thumb2_movsicc_): Likewise.
1040 (*thumb2_movsfcc_soft_insn): Likewise.
1041 (*thumb2_indirect_jump): Likewise.
1042 (*thumb2_and_scc): Likewise.
1043 (*thumb2_ior_scc): Likewise.
1044 (*thumb2_ior_scc_strict_it): Likewise.
1045 (*thumb2_cond_move): Likewise.
1046 (*thumb2_cond_arith): Likewise.
1047 (*thumb2_cond_ari): Likewise.
1048 (*thumb2_cond_sub): Likewise.
1049 (*thumb2_negscc): Likewise.
1050 (*thumb2_movcond): Likewise.
1051 (thumb2_casesi_internal): Likewise.
1052 (thumb2_casesi_internal_pic): Likewise.
1053 (*thumb2_alusi3_short): Likewise.
1054 (*thumb2_mov<mode>_shortim): Likewise.
1055 (*thumb2_addsi_short): Likewise.
1056 (*thumb2_subsi_short): Likewise.
1057 (thumb2_addsi3_compare0): Likewise.
1058 (*thumb2_cbz): Likewise.
1059 (*thumb2_cbnz): Likewise.
1060 (*thumb2_one_cmplsi2_short): Likewise.
1061 (*thumb2_negsi2_short): Likewise.
1062 (*orsi_notsi_si): Likewise.
1063 * config/arm/arm1020e.md: Update with new attributes.
1064 * config/arm/arm1026ejs.md: Update with new attributes.
1065 * config/arm/arm1136jfs.md: Update with new attributes.
1066 * config/arm/arm926ejs.md: Update with new attributes.
1067 * config/arm/cortex-a15.md: Update with new attributes.
1068 * config/arm/cortex-a5.md: Update with new attributes.
1069 * config/arm/cortex-a53.md: Update with new attributes.
1070 * config/arm/cortex-a7.md: Update with new attributes.
1071 * config/arm/cortex-a8.md: Update with new attributes.
1072 * config/arm/cortex-a9.md: Update with new attributes.
1073 * config/arm/cortex-m4.md: Update with new attributes.
1074 * config/arm/cortex-r4.md: Update with new attributes.
1075 * config/arm/fa526.md: Update with new attributes.
1076 * config/arm/fa606te.md: Update with new attributes.
1077 * config/arm/fa626te.md: Update with new attributes.
1078 * config/arm/fa726te.md: Update with new attributes.
1080 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1082 * config/aarch64/aarch64-simd.md
1083 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1084 <vwx> iterator to ensure correct register choice.
1085 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1086 (aarch64_sqdmull_n<mode>): Likewise.
1087 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1088 * config/aarch64/arm_neon.h
1089 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1090 (vml<as><q>_n_<su>16): Likewise.
1091 (vml<as>l_high_lane<q>_<su>16): Likewise.
1092 (vml<as>l_high_n_<su>16): Likewise.
1093 (vml<as>l_lane<q>_<su>16): Likewise.
1094 (vml<as>l_n_<su>16): Likewise.
1095 (vmul<q>_lane<q>_<su>16): Likewise.
1096 (vmul<q>_n_<su>16): Likewise.
1097 (vmull_lane<q>_<su>16): Likewise.
1098 (vmull_n_<su>16): Likewise.
1099 (vmull_high_lane<q>_<su>16): Likewise.
1100 (vmull_high_n_<su>16): Likewise.
1101 (vqrdmulh<q>_n_s16): Likewise.
1103 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1105 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1106 have the correct lane parameter.
1108 2013-09-06 Richard Biener <rguenther@suse.de>
1110 * cfganal.c (control_dependences::~control_dependences):
1111 Properly free all of the vector.
1113 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1116 * config/i386/i386.c (ix86_conditional_register_usage):
1117 Proper initialize extended SSE registers.
1119 2013-09-06 Jan Hubicka <jh@suse.cz>
1121 PR tree-optimization/58311
1122 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1124 2013-09-06 Jan Hubicka <jh@suse.cz>
1126 * Makefile.in (tree-sra.o): Update dependencies.
1127 * tree-sra.c: Include ipa-utils.h
1128 (scan_function): Use recursive_call_p.
1129 (has_caller_p): New function.
1130 (cgraph_for_node_and_aliases): Count also callers of aliases.
1132 2013-09-06 Jan Hubicka <jh@suse.cz>
1135 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1136 * tree-tailcall.c: Include ipa-utils.h.
1137 (find_tail_calls): Use it.
1138 * ipa-pure-const.c (check_call): Likewise.
1139 * ipa-utils.c (recursive_call_p): New function.
1140 * ipa-utils.h (recursive_call_p): Dclare.
1141 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1142 (symtab_semantically_equivalent_p): New function.
1143 * Makefile.in (tree-tailcall.o): Update dependencies.
1145 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1147 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1150 2013-09-06 Richard Biener <rguenther@suse.de>
1152 * lto-streamer.h (lto_global_var_decls): Remove.
1153 * Makefile.in (OBJS): Remove lto-symtab.o.
1154 (lto-symtab.o): Remove.
1155 (GTFILES): Remove lto-symtab.c
1156 * lto-symtab.c: Move to lto/
1158 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1160 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1161 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1162 (UNSPEC_FPINT_RINT): New constant definitions.
1163 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1164 definition with 2 attributes.
1165 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1166 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1169 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1171 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
1172 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1173 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1174 attribute to "z196_alone".
1175 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1178 2013-09-06 Richard Biener <rguenther@suse.de>
1180 * basic-block.h (class control_dependences): New.
1181 * tree-ssa-dce.c (control_dependence_map): Remove.
1183 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1184 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1185 find_pdom, find_control_dependence, find_all_control_dependences):
1187 (mark_control_dependent_edges_necessary,
1188 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1189 tree_dce_done, perform_tree_ssa_dce): Adjust.
1190 * cfganal.c (set_control_dependence_map_bit,
1191 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1192 find_all_control_dependences): Move from tree-ssa-dce.c and
1193 implement as methods of control_dependences class.
1194 (control_dependences::control_dependences): New.
1195 (control_dependences::~control_dependences): Likewise.
1196 (control_dependences::get_edges_dependent_on): Likewise.
1197 (control_dependences::get_edge): Likewise.
1199 2013-09-04 Jan Hubicka <jh@suse.cz>
1201 * tree.c (types_same_for_odr): Drop overactive check.
1202 * ipa-devirt.c (hash_type_name): Likewise.
1204 2013-09-04 Jan Hubicka <jh@suse.cz>
1206 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
1207 (analyze_functions): ... here.
1209 2013-09-04 Jan Hubicka <jh@suse.cz>
1212 * cgraphunit.c (analyze_functions): Clear AUX fields
1213 after processing; initialize assembler name has.
1215 2013-09-05 Jeff Law <law@redhat.com>
1217 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1218 from thread_around_empty_block. Record threading path into PATH.
1219 Recurse if threading through the initial block is successful.
1220 (thread_across_edge): Corresponding changes to slightly simplify.
1222 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1224 * config/aarch64/aarch64.md
1225 (type): Remove frecpe, frecps, frecpx.
1226 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1227 fix to be a TARGET_SIMD instruction.
1228 (aarch64_frecps): Remove.
1229 * config/aarch64/aarch64-simd.md
1230 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1231 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1233 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1234 Sofiane Naci <sofiane.naci@arm.com>
1236 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
1237 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
1238 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1239 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1240 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1241 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1242 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1243 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1244 into "clz, "rbit". Rename "shift" to "shift_imm".
1245 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
1246 changes. Update for attribute changes all occurrences of arlo_* and
1248 * config/arm/arm-fixed.md: Update for attribute changes
1249 all occurrences of arlo_* types.
1250 * config/arm/thumb2.md: Update for attribute changes all occurrences
1252 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1253 (cortexa7_older_only): Likewise.
1254 (cortexa7_younger): Likewise.
1255 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1256 (1020alu_shift_op): Likewise.
1257 (1020alu_shift_reg_op): Likewise.
1258 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1259 (alu_shift_op): Likewise.
1260 (alu_shift_reg_op): Likewise.
1261 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
1262 (11_alu_shift_op): Likewise.
1263 (11_alu_shift_reg_op): Likewise.
1264 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1265 (9_alu_shift_reg_op): Likewise.
1266 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1268 (cortex_a15_alu_shift): Likewise.
1269 (cortex_a15_alu_shift_reg): Likewise.
1270 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1272 (cortex_a5_alu_shift): Likewise.
1273 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1275 (cortex_a53_alu_shift): Likewise.
1276 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1278 (cortex_a7_alu_reg): Likewise.
1279 (cortex_a7_alu_shift): Likewise.
1280 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1282 (cortex_a8_alu_shift): Likewise.
1283 (cortex_a8_alu_shift_reg): Likewise.
1284 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
1285 (cortex_a9_dp_shift): Likewise.
1286 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1288 * config/arm/cortex-r4.md
1289 (cortex_r4_alu): Update for attribute changes.
1290 (cortex_r4_mov): Likewise.
1291 (cortex_r4_alu_shift_reg): Likewise.
1292 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
1293 (526_alu_shift_op): Likewise.
1294 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1295 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
1296 (626te_alu_shift_op): Likewise.
1297 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
1298 (726te_alu_shift_op): Likewise.
1299 (726te_alu_shift_reg_op): Likewise.
1300 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1301 (mp626_alu_shift_op): Likewise.
1302 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1303 (pj4_alu_conds): Likewise.
1304 (pj4_shift): Likewise.
1305 (pj4_shift_conds): Likewise.
1306 (pj4_alu_shift): Likewise.
1307 (pj4_alu_shift_conds): Likewise.
1308 * config/aarch64/aarch64.md: Update for attribute change
1309 all occurrences of arlo_* and shift* types.
1311 2013-09-05 Mike Stump <mikestump@comcast.net>
1313 * tree.h: Move documentation for tree_function_decl to tree-core.h
1314 with the declaration.
1316 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1319 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1320 looking for widest mode.
1322 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1324 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1326 2013-09-05 Richard Biener <rguenther@suse.de>
1328 PR tree-optimization/58137
1329 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1330 Do not create vectors of pointers.
1331 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1332 types for the components of the vector initializer.
1333 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1334 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1336 2013-09-05 Martin Jambor <mjambor@suse.cz>
1338 * ipa-prop.c (remove_described_reference): Accept missing references,
1339 return false if that hppens, otherwise return true.
1340 (cgraph_node_for_jfunc): New function.
1341 (try_decrement_rdesc_refcount): Likewise.
1342 (try_make_edge_direct_simple_call): Use them.
1343 (ipa_edge_removal_hook): Remove references from rdescs.
1344 (ipa_edge_duplication_hook): Clone rdescs and their references
1345 when the new edge has the same caller as the old one.
1346 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1347 reference before removing any edges.
1349 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
1351 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1353 * thumb2.md (thumb2_storewb_parisi): New pattern.
1355 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1357 * config/aarch64/aarch64-option-extensions.def: Add
1358 AARCH64_OPT_EXTENSION of 'crc'.
1359 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1360 (AARCH64_ISA_CRC): Ditto.
1361 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1362 description of the CRC extension.
1364 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1366 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1367 * config/rs6000/linux.h: Ditto.
1368 * alpha/linux.h: Ditto.
1369 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1370 no_c99_libc_has_function.
1371 * config/c6x/uclinux-elf.h: Ditto.
1372 * config/lm32/uclinux-elf.h: Ditto.
1373 * config/m68k/uclinux.h: Ditto.
1374 * config/moxie/uclinux.h: Ditto.
1375 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1376 (crisv32-*-linux*, cris-*-linux*): Ditto.
1377 * config/bfin/bfin.c: Include "tm_p.h".
1379 2013-09-05 Richard Biener <rguenther@suse.de>
1381 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1382 check for a definition without a basic-block.
1384 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1385 Sofiane Naci <sofiane.naci@arm.com>
1387 * config/aarch64/aarch64.md
1388 (*movti_aarch64): Rename r_2_f and f_2_r.
1389 (*movsf_aarch64): Likewise.
1390 (*movdf_aarch64): Likewise.
1391 (*movtf_aarch64): Likewise.
1392 (aarch64_movdi_<mode>low): Likewise.
1393 (aarch64_movdi_<mode>high): Likewise.
1394 (aarch64_mov<mode>high_di): Likewise.
1395 (aarch64_mov<mode>low_di): Likewise.
1396 (aarch64_movtilow_tilow): Likewise.
1397 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1398 values to config/arm/types.md
1399 (attribute "conds"): Update for attribute change.
1400 (anddi3_insn): Likewise.
1401 (iordi3_insn): Likewise.
1402 (xordi3_insn): Likewise.
1403 (one_cmpldi2): Likewise.
1404 * config/arm/types.md (type): Add Neon types.
1405 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1406 use "type" attribute.
1407 (movmisalign<mode>_neon_store): Likewise.
1408 (movmisalign<mode>_neon_load): Likewise.
1409 (vec_set<mode>_internal): Likewise.
1410 (vec_setv2di_internal): Likewise.
1411 (vec_extract<mode>): Likewise.
1412 (vec_extractv2di): Likewise.
1413 (add<mode>3_neon): Likewise.
1414 (adddi3_neon): Likewise.
1415 (sub<mode>3_neon): Likewise.
1416 (subdi3_neon): Likewise.
1417 (mul<mode>3_neon): Likewise.
1418 (mul<mode>3add<mode>_neon): Likewise.
1419 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1420 (fma<VCVTF:mode>4)): Likewise.
1421 (fma<VCVTF:mode>4_intrinsic): Likewise.
1422 (fmsub<VCVTF:mode>4)): Likewise.
1423 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1424 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1425 (ior<mode>3): Likewise.
1426 (and<mode>3): Likewise.
1427 (anddi3_neon): Likewise.
1428 (orn<mode>3_neon): Likewise.
1429 (orndi3_neon): Likewise.
1430 (bic<mode>3_neon): Likewise.
1431 (bicdi3_neon): Likewise.
1432 (xor<mode>3): Likewise.
1433 (one_cmpl<mode>2): Likewise.
1434 (abs<mode>2): Likewise.
1435 (neg<mode>2): Likewise.
1436 (umin<mode>3_neon): Likewise.
1437 (umax<mode>3_neon): Likewise.
1438 (smin<mode>3_neon): Likewise.
1439 (smax<mode>3_neon): Likewise.
1440 (vashl<mode>3): Likewise.
1441 (vashr<mode>3_imm): Likewise.
1442 (vlshr<mode>3_imm): Likewise.
1443 (ashl<mode>3_signed): Likewise.
1444 (ashl<mode>3_unsigned): Likewise.
1445 (neon_load_count): Likewise.
1446 (ashldi3_neon_noclobber): Likewise.
1447 (signed_shift_di3_neon): Likewise.
1448 (unsigned_shift_di3_neon): Likewise.
1449 (ashrdi3_neon_imm_noclobber): Likewise.
1450 (lshrdi3_neon_imm_noclobber): Likewise.
1451 (widen_ssum<mode>3): Likewise.
1452 (widen_usum<mode>3): Likewise.
1453 (quad_halves_<code>v4si): Likewise.
1454 (quad_halves_<code>v4sf): Likewise.
1455 (quad_halves_<code>v8hi): Likewise.
1456 (quad_halves_<code>v16qi): Likewise.
1457 (reduc_splus_v2di): Likewise.
1458 (neon_vpadd_internal<mode>): Likewise.
1459 (neon_vpsmin<mode>): Likewise.
1460 (neon_vpsmax<mode>): Likewise.
1461 (neon_vpumin<mode>): Likewise.
1462 (neon_vpumax<mode>): Likewise.
1463 (ss_add<mode>_neon): Likewise.
1464 (us_add<mode>_neon): Likewise.
1465 (ss_sub<mode>_neon): Likewise.
1466 (us_sub<mode>_neon): Likewise.
1467 (neon_vadd<mode>_unspec): Likewise.
1468 (neon_vaddl<mode>): Likewise.
1469 (neon_vaddw<mode>): Likewise.
1470 (neon_vhadd<mode>): Likewise.
1471 (neon_vqadd<mode>): Likewise.
1472 (neon_vaddhn<mode>): Likewise.
1473 (neon_vmul<mode>): Likewise.
1474 (neon_vmla<mode>): Likewise.
1475 (neon_vmlal<mode>): Likewise.
1476 (neon_vmls<mode>): Likewise.
1477 (neon_vmlsl<mode>): Likewise.
1478 (neon_vqdmulh<mode>): Likewise.
1479 (neon_vqdmlal<mode>): Likewise.
1480 (neon_vqdmlsl<mode>): Likewise.
1481 (neon_vmull<mode>): Likewise.
1482 (neon_vqdmull<mode>): Likewise.
1483 (neon_vsub<mode>_unspec): Likewise.
1484 (neon_vsubl<mode>): Likewise.
1485 (neon_vsubw<mode>): Likewise.
1486 (neon_vqsub<mode>): Likewise.
1487 (neon_vhsub<mode>): Likewise.
1488 (neon_vsubhn<mode>): Likewise.
1489 (neon_vceq<mode>): Likewise.
1490 (neon_vcge<mode>): Likewise.
1491 (neon_vcgeu<mode>): Likewise.
1492 (neon_vcgt<mode>): Likewise.
1493 (neon_vcgtu<mode>): Likewise.
1494 (neon_vcle<mode>): Likewise.
1495 (neon_vclt<mode>): Likewise.
1496 (neon_vcage<mode>): Likewise.
1497 (neon_vcagt<mode>): Likewise.
1498 (neon_vtst<mode>): Likewise.
1499 (neon_vabd<mode>): Likewise.
1500 (neon_vabdl<mode>): Likewise.
1501 (neon_vaba<mode>): Likewise.
1502 (neon_vabal<mode>): Likewise.
1503 (neon_vmax<mode>): Likewise.
1504 (neon_vmin<mode>): Likewise.
1505 (neon_vpaddl<mode>): Likewise.
1506 (neon_vpadal<mode>): Likewise.
1507 (neon_vpmax<mode>): Likewise.
1508 (neon_vpmin<mode>): Likewise.
1509 (neon_vrecps<mode>): Likewise.
1510 (neon_vrsqrts<mode>): Likewise.
1511 (neon_vqabs<mode>): Likewise.
1512 (neon_vqneg<mode>): Likewise.
1513 (neon_vcls<mode>): Likewise.
1514 (clz<mode>2): Likewise.
1515 (popcount<mode>2): Likewise.
1516 (neon_vrecpe): Likewise.
1517 (neon_vrsqrte): Likewise.
1518 (neon_vget_lane<mode>_sext_internal): Likewise.
1519 (neon_vget_lane<mode>_zext_internal): Likewise.
1520 (neon_vdup_n<mode>): Likewise.
1521 (neon_vdup_nv2di): Likewise.
1522 (neon_vdpu_lane<mode>_internal): Likewise.
1523 (neon_vswp<mode>): Likewise.
1524 (float<mode><V_cvtto>2): Likewise.
1525 (floatuns<mode><V_cvtto>2): Likewise.
1526 (fix_trunc<mode><V_cvtto>)2): Likewise
1527 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1528 (neon_vcvt<mode>): Likewise.
1529 (neon_vcvtv4sfv4hf): Likewise.
1530 (neon_vcvtv4hfv4sf): Likewise.
1531 (neon_vcvt_n<mode>): Likewise.
1532 (neon_vmovn<mode>): Likewise.
1533 (neon_vqmovn<mode>): Likewise.
1534 (neon_vqmovun<mode>): Likewise.
1535 (neon_vmovl<mode>): Likewise.
1536 (neon_vmul_lane<mode>): Likewise.
1537 (neon_vmull_lane<mode>): Likewise.
1538 (neon_vqdmull_lane<mode>): Likewise.
1539 (neon_vqdmulh_lane<mode>): Likewise.
1540 (neon_vmla_lane<mode>): Likewise.
1541 (neon_vmlal_lane<mode>): Likewise.
1542 (neon_vqdmlal_lane<mode>): Likewise.
1543 (neon_vmls_lane<mode>): Likewise.
1544 (neon_vmlsl_lane<mode>): Likewise.
1545 (neon_vqdmlsl_lane<mode>): Likewise.
1546 (neon_vext<mode>): Likewise.
1547 (neon_vrev64<mode>): Likewise.
1548 (neon_vrev32<mode>): Likewise.
1549 (neon_vrev16<mode>): Likewise.
1550 (neon_vbsl<mode>_internal): Likewise.
1551 (neon_vshl<mode>): Likewise.
1552 (neon_vqshl<mode>): Likewise.
1553 (neon_vshr_n<mode>): Likewise.
1554 (neon_vshrn_n<mode>): Likewise.
1555 (neon_vqshrn_n<mode>): Likewise.
1556 (neon_vqshrun_n<mode>): Likewise.
1557 (neon_vshl_n<mode>): Likewise.
1558 (neon_vqshl_n<mode>): Likewise.
1559 (neon_vqshlu_n<mode>): Likewise.
1560 (neon_vshll_n<mode>): Likewise.
1561 (neon_vsra_n<mode>): Likewise.
1562 (neon_vsri_n<mode>): Likewise.
1563 (neon_vsli_n<mode>): Likewise.
1564 (neon_vtbl1v8qi): Likewise.
1565 (neon_vtbl2v8qi): Likewise.
1566 (neon_vtbl3v8qi): Likewise.
1567 (neon_vtbl4v8qi): Likewise.
1568 (neon_vtbx1v8qi): Likewise.
1569 (neon_vtbx2v8qi): Likewise.
1570 (neon_vtbx3v8qi): Likewise.
1571 (neon_vtbx4v8qi): Likewise.
1572 (neon_vtrn<mode>_internal): Likewise.
1573 (neon_vzip<mode>_internal): Likewise.
1574 (neon_vuzp<mode>_internal): Likewise.
1575 (neon_vld1<mode>): Likewise.
1576 (neon_vld1_lane<mode>): Likewise.
1577 (neon_vld1_dup<mode>): Likewise.
1578 (neon_vld1_dupv2di): Likewise.
1579 (neon_vst1<mode>): Likewise.
1580 (neon_vst1_lane<mode>): Likewise.
1581 (neon_vld2<mode>): Likewise.
1582 (neon_vld2_lane<mode>): Likewise.
1583 (neon_vld2_dup<mode>): Likewise.
1584 (neon_vst2<mode>): Likewise.
1585 (neon_vst2_lane<mode>): Likewise.
1586 (neon_vld3<mode>): Likewise.
1587 (neon_vld3qa<mode>): Likewise.
1588 (neon_vld3qb<mode>): Likewise.
1589 (neon_vld3_lane<mode>): Likewise.
1590 (neon_vld3_dup<mode>): Likewise.
1591 (neon_vst3<mode>): Likewise.
1592 (neon_vst3qa<mode>): Likewise.
1593 (neon_vst3qb<mode>): Likewise.
1594 (neon_vst3_lane<mode>): Likewise.
1595 (neon_vld4<mode>): Likewise.
1596 (neon_vld4qa<mode>): Likewise.
1597 (neon_vld4qb<mode>): Likewise.
1598 (neon_vld4_lane<mode>): Likewise.
1599 (neon_vld4_dup<mode>): Likewise.
1600 (neon_vst4<mode>): Likewise.
1601 (neon_vst4qa<mode>): Likewise.
1602 (neon_vst4qb<mode>): Likewise.
1603 (neon_vst4_lane<mode>): Likewise.
1604 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1605 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1606 (neon_vec_<US>mult_lo_<mode>): Likewise.
1607 (neon_vec_<US>mult_hi_<mode>): Likewise.
1608 (neon_vec_<US>shiftl_<mode>): Likewise.
1609 (neon_unpack<US>_<mode>): Likewise.
1610 (neon_vec_<US>mult_<mode>): Likewise.
1611 (vec_pack_trunc_<mode>): Likewise.
1612 (neon_vec_pack_trunk_<mode>): Likewise.
1613 (neon_vabd<mode>_2): Likewise.
1614 (neon_vabd<mode>_3): Likewise.
1615 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1616 (thumb2_movsi_vfp): Likewise.
1617 (movdi_vfp): Likewise.
1618 (movdi_vfp_cortexa8): Likewise.
1619 (movhf_vfp_neon): Likewise.
1620 (movhf_vfp): Likewiwse.
1621 (movsf_vfp): Likewiwse.
1622 (thumb2_movsf_vfp): Likewiwse.
1623 (movdf_vfp): Likewise.
1624 (thumb2_movdf_vfp): Likewise.
1625 (movsfcc_vfp): Likewise.
1626 (thumb2_movsfcc_vfp): Likewise.
1627 (movdfcc_vfp): Likewise.
1628 (thumb2_movdfcc_vfp): Likewise.
1629 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1630 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1631 (v10_v2c): Likewise.
1632 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1634 (cortex_a15_neon_int_2): Likewise.
1635 (cortex_a15_neon_int_3): Likewise.
1636 (cortex_a15_neon_int_4): Likewise.
1637 (cortex_a15_neon_int_5): Likewise.
1638 (cortex_a15_neon_vqneg_vqabs): Likewise.
1639 (cortex_a15_neon_vmov): Likewise.
1640 (cortex_a15_neon_vaba): Likewise.
1641 (cortex_a15_neon_vaba_qqq): Likewise.
1642 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1643 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1644 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1645 scalar_64_32_long_scalar): Likewise.
1646 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1647 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1648 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1649 lotype_qdd_64_32_long): Likewise.
1650 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1651 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1652 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1653 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1654 (cortex_a15_neon_shift_1): Likewise.
1655 (cortex_a15_neon_shift_2): Likewise.
1656 (cortex_a15_neon_shift_3): Likewise.
1657 (cortex_a15_neon_vshl_ddd): Likewise.
1658 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1659 (cortex_a15_neon_vsra_vrsra): Likewise.
1660 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1661 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1662 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1663 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1664 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1665 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1666 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1667 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1668 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1669 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1670 (cortex_a15_neon_bp_simple): Likewise.
1671 (cortex_a15_neon_bp_2cycle): Likewise.
1672 (cortex_a15_neon_bp_3cycle): Likewise.
1673 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1674 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1675 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1676 (cortex_a15_neon_vld2_4_regs): Likewise.
1677 (cortex_a15_neon_vld3_vld4): Likewise.
1678 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1679 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1680 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1681 (cortex_a15_neon_vst3_vst4): Likewise.
1682 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1683 (cortex_a15_neon_vld3_vld4_lane" 10
1684 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1685 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1686 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1687 (cortex_a15_neon_ldm_2): Likewise.0
1688 (cortex_a15_neon_stm_2): Likewise.
1689 (cortex_a15_neon_mcr): Likewise.
1690 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1691 (cortex_a15_neon_mrc): Likewise.
1692 (cortex_a15_neon_mrrc): Likewise.
1693 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1695 (cortex_a15_alu_shift): Likewise.
1696 (cortex_a15_alu_shift_reg): Likewise.
1697 (cortex_a15_mult32): Likewise.
1698 (cortex_a15_mult64): Likewise.
1699 (cortex_a15_block): Likewise.
1700 (cortex_a15_branch): Likewise.
1701 (cortex_a15_load1): Likewise.
1702 (cortex_a15_load3): Likewise.
1703 (cortex_a15_store1): Likewise.
1704 (cortex_a15_store3): Likewise.
1705 (cortex_a15_call): Likewise.
1706 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
1707 (cortex_a5_f2r): Likewise.
1708 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1710 (cortex_a53_f2r): Likewise.
1711 * config/arm/cortex-a7.md
1712 (cortex_a7_branch): Update for attribute change.
1713 (cortex_a7_call): Likewise.
1714 (cortex_a7_alu_imm): Likewise.
1715 (cortex_a7_alu_reg): Likewise.
1716 (cortex_a7_alu_shift): Likewise.
1717 (cortex_a7_mul): Likewise.
1718 (cortex_a7_load1): Likewise.
1719 (cortex_a7_store1): Likewise.
1720 (cortex_a7_load2): Likewise.
1721 (cortex_a7_store2): Likewise.
1722 (cortex_a7_load3): Likewise.
1723 (cortex_a7_store3): Likewise.
1724 (cortex_a7_load4): Likewise.
1725 (cortex_a7_store4): Likewise.
1726 (cortex_a7_fpalu): Likewise.
1727 (cortex_a7_fconst): Likewise.
1728 (cortex_a7_fpmuls): Likewise.
1729 (cortex_a7_neon_mul): Likewise.
1730 (cortex_a7_fpmacs): Likewise.
1731 (cortex_a7_neon_mla: Likewise.
1732 (cortex_a7_fpmuld: Likewise.
1733 (cortex_a7_fpmacd: Likewise.
1734 (cortex_a7_fpfmad: Likewise.
1735 (cortex_a7_fdivs: Likewise.
1736 (cortex_a7_fdivd: Likewise.
1737 (cortex_a7_r2f: Likewise.
1738 (cortex_a7_f2r: Likewise.
1739 (cortex_a7_f_flags: Likewise.
1740 (cortex_a7_f_loads: Likewise.
1741 (cortex_a7_f_loadd: Likewise.
1742 (cortex_a7_f_stores: Likewise.
1743 (cortex_a7_f_stored: Likewise.
1744 (cortex_a7_neon): Likewise.
1745 * config/arm/cortex-a8-neon.md
1746 (cortex_a8_neon_mrc): Update for attribute change.
1747 (cortex_a8_neon_mrrc): Likewise.
1748 (cortex_a8_neon_int_1): Likewise.
1749 (cortex_a8_neon_int_2): Likewise.
1750 (cortex_a8_neon_int_3): Likewise.
1751 (cortex_a8_neon_int_4): Likewise.
1752 (cortex_a8_neon_int_5): Likewise.
1753 (cortex_a8_neon_vqneg_vqabs): Likewise.
1754 (cortex_a8_neon_vmov): Likewise.
1755 (cortex_a8_neon_vaba): Likewise.
1756 (cortex_a8_neon_vaba_qqq): Likewise.
1757 (cortex_a8_neon_vsma): Likewise.
1758 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1759 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1760 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1761 long_scalar): Likewise.
1762 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1763 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1764 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1765 long_scalar_qdd_64_32_long): Likewise.
1766 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1767 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1768 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1769 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1770 (cortex_a8_neon_shift_1): Likewise.
1771 (cortex_a8_neon_shift_2): Likewise.
1772 (cortex_a8_neon_shift_3): Likewise.
1773 (cortex_a8_neon_vshl_ddd): Likewise.
1774 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1775 (cortex_a8_neon_vsra_vrsra): Likewise.
1776 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1777 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1778 (cortex_a8_neon_fp_vsum): Likewise.
1779 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1780 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1781 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1782 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1783 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1784 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1785 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1786 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1787 (cortex_a8_neon_bp_simple): Likewise.
1788 (cortex_a8_neon_bp_2cycle): Likewise.
1789 (cortex_a8_neon_bp_3cycle): Likewise.
1790 (cortex_a8_neon_ldr): Likewise.
1791 (cortex_a8_neon_str): Likewise.
1792 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1793 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1794 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1795 (cortex_a8_neon_vld2_4_regs): Likewise.
1796 (cortex_a8_neon_vld3_vld4): Likewise.
1797 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1798 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1799 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1800 (cortex_a8_neon_vst3_vst4): Likewise.
1801 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1802 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1803 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1804 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1805 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1806 (cortex_a8_neon_mcr): Likewise.
1807 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1808 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
1809 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
1811 (ca9_neon_mrrc): Likewise.
1812 (cortex_a9_neon_int_1): Likewise.
1813 (cortex_a9_neon_int_2): Likewise.
1814 (cortex_a9_neon_int_3): Likewise.
1815 (cortex_a9_neon_int_4): Likewise.
1816 (cortex_a9_neon_int_5): Likewise.
1817 (cortex_a9_neon_vqneg_vqabs): Likewise.
1818 (cortex_a9_neon_vmov): Likewise.
1819 (cortex_a9_neon_vaba): Likewise.
1820 (cortex_a9_neon_vaba_qqq): Likewise.
1821 (cortex_a9_neon_vsma): Likewise.
1822 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1823 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1824 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1825 long_scalar): Likewise.
1826 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1827 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1828 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1829 long_scalar_qdd_64_32_long): Likewise.
1830 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1831 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1832 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1833 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1834 (cortex_a9_neon_shift_1): Likewise.
1835 (cortex_a9_neon_shift_2): Likewise.
1836 (cortex_a9_neon_shift_3): Likewise.
1837 (cortex_a9_neon_vshl_ddd): Likewise.
1838 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1839 (cortex_a9_neon_vsra_vrsra): Likewise.
1840 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1841 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1842 (cortex_a9_neon_fp_vsum): Likewise.
1843 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1844 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1845 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1846 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1847 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1848 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1849 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1850 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1851 (cortex_a9_neon_bp_simple): Likewise.
1852 (cortex_a9_neon_bp_2cycle): Likewise.
1853 (cortex_a9_neon_bp_3cycle): Likewise.
1854 (cortex_a9_neon_ldr): Likewise.
1855 (cortex_a9_neon_str): Likewise.
1856 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1857 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1858 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1859 (cortex_a9_neon_vld2_4_regs): Likewise.
1860 (cortex_a9_neon_vld3_vld4): Likewise.
1861 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1862 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1863 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1864 (cortex_a9_neon_vst3_vst4): Likewise.
1865 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1866 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1867 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1868 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1869 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1870 (cortex_a9_neon_mcr): Likewise.
1871 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1872 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
1873 (cortex_a9_fps): Likewise.
1874 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
1876 (cortex_m4_fmuls): Likewise.
1877 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
1879 (cortex_r4_mrc): Likewise.
1880 * config/arm/iterators.md: Update comment referring to neon_type.
1881 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
1882 (iwmmxt_movsi_insn): Likewise.
1883 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
1885 (pj4_core_to_vfp): Likewise.
1886 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
1888 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
1889 (vfp_fstore): Likewise.
1890 * doc/md.texi: Change references to neon_type to refer to type.
1892 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1894 * tree.h (DECL_BUILT_IN): Fix typo in comment.
1896 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
1898 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
1901 2013-09-04 Easwaran Raman <eraman@google.com>
1904 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
1905 (build_and_add_sum): Use it.
1906 (appears_later_in_bb): Simplify code.
1908 2013-09-04 Teresa Johnson <tejohnson@google.com>
1910 * dumpfile.c (dump_finish): Don't close stderr/stdout.
1912 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
1914 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
1916 2013-09-04 Jan Hubicka <jh@suse.cz>
1918 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
1919 * ipa-devirt.c: Include diganostic.h
1920 (odr_type_d): Add types and types_set.
1921 (hash_type_name): Work for types with vtables during LTO.
1922 (odr_hasher::remove): Fix comment; destroy types_set.
1923 (add_type_duplicate): New function,
1924 (get_odr_type): Use it.
1925 (dump_type_inheritance_graph): Dump type duplicates.
1926 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
1928 * tree.c (types_same_for_odr): Give exact answers on types with
1931 2013-09-04 Dodji Seketeli <dodji@redhat.com>
1933 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
1934 explaining their differences.
1936 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
1938 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
1940 2013-09-03 Jeff Law <law@redhat.com>
1942 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
1943 when not threading through a joiner block. Pass joiner/no joiner
1944 state to register_jump_thread.
1945 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
1946 state from argument rather than implying on path length.
1947 Dump the entire jump thread path into debugging dump.
1948 * tree-flow.h (register_jump_thread): Update prototype.
1950 2013-08-29 Xinliang David Li <davidxl@google.com>
1952 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1953 Remove a trivial gcc_assert.
1955 2013-08-29 Xinliang David Li <davidxl@google.com>
1957 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
1958 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
1959 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
1960 Delay base decl alignment adjustment.
1961 * tree-vectorizer.c (vect_destroy_datarefs): New function.
1962 * tree-vectorizer.h: New data structure.
1963 (set_dr_misalignment): New function.
1964 (dr_misalignment): Ditto.
1965 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
1966 (vectorizable_load): Ditto.
1967 (ensure_base_align): New function.
1968 (vectorize_loops): Add dbg_cnt support.
1969 (execute_vect_slp): Ditto.
1970 * dbgcnt.def: New debug counter.
1971 * Makefile: New dependency.
1973 2013-09-03 Meador Inge <meadori@codesourcery.com>
1977 2013-08-30 Meador Inge <meadori@codesourcery.com>
1979 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1981 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
1983 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
1984 function descriptor.
1986 2013-09-03 Richard Biener <rguenther@suse.de>
1988 * tree-affine.c (add_elt_to_tree): Fix association issue,
1989 avoid useless converts and make sure to always return a
1990 properly typed result.
1992 2013-09-03 Richard Biener <rguenther@suse.de>
1995 * fold-const.c (negate_expr_p): Fix division case.
1996 (negate_expr): Likewise.
1998 2013-09-03 Richard Biener <rguenther@suse.de>
2001 * tree-streamer-out.c: Include tm.h.
2002 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
2004 2013-09-03 Jan Hubicka <jh@suse.cz>
2006 * tree-profile.c (tree_profiling): Cleanup CFG when done.
2008 2013-09-03 Alan Modra <amodra@gmail.com>
2010 * config.gcc (powerpc*-*-linux*): Add support for little-endian
2011 multilibs to big-endian target and vice versa.
2012 * config/rs6000/t-linux64: Use := assignment on all vars.
2013 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2014 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2015 * config/rs6000/t-linux64le: New file.
2016 * config/rs6000/t-linux64bele: New file.
2017 * config/rs6000/t-linux64lebe: New file.
2019 2013-09-02 Jan Hubicka <jh@suse.cz>
2021 * ipa-inline-transform.c (inline_transform): Do not
2022 optimize_inline_calls when not optimizing.
2024 2013-09-02 Jan Hubicka <jh@suse.cz>
2026 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2027 duplicated nodes for assembler names.
2028 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2029 hard registers from assembler name hash.
2031 2013-09-02 Jan Hubicka <jh@suse.cz>
2033 * ipa-split.c (execute_split_functions): Split externally visible
2034 functions called once.
2036 2013-09-02 Martin Jambor <mjambor@suse.cz>
2039 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2040 linked list. When finding the correct duplicate, also consider also
2041 the caller in additon to its inlined_to node.
2043 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2045 * config/aarch64/aarch64-simd-builtins.def
2046 (dup_lane_scalar): Remove.
2047 * config/aarch64/aarch64-simd.md
2048 (aarch64_simd_dup): Add 'w->w' alternative.
2049 (aarch64_dup_lane<mode>): Allow for VALL.
2050 (aarch64_dup_lane_scalar<mode>): Remove.
2051 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2052 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2053 (aarch64_get_lane_unsigned<mode>): Likewise.
2054 (aarch64_get_lane<mode>): Likewise.
2055 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2056 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2057 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2058 (VCON): Change container of V2SF.
2059 (vswap_width_name): Likewise.
2060 * config/aarch64/arm_neon.h
2061 (__aarch64_vdup_lane_any): New.
2062 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2063 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2064 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2066 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2069 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2070 the source or the destination is a single hard register.
2072 2013-09-02 Richard Biener <rguenther@suse.de>
2075 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2078 2013-09-02 Richard Biener <rguenther@suse.de>
2080 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2081 arithmetic to sizetype.
2083 2013-09-02 Bin Cheng <bin.cheng@arm.com>
2085 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2086 Find auto-increment use both before and after candidate.
2088 2013-09-02 Marek Polacek <polacek@redhat.com>
2090 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2092 2013-09-01 Jan Hubicka <jh@suse.cz>
2094 * Makefile.in: Add ipa-profile.o
2095 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2096 * cgraph.c (struct cgraph_propagate_frequency_data,
2097 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2098 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2099 * cgraph.h (cgraph_propagate_frequency): Remove.
2100 * ipa-inline-analysis.c: Include ipa-utils.h;
2101 drop duplicated cfgloop.h.
2102 (inline_update_callee_summaries): Update.
2103 * ipa-profile.c: New file.
2104 * ipa-utils.h (ipa_propagate_frequency): Declare.
2105 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
2106 data-streamer.h, value-prof.h.
2107 (symtab_remove_unreachable_nodes): Update profile.
2108 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2109 account_time_size, cmp_counts, dump_histogram,
2110 ipa_profile_generate_summary, ipa_profile_write_summary,
2111 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2112 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
2113 Move to ipa-profile.c.
2115 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
2117 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2119 2013-09-01 Jan Hubicka <jh@suse.cz>
2121 * common.opt (fdevirtualize-speculatively): New function.
2122 * invoke.texi (fdevirtualize-speculatively): Document.
2123 * ipa-devirt.c: Include ipa-inline.h
2124 (likely_target_p): New function.
2125 (ipa_devirt): New function.
2126 (gate_ipa_devirt): New function.
2127 (pass_data_ipa_devirt): New static var.
2128 (pass_ipa_devirt): Likewise.
2129 (make_pass_ipa_devirt): New function.
2130 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2131 (common_handle_option): Disable devirtualization when
2132 value range profiling is available.
2133 * passes.def (pass_ipa_devirt): Add.
2134 * timever.def (TV_IPA_DEVIRT): New timevar.
2135 * tree-pass.h (make_pass_ipa_devirt):
2137 2013-09-01 Iain Sandoe <iain@codesourcery.com>
2139 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2140 include sanitize(undefined).
2142 2013-08-31 Diego Novillo <dnovillo@google.com>
2144 * Makefile.in (TREE_CORE_H): Define.
2146 (GTFILES): Add tree-core.h.
2147 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2149 * tree-core.h: New file.
2150 Move all data structures, enum, typedefs, global
2151 declarations and constants from ...
2154 2013-08-31 Jan Hubicka <jh@suse.cz>
2156 * bulitins.c (expand_builtin): Do not early exit for gcov
2157 instrumented functions.
2159 2013-08-31 Marek Polacek <polacek@redhat.com>
2161 * ubsan.c: Include tm_p.h.
2163 2013-08-31 Jan Hubicka <jh@suse.cz>
2165 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2168 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2169 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
2170 * tree-cfg.c (verify_gimple_label): ... here.
2171 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2172 (ipa_merge_profiles): New function.
2173 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2174 (lto_input_function_body): Likewise.
2175 * ipa-utils.h (ipa_merge_profiles): Declare.
2176 * lto-streamer.h (lto_input_function_body): Update prototype.
2177 (emit_label_in_global_context_p): Remove.
2178 * lto-symtab.c: Include ipa-utils.h
2179 (lto_cgraph_replace_node): Use ipa_merge_profiles.
2181 2013-08-31 Jan Hubicka <jh@suse.cz>
2183 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2185 2013-08-31 Jan Hubicka <jh@suse.cz>
2187 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2189 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
2191 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2192 "cmp" RTX before signed_comparison_operator check to account
2195 2013-08-30 Jan Hubicka <jh@suse.cz>
2197 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
2198 (detect_type_change_1): Rename to ...
2199 (detect_type_change): ... this one; early return on non-polymorphic
2201 (detect_type_change_ssa): Add comp_type parameter; update
2202 use of detect_type_change.
2203 (compute_complex_assign_jump_func): Add param_type parameter;
2204 update use of detect_type_change_ssa.
2205 (compute_complex_ancestor_jump_func): Likewise.
2206 (ipa_get_callee_param_type): New function.
2207 (ipa_compute_jump_functions_for_edge): Compute parameter type;
2208 update calls to the jump function computation functions.
2210 2013-08-30 Teresa Johnson <tejohnson@google.com>
2211 Steven Bosscher <steven@gcc.gnu.org>
2213 * cfgrtl.c (fixup_new_cold_bb): New routine.
2214 (commit_edge_insertions): Invoke fixup_partitions.
2215 (find_partition_fixes): New routine.
2216 (fixup_partitions): Ditto.
2217 (verify_hot_cold_block_grouping): Update comments.
2218 (rtl_verify_edges): Invoke find_partition_fixes.
2219 (rtl_verify_bb_pointers): Update comments.
2220 (rtl_verify_bb_layout): Ditto.
2221 * basic-block.h (probably_never_executed_edge_p): Declare.
2222 (fixup_partitions): Ditto.
2223 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2224 * bb-reorder.c (sanitize_hot_paths): New function.
2225 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2227 * predict.c (probably_never_executed_edge_p): New routine.
2228 * cfg.c (check_bb_profile): Add partition insanity warnings.
2230 2013-08-30 Meador Inge <meadori@codesourcery.com>
2232 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2234 2013-08-30 Marek Polacek <polacek@redhat.com>
2236 * Makefile.in (ubsan.o): Add.
2237 (c-family/c-ubsan.o): Add.
2238 (builtins.o): Add ubsan.h dependency.
2239 * ubsan.h: New file.
2240 * ubsan.c: New file.
2241 * common.opt: Add -fsanitize=undefined option.
2242 (flag_sanitize): Add variable.
2243 (fsanitize=): Add option. Add Driver.
2244 (fsanitize=thread): Remove option.
2245 (fsanitize=address): Likewise.
2246 (static-libubsan): New option.
2247 * doc/invoke.texi: Document the new flag and -static-libubsan.
2248 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2249 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2250 * builtin-attrs.def (ATTR_COLD): Define.
2251 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2252 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2253 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2254 * flag-types.h (sanitize_code): New enum.
2255 * opts.c (common_handle_option): Parse command line arguments
2256 of -fsanitize=. Add -fsanitize=unreachable option.
2257 * varasm.c (get_variable_section): Adjust.
2258 (assemble_noswitch_variable): Likewise.
2259 (assemble_variable): Likewise.
2260 (output_constant_def_contents): Likewise.
2261 (categorize_decl_for_section): Likewise.
2262 (place_block_symbol): Likewise.
2263 (output_object_block): Likewise.
2264 * builtins.def: Likewise.
2265 * toplev.c (compile_file): Likewise.
2266 (process_options): Likewise.
2267 * cppbuiltin.c: Likewise.
2268 * tsan.c (tsan_pass): Likewise.
2269 (tsan_gate): Likewise.
2270 (tsan_gate_O0): Likewise.
2271 * cfgexpand.c (partition_stack_vars): Likewise.
2272 (expand_stack_vars): Likewise.
2273 (defer_stack_allocation): Likewise.
2274 (expand_used_vars): Likewise.
2275 * cfgcleanup.c (old_insns_match_p): Likewise.
2276 * asan.c (asan_finish_file): Likewise.
2277 (asan_instrument): Likewise.
2278 (gate_asan): Likewise.
2279 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2280 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2281 (asan_global_struct): Use pointer_sized_int_node instead
2282 calling build_nonstandard_integer_type.
2283 (initialize_sanitizer_builtins): Likewise.
2284 (asan_finish_file): Likewise.
2285 * gcc.c: Document %{%:function(args):X}.
2286 (static_spec_functions): Add sanitize.
2287 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2288 store funcval != NULL there.
2289 (do_spec_1): Adjust handle_spec_function caller.
2290 (handle_braces): Allow %:function(args) as condition.
2291 (sanitize_spec_function): New function.
2292 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2293 (LIBUBSAN_SPEC): Likewise.
2294 (LIBUBSAN_EARLY_SPEC): Likewise.
2295 (SANITIZER_SPEC): Handle libubsan.
2296 (SANITIZER_EARLY_SPEC): Likewise.
2297 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2298 instead of fsanitize=address.
2299 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2300 instead of fsanitize=address*.
2301 * builtins.c: Include ubsan.h.
2302 (fold_builtin_0): Instrument __builtin_unreachable.
2303 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2304 instead of flag_asan.
2305 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2306 (pointer_sized_int_node): Define.
2307 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
2309 2013-08-30 Mike Stump <mikestump@comcast.net>
2311 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2314 2013-08-29 Jan Hubicka <jh@suse.cz>
2316 * cgraph.c (cgraph_function_body_availability): Handle weakref
2318 * passes.def: Remove pass_fixup_cfg.
2319 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2320 track when we need to remove functions.
2321 (gate_ipa_inline): Execute inlining always; add comment why.
2322 (pass_data_ipa_inline): Remove TODO_remove_functions.
2323 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2324 do not produce summaries.
2325 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2326 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2327 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2328 constant pool and vtable.
2330 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2332 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2333 New arm_neon.h's internal macros to specify 64-bit constants.
2334 Avoid using stdint.h's macros.
2336 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2338 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2339 before applying REG_P.
2341 2013-08-30 Jakub Jelinek <jakub@redhat.com>
2343 PR tree-optimization/58277
2344 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2345 after seeing too many stmts with vdef in between dombb and current
2346 bb, invalidate everything.
2348 2013-08-30 Richard Biener <rguenther@suse.de>
2350 * fold-const.c (fold_single_bit_test): Fix overflow test.
2352 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2354 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2355 and which can live in a register, always retrieve the value on entry.
2356 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2357 passed by invisible reference specially.
2358 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2359 (vt_add_function_parameter): Correctly deal with a parameter passed by
2360 invisible reference.
2362 2013-08-30 Jan Hubicka <jh@suse.cz>
2364 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2366 2013-08-30 Richard Biener <rguenther@suse.de>
2368 PR tree-optimization/58228
2369 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2370 allow invariant loads in nested loop vectorization.
2372 2013-08-30 Richard Biener <rguenther@suse.de>
2374 PR tree-optimization/58223
2375 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2376 (has_anti_or_output_dependence): ... this and adjust to also
2377 look for output dependences.
2378 (mark_nodes_having_upstream_mem_writes): Adjust.
2379 (rdg_flag_uses): Likewise.
2381 2013-08-30 Richard Biener <rguenther@suse.de>
2383 PR tree-optimization/58010
2384 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2385 assert that we have a loop-closed PHI.
2387 2013-08-29 Jan Hubicka <jh@suse.cz>
2389 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2390 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2391 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2392 (lto_free_function_in_decl_state_for_node): New function.
2394 2013-08-29 Xinliang David Li <davidxl@google.com>
2396 * loop-unroll.c (report_unroll_peel): Minor message change.
2397 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2398 Emit alignment peeling message with default -fopt-info.
2399 (vect_loop_versioning): Emit loop version info message.
2400 * tree-vectorizer.c (vectorize_loops): Minor message change.
2401 (execute_vect_slp): Ditto.
2403 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2405 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2406 of the clone from the DECL_NAME of the original function.
2408 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2410 * passes.c (register_pass): Add overload.
2411 * tree-pass.h (register_pass): Forward declare it. Add comment.
2413 2013-08-29 Jan Hubicka <jh@suse.cz>
2415 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2416 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2417 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2418 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2419 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2420 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
2421 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2422 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2424 2013-08-29 Teresa Johnson <tejohnson@google.com>
2426 * dumpfile.c (dump_loc): Output column number.
2427 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2428 * doc/invoke.texi: Document optall -fopt-info flag.
2429 * profile.c (read_profile_edge_counts): Use new dump framework.
2430 (compute_branch_probabilities): Ditto.
2431 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2432 when pass not in any opt group.
2433 * pass_manager.h (pass_manager::get_pass_profile): New method.
2434 * value-prof.c (check_counter): Use new dump framework.
2435 (check_ic_target): Ditto.
2436 * coverage.c (get_coverage_counts): Ditto.
2437 (coverage_init): Setup new dump framework.
2439 2013-08-29 Richard Biener <rguenther@suse.de>
2441 PR tree-optimization/58246
2442 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2443 handle the dominance check inside a basic-block.
2445 2013-08-29 Richard Biener <rguenther@suse.de>
2448 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2449 of default defs that appear in abnormal PHI nodes.
2451 2013-08-29 Richard Biener <rguenther@suse.de>
2453 PR tree-optimization/57685
2454 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2455 single-use operands to avoid exponential complexity.
2457 2013-08-28 Dehao Chen <dehao@google.com>
2459 * ipa-inline.c (edge_badness): Fix integer underflow.
2461 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
2463 * gtm-builtins.def (_ITM_free): Declare leaf.
2465 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2468 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2469 (*tls_local_dynamic_base_64_largepic): Likewise.
2470 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2471 Remove predicate from call operand.
2472 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2473 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2475 2013-08-28 Jeff Law <law@redhat.com>
2477 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2478 checks for the number of predecessors and successors allowed.
2479 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2480 which require copying a joiner block if there is a request which
2481 is a subpath that requires no joiner block copying.
2483 2013-08-28 Jan Hubicka <jh@suse.cz>
2485 * lto-streamer-out.c (DFS_write_tree_body): Drop
2486 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2487 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2488 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2489 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2490 Do not read DECL_ERROR_ISSUED.
2491 (unpack_ts_decl_with_vis_value_fields): Do not read
2493 (lto_input_ts_binfo_tree_pointers): Do not read
2494 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
2495 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2496 write DECL_ERROR_ISSUED..
2497 (pack_ts_decl_with_vis_value_fields): Do not write
2499 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
2500 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
2501 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2502 * tree.h (tree_decl_common): Update comment.
2503 (DECL_ERROR_ISSUED): Remove.
2505 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2508 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2510 2013-08-28 Jan Hubicka <jh@suse.cz>
2512 * builtins.def (free): Declare leaf.
2514 2013-08-27 David Malcolm <dmalcolm@redhat.com>
2517 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2518 * configure: Regenerate.
2520 2013-08-27 Martin Jambor <mjambor@suse.cz>
2522 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2523 (ipa_ancestor_jf_data): Likewise.
2524 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2525 (ipa_get_jf_pass_through_type_preserved): New function.
2526 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2527 (ipa_get_jf_ancestor_type_preserved): New function.
2528 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
2529 (ipa_get_jf_ancestor_result): Likewise.
2530 (propagate_vals_accross_pass_through): Use
2531 ipa_get_jf_pass_through_result to do all the value mappings.
2532 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2533 type_preserved flag.
2534 (ipa_set_jf_cst_copy): New function.
2535 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2536 (ipa_set_jf_arith_pass_through): Likewise.
2537 (ipa_set_ancestor_jf): Likewise.
2538 (compute_complex_assign_jump_func): Set type_preserved instead of
2540 (ipa_compute_jump_functions_for_edge): Likewise.
2541 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2542 (update_jump_functions_after_inlining): Update type_preserved.
2543 Explicitely create jump functions when combining one with pass_through.
2544 (ipa_write_jump_function): Stream the type_preserved flags.
2545 (ipa_read_jump_function): Likewise.
2547 2013-08-27 Jakub Jelinek <jakub@redhat.com>
2548 Aldy Hernandez <aldyh@redhat.com>
2550 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2551 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2552 * function.h (struct function): Add has_force_vect_loops and
2554 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
2555 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2557 * gimple.def: Update CLAUSES comments.
2558 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2559 (gimple_build_omp_for): Add argument to prototype.
2560 (gimple_omp_for_kind): New.
2561 (gimple_omp_for_set_kind): New.
2562 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2563 GOVD_DATA_SHARE_CLASS.
2564 (enum omp_region_type): Add ORT_SIMD.
2565 (gimple_add_tmp_var): Handle ORT_SIMD.
2566 (gimplify_var_or_parm_decl): Same.
2567 (is_gimple_stmt): Same.
2568 (omp_firstprivatize_variable): Same.
2569 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2570 (omp_notice_variable): Handle ORT_SIMD.
2571 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
2572 (omp_check_private): Handle ORT_SIMD.
2573 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2575 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2576 Handle OMP_CLAUSE_LASTPRIVATE.
2577 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2579 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2580 (gimplify_expr): Handle OMP_SIMD.
2581 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2582 (expand_GOMP_SIMD_VF): New.
2583 (expand_GOMP_SIMD_LAST_LANE): New.
2584 * internal-fn.def (GOMP_SIMD_LANE): New.
2585 (GOMP_SIMD_VF): New.
2586 (GOMP_SIMD_LAST_LANE): New.
2587 * omp-low.c: Include target.h.
2588 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2590 (check_omp_nesting_restrictions): Same.
2592 (lower_rec_simd_input_clauses): New.
2593 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2595 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2596 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2597 (expand_omp_build_assign): New.
2598 (expand_omp_for_init_counts): New.
2599 (expand_omp_for_init_vars): New.
2600 (extract_omp_for_update_vars): New.
2601 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2602 and rewrite accordingly.
2603 (expand_omp_simd): New.
2604 (expand_omp_for): Use expand_omp_simd.
2605 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2606 (lower_omp_for): Do not lower the body.
2607 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2609 * tree-flow.h (find_omp_clause): Remove prototype.
2610 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2611 forcing vectorization of the loop, or if flag_tree_vectorize.
2612 (gate_tree_if_conversion): Similarly.
2613 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2614 gimple_build_omp_for.
2615 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2616 * tree-parloops (create_parallel_loop): Pass kind argument to
2617 gimple_build_omp_for.
2618 * tree-pretty-print.c (dump_omp_clause): Add cases for
2619 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2620 OMP_CLAUSE__SIMDUID_.
2621 (dump_generic_node): Handle OMP_SIMD.
2622 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2623 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2624 unroll OMP_SIMD loops here.
2625 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
2626 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
2628 (vect_analyze_data_refs): Handle simd loops.
2629 * tree-vect-loop.c (vectorizable_live_operation): Handle
2631 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
2632 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2633 (vectorizable_load): Same.
2634 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
2635 (struct simduid_to_vf): New.
2636 (simduid_to_vf::hash): New.
2637 (simduid_to-vf::equal): New.
2638 (struct simd_array_to_simduid): New.
2639 (simd_array_to_simduid::hash): New.
2640 (simd_array_to_simduid::equal): New.
2641 (adjust_simduid_builtins): New.
2642 (struct note_simd_array_uses_struct): New.
2643 (note_simd_array_uses_cb): New.
2644 (note_simd_array_uses): New.
2645 (vectorize_loops): Handle simd hints and adjust simd builtins
2647 * tree-vectorizer.h (struct _stmt_vec_info): Add
2648 simd_lane_access_p field.
2649 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2650 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2651 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2652 (omp_clause_code_name): Same.
2653 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2654 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2655 * tree.def (OMP_SIMD): New entry.
2656 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2657 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
2658 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2659 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2660 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2661 (OMP_CLAUSE_LINEAR_STEP): New.
2662 (OMP_CLAUSE_SAFELEN_EXPR): New.
2663 (OMP_CLAUSE__SIMDUID__DECL): New.
2664 (find_omp_clause): New prototype.
2666 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2668 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2669 Haswell processor detection.
2671 2013-08-27 Christian Widmer <shadow@umbrox.de>
2674 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2675 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2676 AVX2 capable processors.
2678 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2680 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2681 of vget_low_* with implementations in terms of other intrinsics.
2683 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
2686 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2687 values to -1, 0 and 1.
2689 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2691 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2692 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2693 (<optab><mode>3_insn): Remove.
2694 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2695 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2696 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2697 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2698 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2699 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2700 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2701 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2702 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2703 (ror<mode>3_insn): Likewise.
2704 * config/aarch64/predicates.md (aarch64_simd_register): New.
2706 2013-08-27 Richard Biener <rguenther@suse.de>
2708 PR tree-optimization/57521
2709 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2710 one edge is non-critical.
2711 (find_phi_replacement_condition): Make sure to use a non-critical
2712 edge. Cleanup and remove old bug workarounds.
2713 (bb_postdominates_preds): Remove.
2714 (if_convertible_loop_p_1): Do not compute post-dominators.
2715 (combine_blocks): Do not free post-dominators.
2716 (main_tree_if_conversion): Likewise.
2717 (pass_data_if_conversion): Add TODO_verify_ssa.
2719 2013-08-27 DJ Delorie <dj@redhat.com>
2721 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2723 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2725 * function.c (assign_parm_find_data_types): Set passed_mode and
2726 nominal_mode to the TYPE_MODE of nominal_type for the built
2727 pointer type in case of the struct-pass-by-reference.
2729 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2731 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2732 (UINT16_TYPE): Change default to "unsigned int".
2734 * config/avr/avr.opt (mfract-convert-truncate): New option.
2735 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2736 is set, round negative fractional integers according to n1169
2737 when converting to integer types.
2739 2013-08-26 Jan Hubicka <jh@suse.cz>
2741 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2742 methods can not be called indirectly when their address is not taken.
2744 2013-08-26 Jan Hubicka <jh@suse.cz>
2746 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2749 2013-08-26 Jan Hubicka <jh@suse.cz>
2751 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2754 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2756 * reload.c (find_valid_class): Allow classes that do not include
2757 FIRST_PSEUDO_REGISTER - 1.
2759 2013-08-26 Jan Hubicka <jh@suse.cz>
2761 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2762 fix edge count/frequency when speculation failed; fix type check
2763 for the direct call.
2765 2013-08-26 Jan Hubicka <jh@suse.cz>
2767 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2769 2013-08-26 Jan Hubicka <jh@suse.cz>
2771 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2772 to be changed by edge redirection.
2774 2013-08-26 Jan Hubicka <jh@suse.cz>
2776 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2777 formating; add sanity check.
2778 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2779 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2780 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2781 (resolve_noninline_speculation): Update callee keys, too.
2783 2013-08-26 Jan Hubicka <jh@suse.cz>
2785 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2786 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2788 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2790 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2793 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
2795 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2796 (x86_64_elf_select_section): Ditto.
2797 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2799 (ix86_pass_by_reference): Ditto.
2800 (output_set_got): Ditto.
2801 (ix86_unary_operator_ok): Ditto.
2802 (ix86_expand_builtin): Ditto.
2804 2013-08-23 Jan Hubicka <jh@suse.cz>
2806 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
2808 2013-08-23 Jan Hubicka <jh@suse.cz>
2810 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
2811 (tree_decl_with_vis): Add FINAL field.
2813 2013-08-23 Jeff Law <law@redhat.com>
2815 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
2816 the debugging dump when the expression is fully redundant.
2818 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2820 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
2821 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
2822 * pretty-print.c (pp_formatted_text_data): Likewise.
2823 (pp_write_text_to_stream): Likewise.
2824 (pp_write_text_as_dot_label_to_stream): Likewise.
2825 (pp_append_r): Likewise.
2826 (pp_format): Likewise.
2827 (pp_flush): Likewise.
2828 (pp_clear_output_area): Likewise.
2829 (pp_append_text): Likewise.
2830 (pp_formatted_text): Likewise.
2831 (pp_remaining_character_count_for_line): Likewise.
2832 (pp_newline): Likewise.
2833 (pp_character): Likewise.
2834 (output_buffer::~output_buffer): Define.
2835 (pretty_printer::~pretty_printer): Destruct output buffer.
2836 * pretty-print.h (output_buffer::~output_buffer): Declare.
2837 (pretty_printer::~pretty_printer): Declare virtual.
2839 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
2842 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
2843 HOST_WIDE_INT_M1U): New macros.
2844 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
2845 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
2846 unsigned -1 for lshift.
2847 * cse.c (cse_insn): Likewise.
2848 * double-int.c (rshift_double, lshift_double): Likewise.
2849 * builtins.c (fold_builtin_bitop): Likewise.
2850 * combine.c (force_to_mode): Likewise.
2851 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
2852 * simplify-rtx.c (simplify_const_unary_operation,
2853 simplify_const_binary_operation): Likewise.
2854 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
2855 check_va_list_escapes): Likewise.
2856 * rtlanal.c (nonzero_bits1): Likewise.
2857 * expmed.c (expand_smod_pow2): Likewise.
2858 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
2860 2013-08-23 Jan Hubicka <jh@suse.cz>
2862 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
2863 as having address taken.
2865 2013-08-23 Jan Hubicka <jh@suse.cz>
2867 * ipa-utils.h (method_class_type): Declare.
2868 * ipa-devirt.c (method_class_type): Export.
2870 * cgraphunit.c (analyze_functions): Do basic devirtualization;
2871 do not walk base classes of anonymous types.
2873 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
2875 PR rtl-optimization/58220
2877 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
2878 handle SEQUENCE insns properly.
2880 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2882 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
2884 (pp_newline_and_indent): Likewise.
2885 (pp_separate_with): Likewise.
2886 * pretty-print.c (pp_newline_and_flush): Define.
2887 (pp_newline_and_indent): Likewise.
2888 (pp_separate_with): Likewise.
2890 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2893 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
2894 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
2896 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
2898 * config/i386/predicates.md (ext_sse_reg_operand): New.
2899 * config/i386/i386.md (*movti_internal): Use
2900 predicate to determine if EVEX is needed.
2901 (*movsi_internal): Ditto.
2902 (*movdf_internal): Ditto.
2903 (*movsf_internal): Ditto.
2904 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2906 2013-08-23 Jakub Jelinek <jakub@redhat.com>
2908 PR tree-optimization/58209
2909 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
2910 (find_tail_calls): Give up for pointer result types if m is non-NULL.
2911 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
2912 emit POINTER_PLUS_EXPR.
2913 (create_tailcall_accumulator): For pointer result type accumulate in
2916 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
2918 * configure.ac: Add backslashes missing from the last change.
2919 * configure: Regenerate.
2921 2013-08-22 Jan Hubicka <jh@susue.cz>
2923 * ipa.c (function_and_variable_visibility): First remember function
2924 was global and then make it local.
2926 2013-08-22 Julian Brown <julian@codesourcery.com>
2928 * configure.ac: Add aarch64 to list of arches which use "nop" in
2930 * configure: Regenerate.
2932 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2934 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
2935 gnu_libc_has_function.
2936 * config/s390/tpf.h: Likewise.
2938 2013-08-22 Jan Hubicka <jh@susue.cz>
2940 * timevar.c (validate_phases): Add cast.
2942 2013-08-22 Jan Hubicka <jh@susue.cz>
2944 * timevar.c (validate_phases): Use size_t for memory.
2945 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
2947 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
2949 * pretty-print.h (output_buffer::output_buffer): Declare.
2950 (pretty_printer::pretty_printer): Likewise.
2951 (pp_construct): Remove.
2952 * pretty-print.c (output_buffer::output_buffer): Define.
2953 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
2954 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
2955 (print_gimple_expr): Likewise.
2956 (print_gimple_seq): Likewise.
2957 (gimple_dump_bb): Likewise.
2958 * sched-vis.c (dump_value_slim): Likewise.
2959 (dump_insn_slim): Likewise.
2960 (dump_rtl_slim): Likewise.
2961 (str_pattern_slim): Likewise.
2962 * tree-mudflap.c (mf_varname_tree): Likewise.
2963 * graph.c (print_graph_cfg): Likewise.
2964 (start_graph_dump): Likewise.
2965 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
2967 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
2968 pretty printer initialization.
2969 * coretypes.h (diagnostic_context): Remove superflous type alias
2971 (pretty_printer): Likewise. Declare directly as a class.
2972 (pretty_print_info): Remove declaration as class.
2973 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
2974 and pp_clear_output_area.
2975 (asan_add_global): Likewise.
2977 2013-08-22 Jan Hubicka <jh@suse.cz>
2979 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
2980 * ipa-utils.h (update_type_inheritance_graph): Declare.
2981 (possible_polymorphic_call_target_p): Declare.
2982 (possible_polymorphic_call_target_p): New.
2983 * ipa-devirt.c: Update toplevel comments.
2984 (cached_polymorphic_call_targets): Move up.
2985 (odr_type_d): Move ID down.
2986 (polymorphic_type_binfo_p): Update comment.
2987 (odr_hasher::remove): Likewise;
2988 (get_odr_type): Set anonymous_namespace.
2989 (dump_odr_type): Dump it.
2990 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
2991 (maybe_record_node): Record node in cached_polymorphic_call_targets.
2992 (record_binfo): Add comment.
2993 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
2995 (devirt_node_removal_hook): Do not iCE when cache is freed.
2996 (possible_polymorphic_call_target_p): New predicate.
2997 (update_type_inheritance_graph): New function.
2999 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
3000 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3001 Sergey Lega <sergey.s.lega@intel.com>
3002 Anna Tikhonova <anna.tikhonova@intel.com>
3003 Ilya Tocar <ilya.tocar@intel.com>
3004 Andrey Turetskiy <andrey.turetskiy@intel.com>
3005 Ilya Verbin <ilya.verbin@intel.com>
3006 Kirill Yukhin <kirill.yukhin@intel.com>
3007 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3009 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
3010 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
3011 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
3012 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3013 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3014 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3015 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3016 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3017 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3018 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3019 OPT_mavx512pf, OPT_mavx512er cases.
3020 * config/i386/constraints.md (v): New constraint.
3021 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3022 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3023 (bit_AVX512CD): New.
3024 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3025 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3026 * config/i386/i386-c.c (ix86_target_macros_internal):
3027 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3029 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3030 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3031 * config/i386/i386.c (regclass_map, dbx_register_map)
3032 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3033 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3034 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3036 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3037 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3038 -mavx512pf options. Fix formatting.
3039 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3040 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3041 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3042 -mavx512cd, -mavx512pf options.
3043 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3044 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3045 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3047 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3048 (ix86_set_reg_reg_cost): Ditto.
3049 (x86_order_regs_for_local_alloc): Ditto.
3050 (MAX_VECT_LEN): Extend to 64-byte.
3051 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3052 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3053 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3054 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3055 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3056 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3057 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3058 (SSE_REG_MODE_P): Support new modes.
3059 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3060 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3061 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3062 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3063 (REG_CLASS_CONTENTS): Add new registers.
3064 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3065 (EXT_REX_SSE_REGNO_P): New.
3066 (HI_REGISTER_NAMES): Add new registers.
3067 * config/i386/i386.md: Define constants for new registers.
3068 (mode): Add new 512-bit modes.
3069 (prefix): Support evex prefix.
3070 (isa): Support avx512f, noavx512f, fma_avx512f.
3071 (ssemodesuffix): Add new 512-bit modes.
3073 (*movxi_internal_avx512f): Ditto.
3074 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3076 (*movsi_internal): Likewise.
3077 (*movdf_internal): Likewise.
3078 (*movsf_internal): Likewise.
3079 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3080 (<code><mode>3): Likewise.
3081 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3083 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3084 with the new constraint "v".
3085 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3087 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3088 with the new constraint "v".
3089 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3090 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3091 (*<plusminus_insn><mode>3): Likewise.
3092 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3093 (*mul<mode>3): Likewise.
3094 (<sse>_vmmul<mode>3): Likewise.
3095 (<sse>_div<mode>3): Likewise.
3096 (<sse>_vmdiv<mode>3): Likewise.
3097 (<sse>_sqrt<mode>2): Likewise.
3098 (<sse>_vmsqrt<mode>2): Likewise.
3099 (*<code><mode>3_finite): Likewise.
3100 (*<code><mode>3) <smaxmin>: Likewise.
3101 (<sse>_vm<code><mode>3): Likewise.
3102 (*<code><mode>3) <any_logic>: Likewise.
3103 (*fma_fmadd_<mode>): Likewise.
3104 (*fma_fmsub_<mode>): Likewise.
3105 (*fma_fnmadd_<mode>): Likewise.
3106 (*fma_fnmsub_<mode>): Likewise.
3107 (*fma_fmaddsub_<mode>): Likewise.
3108 (*fma_fmsubadd_<mode>): Likewise.
3109 (*fmai_fmadd_<mode>): Likewise.
3110 (*fmai_fmsub_<mode>): Likewise.
3111 (*fmai_fnmadd_<mode>): Likewise.
3112 (*fmai_fnmsub_<mode>): Likewise.
3113 (sse_cvtsi2ss): Likewise.
3114 (sse_cvtsi2ssq): Likewise.
3115 (sse_cvtss2si): Likewise.
3116 (sse_cvtss2si_2): Likewise.
3117 (sse_cvtss2siq): Likewise.
3118 (sse_cvtss2siq_2): Likewise.
3119 (sse_cvttss2si): Likewise.
3120 (sse_cvtss2siq_2): Likewise.
3121 (float<sseintvecmodelower><mode>2): Likewise.
3122 (sse2_cvtsd2si_2): Likewise.
3123 (sse2_cvtsd2siq_2): Likewise.
3124 (*<plusminus_insn><mode>3): Likewise.
3125 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3126 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3127 (ashr<mode>3): Likewise.
3128 (<shift_insn><mode>3): Likewise.
3129 (avx2_<code><mode>3): Likewise.
3130 (*avx2_<code><mode>3): Likewise.
3131 (*andnot<mode>3): Likewise.
3132 (*<code><mode>3) <any_logic>: Likewise.
3133 (abs<mode>2): Likewise.
3134 (avx2_permvar<mode>): Likewise.
3135 (avx2_perm<mode>_1): Likewise.
3136 (*avx_vpermilp<mode>): Likewise.
3137 (avx_vpermilvar<mode>3): Likewise.
3138 (avx2_ashrv<mode>): Likewise.
3139 (avx2_<shift_insn>v<mode>): Likewise.
3140 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3142 * doc/rtl.texi: Document XImode.
3144 2013-08-21 Jeff Law <law@redhat.com>
3146 * tree-flow.h (register_jump_thread): Pass vector of edges
3147 instead of each important edge.
3148 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3149 thread path into a vector and pass that to register_jump_thread.
3150 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3151 passed in edge vector to the current 3-edge form.
3154 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3156 * dce.c (fini_dce): Call df_analyze again just in case
3157 delete_unmarked_insns removed anything.
3159 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3161 * reload.h (struct reg_equivs): Rename to ..
3162 (struct reg_equivs_s): .. this.
3164 2013-08-20 Martin Liska <marxin.liska@gmail.com>
3166 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3168 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3170 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
3172 2013-08-21 Jeff Law <law@redhat.com>
3174 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3175 simplify assignments too. If the RHS collapses to a singleton
3176 range, then return the value for the range.
3178 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3180 * config/i386/sse.md (V16): Rename to...
3182 (mov<mode>): Update iterator name.
3183 (*mov<mode>_internal): Ditto.
3184 (push<mode>1): Ditto.
3185 (movmisalign<mode>): Ditto.
3187 2013-08-20 Jan Hubicka <jh@suse.cz>
3190 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3191 entry for direct edges.
3192 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3194 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3196 Revert my last two changes, r201865 and r201864:
3199 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3201 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3202 instances can own GC refs.
3204 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3205 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3206 (gcc::context::gt_pch_nx): Likewise.
3207 (gcc::context::gt_pch_nx): Likewise.
3208 * ggc.h (gt_ggc_mx <T>): New.
3209 (gt_pch_nx_with_op <T>): New.
3210 (gt_pch_nx <T>): New.
3211 * passes.c (opt_pass::gt_ggc_mx): New.
3212 (opt_pass::gt_pch_nx): New.
3213 (opt_pass::gt_pch_nx_with_op): New.
3214 (pass_manager::gt_ggc_mx): New.
3215 (pass_manager::gt_pch_nx): New.
3216 (pass_manager::gt_pch_nx_with_op): New.
3217 (pass_manager::operator new): Use
3218 ggc_internal_cleared_alloc_stat rather than xcalloc.
3219 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3220 (pass_manager::gt_ggc_mx): New.
3221 (pass_manager::gt_pch_nx): New.
3222 (pass_manager::gt_pch_nx_with_op): New.
3223 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3224 (opt_pass::operator new): New.
3225 (opt_pass::gt_ggc_mx): New.
3226 (opt_pass::gt_pch_nx): New.
3227 (opt_pass::gt_pch_nx_with_op): New.
3230 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3232 * Makefile.in (GTFILES): Add context.h.
3233 * context.c (gcc::context::operator new): New.
3234 (gcc::context::gt_ggc_mx): New.
3235 (gcc::context::gt_pch_nx): New.
3236 (gcc::context::gt_pch_nx): New.
3237 * context.h (gcc::context): Add GTY((user)) marking.
3238 (gcc::context::operator new): New.
3239 (gcc::context::gt_ggc_mx): New.
3240 (gcc::context::gt_pch_nx): New.
3241 (gcc::context::gt_pch_nx): New.
3242 (g): Add GTY marking.
3243 (gt_ggc_mx (gcc::context *)): New.
3244 (gt_pch_nx (gcc::context *)): New.
3245 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3246 void *cookie)): New.
3247 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3249 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3251 * dce.c (fini_dce): Call df_analyze again just in case
3252 delete_unmarked_insns removed anything.
3254 2013-08-20 Teresa Johnson <tejohnson@google.com>
3256 PR rtl-optimizations/57451
3257 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3258 from crossing split section boundaries.
3260 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3262 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3263 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3264 with MULTLIB_DEFAULTS.
3266 2013-08-20 Nick Clifton <nickc@redhat.com>
3268 * target.def (narrow_volatile_bitfield): Note that the default
3269 value is false, not !TARGET_STRICT_ALIGN.
3270 * doc/tm.texi: Regenerate.
3272 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3274 Fix LIB_SPEC for systems without libpthread.
3276 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3277 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3279 * config/i386/linux-common.h: Likewise.
3280 * config/mips/linux-common.h: Likewise.
3282 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3284 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3287 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3289 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3290 instances can own GC refs.
3292 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3293 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3294 (gcc::context::gt_pch_nx): Likewise.
3295 (gcc::context::gt_pch_nx): Likewise.
3296 * ggc.h (gt_ggc_mx <T>): New.
3297 (gt_pch_nx_with_op <T>): New.
3298 (gt_pch_nx <T>): New.
3299 * passes.c (opt_pass::gt_ggc_mx): New.
3300 (opt_pass::gt_pch_nx): New.
3301 (opt_pass::gt_pch_nx_with_op): New.
3302 (pass_manager::gt_ggc_mx): New.
3303 (pass_manager::gt_pch_nx): New.
3304 (pass_manager::gt_pch_nx_with_op): New.
3305 (pass_manager::operator new): Use
3306 ggc_internal_cleared_alloc_stat rather than xcalloc.
3307 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3308 (pass_manager::gt_ggc_mx): New.
3309 (pass_manager::gt_pch_nx): New.
3310 (pass_manager::gt_pch_nx_with_op): New.
3311 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3312 (opt_pass::operator new): New.
3313 (opt_pass::gt_ggc_mx): New.
3314 (opt_pass::gt_pch_nx): New.
3315 (opt_pass::gt_pch_nx_with_op): New.
3317 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3319 * Makefile.in (GTFILES): Add context.h.
3320 * context.c (gcc::context::operator new): New.
3321 (gcc::context::gt_ggc_mx): New.
3322 (gcc::context::gt_pch_nx): New.
3323 (gcc::context::gt_pch_nx): New.
3324 * context.h (gcc::context): Add GTY((user)) marking.
3325 (gcc::context::operator new): New.
3326 (gcc::context::gt_ggc_mx): New.
3327 (gcc::context::gt_pch_nx): New.
3328 (gcc::context::gt_pch_nx): New.
3329 (g): Add GTY marking.
3330 (gt_ggc_mx (gcc::context *)): New.
3331 (gt_pch_nx (gcc::context *)): New.
3332 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3333 void *cookie)): New.
3334 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3336 2013-08-20 Alan Modra <amodra@gmail.com>
3339 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3340 (rs6000_emit_epilogue): Likewise.
3342 2013-08-19 Dehao Chen <dehao@google.com>
3344 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3346 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3347 Jakub Jelinek <jakub@redhat.com>
3349 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3350 (BUILT_IN_FABSD64): Likewise.
3351 (BUILT_IN_FABSD128): Likewise.
3352 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3353 (fold_builtin_1): Likewise.
3354 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3355 destination and source operands.
3356 (*abstd2_fpr): Likewise.
3357 (*nabstd2_fpr): Likewise.
3359 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3361 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3364 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
3366 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3369 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3371 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3372 * builtins.c (default_libc_has_function): New.
3373 (gnu_libc_has_function): Ditto.
3374 (no_c99_libc_has_function): Ditto.
3375 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3376 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3377 (fold_builtin_sincos): Likewise.
3378 (fold_builtin_cexp): Likewise.
3379 * builtins.def (DEF_C94_BUILTIN): Likewise.
3380 (DEF_C99_BUILTIN): Likewise.
3381 (DEF_C99_C90RES_BUILTIN): Likewise.
3382 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3383 definitions to using this define.
3384 * config/darwin-protos.h (darwin_libc_has_function): New.
3385 * config/darwin.c (darwin_libc_has_function): Ditto.
3386 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3387 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3388 * config/darwin.h: Ditto.
3389 * config/elfos.h: Ditto.
3390 * config/freebsd.h: Ditto.
3391 * config/i386/cygming.h: Ditto.
3392 * config/i386/djgpp.h: Ditto.
3393 * config/i386/i386-interix.h: Ditto.
3394 * config/microblaze/microblaze.h: Ditto.
3395 * config/mmix/mmix.h: Ditto.
3396 * config/gnu-user.h: Ditto.
3397 * config/ia64/hpux.h: Ditto.
3398 * config/pa/pa-hpux.h: Ditto.
3399 * config/pdp11/pdp11.h: Ditto.
3400 * config/picochip/picochip.h: Ditto.
3401 * config/linux.h: Ditto.
3402 * config/netbsd.h: Ditto.
3403 * config/openbsd.h: Ditto.
3404 * config/rs6000/aix43.h: Ditto.
3405 * config/rs6000/aix51.h: Ditto.
3406 * config/rs6000/aix52.h: Ditto.
3407 * config/rs6000/aix53.h: Ditto.
3408 * config/rs6000/aix61.h: Ditto.
3409 * config/rs6000/darwin.h: Ditto.
3410 * config/rs6000/linux.h: Ditto.
3411 * config/rs6000/linux64.h: Ditto.
3412 * config/s390/tpf.h: Ditto.
3413 * config/sol2-10.h: Ditto.
3414 * config/sol2.h: Ditto.
3415 * config/vms/vms.h: Ditto.
3416 * config/vxworks.h: Ditto.
3417 * config/linux-android.c (linux_android_libc_has_function):
3418 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3419 * config/linux-protos.h (linux_android_libc_has_function):
3421 * config/i386/i386.c (ix86_libc_has_function): New.
3422 * config/i386/i386-protos.h
3423 (ix86_libc_has_function): New declaration.
3424 * config/i386/i386.md
3425 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3426 ("isinf<mode>2): Likewise.
3427 * convert.c (convert_to_integer): Using new target hook
3428 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3429 TARGET_C99_FUNCTIONS.
3430 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3431 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3432 * coretypes.h (function_class): New enum for different
3433 classes of functions.
3434 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3435 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3436 (TARGET_HAS_SINCOS): Likewise.
3437 (TARGET_LIBC_HAS_FUNCTION): New.
3438 * doc/tm.texi: Regenerated.
3439 * targhooks.h (default_libc_has_function): New declaration.
3440 (no_c99_libc_has_function): Ditto.
3441 (gnu_libc_has_function): Ditto.
3442 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3443 and TARGET_HAS_SINCOS.
3445 2013-08-18 Jan Hubicka <jh@suse.cz>
3447 * Makeifle-in (ipa-devirt.o): New.
3448 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3449 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3450 (analyze_functions): Look into possible targets of polymorphic call.
3451 * dumpfile.c (dump_files): Add type-inheritance dump.
3452 * dumpfile.h (TDI_inheritance): New.
3453 * ipa-devirt.c: New file.
3454 * ipa-utils.h (odr_type_d): Forward declare.
3455 (odr_type): New type.
3456 (build_type_inheritance_graph): Declare.
3457 (possible_polymorphic_call_targets): Declare and introduce inline
3458 variant when only edge is pased.
3459 (dump_possible_polymorphic_call_targets): Likewise.
3460 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3461 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3462 (types_same_for_odr): ... here.
3463 * tree.h (type_in_anonymous_namespace_p): Declare.
3465 2013-08-18 Jakub Jelinek <jakub@redhat.com>
3467 PR tree-optimization/58006
3468 * tree-parloops.c (take_address_of): Don't ICE if get_name
3470 (eliminate_local_variables_stmt): Remove clobber stmts.
3472 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3474 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3475 error message is issued for an alias to undefined symbol.
3477 2013-08-18 Jan Hubicka <jh@suse.cz>
3479 * cgraph.c (cgraph_create_indirect_edge): Discover
3480 polymorphic calls and record basic info into indirect_info.
3481 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3482 devirtualization, ignore objc function calls.
3483 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3484 call with no parm index info.
3485 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3486 * tree.c (virtual_method_call_p): New function.
3487 * tree.h (virtual_method_call_p): Declare.
3489 2013-08-16 Jan Hubicka <jh@suse.cz>
3492 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3494 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
3496 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3499 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3501 * gengtype.c (type_for_name): Add special-case support for
3502 locating types within the "gcc::" namespace.
3503 (open_base_files): Emit a "using namespace gcc" directive.
3505 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3508 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3509 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3511 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3512 array instead of each individual operand as a separate argument.
3513 (emit_fusion_gpr_load): Likewise.
3514 (expand_fusion_gpr_load): Add new function declaration.
3516 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3517 signature to have the operands passed as an array, instead of as
3518 separate arguments. Allow ZERO_EXTEND to be in the memory
3519 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3520 depend on the register live/dead flags when peepholes are run.
3521 (expand_fusion_gpr_load): New function to be called from the
3522 peephole2 pass, to change the register that addis sets to be the
3524 (emit_fusion_gpr_load): Change the calling signature to have the
3525 operands passed as an array, instead of as separate arguments.
3526 Allow ZERO_EXTEND to be in the memory address, and also
3527 SIGN_EXTEND if -mpower8-fusion-sign.
3529 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3531 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3532 adjust the register addis loads up in the peephole2 pass. Do not
3533 depend on the register live/dead state when the peephole pass is done.
3535 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3537 * gengtype.c (create_user_defined_type): Ensure that the kind
3538 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3539 declaration is seen before the GTY((user)) marking.
3541 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3544 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3546 2013-08-16 Jan Hubicka <jh@suse.cz>
3548 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3549 arugment expected_type.
3550 (gimple_fold_call): Use it.
3551 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3552 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
3553 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
3554 (try_make_edge_direct_virtual_call): Likewise.
3555 * tree.c (obj_type_ref_class): New.
3556 * tree.h (obj_type_ref_class): Use it.
3558 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3560 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3561 (rtl_slim_pp): Likewise.
3562 (init_rtl_slim_pretty_print): Likewise.
3563 (dump_value_slim): Don't call it. Use local pretty printer.
3564 (dump_insn_slim): Likewise.
3565 (dump_rtl_slim): Likewise.
3566 (str_pattern_slim): Likewise.
3567 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3570 2013-08-16 Jakub Jelinek <jakub@redhat.com>
3572 PR tree-optimization/58164
3573 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3574 walk gimple_goto_dest of GIMPLE_GOTO.
3576 PR tree-optimization/58165
3577 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3578 bi_call must be the last stmt in a bb, don't split_block, instead
3579 use fallthru edge from it and give up if there is none.
3580 Release conds vector when returning early.
3582 2013-08-14 Xinliang David Li <davidxl@google.com>
3584 * config/i386/i386.c (ix86_option_override_internal):
3585 Remove unused variable and field.
3587 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3590 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
3591 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3592 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3593 and Linux, correct BLKmode alignment when 128-bit alignment is
3594 required and compatibility flag is not set.
3595 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3596 for zero-size arguments when compatibility flag is not set.
3598 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3600 PR tree-optimization/58145
3601 * tree-sra.c (build_ref_for_offset): If prev_base has
3602 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3604 2013-08-14 Xinliang David Li <davidxl@google.com>
3606 * config/i386/i386.c (ix86_option_override_internal):
3607 Fix uninitialized variable error.
3609 2013-08-14 Xinliang David Li <davidxl@google.com>
3611 * config/i386/i386.opt: Define two new options.
3612 * config/i386/x86-tune.def: Add arch selector field in macros.
3613 * config/i386/i386.h: Adjust macro definition.
3614 * config/i386/i386.c (ix86_option_override_internal):
3616 (parse_mtune_ctrl_str): New function.
3617 (set_ix86_tune_features): New function.
3618 (ix86_function_specific_restore): Call the new helper function.
3620 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
3622 PR rtl-optimization/57662
3623 * sel-sched.c (code_motion_process_successors): When the current insn
3624 is removed after the recursive traversal, break from the loop.
3625 Add comments and debug printouts.
3627 2013-08-14 Jakub Jelinek <jakub@redhat.com>
3628 Alexandre Oliva <aoliva@redhat.com>
3631 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3632 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3633 there also UNSPEC_PLTOFF.
3635 2013-08-14 Marek Polacek <polacek@redhat.com>
3637 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3638 NUM_CONDITIONS bit positions.
3640 2013-08-13 Cary Coutant <ccoutant@google.com>
3642 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3643 (attr_checksum): Hash vector contents instead of pointer.
3644 (attr_checksum_ordered): Likewise.
3646 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
3648 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3649 when Pmode != word_mode. Add length_address attribute.
3650 (sse3_monitor_<mode>): Merge from sse3_monitor and
3651 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3652 Pmode != word_mode. Update insn length attribute.
3653 * config/i386/i386.c (ix86_option_override_internal): Update
3654 ix86_gen_monitor selection for merged sse3_monitor insn.
3656 2013-08-13 Julian Brown <julian@codesourcery.com>
3658 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3659 perform invalid legitimization on greater-than-word-size modes for
3662 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3664 * ira.c (setup_class_translate_array): Use aclass instead of cl
3665 for classes not fully covered by allocno classes.
3667 2013-08-13 Jakub Jelinek <jakub@redhat.com>
3669 PR tree-optimization/57661
3670 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3671 * tree-inline.c (tree_function_versioning): Initialize it.
3672 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3673 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3674 that is not being copied.
3677 * asan.c (instrument_strlen_call): Fix typo in comment.
3678 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3680 2013-08-13 Steve Ellcey <sellcey@mips.com>
3682 * config/mips/mips.md (prefetch): Use lw instead of ld on
3683 loongson in 32bit mode.
3685 2013-08-13 Nick Clifton <nickc@redhat.com>
3687 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3689 2013-08-13 Jan Hubicka <jh@suse.cz>
3691 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3692 introduced edge; fix typo in sanity check.
3693 (cgraph_resolve_speculation): Export; improve diagnostic.
3694 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3695 speculation at type mismatch.
3696 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3697 (cgraph_resolve_speculation): Declare.
3698 (symtab_can_be_discarded): New function.
3699 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3700 * ipa-inline-transform.c (speculation_removed): New global var.
3701 (clone_inlined_nodes): See if speculation can be removed.
3702 (inline_call): If speculations was removed, we growths may not match.
3703 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3704 (speculation_useful_p): New function.
3705 (resolve_noninline_speculation): New function.
3706 (inline_small_functions): Resolve useless speculations.
3707 * ipa-inline.h (speculation_useful_p): Declare
3708 * ipa.c (can_replace_by_local_alias): Simplify.
3709 (ipa_profile): Produce speculative calls in non-lto, too;
3710 add simple cost model; produce local aliases.
3712 2013-08-13 David Malcolm <dmalcolm@redhat.com>
3714 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
3717 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3719 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3721 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
3723 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3724 iterator instead of X87MODEF.
3726 2013-08-12 Perez Read <netfirewall@gmail.com>
3729 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3730 operand 0 for intel asm alternative.
3731 (*movabs<mode>_2): Ditto for operand 1.
3733 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3735 * config/aarch64/arm_none.h
3736 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3738 2013-08-12 Nick Clifton <nickc@redhat.com>
3740 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3742 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3744 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3745 expand for QI/HImode operand to produce more effictive code for
3746 unsigned char(short) --> float(double) conversion.
3748 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
3750 * doc/invoke.texi: Mention that -ftls-model does not force the final
3753 2013-08-12 Marek Polacek <polacek@redhat.com>
3754 Marc Glisse <marc.glisse@inria.fr>
3756 PR tree-optimization/57980
3757 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3758 when creating -1 constant.
3760 2013-08-10 Jan Hubicka <jh@suse.cz>
3762 Workaround binutils PR14342.
3763 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3764 (gimple_init_edge_profiler): Likewise.
3765 (gimple_gen_ic_func_profiler): Likewise.
3767 2013-08-09 Jan Hubicka <jh@suse.cz>
3769 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3771 2013-08-09 Xinliang David Li <davidxl@google.com>
3773 * config/i386/stringop.def: New file.
3774 * config/i386/stringop.opt: New file.
3775 * config/i386/i386-opts.h: Include stringopt.def.
3776 * config/i386/i386.opt: Include stringopt.opt.
3777 * config/i386/i386.c (ix86_option_override_internal):
3778 Override default size based stringop inline strategies with options.
3779 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3782 2013-08-09 Jan Hubicka <jh@suse.cz>
3784 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
3786 2013-08-09 Jan Hubicka <jh@suse.cz>
3788 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3790 (dump_cgraph_node): Dump profile-id.
3791 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3792 and common_target_probability.
3793 * lto-cgraph.c (lto_output_edge): Stream common targets.
3794 (lto_output_node): Stream profile ids.
3795 (input_node): Stream profile ids.
3796 (input_edge): Stream common targets.
3797 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3798 * ipa.c: Include value-prof.h
3799 (ipa_profile_generate_summary): Turn indirect call statement histograms
3800 into common targets.
3801 (ipa_profile): Turn common targets into speculative edges.
3803 2013-08-09 Jan Hubicka <jh@suse.cz>
3805 * cgraph.h (cgraph_node): Add profile_id.
3806 * value-prof.c (cgraph_node_map): Turn into pointer_map.
3807 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
3808 (del_node_map): Update.
3809 (find_func_by_funcdef_no): Replace by ...
3810 (find_func_by_profile_id): ... this one.
3811 (gimple_ic_transform): Do not remove useful histograms when
3812 speculation is not done; dump info when indirect call removal
3814 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
3815 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
3816 (__gcov_indirect_call_profiler_v2): .. this one.
3817 * profile.h (init_node_map): Update.
3818 * coverage.c (coverage_compute_profile_id): New function.
3819 * coverage.h (coverage_compute_profile_id): Declare.
3820 * tree-profile.c (init_ic_make_global_vars): Make
3821 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
3822 (gimple_init_edge_profiler): Update prototype of
3823 __gcov_indirect_call_profiler.
3824 (gimple_gen_ic_func_profiler): Simplify.
3825 (tree_profiling): Use init_node_map
3827 2013-08-09 Jan Hubicka <jh@suse.cz>
3829 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
3830 non-speculative refs.
3831 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
3832 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
3833 (cgraph_set_call_stmt): Likewise.
3834 (cgraph_create_edge_1): Fix release checking compilatoin;
3836 (cgraph_free_edge): Free indirect info.
3837 (cgraph_turn_edge_to_speculative): New function.
3838 (cgraph_speculative_call_info): New function.
3839 (cgraph_make_edge_direct): Return direct edge; handle speculation.
3840 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
3841 (dump_cgraph_node): Dump speculation.
3842 (verify_edge_count_and_frequency): Accept speculative edges.
3843 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
3844 (verify_cgraph_node): Handle speculation.
3845 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
3846 (cgraph_set_call_stmt): Update prototype.
3847 (cgraph_make_edge_direct): Update prototype.
3848 (cgraph_speculative_call_info): Declare.
3849 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
3850 to change; update call of ipa_find_references.
3851 * ipa-ref.c (ipa_record_reference): Fix return value; clear
3852 lto_stmt_uid and speculative flags.
3853 (ipa_dump_references): Dump speculation.
3854 (ipa_clone_references): Clone speculative flag.
3855 (ipa_clone_referring): Likewise.
3856 (ipa_clone_ref): New function.
3857 (ipa_find_reference): Look into lto_stmt_uids
3858 (ipa_clear_stmts_in_references): Do not clear speculative calls.
3859 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
3860 (ipa_find_reference): Update declaration.
3861 (ipa_clone_ref): Declare.
3862 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
3863 stream speculative flag.
3864 (lto_output_ref): Stream statements uids and speculation.
3865 (input_ref): Likewise.
3866 (input_edge): Stream speuclation.
3867 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
3868 (cgraph_set_call_stmt_including_clones): Handle speculation.
3869 * ipa-inline.c (heap_edge_removal_hook): New function.
3870 (inline_small_functions): Register it.
3871 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
3872 also initialize refs.
3873 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
3875 (try_make_edge_direct_simple_call): Likewise.
3876 (try_make_edge_direct_simple_call): Likewise.
3877 (update_indirect_edges_after_inlining): Likewise.
3878 (remove_described_reference): Look proper lto_stmt_uid.
3879 (propagate_controlled_uses): Likewise.
3880 (propagate_controlled_uses): Liekwise.
3881 * tree-inline.c (copy_bb): Copy speculative edges.
3882 (redirect_all_calls): New function.
3883 (copy_cfg_body): Do redirection after loop info is updated.
3884 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
3886 2013-08-09 Jan Hubicka <jh@suse.cz>
3888 * lto-streamer-out.c (output_function): Renumber PHIs.
3889 * lto-streamer-in.c (input_function): Likewise.
3891 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
3893 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
3894 (get_lane_unsigned): Likewise.
3895 (dup_lane_scalar): Likewise.
3896 (get_lane): enable for VALL.
3897 * config/aarch64/aarch64-simd.md
3898 (aarch64_dup_lane_scalar<mode>): Remove.
3899 (aarch64_get_lane_signed<mode>): Likewise.
3900 (aarch64_get_lane_unsigned<mode>): Likewise.
3901 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
3902 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
3903 (aarch64_get_lane<mode>): Enable for all vector modes.
3904 (aarch64_get_lanedi): Remove misleading constraints.
3905 * config/aarch64/arm_neon.h
3906 (__aarch64_vget_lane_any): Define.
3907 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
3908 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
3909 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
3910 * config/aarch64/iterators.md (VDQQH): New.
3914 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
3916 * configure.ac: Add GAS check for LEON instructions on SPARC.
3917 * configure: Regenerate.
3918 * config.in: Likewise.
3919 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
3921 * config/sparc/sparc.opt (LEON, LEON3): New masks.
3922 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
3924 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
3925 (AS_LEON_FLAG): New macro.
3926 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
3927 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
3928 Deal with LEON and LEON3 for the memory model.
3929 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
3930 (atomic_compare_and_swap<mode>_1): Likewise.
3931 (*atomic_compare_and_swap<mode>_1): Likewise.
3933 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3935 * config/arm/neon.md (vcond): Fix floating-point vector
3936 comparisons against 0.
3938 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
3940 * lra-constraints.c (emit_spill_move): Remove assert.
3941 (process_alt_operands): Add more debugging
3942 output. Increase reject for spilling into memory. Decrease
3943 reject for reloading scratch.
3944 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
3946 2013-08-08 Steve Ellcey <sellcey@mips.com>
3948 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
3949 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
3950 micromips incompatible. Add nan2008.
3951 (MULTILIB_DIRNAMES): Add nan2008.
3952 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3953 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
3954 and micromips incompatible. Add nan2008.
3955 (MULTILIB_DIRNAMES): Add nan2008.
3956 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
3958 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
3960 PR rtl-optimization/58079
3961 * combine.c (combine_simplify_rtx): Avoid using SUBST if
3962 simplify_comparison has widened a comparison with an integer.
3964 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3966 * config/arm/neon.md (movmisalign<mode>): Disable when we
3967 don't allow unaligned accesses.
3968 (*movmisalign<mode>_neon_store): Likewise.
3969 (*movmisalign<mode>_neon_load): Likewise.
3970 (*movmisalign<mode>_neon_store): Likewise.
3971 (*movmisalign<mode>_neon_load): Likewise.
3973 2013-08-08 Jan Hubicka <jh@suse.cz>
3975 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
3976 (make_pass_rebuild_cgraph_edges): Also clear references.
3977 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
3978 * ipa-inline-transform.c (inline_transform): Remove all references
3980 * cgraphunit.c (expand_function): Remove all references after
3982 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
3983 (ipa_find_reference): Rewrite to iterator.
3984 (remove_stmt_references): Likewise.
3985 (ipa_clear_stmts_in_references): New function.
3986 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
3987 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
3989 * ipa-split.c (split_function): Remove references in split function.
3991 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
3994 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
3995 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
3997 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4000 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
4001 suggested mode for the assignment isn't compatible with the
4004 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
4007 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
4009 2013-08-07 Xinliang David Li <davidxl@google.com>
4011 * config/i386/i386.opt: New option -mtune-ctrl=.
4012 * config/i386/x86-tune.def: New file.
4013 * config/i386/i386.h: include x86-tune.def.
4014 * config/i386/i386.c (ix86_option_override_internal):
4015 Parsing -mtune-ctrl= option and set tune features.
4017 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
4020 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
4021 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4024 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4026 * rtl.h (update_alignments): Declare.
4027 * final.c (grow_label_align): New function extracted from...
4028 (shorten_branches): ...here. Call it.
4029 (update_alignments): New function.
4030 * reorg.c (sibling_labels): New variable.
4031 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4032 the new label along with it onto the sibling_labels vector.
4033 (fill_simple_delay_slots): Adjust call to get_label_before.
4034 (fill_slots_from_thread): Likewise.
4035 (relax_delay_slots): Likewise.
4036 (make_return_insns): Likewise.
4037 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4039 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4041 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4042 document its semantics.
4043 (diagnostic_report_diagnostic): Adjust accordingly.
4045 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4047 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
4048 (sparc_option_override): ...and port to new C++ pass API.
4049 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4051 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4053 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4055 2013-08-06 Caroline Tice <cmtice@google.com>
4057 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4058 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4059 * tree-pass.h: Add pass_vtable_verify.
4060 * varasm.c (assemble_variable): Add code to properly set the comdat
4061 section and name for the .vtable_map_vars section.
4062 (assemble_vtyv_preinit_initializer): New function.
4063 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4065 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4066 * vtable-verify.c: New file.
4067 * vtable-verify.h: New file.
4068 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4069 initialiation levels.
4070 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4071 * passes.def: Insert pass_vtable_verify.
4072 * aclocal.m4: Reorder includes.
4073 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4074 -fvtv-counts options.
4075 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4076 as appropriate, if -fvtable-verify=... is used.
4077 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4078 -fvtable-verify=... is used.
4079 * Makefile.in (OBJS): Add vtable-verify.o to list.
4080 (vtable-verify.o): Add new build rule.
4081 (GTFILES): Add vtable-verify.c to list.
4082 * common.opt (fvtable-verify=): New flag.
4083 (vtv_priority): Values for fvtable-verify= flag.
4084 (fvtv-counts): New flag.
4085 (fvtv-debug): New flag.
4086 * tree.h (save_vtable_map_decl): New extern function decl.
4088 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4090 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4091 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4092 (pass_data_rl78_devirt): ...new pass_data instance and...
4093 (make_pass_rl78_devirt): ...new function.
4094 (rl78_asm_file_start): Port pass registration to new C++ API.
4096 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4098 * coretypes.h (rtl_opt_pass): Add.
4099 (gcc::context): Add.
4100 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4101 (epiphany_init): Port to new C++ pass API.
4102 (epiphany_optimize_mode_switching): Likewise.
4103 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4104 (pass_manager::get_pass_mode_switching): New.
4105 (pass_manager::get_pass_peephole2): New.
4106 * mode-switching.c (pass_mode_switching): Add clone method.
4107 * recog.c (pass_peephole2): Add clone method.
4108 (pass_split_all_insns): Add clone method.
4110 2013-08-06 David Malcolm <dmalcolm@redhat.com>
4112 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4113 (mips_option_override): ...here, porting to new C++ API for passes.
4115 2013-08-06 Jan Hubicka <jh@suse.cz>
4117 * cgraph.c (cgraph_get_body): New function based on lto.c
4119 * cgraph.h (cgraph_get_body): Declare.
4120 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4122 * cgraphunit.c (expand_function): Get body prior expanding.
4123 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
4124 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4126 * passes.c (do_per_function_toporder): Get body.
4127 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4128 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4130 2013-08-06 Martin Jambor <mjambor@suse.cz>
4133 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4134 is not re-finalized. Rename second parameter to no_collect.
4136 2013-08-06 Martin Jambor <mjambor@suse.cz>
4139 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4140 MEM_REF has proper alignment information.
4142 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
4145 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4147 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4148 rtx (*) (rtx, ...) with insn_gen_fn.
4149 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4150 insn_gen_fn::stored_funcptr. Add initializer braces.
4152 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4154 Rewrite how instances of passes are cloned to remove assumptions
4155 about their sizes (thus allowing pass subclasses to have
4156 additional data fields, albeit non-GC-managed ones at this point).
4158 * passes.c (make_pass_instance): Now that passes have clone
4159 methods, rewrite this function to eliminate XNEW and memcpy
4160 calls that used hardcoded sizes. Since this function no longer
4161 creates pass instances, rename it to...
4162 (add_pass_instance): ...this. Document the old way that passes were
4163 numbered and flagged, and rework this function to continue using it.
4164 (next_pass_1): Add an initial_pass argument for use by
4166 (position_pass): When adding multiple instances of a pass, use
4167 the pass's clone method, rather than relying on the XNEW/memcpy
4168 within the former make_pass_instance (now add_pass_instance).
4169 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4170 the initial instance of the current pass within the pass manager.
4172 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4174 This is the automated part of the conversion of passes from C
4175 structs to C++ classes.
4177 Patch autogenerated by refactor_passes.py from
4178 https://github.com/davidmalcolm/gcc-refactoring-scripts
4179 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4181 * asan.c (pass_asan): Convert from a global struct to a subclass of
4182 gimple_opt_pass along with...
4183 (pass_data_asan): ...new pass_data instance and...
4184 (make_pass_asan): ...new function.
4185 (pass_asan_O0): Convert from a global struct to a subclass of
4186 gimple_opt_pass along with...
4187 (pass_data_asan_O0): ...new pass_data instance and...
4188 (make_pass_asan_O0): ...new function.
4189 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4190 subclass of rtl_opt_pass along with...
4191 (pass_data_inc_dec): ...new pass_data instance and...
4192 (make_pass_inc_dec): ...new function.
4193 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4194 a subclass of rtl_opt_pass along with...
4195 (pass_data_reorder_blocks): ...new pass_data instance and...
4196 (make_pass_reorder_blocks): ...new function.
4197 (pass_duplicate_computed_gotos): Convert from a global struct to a
4198 subclass of rtl_opt_pass along with...
4199 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4200 (make_pass_duplicate_computed_gotos): ...new function.
4201 (pass_partition_blocks): Convert from a global struct to a subclass of
4202 rtl_opt_pass along with...
4203 (pass_data_partition_blocks): ...new pass_data instance and...
4204 (make_pass_partition_blocks): ...new function.
4205 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
4206 struct to a subclass of rtl_opt_pass along with...
4207 (pass_data_branch_target_load_optimize1): ...new pass_data instance
4209 (make_pass_branch_target_load_optimize1): ...new function.
4210 (pass_branch_target_load_optimize2): Convert from a global struct to a
4211 subclass of rtl_opt_pass along with...
4212 (pass_data_branch_target_load_optimize2): ...new pass_data instance
4214 (make_pass_branch_target_load_optimize2): ...new function.
4215 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4216 of rtl_opt_pass along with...
4217 (pass_data_jump): ...new pass_data instance and...
4218 (make_pass_jump): ...new function.
4219 (pass_jump2): Convert from a global struct to a subclass of
4220 rtl_opt_pass along with...
4221 (pass_data_jump2): ...new pass_data instance and...
4222 (make_pass_jump2): ...new function.
4223 * cfgexpand.c (pass_expand): Convert from a global struct to a
4224 subclass of rtl_opt_pass along with...
4225 (pass_data_expand): ...new pass_data instance and...
4226 (make_pass_expand): ...new function.
4227 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4228 of rtl_opt_pass along with...
4229 (pass_data_free_cfg): ...new pass_data instance and...
4230 (make_pass_free_cfg): ...new function.
4231 (pass_into_cfg_layout_mode): Convert from a global struct to a
4232 subclass of rtl_opt_pass along with...
4233 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4234 (make_pass_into_cfg_layout_mode): ...new function.
4235 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4236 subclass of rtl_opt_pass along with...
4237 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4238 (make_pass_outof_cfg_layout_mode): ...new function.
4239 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4240 struct to a subclass of gimple_opt_pass along with...
4241 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4242 (make_pass_build_cgraph_edges): ...new function.
4243 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4244 subclass of gimple_opt_pass along with...
4245 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4246 (make_pass_rebuild_cgraph_edges): ...new function.
4247 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4248 subclass of gimple_opt_pass along with...
4249 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4251 (make_pass_remove_cgraph_callee_edges): ...new function.
4252 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4253 struct to a subclass of rtl_opt_pass along with...
4254 (pass_data_stack_adjustments): ...new pass_data instance and...
4255 (make_pass_stack_adjustments): ...new function.
4256 * combine.c (pass_combine): Convert from a global struct to a subclass
4257 of rtl_opt_pass along with...
4258 (pass_data_combine): ...new pass_data instance and...
4259 (make_pass_combine): ...new function.
4260 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4261 global struct to a subclass of rtl_opt_pass along with...
4262 (pass_data_compare_elim_after_reload): ...new pass_data instance
4264 (make_pass_compare_elim_after_reload): ...new function.
4265 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4266 of rtl_opt_pass along with...
4267 (pass_data_rtl_cprop): ...new pass_data instance and...
4268 (make_pass_rtl_cprop): ...new function.
4269 * cse.c (pass_cse): Convert from a global struct to a subclass of
4270 rtl_opt_pass along with...
4271 (pass_data_cse): ...new pass_data instance and...
4272 (make_pass_cse): ...new function.
4273 (pass_cse2): Convert from a global struct to a subclass of
4274 rtl_opt_pass along with...
4275 (pass_data_cse2): ...new pass_data instance and...
4276 (make_pass_cse2): ...new function.
4277 (pass_cse_after_global_opts): Convert from a global struct to a
4278 subclass of rtl_opt_pass along with...
4279 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4280 (make_pass_cse_after_global_opts): ...new function.
4281 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4282 of rtl_opt_pass along with...
4283 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4284 (make_pass_ud_rtl_dce): ...new function.
4285 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4286 rtl_opt_pass along with...
4287 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4288 (make_pass_fast_rtl_dce): ...new function.
4289 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4290 a subclass of rtl_opt_pass along with...
4291 (pass_data_df_initialize_opt): ...new pass_data instance and...
4292 (make_pass_df_initialize_opt): ...new function.
4293 (pass_df_initialize_no_opt): Convert from a global struct to a
4294 subclass of rtl_opt_pass along with...
4295 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4296 (make_pass_df_initialize_no_opt): ...new function.
4297 (pass_df_finish): Convert from a global struct to a subclass of
4298 rtl_opt_pass along with...
4299 (pass_data_df_finish): ...new pass_data instance and...
4300 (make_pass_df_finish): ...new function.
4301 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4302 rtl_opt_pass along with...
4303 (pass_data_rtl_dse1): ...new pass_data instance and...
4304 (make_pass_rtl_dse1): ...new function.
4305 (pass_rtl_dse2): Convert from a global struct to a subclass of
4306 rtl_opt_pass along with...
4307 (pass_data_rtl_dse2): ...new pass_data instance and...
4308 (make_pass_rtl_dse2): ...new function.
4309 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4310 subclass of rtl_opt_pass along with...
4311 (pass_data_dwarf2_frame): ...new pass_data instance and...
4312 (make_pass_dwarf2_frame): ...new function.
4313 * except.c (pass_set_nothrow_function_flags): Convert from a global
4314 struct to a subclass of rtl_opt_pass along with...
4315 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4317 (make_pass_set_nothrow_function_flags): ...new function.
4318 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4319 subclass of rtl_opt_pass along with...
4320 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4322 (make_pass_convert_to_eh_region_ranges): ...new function.
4323 * final.c (pass_compute_alignments): Convert from a global struct to a
4324 subclass of rtl_opt_pass along with...
4325 (pass_data_compute_alignments): ...new pass_data instance and...
4326 (make_pass_compute_alignments): ...new function.
4327 (pass_final): Convert from a global struct to a subclass of
4328 rtl_opt_pass along with...
4329 (pass_data_final): ...new pass_data instance and...
4330 (make_pass_final): ...new function.
4331 (pass_shorten_branches): Convert from a global struct to a subclass of
4332 rtl_opt_pass along with...
4333 (pass_data_shorten_branches): ...new pass_data instance and...
4334 (make_pass_shorten_branches): ...new function.
4335 (pass_clean_state): Convert from a global struct to a subclass of
4336 rtl_opt_pass along with...
4337 (pass_data_clean_state): ...new pass_data instance and...
4338 (make_pass_clean_state): ...new function.
4339 * function.c (pass_instantiate_virtual_regs): Convert from a global
4340 struct to a subclass of rtl_opt_pass along with...
4341 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4342 (make_pass_instantiate_virtual_regs): ...new function.
4343 (pass_leaf_regs): Convert from a global struct to a subclass of
4344 rtl_opt_pass along with...
4345 (pass_data_leaf_regs): ...new pass_data instance and...
4346 (make_pass_leaf_regs): ...new function.
4347 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4348 subclass of rtl_opt_pass along with...
4349 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4351 (make_pass_thread_prologue_and_epilogue): ...new function.
4352 (pass_match_asm_constraints): Convert from a global struct to a
4353 subclass of rtl_opt_pass along with...
4354 (pass_data_match_asm_constraints): ...new pass_data instance and...
4355 (make_pass_match_asm_constraints): ...new function.
4356 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4357 subclass of rtl_opt_pass along with...
4358 (pass_data_rtl_fwprop): ...new pass_data instance and...
4359 (make_pass_rtl_fwprop): ...new function.
4360 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4361 rtl_opt_pass along with...
4362 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4363 (make_pass_rtl_fwprop_addr): ...new function.
4364 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4365 rtl_opt_pass along with...
4366 (pass_data_rtl_pre): ...new pass_data instance and...
4367 (make_pass_rtl_pre): ...new function.
4368 (pass_rtl_hoist): Convert from a global struct to a subclass of
4369 rtl_opt_pass along with...
4370 (pass_data_rtl_hoist): ...new pass_data instance and...
4371 (make_pass_rtl_hoist): ...new function.
4372 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4373 subclass of gimple_opt_pass along with...
4374 (pass_data_lower_cf): ...new pass_data instance and...
4375 (make_pass_lower_cf): ...new function.
4376 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4377 from a global struct to a subclass of gimple_opt_pass along with...
4378 (pass_data_strength_reduction): ...new pass_data instance and...
4379 (make_pass_strength_reduction): ...new function.
4380 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4381 of rtl_opt_pass along with...
4382 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4383 (make_pass_rtl_ifcvt): ...new function.
4384 (pass_if_after_combine): Convert from a global struct to a subclass of
4385 rtl_opt_pass along with...
4386 (pass_data_if_after_combine): ...new pass_data instance and...
4387 (make_pass_if_after_combine): ...new function.
4388 (pass_if_after_reload): Convert from a global struct to a subclass of
4389 rtl_opt_pass along with...
4390 (pass_data_if_after_reload): ...new pass_data instance and...
4391 (make_pass_if_after_reload): ...new function.
4392 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4393 a subclass of rtl_opt_pass along with...
4394 (pass_data_initialize_regs): ...new pass_data instance and...
4395 (make_pass_initialize_regs): ...new function.
4396 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4397 of ipa_opt_pass_d along with...
4398 (pass_data_ipa_cp): ...new pass_data instance and...
4399 (make_pass_ipa_cp): ...new function.
4400 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4401 global struct to a subclass of gimple_opt_pass along with...
4402 (pass_data_inline_parameters): ...new pass_data instance and...
4403 (make_pass_inline_parameters): ...new function.
4404 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4405 subclass of gimple_opt_pass along with...
4406 (pass_data_early_inline): ...new pass_data instance and...
4407 (make_pass_early_inline): ...new function.
4408 (pass_ipa_inline): Convert from a global struct to a subclass of
4409 ipa_opt_pass_d along with...
4410 (pass_data_ipa_inline): ...new pass_data instance and...
4411 (make_pass_ipa_inline): ...new function.
4412 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4413 struct to a subclass of gimple_opt_pass along with...
4414 (pass_data_local_pure_const): ...new pass_data instance and...
4415 (make_pass_local_pure_const): ...new function.
4416 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4417 ipa_opt_pass_d along with...
4418 (pass_data_ipa_pure_const): ...new pass_data instance and...
4419 (make_pass_ipa_pure_const): ...new function.
4420 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4421 to a subclass of ipa_opt_pass_d along with...
4422 (pass_data_ipa_reference): ...new pass_data instance and...
4423 (make_pass_ipa_reference): ...new function.
4424 * ipa-split.c (pass_split_functions): Convert from a global struct to
4425 a subclass of gimple_opt_pass along with...
4426 (pass_data_split_functions): ...new pass_data instance and...
4427 (make_pass_split_functions): ...new function.
4428 (pass_feedback_split_functions): Convert from a global struct to a
4429 subclass of gimple_opt_pass along with...
4430 (pass_data_feedback_split_functions): ...new pass_data instance and...
4431 (make_pass_feedback_split_functions): ...new function.
4432 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4433 global struct to a subclass of simple_ipa_opt_pass along with...
4434 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4436 (make_pass_ipa_function_and_variable_visibility): ...new function.
4437 (pass_ipa_free_inline_summary): Convert from a global struct to a
4438 subclass of simple_ipa_opt_pass along with...
4439 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4440 (make_pass_ipa_free_inline_summary): ...new function.
4441 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4442 subclass of ipa_opt_pass_d along with...
4443 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4445 (make_pass_ipa_whole_program_visibility): ...new function.
4446 (pass_ipa_profile): Convert from a global struct to a subclass of
4447 ipa_opt_pass_d along with...
4448 (pass_data_ipa_profile): ...new pass_data instance and...
4449 (make_pass_ipa_profile): ...new function.
4450 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4451 ipa_opt_pass_d along with...
4452 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4453 (make_pass_ipa_cdtor_merge): ...new function.
4454 * ira.c (pass_ira): Convert from a global struct to a subclass of
4455 rtl_opt_pass along with...
4456 (pass_data_ira): ...new pass_data instance and...
4457 (make_pass_ira): ...new function.
4458 (pass_reload): Convert from a global struct to a subclass of
4459 rtl_opt_pass along with...
4460 (pass_data_reload): ...new pass_data instance and...
4461 (make_pass_reload): ...new function.
4462 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4463 subclass of rtl_opt_pass along with...
4464 (pass_data_cleanup_barriers): ...new pass_data instance and...
4465 (make_pass_cleanup_barriers): ...new function.
4466 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4467 of rtl_opt_pass along with...
4468 (pass_data_loop2): ...new pass_data instance and...
4469 (make_pass_loop2): ...new function.
4470 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4471 rtl_opt_pass along with...
4472 (pass_data_rtl_loop_init): ...new pass_data instance and...
4473 (make_pass_rtl_loop_init): ...new function.
4474 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4475 rtl_opt_pass along with...
4476 (pass_data_rtl_loop_done): ...new pass_data instance and...
4477 (make_pass_rtl_loop_done): ...new function.
4478 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4479 subclass of rtl_opt_pass along with...
4480 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4481 (make_pass_rtl_move_loop_invariants): ...new function.
4482 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4483 rtl_opt_pass along with...
4484 (pass_data_rtl_unswitch): ...new pass_data instance and...
4485 (make_pass_rtl_unswitch): ...new function.
4486 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4487 subclass of rtl_opt_pass along with...
4488 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4490 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4491 (pass_rtl_doloop): Convert from a global struct to a subclass of
4492 rtl_opt_pass along with...
4493 (pass_data_rtl_doloop): ...new pass_data instance and...
4494 (make_pass_rtl_doloop): ...new function.
4495 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4496 a subclass of rtl_opt_pass along with...
4497 (pass_data_lower_subreg): ...new pass_data instance and...
4498 (make_pass_lower_subreg): ...new function.
4499 (pass_lower_subreg2): Convert from a global struct to a subclass of
4500 rtl_opt_pass along with...
4501 (pass_data_lower_subreg2): ...new pass_data instance and...
4502 (make_pass_lower_subreg2): ...new function.
4503 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4504 struct to a subclass of ipa_opt_pass_d along with...
4505 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4506 (make_pass_ipa_lto_gimple_out): ...new function.
4507 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4508 of ipa_opt_pass_d along with...
4509 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4510 (make_pass_ipa_lto_finish_out): ...new function.
4511 * mode-switching.c (pass_mode_switching): Convert from a global struct
4512 to a subclass of rtl_opt_pass along with...
4513 (pass_data_mode_switching): ...new pass_data instance and...
4514 (make_pass_mode_switching): ...new function.
4515 * modulo-sched.c (pass_sms): Convert from a global struct to a
4516 subclass of rtl_opt_pass along with...
4517 (pass_data_sms): ...new pass_data instance and...
4518 (make_pass_sms): ...new function.
4519 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4520 subclass of gimple_opt_pass along with...
4521 (pass_data_expand_omp): ...new pass_data instance and...
4522 (make_pass_expand_omp): ...new function.
4523 (pass_lower_omp): Convert from a global struct to a subclass of
4524 gimple_opt_pass along with...
4525 (pass_data_lower_omp): ...new pass_data instance and...
4526 (make_pass_lower_omp): ...new function.
4527 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4528 of gimple_opt_pass along with...
4529 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4530 (make_pass_diagnose_omp_blocks): ...new function.
4531 * passes.c (pass_early_local_passes): Convert from a global struct to
4532 a subclass of simple_ipa_opt_pass along with...
4533 (pass_data_early_local_passes): ...new pass_data instance and...
4534 (make_pass_early_local_passes): ...new function.
4535 (pass_all_early_optimizations): Convert from a global struct to a
4536 subclass of gimple_opt_pass along with...
4537 (pass_data_all_early_optimizations): ...new pass_data instance and...
4538 (make_pass_all_early_optimizations): ...new function.
4539 (pass_all_optimizations): Convert from a global struct to a subclass
4540 of gimple_opt_pass along with...
4541 (pass_data_all_optimizations): ...new pass_data instance and...
4542 (make_pass_all_optimizations): ...new function.
4543 (pass_all_optimizations_g): Convert from a global struct to a subclass
4544 of gimple_opt_pass along with...
4545 (pass_data_all_optimizations_g): ...new pass_data instance and...
4546 (make_pass_all_optimizations_g): ...new function.
4547 (pass_rest_of_compilation): Convert from a global struct to a subclass
4548 of rtl_opt_pass along with...
4549 (pass_data_rest_of_compilation): ...new pass_data instance and...
4550 (make_pass_rest_of_compilation): ...new function.
4551 (pass_postreload): Convert from a global struct to a subclass of
4552 rtl_opt_pass along with...
4553 (pass_data_postreload): ...new pass_data instance and...
4554 (make_pass_postreload): ...new function.
4555 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4556 subclass of rtl_opt_pass along with...
4557 (pass_data_gcse2): ...new pass_data instance and...
4558 (make_pass_gcse2): ...new function.
4559 * postreload.c (pass_postreload_cse): Convert from a global struct to
4560 a subclass of rtl_opt_pass along with...
4561 (pass_data_postreload_cse): ...new pass_data instance and...
4562 (make_pass_postreload_cse): ...new function.
4563 * predict.c (pass_profile): Convert from a global struct to a subclass
4564 of gimple_opt_pass along with...
4565 (pass_data_profile): ...new pass_data instance and...
4566 (make_pass_profile): ...new function.
4567 (pass_strip_predict_hints): Convert from a global struct to a subclass
4568 of gimple_opt_pass along with...
4569 (pass_data_strip_predict_hints): ...new pass_data instance and...
4570 (make_pass_strip_predict_hints): ...new function.
4571 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4572 of rtl_opt_pass along with...
4573 (pass_data_peephole2): ...new pass_data instance and...
4574 (make_pass_peephole2): ...new function.
4575 (pass_split_all_insns): Convert from a global struct to a subclass of
4576 rtl_opt_pass along with...
4577 (pass_data_split_all_insns): ...new pass_data instance and...
4578 (make_pass_split_all_insns): ...new function.
4579 (pass_split_after_reload): Convert from a global struct to a subclass
4580 of rtl_opt_pass along with...
4581 (pass_data_split_after_reload): ...new pass_data instance and...
4582 (make_pass_split_after_reload): ...new function.
4583 (pass_split_before_regstack): Convert from a global struct to a
4584 subclass of rtl_opt_pass along with...
4585 (pass_data_split_before_regstack): ...new pass_data instance and...
4586 (make_pass_split_before_regstack): ...new function.
4587 (pass_split_before_sched2): Convert from a global struct to a subclass
4588 of rtl_opt_pass along with...
4589 (pass_data_split_before_sched2): ...new pass_data instance and...
4590 (make_pass_split_before_sched2): ...new function.
4591 (pass_split_for_shorten_branches): Convert from a global struct to a
4592 subclass of rtl_opt_pass along with...
4593 (pass_data_split_for_shorten_branches): ...new pass_data instance
4595 (make_pass_split_for_shorten_branches): ...new function.
4596 * ree.c (pass_ree): Convert from a global struct to a subclass of
4597 rtl_opt_pass along with...
4598 (pass_data_ree): ...new pass_data instance and...
4599 (make_pass_ree): ...new function.
4600 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4601 subclass of rtl_opt_pass along with...
4602 (pass_data_stack_regs): ...new pass_data instance and...
4603 (make_pass_stack_regs): ...new function.
4604 (pass_stack_regs_run): Convert from a global struct to a subclass of
4605 rtl_opt_pass along with...
4606 (pass_data_stack_regs_run): ...new pass_data instance and...
4607 (make_pass_stack_regs_run): ...new function.
4608 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4609 subclass of rtl_opt_pass along with...
4610 (pass_data_cprop_hardreg): ...new pass_data instance and...
4611 (make_pass_cprop_hardreg): ...new function.
4612 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4613 subclass of rtl_opt_pass along with...
4614 (pass_data_reginfo_init): ...new pass_data instance and...
4615 (make_pass_reginfo_init): ...new function.
4616 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4617 of rtl_opt_pass along with...
4618 (pass_data_regmove): ...new pass_data instance and...
4619 (make_pass_regmove): ...new function.
4620 * regrename.c (pass_regrename): Convert from a global struct to a
4621 subclass of rtl_opt_pass along with...
4622 (pass_data_regrename): ...new pass_data instance and...
4623 (make_pass_regrename): ...new function.
4624 * reorg.c (pass_delay_slots): Convert from a global struct to a
4625 subclass of rtl_opt_pass along with...
4626 (pass_data_delay_slots): ...new pass_data instance and...
4627 (make_pass_delay_slots): ...new function.
4628 (pass_machine_reorg): Convert from a global struct to a subclass of
4629 rtl_opt_pass along with...
4630 (pass_data_machine_reorg): ...new pass_data instance and...
4631 (make_pass_machine_reorg): ...new function.
4632 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4633 of rtl_opt_pass along with...
4634 (pass_data_sched): ...new pass_data instance and...
4635 (make_pass_sched): ...new function.
4636 (pass_sched2): Convert from a global struct to a subclass of
4637 rtl_opt_pass along with...
4638 (pass_data_sched2): ...new pass_data instance and...
4639 (make_pass_sched2): ...new function.
4640 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4641 to a subclass of rtl_opt_pass along with...
4642 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4643 (make_pass_stack_ptr_mod): ...new function.
4644 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4645 to a subclass of rtl_opt_pass along with...
4646 (pass_data_rtl_store_motion): ...new pass_data instance and...
4647 (make_pass_rtl_store_motion): ...new function.
4648 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4649 of gimple_opt_pass along with...
4650 (pass_data_tracer): ...new pass_data instance and...
4651 (make_pass_tracer): ...new function.
4652 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4653 to a subclass of gimple_opt_pass along with...
4654 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4655 (make_pass_diagnose_tm_blocks): ...new function.
4656 (pass_lower_tm): Convert from a global struct to a subclass of
4657 gimple_opt_pass along with...
4658 (pass_data_lower_tm): ...new pass_data instance and...
4659 (make_pass_lower_tm): ...new function.
4660 (pass_tm_init): Convert from a global struct to a subclass of
4661 gimple_opt_pass along with...
4662 (pass_data_tm_init): ...new pass_data instance and...
4663 (make_pass_tm_init): ...new function.
4664 (pass_tm_mark): Convert from a global struct to a subclass of
4665 gimple_opt_pass along with...
4666 (pass_data_tm_mark): ...new pass_data instance and...
4667 (make_pass_tm_mark): ...new function.
4668 (pass_tm_edges): Convert from a global struct to a subclass of
4669 gimple_opt_pass along with...
4670 (pass_data_tm_edges): ...new pass_data instance and...
4671 (make_pass_tm_edges): ...new function.
4672 (pass_tm_memopt): Convert from a global struct to a subclass of
4673 gimple_opt_pass along with...
4674 (pass_data_tm_memopt): ...new pass_data instance and...
4675 (make_pass_tm_memopt): ...new function.
4676 (pass_ipa_tm): Convert from a global struct to a subclass of
4677 simple_ipa_opt_pass along with...
4678 (pass_data_ipa_tm): ...new pass_data instance and...
4679 (make_pass_ipa_tm): ...new function.
4680 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4681 subclass of gimple_opt_pass along with...
4682 (pass_data_call_cdce): ...new pass_data instance and...
4683 (make_pass_call_cdce): ...new function.
4684 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4685 subclass of gimple_opt_pass along with...
4686 (pass_data_build_cfg): ...new pass_data instance and...
4687 (make_pass_build_cfg): ...new function.
4688 (pass_split_crit_edges): Convert from a global struct to a subclass of
4689 gimple_opt_pass along with...
4690 (pass_data_split_crit_edges): ...new pass_data instance and...
4691 (make_pass_split_crit_edges): ...new function.
4692 (pass_warn_function_return): Convert from a global struct to a
4693 subclass of gimple_opt_pass along with...
4694 (pass_data_warn_function_return): ...new pass_data instance and...
4695 (make_pass_warn_function_return): ...new function.
4696 (pass_warn_function_noreturn): Convert from a global struct to a
4697 subclass of gimple_opt_pass along with...
4698 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4699 (make_pass_warn_function_noreturn): ...new function.
4700 (pass_warn_unused_result): Convert from a global struct to a subclass
4701 of gimple_opt_pass along with...
4702 (pass_data_warn_unused_result): ...new pass_data instance and...
4703 (make_pass_warn_unused_result): ...new function.
4704 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4705 a subclass of gimple_opt_pass along with...
4706 (pass_data_merge_phi): ...new pass_data instance and...
4707 (make_pass_merge_phi): ...new function.
4708 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4709 a subclass of gimple_opt_pass along with...
4710 (pass_data_lower_complex): ...new pass_data instance and...
4711 (make_pass_lower_complex): ...new function.
4712 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4713 gimple_opt_pass along with...
4714 (pass_data_lower_complex_O0): ...new pass_data instance and...
4715 (make_pass_lower_complex_O0): ...new function.
4716 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4717 subclass of gimple_opt_pass along with...
4718 (pass_data_lower_eh): ...new pass_data instance and...
4719 (make_pass_lower_eh): ...new function.
4720 (pass_refactor_eh): Convert from a global struct to a subclass of
4721 gimple_opt_pass along with...
4722 (pass_data_refactor_eh): ...new pass_data instance and...
4723 (make_pass_refactor_eh): ...new function.
4724 (pass_lower_resx): Convert from a global struct to a subclass of
4725 gimple_opt_pass along with...
4726 (pass_data_lower_resx): ...new pass_data instance and...
4727 (make_pass_lower_resx): ...new function.
4728 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4729 of gimple_opt_pass along with...
4730 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4731 (make_pass_lower_eh_dispatch): ...new function.
4732 (pass_cleanup_eh): Convert from a global struct to a subclass of
4733 gimple_opt_pass along with...
4734 (pass_data_cleanup_eh): ...new pass_data instance and...
4735 (make_pass_cleanup_eh): ...new function.
4736 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4737 to a subclass of simple_ipa_opt_pass along with...
4738 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4739 (make_pass_ipa_lower_emutls): ...new function.
4740 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4741 a subclass of gimple_opt_pass along with...
4742 (pass_data_if_conversion): ...new pass_data instance and...
4743 (make_pass_if_conversion): ...new function.
4744 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4745 subclass of gimple_opt_pass along with...
4746 (pass_data_build_ssa): ...new pass_data instance and...
4747 (make_pass_build_ssa): ...new function.
4748 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4749 global struct to a subclass of gimple_opt_pass along with...
4750 (pass_data_loop_distribution): ...new pass_data instance and...
4751 (make_pass_loop_distribution): ...new function.
4752 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4753 subclass of gimple_opt_pass along with...
4754 (pass_data_mudflap_1): ...new pass_data instance and...
4755 (make_pass_mudflap_1): ...new function.
4756 (pass_mudflap_2): Convert from a global struct to a subclass of
4757 gimple_opt_pass along with...
4758 (pass_data_mudflap_2): ...new pass_data instance and...
4759 (make_pass_mudflap_2): ...new function.
4760 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4761 subclass of gimple_opt_pass along with...
4762 (pass_data_mudflap_1): ...new pass_data instance and...
4763 (make_pass_mudflap_1): ...new function.
4764 (pass_mudflap_2): Convert from a global struct to a subclass of
4765 gimple_opt_pass along with...
4766 (pass_data_mudflap_2): ...new pass_data instance and...
4767 (make_pass_mudflap_2): ...new function.
4768 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4769 gimple_opt_pass along with...
4770 (pass_data_nrv): ...new pass_data instance and...
4771 (make_pass_nrv): ...new function.
4772 (pass_return_slot): Convert from a global struct to a subclass of
4773 gimple_opt_pass along with...
4774 (pass_data_return_slot): ...new pass_data instance and...
4775 (make_pass_return_slot): ...new function.
4776 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4777 to a subclass of gimple_opt_pass along with...
4778 (pass_data_object_sizes): ...new pass_data instance and...
4779 (make_pass_object_sizes): ...new function.
4780 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4781 global struct to a subclass of gimple_opt_pass along with...
4782 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4784 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4785 (pass_fixup_cfg): Convert from a global struct to a subclass of
4786 gimple_opt_pass along with...
4787 (pass_data_fixup_cfg): ...new pass_data instance and...
4788 (make_pass_fixup_cfg): ...new function.
4789 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4790 (make_pass_mudflap_1): ...new function.
4791 (pass_mudflap_2): Replace declaration with that of...
4792 (make_pass_mudflap_2): ...new function.
4793 (pass_asan): Replace declaration with that of...
4794 (make_pass_asan): ...new function.
4795 (pass_asan_O0): Replace declaration with that of...
4796 (make_pass_asan_O0): ...new function.
4797 (pass_tsan): Replace declaration with that of...
4798 (make_pass_tsan): ...new function.
4799 (pass_tsan_O0): Replace declaration with that of...
4800 (make_pass_tsan_O0): ...new function.
4801 (pass_lower_cf): Replace declaration with that of...
4802 (make_pass_lower_cf): ...new function.
4803 (pass_refactor_eh): Replace declaration with that of...
4804 (make_pass_refactor_eh): ...new function.
4805 (pass_lower_eh): Replace declaration with that of...
4806 (make_pass_lower_eh): ...new function.
4807 (pass_lower_eh_dispatch): Replace declaration with that of...
4808 (make_pass_lower_eh_dispatch): ...new function.
4809 (pass_lower_resx): Replace declaration with that of...
4810 (make_pass_lower_resx): ...new function.
4811 (pass_build_cfg): Replace declaration with that of...
4812 (make_pass_build_cfg): ...new function.
4813 (pass_early_tree_profile): Replace declaration with that of...
4814 (make_pass_early_tree_profile): ...new function.
4815 (pass_cleanup_eh): Replace declaration with that of...
4816 (make_pass_cleanup_eh): ...new function.
4817 (pass_sra): Replace declaration with that of...
4818 (make_pass_sra): ...new function.
4819 (pass_sra_early): Replace declaration with that of...
4820 (make_pass_sra_early): ...new function.
4821 (pass_early_ipa_sra): Replace declaration with that of...
4822 (make_pass_early_ipa_sra): ...new function.
4823 (pass_tail_recursion): Replace declaration with that of...
4824 (make_pass_tail_recursion): ...new function.
4825 (pass_tail_calls): Replace declaration with that of...
4826 (make_pass_tail_calls): ...new function.
4827 (pass_tree_loop): Replace declaration with that of...
4828 (make_pass_tree_loop): ...new function.
4829 (pass_tree_loop_init): Replace declaration with that of...
4830 (make_pass_tree_loop_init): ...new function.
4831 (pass_lim): Replace declaration with that of...
4832 (make_pass_lim): ...new function.
4833 (pass_tree_unswitch): Replace declaration with that of...
4834 (make_pass_tree_unswitch): ...new function.
4835 (pass_predcom): Replace declaration with that of...
4836 (make_pass_predcom): ...new function.
4837 (pass_iv_canon): Replace declaration with that of...
4838 (make_pass_iv_canon): ...new function.
4839 (pass_scev_cprop): Replace declaration with that of...
4840 (make_pass_scev_cprop): ...new function.
4841 (pass_empty_loop): Replace declaration with that of...
4842 (make_pass_empty_loop): ...new function.
4843 (pass_record_bounds): Replace declaration with that of...
4844 (make_pass_record_bounds): ...new function.
4845 (pass_graphite): Replace declaration with that of...
4846 (make_pass_graphite): ...new function.
4847 (pass_graphite_transforms): Replace declaration with that of...
4848 (make_pass_graphite_transforms): ...new function.
4849 (pass_if_conversion): Replace declaration with that of...
4850 (make_pass_if_conversion): ...new function.
4851 (pass_loop_distribution): Replace declaration with that of...
4852 (make_pass_loop_distribution): ...new function.
4853 (pass_vectorize): Replace declaration with that of...
4854 (make_pass_vectorize): ...new function.
4855 (pass_slp_vectorize): Replace declaration with that of...
4856 (make_pass_slp_vectorize): ...new function.
4857 (pass_complete_unroll): Replace declaration with that of...
4858 (make_pass_complete_unroll): ...new function.
4859 (pass_complete_unrolli): Replace declaration with that of...
4860 (make_pass_complete_unrolli): ...new function.
4861 (pass_parallelize_loops): Replace declaration with that of...
4862 (make_pass_parallelize_loops): ...new function.
4863 (pass_loop_prefetch): Replace declaration with that of...
4864 (make_pass_loop_prefetch): ...new function.
4865 (pass_iv_optimize): Replace declaration with that of...
4866 (make_pass_iv_optimize): ...new function.
4867 (pass_tree_loop_done): Replace declaration with that of...
4868 (make_pass_tree_loop_done): ...new function.
4869 (pass_ch): Replace declaration with that of...
4870 (make_pass_ch): ...new function.
4871 (pass_ccp): Replace declaration with that of...
4872 (make_pass_ccp): ...new function.
4873 (pass_phi_only_cprop): Replace declaration with that of...
4874 (make_pass_phi_only_cprop): ...new function.
4875 (pass_build_ssa): Replace declaration with that of...
4876 (make_pass_build_ssa): ...new function.
4877 (pass_build_alias): Replace declaration with that of...
4878 (make_pass_build_alias): ...new function.
4879 (pass_build_ealias): Replace declaration with that of...
4880 (make_pass_build_ealias): ...new function.
4881 (pass_dominator): Replace declaration with that of...
4882 (make_pass_dominator): ...new function.
4883 (pass_dce): Replace declaration with that of...
4884 (make_pass_dce): ...new function.
4885 (pass_dce_loop): Replace declaration with that of...
4886 (make_pass_dce_loop): ...new function.
4887 (pass_cd_dce): Replace declaration with that of...
4888 (make_pass_cd_dce): ...new function.
4889 (pass_call_cdce): Replace declaration with that of...
4890 (make_pass_call_cdce): ...new function.
4891 (pass_merge_phi): Replace declaration with that of...
4892 (make_pass_merge_phi): ...new function.
4893 (pass_split_crit_edges): Replace declaration with that of...
4894 (make_pass_split_crit_edges): ...new function.
4895 (pass_pre): Replace declaration with that of...
4896 (make_pass_pre): ...new function.
4897 (pass_profile): Replace declaration with that of...
4898 (make_pass_profile): ...new function.
4899 (pass_strip_predict_hints): Replace declaration with that of...
4900 (make_pass_strip_predict_hints): ...new function.
4901 (pass_lower_complex_O0): Replace declaration with that of...
4902 (make_pass_lower_complex_O0): ...new function.
4903 (pass_lower_complex): Replace declaration with that of...
4904 (make_pass_lower_complex): ...new function.
4905 (pass_lower_vector): Replace declaration with that of...
4906 (make_pass_lower_vector): ...new function.
4907 (pass_lower_vector_ssa): Replace declaration with that of...
4908 (make_pass_lower_vector_ssa): ...new function.
4909 (pass_lower_omp): Replace declaration with that of...
4910 (make_pass_lower_omp): ...new function.
4911 (pass_diagnose_omp_blocks): Replace declaration with that of...
4912 (make_pass_diagnose_omp_blocks): ...new function.
4913 (pass_expand_omp): Replace declaration with that of...
4914 (make_pass_expand_omp): ...new function.
4915 (pass_expand_omp_ssa): Replace declaration with that of...
4916 (make_pass_expand_omp_ssa): ...new function.
4917 (pass_object_sizes): Replace declaration with that of...
4918 (make_pass_object_sizes): ...new function.
4919 (pass_strlen): Replace declaration with that of...
4920 (make_pass_strlen): ...new function.
4921 (pass_fold_builtins): Replace declaration with that of...
4922 (make_pass_fold_builtins): ...new function.
4923 (pass_stdarg): Replace declaration with that of...
4924 (make_pass_stdarg): ...new function.
4925 (pass_early_warn_uninitialized): Replace declaration with that of...
4926 (make_pass_early_warn_uninitialized): ...new function.
4927 (pass_late_warn_uninitialized): Replace declaration with that of...
4928 (make_pass_late_warn_uninitialized): ...new function.
4929 (pass_cse_reciprocals): Replace declaration with that of...
4930 (make_pass_cse_reciprocals): ...new function.
4931 (pass_cse_sincos): Replace declaration with that of...
4932 (make_pass_cse_sincos): ...new function.
4933 (pass_optimize_bswap): Replace declaration with that of...
4934 (make_pass_optimize_bswap): ...new function.
4935 (pass_optimize_widening_mul): Replace declaration with that of...
4936 (make_pass_optimize_widening_mul): ...new function.
4937 (pass_warn_function_return): Replace declaration with that of...
4938 (make_pass_warn_function_return): ...new function.
4939 (pass_warn_function_noreturn): Replace declaration with that of...
4940 (make_pass_warn_function_noreturn): ...new function.
4941 (pass_cselim): Replace declaration with that of...
4942 (make_pass_cselim): ...new function.
4943 (pass_phiopt): Replace declaration with that of...
4944 (make_pass_phiopt): ...new function.
4945 (pass_forwprop): Replace declaration with that of...
4946 (make_pass_forwprop): ...new function.
4947 (pass_phiprop): Replace declaration with that of...
4948 (make_pass_phiprop): ...new function.
4949 (pass_tree_ifcombine): Replace declaration with that of...
4950 (make_pass_tree_ifcombine): ...new function.
4951 (pass_dse): Replace declaration with that of...
4952 (make_pass_dse): ...new function.
4953 (pass_nrv): Replace declaration with that of...
4954 (make_pass_nrv): ...new function.
4955 (pass_rename_ssa_copies): Replace declaration with that of...
4956 (make_pass_rename_ssa_copies): ...new function.
4957 (pass_sink_code): Replace declaration with that of...
4958 (make_pass_sink_code): ...new function.
4959 (pass_fre): Replace declaration with that of...
4960 (make_pass_fre): ...new function.
4961 (pass_check_data_deps): Replace declaration with that of...
4962 (make_pass_check_data_deps): ...new function.
4963 (pass_copy_prop): Replace declaration with that of...
4964 (make_pass_copy_prop): ...new function.
4965 (pass_vrp): Replace declaration with that of...
4966 (make_pass_vrp): ...new function.
4967 (pass_uncprop): Replace declaration with that of...
4968 (make_pass_uncprop): ...new function.
4969 (pass_return_slot): Replace declaration with that of...
4970 (make_pass_return_slot): ...new function.
4971 (pass_reassoc): Replace declaration with that of...
4972 (make_pass_reassoc): ...new function.
4973 (pass_rebuild_cgraph_edges): Replace declaration with that of...
4974 (make_pass_rebuild_cgraph_edges): ...new function.
4975 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
4976 (make_pass_remove_cgraph_callee_edges): ...new function.
4977 (pass_build_cgraph_edges): Replace declaration with that of...
4978 (make_pass_build_cgraph_edges): ...new function.
4979 (pass_local_pure_const): Replace declaration with that of...
4980 (make_pass_local_pure_const): ...new function.
4981 (pass_tracer): Replace declaration with that of...
4982 (make_pass_tracer): ...new function.
4983 (pass_warn_unused_result): Replace declaration with that of...
4984 (make_pass_warn_unused_result): ...new function.
4985 (pass_diagnose_tm_blocks): Replace declaration with that of...
4986 (make_pass_diagnose_tm_blocks): ...new function.
4987 (pass_lower_tm): Replace declaration with that of...
4988 (make_pass_lower_tm): ...new function.
4989 (pass_tm_init): Replace declaration with that of...
4990 (make_pass_tm_init): ...new function.
4991 (pass_tm_mark): Replace declaration with that of...
4992 (make_pass_tm_mark): ...new function.
4993 (pass_tm_memopt): Replace declaration with that of...
4994 (make_pass_tm_memopt): ...new function.
4995 (pass_tm_edges): Replace declaration with that of...
4996 (make_pass_tm_edges): ...new function.
4997 (pass_split_functions): Replace declaration with that of...
4998 (make_pass_split_functions): ...new function.
4999 (pass_feedback_split_functions): Replace declaration with that of...
5000 (make_pass_feedback_split_functions): ...new function.
5001 (pass_strength_reduction): Replace declaration with that of...
5002 (make_pass_strength_reduction): ...new function.
5003 (pass_ipa_lower_emutls): Replace declaration with that of...
5004 (make_pass_ipa_lower_emutls): ...new function.
5005 (pass_ipa_function_and_variable_visibility): Replace declaration with
5007 (make_pass_ipa_function_and_variable_visibility): ...new function.
5008 (pass_ipa_tree_profile): Replace declaration with that of...
5009 (make_pass_ipa_tree_profile): ...new function.
5010 (pass_early_local_passes): Replace declaration with that of...
5011 (make_pass_early_local_passes): ...new function.
5012 (pass_ipa_whole_program_visibility): Replace declaration with that
5014 (make_pass_ipa_whole_program_visibility): ...new function.
5015 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5016 (make_pass_ipa_lto_gimple_out): ...new function.
5017 (pass_ipa_increase_alignment): Replace declaration with that of...
5018 (make_pass_ipa_increase_alignment): ...new function.
5019 (pass_ipa_inline): Replace declaration with that of...
5020 (make_pass_ipa_inline): ...new function.
5021 (pass_ipa_free_lang_data): Replace declaration with that of...
5022 (make_pass_ipa_free_lang_data): ...new function.
5023 (pass_ipa_free_inline_summary): Replace declaration with that of...
5024 (make_pass_ipa_free_inline_summary): ...new function.
5025 (pass_ipa_cp): Replace declaration with that of...
5026 (make_pass_ipa_cp): ...new function.
5027 (pass_ipa_reference): Replace declaration with that of...
5028 (make_pass_ipa_reference): ...new function.
5029 (pass_ipa_pure_const): Replace declaration with that of...
5030 (make_pass_ipa_pure_const): ...new function.
5031 (pass_ipa_pta): Replace declaration with that of...
5032 (make_pass_ipa_pta): ...new function.
5033 (pass_ipa_lto_finish_out): Replace declaration with that of...
5034 (make_pass_ipa_lto_finish_out): ...new function.
5035 (pass_ipa_tm): Replace declaration with that of...
5036 (make_pass_ipa_tm): ...new function.
5037 (pass_ipa_profile): Replace declaration with that of...
5038 (make_pass_ipa_profile): ...new function.
5039 (pass_ipa_cdtor_merge): Replace declaration with that of...
5040 (make_pass_ipa_cdtor_merge): ...new function.
5041 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5043 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5044 (pass_init_datastructures): Replace declaration with that of...
5045 (make_pass_init_datastructures): ...new function.
5046 (pass_fixup_cfg): Replace declaration with that of...
5047 (make_pass_fixup_cfg): ...new function.
5048 (pass_expand): Replace declaration with that of...
5049 (make_pass_expand): ...new function.
5050 (pass_instantiate_virtual_regs): Replace declaration with that of...
5051 (make_pass_instantiate_virtual_regs): ...new function.
5052 (pass_rtl_fwprop): Replace declaration with that of...
5053 (make_pass_rtl_fwprop): ...new function.
5054 (pass_rtl_fwprop_addr): Replace declaration with that of...
5055 (make_pass_rtl_fwprop_addr): ...new function.
5056 (pass_jump): Replace declaration with that of...
5057 (make_pass_jump): ...new function.
5058 (pass_jump2): Replace declaration with that of...
5059 (make_pass_jump2): ...new function.
5060 (pass_lower_subreg): Replace declaration with that of...
5061 (make_pass_lower_subreg): ...new function.
5062 (pass_cse): Replace declaration with that of...
5063 (make_pass_cse): ...new function.
5064 (pass_fast_rtl_dce): Replace declaration with that of...
5065 (make_pass_fast_rtl_dce): ...new function.
5066 (pass_ud_rtl_dce): Replace declaration with that of...
5067 (make_pass_ud_rtl_dce): ...new function.
5068 (pass_rtl_dce): Replace declaration with that of...
5069 (make_pass_rtl_dce): ...new function.
5070 (pass_rtl_dse1): Replace declaration with that of...
5071 (make_pass_rtl_dse1): ...new function.
5072 (pass_rtl_dse2): Replace declaration with that of...
5073 (make_pass_rtl_dse2): ...new function.
5074 (pass_rtl_dse3): Replace declaration with that of...
5075 (make_pass_rtl_dse3): ...new function.
5076 (pass_rtl_cprop): Replace declaration with that of...
5077 (make_pass_rtl_cprop): ...new function.
5078 (pass_rtl_pre): Replace declaration with that of...
5079 (make_pass_rtl_pre): ...new function.
5080 (pass_rtl_hoist): Replace declaration with that of...
5081 (make_pass_rtl_hoist): ...new function.
5082 (pass_rtl_store_motion): Replace declaration with that of...
5083 (make_pass_rtl_store_motion): ...new function.
5084 (pass_cse_after_global_opts): Replace declaration with that of...
5085 (make_pass_cse_after_global_opts): ...new function.
5086 (pass_rtl_ifcvt): Replace declaration with that of...
5087 (make_pass_rtl_ifcvt): ...new function.
5088 (pass_into_cfg_layout_mode): Replace declaration with that of...
5089 (make_pass_into_cfg_layout_mode): ...new function.
5090 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5091 (make_pass_outof_cfg_layout_mode): ...new function.
5092 (pass_loop2): Replace declaration with that of...
5093 (make_pass_loop2): ...new function.
5094 (pass_rtl_loop_init): Replace declaration with that of...
5095 (make_pass_rtl_loop_init): ...new function.
5096 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5097 (make_pass_rtl_move_loop_invariants): ...new function.
5098 (pass_rtl_unswitch): Replace declaration with that of...
5099 (make_pass_rtl_unswitch): ...new function.
5100 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5101 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5102 (pass_rtl_doloop): Replace declaration with that of...
5103 (make_pass_rtl_doloop): ...new function.
5104 (pass_rtl_loop_done): Replace declaration with that of...
5105 (make_pass_rtl_loop_done): ...new function.
5106 (pass_web): Replace declaration with that of...
5107 (make_pass_web): ...new function.
5108 (pass_cse2): Replace declaration with that of...
5109 (make_pass_cse2): ...new function.
5110 (pass_df_initialize_opt): Replace declaration with that of...
5111 (make_pass_df_initialize_opt): ...new function.
5112 (pass_df_initialize_no_opt): Replace declaration with that of...
5113 (make_pass_df_initialize_no_opt): ...new function.
5114 (pass_reginfo_init): Replace declaration with that of...
5115 (make_pass_reginfo_init): ...new function.
5116 (pass_inc_dec): Replace declaration with that of...
5117 (make_pass_inc_dec): ...new function.
5118 (pass_stack_ptr_mod): Replace declaration with that of...
5119 (make_pass_stack_ptr_mod): ...new function.
5120 (pass_initialize_regs): Replace declaration with that of...
5121 (make_pass_initialize_regs): ...new function.
5122 (pass_combine): Replace declaration with that of...
5123 (make_pass_combine): ...new function.
5124 (pass_if_after_combine): Replace declaration with that of...
5125 (make_pass_if_after_combine): ...new function.
5126 (pass_ree): Replace declaration with that of...
5127 (make_pass_ree): ...new function.
5128 (pass_partition_blocks): Replace declaration with that of...
5129 (make_pass_partition_blocks): ...new function.
5130 (pass_match_asm_constraints): Replace declaration with that of...
5131 (make_pass_match_asm_constraints): ...new function.
5132 (pass_regmove): Replace declaration with that of...
5133 (make_pass_regmove): ...new function.
5134 (pass_split_all_insns): Replace declaration with that of...
5135 (make_pass_split_all_insns): ...new function.
5136 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5137 (make_pass_fast_rtl_byte_dce): ...new function.
5138 (pass_lower_subreg2): Replace declaration with that of...
5139 (make_pass_lower_subreg2): ...new function.
5140 (pass_mode_switching): Replace declaration with that of...
5141 (make_pass_mode_switching): ...new function.
5142 (pass_sms): Replace declaration with that of...
5143 (make_pass_sms): ...new function.
5144 (pass_sched): Replace declaration with that of...
5145 (make_pass_sched): ...new function.
5146 (pass_ira): Replace declaration with that of...
5147 (make_pass_ira): ...new function.
5148 (pass_reload): Replace declaration with that of...
5149 (make_pass_reload): ...new function.
5150 (pass_clean_state): Replace declaration with that of...
5151 (make_pass_clean_state): ...new function.
5152 (pass_branch_prob): Replace declaration with that of...
5153 (make_pass_branch_prob): ...new function.
5154 (pass_value_profile_transformations): Replace declaration with that
5156 (make_pass_value_profile_transformations): ...new function.
5157 (pass_postreload_cse): Replace declaration with that of...
5158 (make_pass_postreload_cse): ...new function.
5159 (pass_gcse2): Replace declaration with that of...
5160 (make_pass_gcse2): ...new function.
5161 (pass_split_after_reload): Replace declaration with that of...
5162 (make_pass_split_after_reload): ...new function.
5163 (pass_branch_target_load_optimize1): Replace declaration with that
5165 (make_pass_branch_target_load_optimize1): ...new function.
5166 (pass_thread_prologue_and_epilogue): Replace declaration with that
5168 (make_pass_thread_prologue_and_epilogue): ...new function.
5169 (pass_stack_adjustments): Replace declaration with that of...
5170 (make_pass_stack_adjustments): ...new function.
5171 (pass_peephole2): Replace declaration with that of...
5172 (make_pass_peephole2): ...new function.
5173 (pass_if_after_reload): Replace declaration with that of...
5174 (make_pass_if_after_reload): ...new function.
5175 (pass_regrename): Replace declaration with that of...
5176 (make_pass_regrename): ...new function.
5177 (pass_cprop_hardreg): Replace declaration with that of...
5178 (make_pass_cprop_hardreg): ...new function.
5179 (pass_reorder_blocks): Replace declaration with that of...
5180 (make_pass_reorder_blocks): ...new function.
5181 (pass_branch_target_load_optimize2): Replace declaration with that
5183 (make_pass_branch_target_load_optimize2): ...new function.
5184 (pass_leaf_regs): Replace declaration with that of...
5185 (make_pass_leaf_regs): ...new function.
5186 (pass_split_before_sched2): Replace declaration with that of...
5187 (make_pass_split_before_sched2): ...new function.
5188 (pass_compare_elim_after_reload): Replace declaration with that of...
5189 (make_pass_compare_elim_after_reload): ...new function.
5190 (pass_sched2): Replace declaration with that of...
5191 (make_pass_sched2): ...new function.
5192 (pass_stack_regs): Replace declaration with that of...
5193 (make_pass_stack_regs): ...new function.
5194 (pass_stack_regs_run): Replace declaration with that of...
5195 (make_pass_stack_regs_run): ...new function.
5196 (pass_df_finish): Replace declaration with that of...
5197 (make_pass_df_finish): ...new function.
5198 (pass_compute_alignments): Replace declaration with that of...
5199 (make_pass_compute_alignments): ...new function.
5200 (pass_duplicate_computed_gotos): Replace declaration with that of...
5201 (make_pass_duplicate_computed_gotos): ...new function.
5202 (pass_variable_tracking): Replace declaration with that of...
5203 (make_pass_variable_tracking): ...new function.
5204 (pass_free_cfg): Replace declaration with that of...
5205 (make_pass_free_cfg): ...new function.
5206 (pass_machine_reorg): Replace declaration with that of...
5207 (make_pass_machine_reorg): ...new function.
5208 (pass_cleanup_barriers): Replace declaration with that of...
5209 (make_pass_cleanup_barriers): ...new function.
5210 (pass_delay_slots): Replace declaration with that of...
5211 (make_pass_delay_slots): ...new function.
5212 (pass_split_for_shorten_branches): Replace declaration with that of...
5213 (make_pass_split_for_shorten_branches): ...new function.
5214 (pass_split_before_regstack): Replace declaration with that of...
5215 (make_pass_split_before_regstack): ...new function.
5216 (pass_convert_to_eh_region_ranges): Replace declaration with that
5218 (make_pass_convert_to_eh_region_ranges): ...new function.
5219 (pass_shorten_branches): Replace declaration with that of...
5220 (make_pass_shorten_branches): ...new function.
5221 (pass_set_nothrow_function_flags): Replace declaration with that of...
5222 (make_pass_set_nothrow_function_flags): ...new function.
5223 (pass_dwarf2_frame): Replace declaration with that of...
5224 (make_pass_dwarf2_frame): ...new function.
5225 (pass_final): Replace declaration with that of...
5226 (make_pass_final): ...new function.
5227 (pass_rtl_seqabstr): Replace declaration with that of...
5228 (make_pass_rtl_seqabstr): ...new function.
5229 (pass_release_ssa_names): Replace declaration with that of...
5230 (make_pass_release_ssa_names): ...new function.
5231 (pass_early_inline): Replace declaration with that of...
5232 (make_pass_early_inline): ...new function.
5233 (pass_inline_parameters): Replace declaration with that of...
5234 (make_pass_inline_parameters): ...new function.
5235 (pass_update_address_taken): Replace declaration with that of...
5236 (make_pass_update_address_taken): ...new function.
5237 (pass_convert_switch): Replace declaration with that of...
5238 (make_pass_convert_switch): ...new function.
5239 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5240 to a subclass of simple_ipa_opt_pass along with...
5241 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5242 (make_pass_ipa_tree_profile): ...new function.
5243 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5244 subclass of gimple_opt_pass along with...
5245 (pass_data_sra_early): ...new pass_data instance and...
5246 (make_pass_sra_early): ...new function.
5247 (pass_sra): Convert from a global struct to a subclass of
5248 gimple_opt_pass along with...
5249 (pass_data_sra): ...new pass_data instance and...
5250 (make_pass_sra): ...new function.
5251 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5252 gimple_opt_pass along with...
5253 (pass_data_early_ipa_sra): ...new pass_data instance and...
5254 (make_pass_early_ipa_sra): ...new function.
5255 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5256 subclass of gimple_opt_pass along with...
5257 (pass_data_ccp): ...new pass_data instance and...
5258 (make_pass_ccp): ...new function.
5259 (pass_fold_builtins): Convert from a global struct to a subclass of
5260 gimple_opt_pass along with...
5261 (pass_data_fold_builtins): ...new pass_data instance and...
5262 (make_pass_fold_builtins): ...new function.
5263 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5264 subclass of gimple_opt_pass along with...
5265 (pass_data_copy_prop): ...new pass_data instance and...
5266 (make_pass_copy_prop): ...new function.
5267 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5268 global struct to a subclass of gimple_opt_pass along with...
5269 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5270 (make_pass_rename_ssa_copies): ...new function.
5271 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5272 subclass of gimple_opt_pass along with...
5273 (pass_data_dce): ...new pass_data instance and...
5274 (make_pass_dce): ...new function.
5275 (pass_dce_loop): Convert from a global struct to a subclass of
5276 gimple_opt_pass along with...
5277 (pass_data_dce_loop): ...new pass_data instance and...
5278 (make_pass_dce_loop): ...new function.
5279 (pass_cd_dce): Convert from a global struct to a subclass of
5280 gimple_opt_pass along with...
5281 (pass_data_cd_dce): ...new pass_data instance and...
5282 (make_pass_cd_dce): ...new function.
5283 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5284 subclass of gimple_opt_pass along with...
5285 (pass_data_dominator): ...new pass_data instance and...
5286 (make_pass_dominator): ...new function.
5287 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5288 gimple_opt_pass along with...
5289 (pass_data_phi_only_cprop): ...new pass_data instance and...
5290 (make_pass_phi_only_cprop): ...new function.
5291 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5292 subclass of gimple_opt_pass along with...
5293 (pass_data_dse): ...new pass_data instance and...
5294 (make_pass_dse): ...new function.
5295 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5296 a subclass of gimple_opt_pass along with...
5297 (pass_data_forwprop): ...new pass_data instance and...
5298 (make_pass_forwprop): ...new function.
5299 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5300 struct to a subclass of gimple_opt_pass along with...
5301 (pass_data_tree_ifcombine): ...new pass_data instance and...
5302 (make_pass_tree_ifcombine): ...new function.
5303 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5304 subclass of gimple_opt_pass along with...
5305 (pass_data_ch): ...new pass_data instance and...
5306 (make_pass_ch): ...new function.
5307 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5308 subclass of gimple_opt_pass along with...
5309 (pass_data_tree_loop): ...new pass_data instance and...
5310 (make_pass_tree_loop): ...new function.
5311 (pass_tree_loop_init): Convert from a global struct to a subclass of
5312 gimple_opt_pass along with...
5313 (pass_data_tree_loop_init): ...new pass_data instance and...
5314 (make_pass_tree_loop_init): ...new function.
5315 (pass_lim): Convert from a global struct to a subclass of
5316 gimple_opt_pass along with...
5317 (pass_data_lim): ...new pass_data instance and...
5318 (make_pass_lim): ...new function.
5319 (pass_tree_unswitch): Convert from a global struct to a subclass of
5320 gimple_opt_pass along with...
5321 (pass_data_tree_unswitch): ...new pass_data instance and...
5322 (make_pass_tree_unswitch): ...new function.
5323 (pass_predcom): Convert from a global struct to a subclass of
5324 gimple_opt_pass along with...
5325 (pass_data_predcom): ...new pass_data instance and...
5326 (make_pass_predcom): ...new function.
5327 (pass_vectorize): Convert from a global struct to a subclass of
5328 gimple_opt_pass along with...
5329 (pass_data_vectorize): ...new pass_data instance and...
5330 (make_pass_vectorize): ...new function.
5331 (pass_graphite): Convert from a global struct to a subclass of
5332 gimple_opt_pass along with...
5333 (pass_data_graphite): ...new pass_data instance and...
5334 (make_pass_graphite): ...new function.
5335 (pass_graphite_transforms): Convert from a global struct to a subclass
5336 of gimple_opt_pass along with...
5337 (pass_data_graphite_transforms): ...new pass_data instance and...
5338 (make_pass_graphite_transforms): ...new function.
5339 (pass_check_data_deps): Convert from a global struct to a subclass of
5340 gimple_opt_pass along with...
5341 (pass_data_check_data_deps): ...new pass_data instance and...
5342 (make_pass_check_data_deps): ...new function.
5343 (pass_iv_canon): Convert from a global struct to a subclass of
5344 gimple_opt_pass along with...
5345 (pass_data_iv_canon): ...new pass_data instance and...
5346 (make_pass_iv_canon): ...new function.
5347 (pass_scev_cprop): Convert from a global struct to a subclass of
5348 gimple_opt_pass along with...
5349 (pass_data_scev_cprop): ...new pass_data instance and...
5350 (make_pass_scev_cprop): ...new function.
5351 (pass_record_bounds): Convert from a global struct to a subclass of
5352 gimple_opt_pass along with...
5353 (pass_data_record_bounds): ...new pass_data instance and...
5354 (make_pass_record_bounds): ...new function.
5355 (pass_complete_unroll): Convert from a global struct to a subclass of
5356 gimple_opt_pass along with...
5357 (pass_data_complete_unroll): ...new pass_data instance and...
5358 (make_pass_complete_unroll): ...new function.
5359 (pass_complete_unrolli): Convert from a global struct to a subclass of
5360 gimple_opt_pass along with...
5361 (pass_data_complete_unrolli): ...new pass_data instance and...
5362 (make_pass_complete_unrolli): ...new function.
5363 (pass_parallelize_loops): Convert from a global struct to a subclass
5364 of gimple_opt_pass along with...
5365 (pass_data_parallelize_loops): ...new pass_data instance and...
5366 (make_pass_parallelize_loops): ...new function.
5367 (pass_loop_prefetch): Convert from a global struct to a subclass of
5368 gimple_opt_pass along with...
5369 (pass_data_loop_prefetch): ...new pass_data instance and...
5370 (make_pass_loop_prefetch): ...new function.
5371 (pass_iv_optimize): Convert from a global struct to a subclass of
5372 gimple_opt_pass along with...
5373 (pass_data_iv_optimize): ...new pass_data instance and...
5374 (make_pass_iv_optimize): ...new function.
5375 (pass_tree_loop_done): Convert from a global struct to a subclass of
5376 gimple_opt_pass along with...
5377 (pass_data_tree_loop_done): ...new pass_data instance and...
5378 (make_pass_tree_loop_done): ...new function.
5379 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5380 struct to a subclass of gimple_opt_pass along with...
5381 (pass_data_cse_reciprocals): ...new pass_data instance and...
5382 (make_pass_cse_reciprocals): ...new function.
5383 (pass_cse_sincos): Convert from a global struct to a subclass of
5384 gimple_opt_pass along with...
5385 (pass_data_cse_sincos): ...new pass_data instance and...
5386 (make_pass_cse_sincos): ...new function.
5387 (pass_optimize_bswap): Convert from a global struct to a subclass of
5388 gimple_opt_pass along with...
5389 (pass_data_optimize_bswap): ...new pass_data instance and...
5390 (make_pass_optimize_bswap): ...new function.
5391 (pass_optimize_widening_mul): Convert from a global struct to a
5392 subclass of gimple_opt_pass along with...
5393 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5394 (make_pass_optimize_widening_mul): ...new function.
5395 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5396 subclass of gimple_opt_pass along with...
5397 (pass_data_phiopt): ...new pass_data instance and...
5398 (make_pass_phiopt): ...new function.
5399 (pass_cselim): Convert from a global struct to a subclass of
5400 gimple_opt_pass along with...
5401 (pass_data_cselim): ...new pass_data instance and...
5402 (make_pass_cselim): ...new function.
5403 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5404 subclass of gimple_opt_pass along with...
5405 (pass_data_phiprop): ...new pass_data instance and...
5406 (make_pass_phiprop): ...new function.
5407 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5408 subclass of gimple_opt_pass along with...
5409 (pass_data_pre): ...new pass_data instance and...
5410 (make_pass_pre): ...new function.
5411 (pass_fre): Convert from a global struct to a subclass of
5412 gimple_opt_pass along with...
5413 (pass_data_fre): ...new pass_data instance and...
5414 (make_pass_fre): ...new function.
5415 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5416 subclass of gimple_opt_pass along with...
5417 (pass_data_reassoc): ...new pass_data instance and...
5418 (make_pass_reassoc): ...new function.
5419 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5420 subclass of gimple_opt_pass along with...
5421 (pass_data_sink_code): ...new pass_data instance and...
5422 (make_pass_sink_code): ...new function.
5423 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5424 subclass of gimple_opt_pass along with...
5425 (pass_data_strlen): ...new pass_data instance and...
5426 (make_pass_strlen): ...new function.
5427 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5428 struct to a subclass of gimple_opt_pass along with...
5429 (pass_data_build_alias): ...new pass_data instance and...
5430 (make_pass_build_alias): ...new function.
5431 (pass_build_ealias): Convert from a global struct to a subclass of
5432 gimple_opt_pass along with...
5433 (pass_data_build_ealias): ...new pass_data instance and...
5434 (make_pass_build_ealias): ...new function.
5435 (pass_ipa_pta): Convert from a global struct to a subclass of
5436 simple_ipa_opt_pass along with...
5437 (pass_data_ipa_pta): ...new pass_data instance and...
5438 (make_pass_ipa_pta): ...new function.
5439 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5440 subclass of gimple_opt_pass along with...
5441 (pass_data_uncprop): ...new pass_data instance and...
5442 (make_pass_uncprop): ...new function.
5443 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5444 global struct to a subclass of gimple_opt_pass along with...
5445 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5446 (make_pass_late_warn_uninitialized): ...new function.
5447 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5448 to a subclass of gimple_opt_pass along with...
5449 (pass_data_init_datastructures): ...new pass_data instance and...
5450 (make_pass_init_datastructures): ...new function.
5451 (pass_early_warn_uninitialized): Convert from a global struct to a
5452 subclass of gimple_opt_pass along with...
5453 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5454 (make_pass_early_warn_uninitialized): ...new function.
5455 (pass_update_address_taken): Convert from a global struct to a
5456 subclass of gimple_opt_pass along with...
5457 (pass_data_update_address_taken): ...new pass_data instance and...
5458 (make_pass_update_address_taken): ...new function.
5459 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5460 struct to a subclass of gimple_opt_pass along with...
5461 (pass_data_release_ssa_names): ...new pass_data instance and...
5462 (make_pass_release_ssa_names): ...new function.
5463 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5464 subclass of gimple_opt_pass along with...
5465 (pass_data_stdarg): ...new pass_data instance and...
5466 (make_pass_stdarg): ...new function.
5467 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5468 global struct to a subclass of gimple_opt_pass along with...
5469 (pass_data_convert_switch): ...new pass_data instance and...
5470 (make_pass_convert_switch): ...new function.
5471 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5472 to a subclass of gimple_opt_pass along with...
5473 (pass_data_tail_recursion): ...new pass_data instance and...
5474 (make_pass_tail_recursion): ...new function.
5475 (pass_tail_calls): Convert from a global struct to a subclass of
5476 gimple_opt_pass along with...
5477 (pass_data_tail_calls): ...new pass_data instance and...
5478 (make_pass_tail_calls): ...new function.
5479 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5480 struct to a subclass of gimple_opt_pass along with...
5481 (pass_data_lower_vector): ...new pass_data instance and...
5482 (make_pass_lower_vector): ...new function.
5483 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5484 gimple_opt_pass along with...
5485 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5486 (make_pass_lower_vector_ssa): ...new function.
5487 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5488 to a subclass of gimple_opt_pass along with...
5489 (pass_data_slp_vectorize): ...new pass_data instance and...
5490 (make_pass_slp_vectorize): ...new function.
5491 (pass_ipa_increase_alignment): Convert from a global struct to a
5492 subclass of simple_ipa_opt_pass along with...
5493 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5494 (make_pass_ipa_increase_alignment): ...new function.
5495 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5496 gimple_opt_pass along with...
5497 (pass_data_vrp): ...new pass_data instance and...
5498 (make_pass_vrp): ...new function.
5499 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5500 subclass of simple_ipa_opt_pass along with...
5501 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5502 (make_pass_ipa_free_lang_data): ...new function.
5503 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5504 gimple_opt_pass along with...
5505 (pass_data_tsan): ...new pass_data instance and...
5506 (make_pass_tsan): ...new function.
5507 (pass_tsan_O0): Convert from a global struct to a subclass of
5508 gimple_opt_pass along with...
5509 (pass_data_tsan_O0): ...new pass_data instance and...
5510 (make_pass_tsan_O0): ...new function.
5511 * var-tracking.c (pass_variable_tracking): Convert from a global
5512 struct to a subclass of rtl_opt_pass along with...
5513 (pass_data_variable_tracking): ...new pass_data instance and...
5514 (make_pass_variable_tracking): ...new function.
5515 * web.c (pass_web): Convert from a global struct to a subclass of
5516 rtl_opt_pass along with...
5517 (pass_data_web): ...new pass_data instance and...
5518 (make_pass_web): ...new function.
5519 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5520 declaration with that of...
5521 (make_pass_mode_switch_use): ...new function.
5522 (pass_resolve_sw_modes): Replace declaration with that of...
5523 (make_pass_resolve_sw_modes): ...new function.
5524 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5525 from a global struct to a subclass of rtl_opt_pass along with...
5526 (pass_data_mode_switch_use): ...new pass_data instance and...
5527 (make_pass_mode_switch_use): ...new function.
5528 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5529 from a global struct to a subclass of rtl_opt_pass along with...
5530 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5531 (make_pass_resolve_sw_modes): ...new function.
5532 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5533 struct to a subclass of rtl_opt_pass along with...
5534 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5535 (make_pass_insert_vzeroupper): ...new function.
5536 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5537 global struct to a subclass of rtl_opt_pass along with...
5538 (pass_data_work_around_errata): ...new pass_data instance and...
5539 (make_pass_work_around_errata): ...new function.
5540 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5541 struct to a subclass of rtl_opt_pass along with...
5542 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5543 (make_pass_mips_machine_reorg2): ...new function.
5545 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5547 * passes.c (pass_manager::operator new): New.
5549 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5551 Handwritten part of conversion of passes to C++ classes.
5553 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5554 (toplev.o): Add dep on PASS_MANAGER_H.
5555 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5556 of early local pases to reflect this moving from a global to a
5557 member of gcc::pass_manager.
5558 (cgraph_add_new_function): Likewise.
5559 * lto-cgraph.c (lto_output_node): Update for conversion of
5560 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5561 * passes.c (opt_pass::clone): New.
5562 (opt_pass::gate): New.
5563 (opt_pass::execute): New.
5564 (opt_pass::opt_pass): New.
5565 (pass_manager::execute_early_local_passes): New.
5566 (pass_manager::execute_pass_mode_switching): new.
5567 (finish_optimization_passes): Convert to...
5568 (pass_manager::finish_optimization_passes): ...this.
5569 (finish_optimization_passes): Update for conversion of passes to
5571 (register_dump_files_1): Use has_gate since we cannot portably
5572 check a vtable entry against NULL.
5573 (dump_one_pass): Likewise.
5574 (ipa_write_summaries_2): Likewise.
5575 (ipa_write_optimization_summaries_1): Likewise.
5576 (ipa_read_summaries_1): Likewise.
5577 (ipa_read_optimization_summaries_1): Likewise.
5578 (execute_ipa_stmt_fixups): Likewise.
5579 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5580 pass-creation functions rather than wiring up globals, and
5581 storing the results in fields of pass_manager generated using
5583 (pass_manager::dump_profile_report): Update for conversion of
5584 passes to C++ classes.
5585 (pass_manager::execute_ipa_summary_passes): Likewise.
5586 (execute_one_ipa_transform_pass): Likewise.
5587 (execute_one_pass): Use has_gate and has_execute since we cannot
5588 portably check a vtable entry against NULL.
5589 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5590 (pass_manager): Use pass-instances.def to add fields for the
5591 various pass instances.
5592 * toplev.c (finalize): Update for move of
5593 finish_optimization_passes to a method of gcc::pass_manager.
5594 * toplev.h (finish_optimization_passes): Move to method of class
5596 * tree-pass.h (struct pass_data): New.
5597 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
5598 (opt_pass::gate): Convert to virtual function.
5599 (opt_pass::~opt_pass): New.
5600 (opt_pass::clone): New.
5601 (opt_pass::execute): Convert to virtual function.
5602 (opt_pass::opt_pass): New.
5603 (opt_pass::ctxt_): new.
5604 (gimple_opt_pass): Convert to subclass of opt_pass.
5605 (gimple_opt_pass::gimple_opt_pass): New.
5606 (rtl_opt_pass): Convert to subclass of opt_pass.
5607 (rtl_opt_pass::rtl_opt_pass): New.
5608 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5609 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5610 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5611 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5612 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5613 invocation of pass_mode_switching to reflect this moving from a
5614 global to a member of gcc::pass_manager.
5615 (ix86_option_override): Rework how pass_insert_vzeroupper is
5616 added to the pass_manager to reflect autogenerated changes.
5617 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
5619 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
5621 PR rtl-optimization/57708
5622 * recog.c (peep2_find_free_register): Validate all regs in a
5625 2013-08-05 Jan Hubicka <jh@suse.cz>
5628 * cgraph.c (verify_cgraph_node): Accept local flags from other
5630 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5631 (function_and_variable_visibility): Likewise.
5632 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5634 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5636 * graph.c (init_graph_slim_pretty_print): Remove.
5637 (print_graph_cfg): Do not call it. Use local pretty printer.
5638 (start_graph_dump): Likewise.
5640 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5642 * gimple-pretty-print.c (buffer): Remove.
5643 (initialized): Likewise.
5644 (maybe_init_pretty_print): Likewise.
5645 (print_gimple_stmt): Do not call it. Use non-static local
5646 pretty_printer variable.
5647 (print_gimple_expr): Likewise.
5648 (print_gimple_seq): Likewise.
5649 (gimple_dump_bb): Likewise.
5651 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5653 * asan.c (asan_pp): Remove.
5654 (asan_pp_initialized): Likewise.
5655 (asan_pp_initialize): Likewise.
5656 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5657 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5658 (asan_add_global): Likewise.
5660 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5662 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5663 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5664 * pretty-print.c (pp_formatted_text_data): Likewise.
5665 (pp_indent): Rename from pp_base_indent.
5666 (pp_format): Rename from pp_base_format.
5667 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5668 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5669 (pp_flush): Rename from pp_base_flush.
5670 (pp_set_line_maximum_length): Rename from
5671 pp_base_set_line_maximum_length.
5672 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5673 (pp_set_prefix): Rename from pp_base_set_prefix.
5674 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5675 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5676 (pp_append_text): Rename from pp_base_append_text.
5677 (pp_formatted_text): Rename from pp_base_formatted_text.
5678 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5679 (pp_remaining_character_count_for_line): Rename from
5680 pp_base_remaining_character_count_for_line.
5681 (pp_newline): Rename from pp_base_newline.
5682 (pp_character): Rename from pp_base_character.
5683 (pp_string): Rename from pp_base_string.
5684 (pp_maybe_space): Rename from pp_base_maybe_space.
5685 * asan.c (asan_pp_string): Adjust.
5686 (asan_emit_stack_protection): Likewise.
5687 (asan_add_global): Likewise.
5688 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5689 * tree-mudflap.c (mf_varname_tree): Likewise.
5690 * tree-pretty-print.c (pp_tree_identifier): Rename from
5691 pp_base_tree_identifier.
5692 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5693 Declare as function.
5695 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5697 * pretty-print.h (pp_bar_bar): New.
5698 (pp_ampersand_ampersand): Likewise.
5699 (pp_less_equal): Likewise.
5700 (pp_greater_equal): Likewise.
5701 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5702 printer functions instead of pp_string or operators and punctuators.
5703 (dump_gimple_call): Likewise.
5704 (dump_gimple_omp_for): Likewise.
5705 (dump_gimple_transaction): Likewise.
5706 (dump_gimple_phi): Likewise.
5707 (pp_gimple_stmt_1): Likewise.
5708 * sched-vis.c (print_insn): Likewise.
5709 * tree-mudflap.c (mf_varname_tree): Likewise.
5710 * tree-pretty-print.c (dump_block_node): Likewise.
5711 (dump_generic_node): Likewise.
5713 2013-08-02 Jan Hubicka <jh@suse.cz>
5715 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5717 * lto-streamer-out.c (tree_is_indexable): Results decls and
5718 parm decls are not indexable.
5719 (DFS_write_tree_body): Do not follow args and results.
5720 (hash_tree): Likewise.
5721 (output_functions): Rearrange so struct function is needed
5722 only when real body is output; be able to also ouptut abstract
5723 functions; output DECL_ARGUMENTS and DECL_RESULT.
5724 (lto_output): When not in WPA, ale store abstract functions.
5725 (write_symbol): Do not care about RESULT_DECL.
5726 (output_symbol_p): Handle correctly sbtract decls.
5727 * lto-streamer-in.c (input_function): Rearrange so struct
5728 function can be NULL at entry; allow streaming of
5729 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5730 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5731 sanity check during LTO.
5732 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5733 RESULT_DECl and DECL_ARGUMENTS.
5734 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5737 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5739 * pretty-print.h (pp_underscore): New.
5741 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5742 printer functions instead of pp_character.
5743 (dump_binary_rhs): Likewise.
5744 (dump_ternary_rhs): Likewise.
5745 (dump_gimple_call_args): Likewise.
5746 (pp_points_to_solution): Likewise.
5747 (dump_gimple_call): Likewise.
5748 (dump_gimple_switch): Likewise.
5749 (dump_gimple_cond): Likewise.
5750 (dump_gimple_bind): Likewise.
5751 (dump_gimple_try): Likewise.
5752 (dump_gimple_omp_for): Likewise.
5753 (dump_gimple_omp_continue): Likewise.
5754 (dump_gimple_omp_single): Likewise.
5755 (dump_gimple_omp_sections): Likewise.
5756 (dump_gimple_omp_block): Likewise.
5757 (dump_gimple_omp_critical): Likewise.
5758 (dump_gimple_transaction): Likewise.
5759 (dump_gimple_asm): Likewise.
5760 (dump_gimple_phi): Likewise.
5761 (dump_gimple_omp_parallel): Likewise.
5762 (dump_gimple_omp_task): Likewise.
5763 (dump_gimple_omp_atomic_load): Likewise.
5764 (dump_gimple_omp_atomic_store): Likewise.
5765 (dump_gimple_mem_ops): Likewise.
5766 (pp_gimple_stmt_1): Likewise.
5767 (pp_cfg_jump): Likewise.
5768 (dump_implicit_edges): Likewise.
5769 (gimple_dump_bb_for_graph): Likewise.
5770 * graph.c (draw_cfg_node): Likewise.
5771 * langhooks.c (lhd_print_error_function): Likewise.
5772 * sched-vis.c (print_exp): Likewise.
5773 (print_value): Likewise.
5774 (print_pattern): Likewise.
5775 (print_insn): Likewise.
5776 (rtl_dump_bb_for_graph): Likewise.
5777 * tree-pretty-print.c (dump_function_declaration): Likewise.
5778 (dump_array_domain): Likewise.
5779 (dump_omp_clause): Likewise.
5780 (dump_location): Likewise.
5781 (dump_generic_node): Likewise.
5782 (print_struct_decl): Likewise.
5783 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5785 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5787 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5788 candidate table when replacing a candidate statement.
5789 (replace_rhs_if_not_dup): Likewise.
5790 (replace_one_candidate): Likewise.
5792 2013-08-02 Jan Hubicka <jh@suse.cz>
5793 Martin Liska <marxin.liska@gmail.com>
5795 * cgraphunit.c (add_new_function): Fix logic when adding from
5797 (assemble_thunk): Rename to ...
5798 (expand_thunk); .. this one; export; get it working with
5799 general functions; make produced gimple valid.
5800 * cgraph.h (expand_thunk): Declare.
5802 2013-08-02 Jan Hubicka <jh@suse.cz>
5804 * ipa-cp.c (gather_context_independent_values): Use
5805 ipa_get_param_move_cost.
5806 (get_replacement_map): Remove PARAM; move parameter folding
5808 (create_specialized_node): Update.
5809 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
5810 assert that we have gimple body; update move_cost.
5811 (count_formal_params): Assert that we have gimple body.
5812 (ipa_dump_param): New function.
5813 (ipa_alloc_node_params): Break out from ...
5814 (ipa_initialize_node_params): ... here.
5815 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
5816 (ipa_write_node_info): Stream move costs.
5817 (ipa_read_node_info): Read move costs.
5818 (ipa_update_after_lto_read): Do not recompute node params.
5819 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
5820 (ipa_get_param): Check we are not in WPA.
5821 (ipa_get_param_move_cost): New.
5822 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
5823 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
5824 parm numbers to be present.
5826 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5828 PR rtl-optimization/58048
5829 * lra-constraints.c (process_alt_operands): Don't check asm
5830 operand on register.
5832 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
5834 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
5835 the implied StoreLoad barrier for atomic operations if before.
5837 2013-08-02 Jan Hubicka <jh@suse.cz>
5838 Martin Liska <marxin.liska@gmail.com>
5840 * cgraph.c (cgraph_function_body_availability): Do not check
5842 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
5843 symtab_node_availability): Declare.
5844 * ipa.c (can_replace_by_local_alias): New.
5845 (function_and_variable_visibility): Use it.
5846 * symtab.c (symtab_for_node_and_aliases,
5847 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
5849 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
5851 PR rtl-optimization/57963
5852 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
5853 (lra_constraints): Use them.
5855 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
5857 * config/arm/types.md (define_attr "type"): Add "load_acq"
5859 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
5861 (cortex_a53_store1): Likewise.
5863 2013-08-01 Jan Hubicka <jh@suse.cz>
5865 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
5866 partitions are not needed.
5868 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
5870 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
5871 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
5872 MAYBE_NON_Q_CLASS_P where appropriate.
5874 2013-08-01 Jan Hubicka <jh@suse.cz>
5876 * cgraph.h (release_function_body): Declare.
5877 * tree.c (free_lang_data_in_decl): Free, parameters and return values
5878 of unused delcarations.
5880 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5882 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
5883 RTL form when subtracting a constant.
5885 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5887 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
5888 Generate canonical plus rtx with negated immediate instead of minus
5890 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
5892 2013-08-01 Jan Hubicka <jh@suse.cz>
5894 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
5895 (cgraph_release_function_body): Likewise.
5896 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5897 * cgraph.h (cgrpah_node): Rename abstract_and_needed
5898 to used_as_abstract_origin.
5899 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
5900 symbols used as abstract origins.
5901 * cgraphunit.c (analyze_functions): Update.
5902 * ipa.c (symtab_remove_unreachable_nodes): Recompute
5903 used_as_abstract_origin.
5904 * tree-inline.c (tree_function_versioning): Update
5905 used_as_abstract_origin; be ready for DECL_RESULT and
5906 DECL_ARGUMENTS to be NULL.
5908 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
5909 for abstract functions.
5910 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
5913 2013-08-01 Jan Hubicka <jh@suse.cz>
5915 * profile.c (compute_value_histograms): Fix thinko.
5917 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5919 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
5920 aarch-common-protos.h to extra_headers.
5921 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5922 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
5923 * config/aarch64/t-aarch64 (aarch-common.o): Define.
5925 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
5927 * config/aarch64/aarch64.md (define_attr "type"): Delete.
5928 Include "../arm/types.md". Define "type" attribute for all patterns.
5929 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
5932 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
5934 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
5935 to support power8 load fusion.
5936 (fusion_gpr_mem_load): Likewise.
5938 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
5940 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
5941 declarations for power8 load fusion.
5942 (emit_fusion_gpr_load): Likewise.
5944 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
5945 tuning for power8, turn on fusion mode by default. Turn on sign
5946 extending fusion mode if normal fusion mode is on, and we are at
5948 (fusion_gpr_load_p): New function, return true if we can fuse an
5949 addis instruction with a dependent load to a GPR.
5950 (emit_fusion_gpr_load): Emit the instructions for power8 load
5953 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
5954 (VSX load fusion peepholes): New peepholes to fuse together an
5955 addi instruction with a VSX load instruction.
5957 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
5958 peepholes to fuse an addis instruction with a load to a GPR base
5959 register. If we are supporting sign extending fusions, convert
5960 sign extending loads to zero extending loads and add an explicit
5963 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5965 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
5966 aarch-common-protos.h to extra_headers.
5967 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
5968 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
5969 (arm_early_store_addr_dep): Likewise.
5970 (arm_no_early_alu_shift_dep): Likewise.
5971 (arm_no_early_alu_shift_value_dep): Likewise.
5972 (arm_no_early_mul_dep): Likewise.
5973 (arm_no_early_store_addr_dep): Likewise.
5974 (arm_mac_accumulator_is_mul_result): Likewise.
5975 (arm_mac_accumulator_is_result): Likewise.
5976 * config/arm/aarch-common.c: ... here. New file.
5977 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
5979 (arm_early_store_addr_dep): Likewise.
5980 (arm_no_early_alu_shift_dep): Likewise.
5981 (arm_no_early_alu_shift_value_dep): Likewise.
5982 (arm_no_early_mul_dep): Likewise.
5983 (arm_no_early_store_addr_dep): Likewise.
5984 (arm_mac_accumulator_is_mul_result): Likewise.
5985 (arm_mac_accumulator_is_result): Likewise.
5986 * config/arm/aarch-common-protos.h: ... here. New file.
5987 * config/arm/t-arm (aarch-common.o): Define.
5989 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
5991 * config/arm/arm.md: Include new file "types.md".
5992 (define_attr "type"): Move from here to ...
5993 (define_attr "mul32"): Likewise.
5994 (define_attr "mul64"): Likewise.
5995 * config/arm/types.md: ... here. New file.
5997 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
5999 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
6000 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
6002 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6004 * gen-pass-instances.awk: Fix offset of substr().
6006 2013-07-31 David Malcolm <dmalcolm@redhat.com>
6008 * Makefile.in (pass-instances.def): New.
6009 (passes.o): Replace dependency on passes.def with one on
6012 * gen-pass-instances.awk: New.
6014 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6015 rather than passes.def, updating local definition of NEXT_PASS
6016 macro to add an extra NUM parameter (currently unused).
6018 2013-07-30 David Malcolm <dmalcolm@redhat.com>
6020 * Makefile.in (PASS_MANAGER_H): New.
6021 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6022 (passes.o): Likewise.
6023 (statistics.o): Likewise.
6024 (cgraphunit.o): Likewise.
6025 (context.o): Depend on PASS_MANAGER_H.
6027 * pass_manager.h: New.
6029 * cgraphunit.c (cgraph_add_new_function): Update for moves
6030 of globals to fields of pass_manager.
6031 (analyze_function): Likewise.
6032 (expand_function): Likewise.
6033 (ipa_passes): Likewise.
6034 (compile): Likewise.
6036 * context.c (context::context): New.
6037 * context.h (context::context): New.
6038 (context::get_passes): New.
6039 (context::passes_): New.
6041 * lto-cgraph.c (input_node): Update for moves of globals to
6042 fields of pass_manager.
6044 * passes.c (all_passes): Remove, in favor of a field of the
6045 same name within the new class pass_manager.
6046 (all_small_ipa_passes): Likewise.
6047 (all_lowering_passes): Likewise.
6048 (all_regular_ipa_passes): Likewise.
6049 (all_late_ipa_passes): Likewise.
6050 (all_lto_gen_passes): Likewise.
6051 (passes_by_id): Likewise.
6052 (passes_by_id_size): Likewise.
6053 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6054 the new class pass_manager.
6055 (set_pass_for_id): Convert to...
6056 (pass_manager::set_pass_for_id): ...method.
6057 (get_pass_for_id): Convert to...
6058 (pass_manager::get_pass_for_id): ...method.
6059 (register_one_dump_file): Move body of implementation into...
6060 (pass_manager::register_one_dump_file): ...here.
6061 (register_dump_files_1): Convert to...
6062 (pass_manager::register_dump_files_1): ...method.
6063 (register_dump_files): Convert to...
6064 (pass_manager::register_dump_files): ...method.
6065 (create_pass_tab): Update for moves of globals to fields of
6067 (dump_passes): Move body of implementation into...
6068 (pass_manager::dump_passes): ...here.
6069 (register_pass): Move body of implementation into...
6070 (pass_manager::register_pass): ...here.
6071 (init_optimization_passes): Convert into...
6072 (pass_manager::pass_manager): ...constructor for new
6073 pass_manager class, and initialize the pass_lists array.
6074 (check_profile_consistency): Update for moves of globals to
6075 fields of pass_manager.
6076 (dump_profile_report): Move body of implementation into...
6077 (pass_manager::dump_profile_report): ...here.
6078 (ipa_write_summaries_1): Update for moves of pass lists from
6079 being globals to fields of pass_manager.
6080 (ipa_write_optimization_summaries): Likewise.
6081 (ipa_read_summaries): Likewise.
6082 (ipa_read_optimization_summaries): Likewise.
6083 (execute_all_ipa_stmt_fixups): Likewise.
6085 * statistics.c (statistics_fini): Update for moves of globals to
6086 fields of pass_manager.
6088 * toplev.c (general_init): Replace call to
6089 init_optimization_passes with construction of the pass_manager
6092 * tree-pass.h (all_passes): Remove, in favor of a field of the
6093 same name within the new class pass_manager.
6094 (all_small_ipa_passes): Likewise.
6095 (all_lowering_passes): Likewise.
6096 (all_regular_ipa_passes): Likewise.
6097 (all_lto_gen_passes): Likewise.
6098 (all_late_ipa_passes): Likewise.
6099 (passes_by_id): Likewise.
6100 (passes_by_id_size): Likewise.
6101 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6102 the new class pass_manager.
6103 (get_pass_for_id): Remove.
6105 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6107 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6110 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6112 * arm.md (mulhi3): New expand pattern.
6114 2013-07-30 Jan Hubicka <jh@suse.cz>
6115 Martin Liska <marxin.liska@gmail.com>
6117 * profile.c (compute_value_histograms): Do not ICE when
6118 there is mismatch only on some counters.
6120 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6122 PR rtl-optimization/57637
6123 * function.c (move_insn_for_shrink_wrap): Also check the
6124 GEN set of the LIVE problem for the liveness analysis
6125 if it exists, otherwise give up.
6127 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6129 PR tree-optimization/57993
6130 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6131 replaced statement in the candidate table.
6132 (phi_add_costs): Return infinite cost when the hidden basis does
6133 not dominate all phis on which the candidate is dependent.
6134 (replace_one_candidate): Record replaced statement in the
6137 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6139 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6140 (ashlv2si3): New expander.
6141 (*ashlv2si3_i): New define_insn_and_split.
6142 * predicates.md (float_operation): Allow patterns with three
6145 PR rtl-optimization/58021
6146 * mode-switching.c (create_pre_exit): Always split off preceding
6147 insns if we are not at the basic block head.
6149 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6151 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6152 (UCLIBC_DYNAMIC_LINKER): New macro.
6153 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6155 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6156 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6158 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6159 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6160 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6161 for SF and DF modes. Use ieee_quad_format for TF mode.
6162 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6163 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6164 (mips_option_override): Handle `-mnan=legacy'.
6165 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6166 `-mabs=2008' and `-mnan=2008'.
6167 (OPTION_DEFAULT_SPECS): Add "nan" default.
6168 (ASM_SPEC): Handle `-mnan='.
6169 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6170 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6171 comment accordingly.
6172 (neg<mode>2): Likewise.
6173 * config/mips/mips.opt (mabs, mnan): New options.
6174 * doc/install.texi (Configuration): Document `--with-nan=' option.
6175 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6177 (MIPS Options): Document them.
6178 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6179 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6180 * configure: Regenerate.
6181 * config.in: Regenerate.
6183 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6185 * config/i386/i386.md (float post-reload splitters): Do not check
6186 for subregs of SSE registers.
6188 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6189 H.J. Lu <hongjiu.lu@intel.com>
6193 * config/i386/i386.md (post-reload splitter
6194 to avoid partial SSE reg dependency stalls): New pattern.
6196 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6198 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6200 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6202 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6203 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
6205 2013-07-26 Cary Coutant <ccoutant@google.com>
6207 * dwarf2out.c (die_checksum_ordered): Don't include template
6208 instantiations in signature.
6209 (is_template_parameter): New function.
6210 (is_template_instantiation): New function.
6211 (generate_skeleton_bottom_up): Don't include template instantiations
6213 (generate_skeleton): Likewise.
6214 (break_out_comdat_types): Move recursive call to break out nested
6216 (prune_unused_types_mark_generic_parms_dies): Call
6217 is_template_parameter.
6219 2013-07-26 Ian Bolton <ian.bolton@arm.com>
6221 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6222 uses vector registers.
6223 * config/aarch64/iterators.md: Add attributes rtn and vas.
6225 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6226 Richard Earnshaw <richard.earnshaw@arm.com>
6228 * combine.c (simplify_comparison): Re-canonicalize operands
6230 * config/arm/arm.md (movcond_addsi): New splitter.
6232 2013-07-25 Sterling Augustine <saugustine@google.com>
6234 * dwarf2out.c (size_of_pubnames): Move code to...
6235 (include_pubname_in_output): ...here. New.
6236 (want_pubnames): Rearrange.
6237 (output_pubnames): Call include_pubname_in_output. Move assertion.
6239 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6241 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6243 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6246 * doc/extend.texi: Remove obsolete builtins. Fix
6247 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6249 2013-07-25 Jan Hubicka <jh@suse.cz>
6251 * cgraph.c (release_function_body): Break out from ...
6252 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6254 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6255 old_tree in the map.
6256 (create_specialized_node): Update.
6257 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6259 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6260 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
6261 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6262 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6265 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6267 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6268 addsi3_carryin_alt2_<optab>): Correct output template.
6270 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6272 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6273 Adjust for arm_restrict_it.
6274 Remove trailing whitespace.
6276 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6278 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6279 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6281 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6283 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6285 PR rtl-optimization/57960
6286 * lra-constraints.c (process_alt_operands): Use the right mode
6287 when checking strict_low.
6289 2013-07-25 Jan Hubicka <jh@suse.cz>
6291 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6292 * cgraph.c (cgraph_remove_node): Do not release function body
6293 when in cgraph streaming.
6294 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
6295 in other partitions are not considered reachable; fix handling of
6298 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6300 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6302 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6307 * config/arm/arm.md ("*sibcall_insn): Replace use of
6308 Ss with US. Adjust output for v5 and v4t.
6309 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
6311 * config/arm/constraints.md ("Ss"): Rename to US.
6313 2013-07-25 Terry Guo <terry.guo@arm.com>
6315 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6316 shift_add/shift_sub0/shift_sub1 RTXs.
6318 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
6319 Anton Blanchard <anton@au1.ibm.com>
6321 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6322 (altivec_vpks<VI_char>ss): Likewise.
6323 (altivec_vpks<VI_char>us): Likewise.
6324 (altivec_vpku<VI_char>us): Likewise.
6325 (altivec_vpku<VI_char>um): Likewise.
6327 2013-07-24 David Malcolm <dmalcolm@redhat.com>
6329 Introduce context class.
6331 * Makefile.in (CONTEXT_H): New.
6332 (OBJS): Add context.o.
6333 (toplev.o): Add CONTEXT_H to dependencies.
6336 * toplev.c (general_init): Create the singleton gcc::context instance.
6342 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6344 PR rtl-optimization/57968
6345 * mode-switching.c (create_pre_exit): Allow instructions that
6346 don't set a return register to need a non-exit mode.
6348 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6349 Anton Blanchard <anton@au1.ibm.com>
6351 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6352 operands to vperm for little endian.
6353 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6354 of lvsl to create the control mask for a vperm for little endian.
6356 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6357 Anton Blanchard <anton@au1.ibm.com>
6359 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6360 two operands for little-endian.
6362 2013-07-23 Steve Ellcey <sellcey@mips.com>
6364 * config/mips/mips.c (mips_case_values_threshold): New.
6365 (TARGET_CASE_VALUES_THRESHOLD): Define.
6367 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6368 Anton Blanchard <anton@au1.ibm.com>
6370 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6371 selection of field for vector splat in little endian mode.
6373 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6375 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6376 expanders to rs6000.md.
6377 (ior<mode>3): Likewise.
6378 (and<mode>3): Likewise.
6379 (one_cmpl<mode>2): Likewise.
6380 (nor<mode>3): Likewise.
6381 (andc<mode>3): Likewise.
6382 (eqv<mode>3): Likewise.
6383 (nand<mode>3): Likewise.
6384 (orc<mode>3): Likewise.
6386 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6389 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6390 to split multi-word logical operations.
6391 (rs6000_split_logical_di): Likewise.
6392 (rs6000_split_logical): Likewise.
6394 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6395 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6396 and allow TImode operations in 32-bit.
6397 (vsx_and<mode>3_64bit): Likewise.
6398 (vsx_ior<mode>3_32bit): Likewise.
6399 (vsx_ior<mode>3_64bit): Likewise.
6400 (vsx_xor<mode>3_32bit): Likewise.
6401 (vsx_xor<mode>3_64bit): Likewise.
6402 (vsx_one_cmpl<mode>2_32bit): Likewise.
6403 (vsx_one_cmpl<mode>2_64bit): Likewise.
6404 (vsx_nor<mode>3_32bit): Likewise.
6405 (vsx_nor<mode>3_64bit): Likewise.
6406 (vsx_andc<mode>3_32bit): Likewise.
6407 (vsx_andc<mode>3_64bit): Likewise.
6408 (vsx_eqv<mode>3_32bit): Likewise.
6409 (vsx_eqv<mode>3_64bit): Likewise.
6410 (vsx_nand<mode>3_32bit): Likewise.
6411 (vsx_nand<mode>3_64bit): Likewise.
6412 (vsx_orc<mode>3_32bit): Likewise.
6413 (vsx_orc<mode>3_64bit): Likewise.
6415 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6416 logical types in GPRs.
6418 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6419 logical insns to rs6000.md, and allow TImode operations in
6421 (altivec_ior<mode>3): Likewise.
6422 (altivec_xor<mode>3): Likewise.
6423 (altivec_one_cmpl<mode>2): Likewise.
6424 (altivec_nor<mode>3): Likewise.
6425 (altivec_andc<mode>3): Likewise.
6427 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6428 attributes for moving the 128-bit logical operations into
6430 (BOOL_REGS_OUTPUT): Likewise.
6431 (BOOL_REGS_OP1): Likewise.
6432 (BOOL_REGS_OP2): Likewise.
6433 (BOOL_REGS_UNARY): Likewise.
6434 (BOOL_REGS_AND_CR0): Likewise.
6435 (one_cmpl<mode>2): Add support for DI logical operations on
6436 32-bit, splitting the operations to 32-bit.
6440 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6441 changes to combine the 32/64-bit code, allow logical operations on
6442 TI mode in 32-bit, and to use similar match_operator patterns like
6443 scalar mode uses. Combine the Altivec and VSX code for logical
6444 operations, and move it here.
6445 (ior<mode>3, 128-bit types): Likewise.
6446 (xor<mode>3, 128-bit types): Likewise.
6447 (one_cmpl<mode>3, 128-bit types): Likewise.
6448 (nor<mode>3, 128-bit types): Likewise.
6449 (andc<mode>3, 128-bit types): Likewise.
6450 (eqv<mode>3, 128-bit types): Likewise.
6451 (nand<mode>3, 128-bit types): Likewise.
6452 (orc<mode>3, 128-bit types): Likewise.
6453 (and<mode>3_internal): Likewise.
6454 (bool<mode>3_internal): Likewise.
6455 (boolc<mode>3_internal1): Likewise.
6456 (boolc<mode>3_internal2): Likewise.
6457 (boolcc<mode>3_internal1): Likewise.
6458 (boolcc<mode>3_internal2): Likewise.
6459 (eqv<mode>3_internal1): Likewise.
6460 (eqv<mode>3_internal2): Likewise.
6461 (one_cmpl1<mode>3_internal): Likewise.
6463 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6465 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6466 Rename flag_stack_usage to flag_stack_usage_info.
6468 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6470 * config/microblaze/sync.md: New file.
6471 * config/microblaze/microblaze.md: Include sync.md
6472 * config/microblaze/microblaze.c: Add print_operand 'y'.
6473 * config/microblaze/constraints.md: Add memory_contraint
6474 'Q' which is a single register.
6476 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6478 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6480 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6482 * reload.c (find_reloads): Exit loop once we find this operand
6483 cannot be reloaded somehow for this alternative.
6485 * reload.c (find_reloads): Exit loop once we find a hard register.
6487 * rtlanal.c (computed_jump_p): Exit loop once we find label
6490 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6492 * cfgloopmanip.c (remove_path): Exit loop after setting
6495 * gensupport.c (subst_dup): Avoid loop if code is not
6496 MATCH_DUP nor MATCH_OP_DUP.
6498 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6500 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6502 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6504 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6505 true for SP_REGNUM if mode == ptr_mode.
6506 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6507 with value R0_REGNUM + 31.
6509 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6511 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6512 pad pointer-typed argument downward.
6514 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6516 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6517 and __ILP32__ when the ILP32 model is in use.
6519 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6521 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6522 (aarch64_load_symref_appropriately): In the case of
6523 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6524 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6525 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6526 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6527 if 'mode' doesn't equal to 'ptr_mode'.
6528 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6529 'vcall_offset'; change to call aarch64_emit_move differently depending
6530 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6531 to calculate the upper bound of 'vcall_offset'.
6532 (aarch64_cannot_force_const_mem): Change to also return true if
6534 (aarch64_legitimize_reload_address): In the case of large
6535 displacements, add new local variable 'xmode' and an assertion
6536 based on it; change to use 'xmode' to generate the new rtx and
6538 (aarch64_asm_trampoline_template): Change to generate the template
6539 differently depending on TARGET_ILP32 or not; change to use
6540 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6541 (aarch64_trampoline_size): Removed.
6542 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6543 and replace immediate literals with it. Change to use 'ptr_mode'
6544 instead of 'DImode' and call convert_memory_address if the mode
6545 of 'fnaddr' doesn't equal to 'ptr_mode'.
6546 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6548 (aarch64_elf_asm_destructor): Likewise.
6549 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6550 on TARGET_ILP32 instead of aarch64_trampoline_size.
6551 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6552 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6553 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6554 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6555 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6556 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6557 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6558 depending on the value of 'mode'.
6559 (add_losym_<mode>): New.
6560 (ldr_got_small_<mode>): New, based on ldr_got_small.
6561 (ldr_got_small): Remove.
6562 (ldr_got_small_sidi): New.
6563 * config/aarch64/iterators.md (P): New.
6564 (PTR): Change to 'ptr_mode' in the condition.
6566 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6568 * config.gcc (aarch64*-*-*): Support --with-abi.
6569 (aarch64*-*-elf): Support --with-multilib-list.
6570 (aarch64*-*-linux*): Likewise.
6571 (supported_defaults): Add abi to aarch64*-*-*.
6572 * configure.ac: Mention AArch64 for --with-multilib-list.
6573 * configure: Re-generated.
6574 * config/aarch64/biarchilp32.h: New file.
6575 * config/aarch64/biarchlp64.h: New file.
6576 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6578 (MULTILIB_DEFAULTS): Ditto.
6579 (DRIVER_SELF_SPECS): Ditto.
6580 (ASM_SPEC): Update to also substitute -mabi.
6581 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6582 file whose name depends on -mabi= and -mbig-endian.
6583 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6585 (POINTER_SIZE): New define.
6586 (POINTERS_EXTEND_UNSIGNED): Ditto.
6587 (enum aarch64_abi_type): New enumeration tag.
6588 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6589 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6590 (TARGET_ILP32): New define.
6591 * config/aarch64/aarch64.opt (mabi): New.
6593 (ilp32, lp64): New values for -mabi.
6594 * config/aarch64/t-aarch64 (comma): New define.
6595 (MULTILIB_OPTIONS): Ditto.
6596 (MULTILIB_DIRNAMES): Ditto.
6597 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6598 * doc/invoke.texi: Document -mabi for AArch64.
6600 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
6602 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6604 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6605 Anton Blanchard <anton@au1.ibm.com>
6607 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6608 endianness when selecting field to splat.
6610 2013-07-22 Eric Christopher <echristo@gmail.com>
6612 * dwarf2out.c (die_odr_checksum): New function to use
6613 CHECKSUM_ macros and ULEB128 for DIE tag.
6614 (generate_type_signature): Use.
6616 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6618 * config.gcc (sparc*-*-*): Accept leon3 processor.
6619 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6620 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6621 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6622 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6623 (mfix-ut699): Adjust comment.
6624 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6625 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6626 (CPP_CPU_SPEC): Likewise.
6627 (ASM_CPU_SPEC): Likewise.
6628 * config/sparc/sparc.c (leon3_cost): New constant.
6629 (sparc_option_override): Add leon3 support.
6630 (mem_ref): New function.
6631 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6632 (sparc_do_work_around_errata): Look into the instruction in the delay
6633 slot and adjust accordingly. Add fix for the data cache nullify issues
6634 of the UT699. Change insertion position for the NOP.
6635 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6636 (leon3_load): New reservation.
6637 (leon_store): Bump latency to 2.
6638 (grfpu): New automaton.
6639 (grfpu_alu): New unit.
6640 (grfpu_ds): Likewise.
6641 (leon_fp_alu): Adjust.
6642 (leon_fp_mult): Delete.
6643 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6644 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6645 * config/sparc/sparc.md (cpu): Add leon3.
6646 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6648 (atomic_test_and_set): Likewise.
6651 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6653 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6654 default for R5900 targets.
6655 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6656 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6657 * config/mips/mips.c (mips_option_override): Report an error for
6658 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6659 for -march=r5900 -mhard-float.
6661 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6663 * df-problems.c (can_move_insns_across): Exit loop once we
6664 find a non-fixed, non-global register.
6666 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6669 * omega.c (omega_eliminate_red): Break after setting red_found.
6670 (omega_problem_has_red_equations): Similarly after setting found.
6671 (omega_query_variable): Similarly after setting coupled.
6673 2013-07-22 Marek Polacek <polacek@redhat.com>
6675 * gimplify.c: Don't include gimple.h twice.
6677 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6679 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6680 instead of TARGET_THUMB1.
6681 (Pz): New constraint.
6682 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6684 (compare_negsi_si): Likewise.
6685 (compare_addsi2_op0): Likewise.
6686 (compare_addsi2_op1): Likewise.
6687 (addsi3_carryin_<optab>): Likewise.
6688 (addsi3_carryin_alt2_<optab>): Likewise.
6689 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6690 for arm_restrict_it.
6691 (subsi3_carryin): Likewise.
6692 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6693 (minmax_arithsi): Disable for arm_restrict_it.
6694 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6695 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6696 (satsi_<SAT:code>_shift): Likewise.
6697 (arm_shiftsi3): Add alternative for 16-bit encoding.
6698 (arm32_movhf): Disable for arm_restrict_it.
6699 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6700 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6702 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6704 * config/arm/arm.md (attribute "insn"): Delete.
6705 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6706 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6707 (not_shiftsi): Update for attribute change.
6708 (not_shiftsi_compare0): Likewise.
6709 (not_shiftsi_compare0_scratch): Likewise.
6710 (arm_one_cmplsi2): Likewise.
6711 (thumb1_one_cmplsi2): Likewise.
6712 (notsi_compare0): Likewise.
6713 (notsi_compare0_scratch): Likewise.
6714 (thumb1_movdi_insn): Likewise.
6715 (arm_movsi_insn): Likewise.
6716 (movhi_insn_arch4): Likewise.
6717 (movhi_bytes): Likewise.
6718 (arm_movqi_insn): Likewise.
6719 (thumb1_movqi_insn): Likewise.
6720 (arm32_movhf): Likewise.
6721 (thumb1_movhf): Likewise.
6722 (arm_movsf_soft_insn): Likewise.
6723 (thumb1_movsf_insn): Likewise.
6724 (thumb_movdf_insn): Likewise.
6725 (movsicc_insn): Likewise.
6726 (movsfcc_soft_insn): Likewise.
6727 (and_scc): Likewise.
6728 (cond_move): Likewise.
6729 (if_move_not): Likewise.
6730 (if_not_move): Likewise.
6731 (if_shift_move): Likewise.
6732 (if_move_shift): Likewise.
6733 (if_shift_shift): Likewise.
6734 (if_not_arith): Likewise.
6735 (if_arith_not): Likewise.
6736 (cond_move_not): Likewise.
6737 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6738 (neon_mov<mode>): Likewise.
6739 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6740 (thumb2_movsi_vfp): Likewise.
6741 (movsf_vfp): Likewise.
6742 (thumb2_movsf_vfp): Likewise.
6743 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6745 (cortexa7_older_only): Likewise.
6746 (cortexa7_younger): Likewise.
6747 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6748 (1020alu_shift_op): Likewise.
6749 (1020alu_shift_reg_op): Likewise.
6750 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6751 (alu_shift_op): Likewise.
6752 (alu_shift_reg_op): Likewise.
6753 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6754 (11_alu_shift_op): Likewise.
6755 (11_alu_shift_reg_op): Likewise.
6756 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6757 (9_alu_shift_reg_op): Likewise.
6758 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6760 (cortex_a15_alu_shift): Likewise.
6761 (cortex_a15_alu_shift_reg): Likewise.
6762 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6763 (cortex_a5_alu_shift): Likewise.
6764 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6766 (cortex_a53_alu_shift): Likewise.
6767 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6769 (cortex_a7_alu_reg): Likewise.
6770 (cortex_a7_alu_shift): Likewise.
6771 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6772 (cortex_a8_alu_shift): Likewise.
6773 (cortex_a8_alu_shift_reg): Likewise.
6774 (cortex_a8_mov): Likewise.
6775 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6776 (cortex_a9_dp_shift): Likewise.
6777 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6778 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6779 (cortex_r4_mov): Likewise.
6780 (cortex_r4_alu_shift): Likewise.
6781 (cortex_r4_alu_shift_reg): Likewise.
6782 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6783 (526_alu_shift_op): Likewise.
6784 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6785 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6786 (626te_alu_shift_op): Likewise.
6787 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6788 (726te_alu_op): Likewise.
6789 (726te_alu_shift_op): Likewise.
6790 (726te_alu_shift_reg_op): Likewise.
6791 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6792 (mp626_alu_shift_op): Likewise.
6793 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6794 (pj4_alu_e1_conds): Likewise.
6795 (pj4_alu): Likewise.
6796 (pj4_alu_conds): Likewise.
6797 (pj4_shift): Likewise.
6798 (pj4_shift_conds): Likewise.
6799 (pj4_alu_shift): Likewise.
6800 (pj4_alu_shift_conds): Likewise.
6802 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6804 * config/arm/predicates.md (shiftable_operator_strict_it):
6806 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
6807 Disable cond_exec version for arm_restrict_it.
6808 (thumb2_smaxsi3): Convert to generate cond_exec.
6809 (thumb2_sminsi3): Likewise.
6810 (thumb32_umaxsi3): Likewise.
6811 (thumb2_uminsi3): Likewise.
6812 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
6813 (thumb2_neg_abssi2): Likewise.
6814 (thumb2_mov_scc): Add alternative for 16-bit encoding.
6815 (thumb2_movsicc_insn): Adjust alternatives.
6816 (thumb2_mov_negscc): Disable for arm_restrict_it.
6817 (thumb2_mov_negscc_strict_it): New pattern.
6818 (thumb2_mov_notscc_strict_it): New pattern.
6819 (thumb2_mov_notscc): Disable for arm_restrict_it.
6820 (thumb2_ior_scc): Likewise.
6821 (thumb2_ior_scc_strict_it): New pattern.
6822 (thumb2_cond_move): Adjust for arm_restrict_it.
6823 (thumb2_cond_arith): Disable for arm_restrict_it.
6824 (thumb2_cond_arith_strict_it): New pattern.
6825 (thumb2_cond_sub): Adjust for arm_restrict_it.
6826 (thumb2_movcond): Likewise.
6827 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
6828 (thumb2_zero_extendhisi2_v6): Likewise.
6829 (thumb2_zero_extendqisi2_v6): Likewise.
6830 (orsi_notsi_si): Likewise.
6831 (orsi_not_shiftsi_si): Likewise.
6833 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
6835 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
6836 instruction sequence is 1 byte shorter.
6838 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
6840 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
6841 it is not needed after split.
6843 2013-07-20 Iain Sandoe <iain@codesourcery.com>
6846 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
6847 second label for nonlocal goto receivers. Don't output pic base labels
6848 unless we're producing PIC; mark that action unreachable().
6849 (ix86_save_reg): If the function contains a nonlocal label, save the
6851 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
6852 * config/darwin.c (emitted_pic_label_num): New GTY.
6853 (update_pic_label_number_if_needed): New.
6854 (machopic_output_function_base_name): Adjust for nonlocal receiver
6856 (machopic_should_output_picbase_label): New.
6857 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
6858 (nonlocal_goto_receiver): New insn and split.
6860 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
6862 * config/aarch64/aarch64-builtins.c
6863 (aarch64_fold_builtin): Fold abs in all modes.
6864 * config/aarch64/aarch64-simd-builtins.def
6865 (abs): Enable for all modes.
6866 * config/aarch64/arm_neon.h
6867 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
6868 (vabs_f64): Add missing intrinsic.
6870 2013-07-19 Ian Bolton <ian.bolton@arm.com>
6872 * config/aarch64/arm_neon.h (vabs_s64): New function
6874 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
6877 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
6878 * config/avr/avr.md (adjust_len): Add `round'.
6879 * config/avr/avr-protos.h (avr_out_round): New prototype.
6880 (avr_out_plus): Add `out_label' argument.
6881 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
6882 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
6883 Handle the case where `insn' is just a pattern.
6884 (avr_out_bitop): Handle the case where `insn' is just a pattern.
6885 (avr_out_round): New function.
6886 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
6888 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
6890 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6891 Add check for flag_stack_usage to handle -fstack-usage support
6893 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
6895 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
6896 interaction for new Power8 flags and VSX.
6898 2013-07-18 Sriraman Tallam <tmsriram@google.com>
6901 * tree-inline.c (expand_call_inline): Emit errors during
6902 early_inlining only if optimization is not turned on.
6904 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6908 * passes.c (init_optimization_passes): Move the construction of
6909 the pass hierarchy into a new passes.def file.
6911 * Makefile.in (passes.o): Add dependency on passes.def.
6913 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6915 * passes.c (init_optimization_passes): Introduce macros for
6916 constructing the tree of passes (INSERT_PASSES_AFTER,
6917 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
6918 TERMINATE_PASS_LIST).
6920 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
6921 Wei Mi <wmi@google.com>
6923 PR rtl-optimization/57878
6924 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
6926 (reload_pseudo_compare_func): Check nregs first for reload
6929 2013-07-18 David Malcolm <dmalcolm@redhat.com>
6931 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
6933 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
6935 * read-rtl.c (validate_const_int): Once an invalid character is
6936 seen, quit the loop.
6938 * gengtype.c (write_roots): Similarly once we find the "deletable"
6939 or "if_marked" option.
6941 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6943 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
6944 "xtab" and "sat". Move value "clz" from here to ...
6945 (attriubte "type"): ... here.
6946 (satsi_<SAT:code>): Delete "insn" attribute.
6947 (satsi_<SAT:code>_shift): Likewise.
6948 (arm_zero_extendqisi2addsi): Likewise.
6949 (arm_extendqisi2addsi): Likewise.
6950 (clzsi2): Update for attribute changes.
6951 (rbitsi2): Likewise.
6952 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
6954 (arm_usatsihi): Likewise.
6955 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6957 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
6959 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
6960 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
6961 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
6962 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
6963 in alphabetical order.
6964 (attribute "core_cycles"): Update for attribute changes.
6965 (arm_addsi3): Likewise.
6966 (addsi3_compare0): Likewise.
6967 (addsi3_compare0_scratch): Likewise.
6968 (addsi3_compare_op1): Likewise.
6969 (addsi3_compare_op2): Likewise.
6970 (compare_addsi2_op0): Likewise.
6971 (compare_addsi2_op1): Likewise.
6972 (addsi3_carryin_shift_<optab>): Likewise.
6973 (subsi3_carryin_shift): Likewise.
6974 (rsbsi3_carryin_shift): Likewise.
6975 (arm_subsi3_insn): Likewise.
6976 (subsi3_compare0): Likewise.
6977 (subsi3_compare): Likewise.
6978 (arm_andsi3_insn): Likewise.
6979 (thumb1_andsi3_insn): Likewise.
6980 (andsi3_compare0): Likewise.
6981 (andsi3_compare0_scratch): Likewise.
6982 (zeroextractsi_compare0_scratch
6983 (andsi_not_shiftsi_si): Likewise.
6984 (iorsi3_insn): Likewise.
6985 (iorsi3_compare0): Likewise.
6986 (iorsi3_compare0_scratch): Likewise.
6987 (arm_xorsi3): Likewise.
6988 (thumb1_xorsi3_insn): Likewise.
6989 (xorsi3_compare0): Likewise.
6990 (xorsi3_compare0_scratch): Likewise.
6991 (satsi_<SAT:code>_shift): Likewise.
6993 (arm_shiftsi3): Likewise.
6994 (shiftsi3_compare0): Likewise.
6995 (not_shiftsi): Likewise.
6996 (not_shiftsi_compare0): Likewise.
6997 (not_shiftsi_compare0_scratch): Likewise.
6998 (arm_one_cmplsi2): Likewise.
6999 (thumb_one_complsi2): Likewise.
7000 (notsi_compare0): Likewise.
7001 (notsi_compare0_scratch): Likewise.
7002 (thumb1_zero_extendhisi2): Likewise.
7003 (arm_zero_extendhisi2): Likewise.
7004 (arm_zero_extendhisi2_v6): Likewise.
7005 (arm_zero_extendhisi2addsi): Likewise.
7006 (thumb1_zero_extendqisi2): Likewise.
7007 (thumb1_zero_extendqisi2_v6): Likewise.
7008 (arm_zero_extendqisi2): Likewise.
7009 (arm_zero_extendqisi2_v6): Likewise.
7010 (arm_zero_extendqisi2addsi): Likewise.
7011 (thumb1_extendhisi2): Likewise.
7012 (arm_extendhisi2): Likewise.
7013 (arm_extendhisi2_v6): Likewise.
7014 (arm_extendqisi): Likewise.
7015 (arm_extendqisi_v6): Likewise.
7016 (arm_extendqisi2addsi): Likewise.
7017 (thumb1_extendqisi2): Likewise.
7018 (thumb1_movdi_insn): Likewise.
7019 (arm_movsi_insn): Likewise.
7020 (movsi_compare0): Likewise.
7021 (movhi_insn_arch4): Likewise.
7022 (movhi_bytes): Likewise.
7023 (arm_movqi_insn): Likewise.
7024 (thumb1_movqi_insn): Likewise.
7025 (arm32_movhf): Likewise.
7026 (thumb1_movhf): Likewise.
7027 (arm_movsf_soft_insn): Likewise.
7028 (thumb1_movsf_insn): Likewise.
7029 (movdf_soft_insn): Likewise.
7030 (thumb_movdf_insn): Likewise.
7031 (arm_cmpsi_insn): Likewise.
7032 (cmpsi_shiftsi): Likewise.
7033 (cmpsi_shiftsi_swp): Likewise.
7034 (arm_cmpsi_negshiftsi_si): Likewise.
7035 (movsicc_insn): Likewise.
7036 (movsfcc_soft_insn): Likewise.
7037 (arith_shiftsi): Likewise.
7038 (arith_shiftsi_compare0
7039 (arith_shiftsi_compare0_scratch
7040 (sub_shiftsi): Likewise.
7041 (sub_shiftsi_compare0
7042 (sub_shiftsi_compare0_scratch
7043 (and_scc): Likewise.
7044 (cond_move): Likewise.
7045 (if_plus_move): Likewise.
7046 (if_move_plus): Likewise.
7047 (if_move_not): Likewise.
7048 (if_not_move): Likewise.
7049 (if_shift_move): Likewise.
7050 (if_move_shift): Likewise.
7051 (if_shift_shift): Likewise.
7052 (if_not_arith): Likewise.
7053 (if_arith_not): Likewise.
7054 (cond_move_not): Likewise.
7055 (thumb1_ashlsi3): Set type attribute.
7056 (thumb1_ashrsi3): Likewise.
7057 (thumb1_lshrsi3): Likewise.
7058 (thumb1_rotrsi3): Likewise.
7059 (shiftsi3_compare0_scratch): Likewise.
7060 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7061 (neon_mov<mode>): Likewise.
7062 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7064 (thumb2_movsi_insn): Likewise.
7065 (thumb2_cmpsi_neg_shiftsi): Likewise.
7066 (thumb2_extendqisi_v6): Likewise.
7067 (thumb2_zero_extendhisi2_v6): Likewise.
7068 (thumb2_zero_extendqisi2_v6): Likewise.
7069 (thumb2_shiftsi3_short): Likewise.
7070 (thumb2_addsi3_compare0_scratch): Likewise.
7071 (orsi_not_shiftsi_si): Likewise.
7072 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7073 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7075 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7076 (1020alu_shift_op): Likewise.
7077 (1020alu_shift_reg_op): Likewise.
7078 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7079 (alu_shift_op): Likewise.
7080 (alu_shift_reg_op): Likewise.
7081 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7082 (11_alu_shift_op): Likewise.
7083 (11_alu_shift_reg_op): Likewise.
7084 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7085 (9_alu_shift_reg_op): Likewise.
7086 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7088 (cortex_a15_alu_shift): Likewise.
7089 (cortex_a15_alu_shift_reg): Likewise.
7090 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7092 (cortex_a5_alu_shift): Likewise.
7093 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7095 (cortex_a53_alu_shift): Likewise.
7096 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7098 (cortex_a7_alu_reg): Likewise.
7099 (cortex_a7_alu_shift): Likewise.
7100 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7102 (cortex_a8_alu_shift): Likewise.
7103 (cortex_a8_alu_shift_reg): Likewise.
7104 (cortex_a8_mov): Likewise.
7105 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7106 (cortex_a9_dp_shift): Likewise.
7107 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7109 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7111 (cortex_r4_mov): Likewise.
7112 (cortex_r4_alu_shift): Likewise.
7113 (cortex_r4_alu_shift_reg): Likewise.
7114 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7115 (526_alu_shift_op): Likewise.
7116 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7117 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7118 (626te_alu_shift_op): Likewise.
7119 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7120 (726te_alu_op): Likewise.
7121 (726te_alu_shift_op): Likewise.
7122 (726te_alu_shift_reg_op): Likewise.
7123 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7124 (mp626_alu_shift_op): Likewise.
7125 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7126 (pj4_alu_e1_conds): Likewise.
7127 (pj4_alu): Likewise.
7128 (pj4_alu_conds): Likewise.
7129 (pj4_shift): Likewise.
7130 (pj4_shift_conds): Likewise.
7131 (pj4_alu_shift): Likewise.
7132 (pj4_alu_shift_conds): Likewise.
7133 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7135 (cortexa7_older_only): Likewise.
7136 (cortexa7_younger): Likewise.
7138 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7140 * ipa-pure-const.c (generate_summary): Rename to...
7141 (pure_const_generate_summary): ... this.
7143 2013-07-17 Iain Sandoe <iain@codesourcery.com>
7145 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7147 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
7150 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7153 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7155 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7156 enabled without -march=zEC12.
7157 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7160 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7162 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7163 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7164 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7165 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7166 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7167 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7168 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7170 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7171 and ISA_HAS_NMADD3_NMSUB3.
7172 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7173 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7174 (nmsub4<mode>, nmsub3<mode>): Likewise.
7175 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7177 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7179 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7180 TARGET_MIPS5400 checking.
7182 2013-07-16 Jakub Jelinek <jakub@redhat.com>
7183 Peter Bergner <bergner@vnet.ibm.com>
7185 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7186 registers in the comment.
7187 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7188 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7189 rather than FIRST_PSEUDO_REGISTERS.
7191 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
7193 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7194 enable extra ISA flags with TARGET_HTM.
7196 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7198 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7201 2013-07-15 Cong Hou <congh@google.com>
7203 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7204 in compare function for sorting.
7206 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
7208 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
7209 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
7210 * config/rs6000/rs6000.opt: Add -mhtm option.
7211 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
7212 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
7213 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7214 __HTM__ if the HTM instructions are available.
7215 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7216 htm_spr_reg_operand): New define_predicates.
7217 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7218 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7220 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7221 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7222 HTM builtin functions.
7223 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7224 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7225 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7226 (rs6000_builtin_mask_calculate): Likewise.
7227 (rs6000_option_override_internal): Likewise.
7228 (bdesc_htm): Add new HTM builtin support.
7229 (htm_spr_num): New function.
7230 (htm_spr_regno): Likewise.
7231 (rs6000_htm_spr_icode): Likewise.
7232 (htm_expand_builtin): Likewise.
7233 (htm_init_builtins): Likewise.
7234 (rs6000_expand_builtin): Add support for HTM builtin functions.
7235 (rs6000_init_builtins): Likewise.
7236 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7238 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7239 (TARGET_HTM, MASK_HTM): Define macros.
7240 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7241 (FIXED_REGISTERS): Likewise.
7242 (CALL_USED_REGISTERS): Likewise.
7243 (CALL_REALLY_USED_REGISTERS): Likewise.
7244 (REG_ALLOC_ORDER): Likewise.
7245 (enum reg_class): Likewise.
7246 (REG_CLASS_NAMES): Likewise.
7247 (REG_CLASS_CONTENTS): Likewise.
7248 (REGISTER_NAMES): Likewise.
7249 (ADDITIONAL_REGISTER_NAMES): Likewise.
7250 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7251 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7252 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7253 * config/rs6000/htm.md: New file.
7254 * config/rs6000/htmintrin.h: New file.
7255 * config/rs6000/htmxlintrin.h: New file.
7257 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7259 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7260 Define SYMBOL_TINY_GOT, update comment.
7261 * config/aarch64/aarch64.c
7262 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7263 (aarch64_expand_mov_immediate): Likewise.
7264 (aarch64_print_operand): Likewise.
7265 (aarch64_classify_symbol): Likewise.
7266 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7267 (ldr_got_tiny): Define.
7269 2013-07-13 Tobias Grosser <tobias@grosser.es>
7271 PR tree-optimization/54094
7272 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7273 scheduling dimension for the parallelism check from the polyhedral
7274 information in the AST.
7275 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7278 2013-07-13 Jason Merrill <jason@redhat.com>
7280 * print-tree.c (debug_vec_tree): Use debug_raw.
7281 (debug_raw (vec<tree, va_gc> &)): New.
7282 (debug_raw (vec<tree, va_gc> *)): New.
7283 * tree.h: Declare them.
7285 2013-07-13 Bin Cheng <bin.cheng@arm.com>
7287 * ifcvt.c (ifcvt_after_combine): New static variable.
7288 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7290 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7291 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7292 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7294 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7296 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7298 2013-07-12 Michael Matz <matz@suse.de>
7301 * convert.c (convert_to_real): Reject non-float inner types.
7303 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7305 * config/aarch64/aarch64-protos.h
7306 (aarch64_simd_immediate_valid_for_move): Remove.
7307 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7308 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7310 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7312 2013-07-11 Steve Ellcey <sellcey@mips.com>
7314 * config/mips/mips.c (mips_conditional_register_usage): Do not
7315 use t[0-7] registers in MIPS16 mode when optimizing for size.
7317 2013-07-11 Sriraman Tallam <tmsriram@google.com>
7319 * config/i386/i386.c (dispatch_function_versions): Fix array
7320 indexing of function_version_info to match actual_versions.
7322 2013-07-11 Teresa Johnson <tejohnson@google.com>
7324 * vec.h (struct va_gc): Move release out-of-line.
7325 (va_gc::release): Call ggc_free on released vec.
7327 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7329 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7330 Require GOT register as additional operand in UNSPEC.
7331 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7332 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7333 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7334 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7335 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7336 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7337 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7339 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
7342 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7343 name seen by assembler/linker rather if available.
7345 2013-07-11 Andreas Schwab <schwab@suse.de>
7347 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7349 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7351 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
7353 2013-07-10 Joseph Myers <joseph@codesourcery.com>
7355 * doc/tm.texi.in: Move hook documentation to ....
7356 * target.def: ... here.
7358 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7360 * doc/tm.texi: Regenerate.
7362 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7365 * doc/invoke.texi: Document -Wconditionally-supported.
7367 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7370 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7373 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7376 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7377 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7378 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7379 Remove duplicate devices.
7380 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7381 * config/avr/t-multilib: Regenerate.
7382 * config/avr/avr-tables.opt: Regenerate.
7383 * doc/avr-mmcu.texi: Regenerate.
7385 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7388 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7390 2013-07-10 Graham Stott <graham.stott@btinternet.com>
7392 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7393 the cost of MULT when optimizing for size.
7395 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7397 * config/cr16/cr16-protos.h: Don't include target.h.
7399 2013-07-09 Joseph Myers <joseph@codesourcery.com>
7401 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7402 adjust register size for TDmode and TFmode for VSX registers.
7404 2013-07-08 Kai Tietz <ktietz@redhat.com>
7407 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7408 hook_bool_const_tree_true.
7410 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7412 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7413 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7414 * config/s390/s390.md: Define FPR*_REGNUM constants.
7415 Fix FPR2_REGNUM constant (18 -> 17).
7416 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7417 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7418 ("trunc<BFP:mode><DFP_ALL:mode>2")
7419 ("trunc<DFP_ALL:mode><BFP:mode>2")
7420 ("*extend<BFP:mode><DFP_ALL:mode>2")
7421 ("*extend<DFP_ALL:mode><BFP:mode>2")
7422 ("extend<BFP:mode><DFP_ALL:mode>2")
7423 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7426 2013-07-08 Graham Stott <graham.stott@btinternet.com>
7428 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
7430 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7432 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7433 and cfun_fpr_bit_p to cfun_fpr_save_p.
7434 (s390_frame_area, s390_register_info, s390_frame_info)
7435 (s390_emit_prologue, s390_emit_epilogue)
7436 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7438 * config/s390/s390.h: Define *_REGNUM macros for floating point
7441 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7443 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7445 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7447 PR rtl-optimization/57786
7448 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7449 and break out of the loop when it is set to false.
7451 2013-07-08 Jakub Jelinek <jakub@redhat.com>
7454 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7455 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7456 (const_int 63)) 0)).
7457 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7458 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7459 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7461 PR rtl-optimization/57829
7462 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7463 mask bits outside of mode are just sign-extension from mode to HWI.
7465 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7467 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7468 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7469 adjust_address instead of change_address to keep info about alignment.
7470 (emit_strmov): Remove.
7471 (emit_memmov): New function.
7472 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7473 (expand_movmem_epilogue): Likewise and return updated rtx for
7475 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7476 destination and source.
7477 (decide_alignment): Refactor, handle vector_loop.
7478 (ix86_expand_movmem): Likewise.
7479 (ix86_expand_setmem): Likewise.
7480 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7482 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
7484 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7485 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7487 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7489 * config/i386/sse.md (sse_movlhps): Change alternative 3
7490 of operand 2 to "m".
7492 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7495 * config/i386/sse.md (iptr): New mode attribute.
7496 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7497 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7498 (<sse>_vmmul<mode>3): Ditto.
7499 (<sse>_vmdiv<mode>3): Ditto.
7500 (sse_vmrcpv4sf2): Ditto.
7501 (<sse>_vmsqrt<mode>2): Ditto.
7502 (sse_vmrsqrtv4sf2): Ditto.
7503 (<sse>_vm<code><mode>3): Ditto.
7504 (avx_vmcmp<mode>3): Ditto.
7505 (<sse>_vmmaskcmp<mode>3): Ditto.
7506 (<sse>_comi): Ditto.
7507 (<sse>_ucomi): Ditto.
7508 (*xop_vmfrcz_<mode>): Ditto.
7509 (*fmai_fmadd_<mode>): Ditto.
7510 (*fmai_fmsub_<mode>): Ditto.
7511 (*fmai_fnmadd_<mode>): Ditto.
7512 (*fmai_fnmsub_<mode>): Ditto.
7513 (*fma4i_vmfmadd_<mode>): Ditto.
7514 (*fma4i_vmfmsub_<mode>): Ditto.
7515 (*fma4i_vmfnmadd_<mode>): Ditto.
7516 (*fma4i_vmfnmsub_<mode>): Ditto.
7517 (*xop_vmfrcz_<mode>): Ditto.
7518 (sse_cvtps2pi): Ditto.
7519 (sse_cvttps2pi): Ditto.
7520 (sse_cvtss2si): Ditto.
7521 (sse_cvtss2si_2): Ditto.
7522 (sse_cvtss2siq_2): Ditto.
7523 (sse_cvttss2si): Ditto.
7524 (sse_cvttss2siq): Ditto.
7525 (sse_cvtsd2si): Ditto.
7526 (sse_cvtsd2si_2): Ditto.
7527 (sse_cvtsd2siq_2): Ditto.
7528 (sse_cvttsd2si): Ditto.
7529 (sse_cvttsd2siq): Ditto.
7530 (sse_cvtsd2ss): Ditto.
7531 (sse_cvtss2sd): Ditto.
7532 (avx2_pbroadcast<mode>): Ditto.
7533 (avx2_pbroadcast<mode>_1): Ditto.
7534 (*avx_vperm_broadcast_v4sf): Ditto.
7536 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7537 (sse_movlhps): Ditto.
7538 (sse_storehps): Ditto.
7539 (sse_loadhps): Ditto.
7540 (sse_storelps): Ditto.
7541 (sse_loadlps): Ditto.
7542 (*vec_concatv4sf): Ditto.
7543 (*vec_interleave_highv2df): Ditto.
7544 (*vec_interleave_lowv2df): Ditto.
7545 (*vec_extractv2df_1_sse): Ditto.
7546 (*vec_extractv2df_0_sse): Ditto.
7547 (sse2_storelpd): Ditto.
7548 (sse2_loadlpd): Ditto.
7549 (sse2_movsd): Ditto.
7550 (*vec_concatv4si): Ditto.
7551 (vec_concatv2di): Ditto.
7553 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7554 for Intel asm dialect.
7555 (mmx_punpcklwd): Ditto.
7556 (mmx_punpckldq): Ditto.
7558 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7559 for intel assembler dialect.
7561 2013-07-06 Jakub Jelinek <jakub@redhat.com>
7564 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7565 for BUILT_IN_C{LZ,LRSB}*.
7566 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7567 * tree-vrp.c (extract_range_basic): Handle
7568 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7569 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7570 fall thru to code calling set_value*.
7571 * builtins.c (expand_builtin): Remove *IMAX cases.
7572 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7573 if width is bigger than 2*HWI.
7575 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7577 PR rtl-optimization/55342
7578 * lra-int.h (lra_subreg_reload_pseudos): New.
7579 * lra.c: Add undoing optional reloads to the block diagram.
7580 (lra_subreg_reload_pseudos): New.
7581 (lra_optional_reload_pseudos): Change comments.
7582 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7583 lra_optional_reload_pseudos after undo transformations.
7584 * lra-assigns.c (pseudo_prefix_title): New.
7585 (lra_setup_reg_renumber): Use it.
7586 (spill_for): Ditto. Check subreg reload pseudos too.
7587 (assign_by_spills): Consider subreg reload pseudos too.
7588 * lra-constraints.c (simplify_operand_subreg): Use
7589 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7590 (curr_insn_transform): Recognize and do optional reloads.
7591 (undo_optional_reloads): New.
7592 (lra_undo_inheritance): Call undo_optional_reloads.
7594 2013-07-05 Thomas Quinot <quinot@adacore.com>
7596 * tree-complex.c (expand_complex_operations_1): Fix typo.
7598 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7600 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7601 (tune_params): New member 'const vec_costs'.
7602 * config/aarch64/aarch64.c (generic_vector_cost): New.
7603 (generic_tunings): New member 'generic_vector_cost'.
7604 (aarch64_builtin_vectorization_cost): New.
7605 (aarch64_add_stmt_cost): New.
7606 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7607 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7609 2013-07-03 Jakub Jelinek <jakub@redhat.com>
7612 * config/i386/predicates.md (vsib_address_operand): Disallow
7613 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7615 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7618 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7619 expand_builtin_setjmp_receiver.
7620 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7621 with NULL for the label parameter.
7622 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7623 the frame-pointer. Adjust comments.
7624 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7625 only if LABEL is non-NULL.
7627 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7629 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7630 (ARM_ABI_AAPCS64): Ditto.
7632 (ARM_DEFAULT_ABI): Ditto.
7634 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7636 * config/aarch64/aarch64-builtins.c
7637 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7638 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7640 * config/aarch64/aarch64-simd.md
7641 (aarch64_ld1<VALL:mode>): New.
7642 (aarch64_st1<VALL:mode>): Likewise.
7643 * config/aarch64/arm_neon.h
7644 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7646 2013-07-02 Sriraman Tallam <tmsriram@google.com>
7648 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7650 (ix86_option_override_internal):Turn on all -mavx target flags by
7651 default as they are dependent on AVX anyway.
7653 2013-07-02 Cary Coutant <ccoutant@google.com>
7655 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7657 (loc_checksum_ordered): Likewise.
7658 (hash_loc_operands): Remove inline keyword.
7660 2013-07-02 Jakub Jelinek <jakub@redhat.com>
7662 PR tree-optimization/57741
7663 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7664 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7665 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7666 Allow REAL_CST step_exprs if flag_associative_math.
7667 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7669 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7671 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
7673 2013-07-02 Ian Bolton <ian.bolton@arm.com>
7675 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7677 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7679 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7681 (iorsi3_insn): Likewise.
7682 (arm_xorsi3): Likewise.
7684 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7686 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7688 (attribute "type"): ... here, and prefix with "wmmx_".
7689 (attribute "core_cycles"): Update for attribute changes.
7690 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7691 (tbcstv4hi): Likewise.
7692 (tbcstv2si): Likewise.
7693 (iwmmxt_iordi3): Likewise.
7694 (iwmmxt_xordi3): Likewise.
7695 (iwmmxt_anddi3): Likewise.
7696 (iwmmxt_nanddi3): Likewise.
7697 (iwmmxt_arm_movdi): Likewise.
7698 (iwmmxt_movsi_insn): Likewise.
7699 (mov<mode>_internal): Likewise.
7700 (and<mode>3_iwmmxt): Likewise.
7701 (ior<mode>3_iwmmxt): Likewise.
7702 (xor<mode>3_iwmmxt): Likewise.
7703 (add<mode>3_iwmmxt): Likewise.
7704 (ssaddv8qi3): Likewise.
7705 (ssaddv4hi3): Likewise.
7706 (ssaddv2si3): Likewise.
7707 (usaddv8qi3): Likewise.
7708 (usaddv4hi3): Likewise.
7709 (usaddv2si3): Likewise.
7710 (sub<mode>3_iwmmxt): Likewise.
7711 (sssubv8qi3): Likewise.
7712 (sssubv4hi3): Likewise.
7713 (sssubv2si3): Likewise.
7714 (ussubv8qi3): Likewise.
7715 (ussubv4hi3): Likewise.
7716 (ussubv2si3): Likewise.
7717 (mulv4hi3_iwmmxt): Likewise.
7718 (smulv4hi3_highpart): Likewise.
7719 (umulv4hi3_highpart): Likewise.
7720 (iwmmxt_wmacs): Likewise.
7721 (iwmmxt_wmacsz): Likewise.
7722 (iwmmxt_wmacu): Likewise.
7723 (iwmmxt_wmacuz): Likewise.
7724 (iwmmxt_clrdi): Likewise.
7725 (iwmmxt_clrv8qi): Likewise.
7726 (iwmmxt_clr4hi): Likewise.
7727 (iwmmxt_clr2si): Likewise.
7728 (iwmmxt_uavgrndv8qi3): Likewise.
7729 (iwmmxt_uavgrndv4hi3): Likewise.
7730 (iwmmxt_uavgv8qi3): Likewise.
7731 (iwmmxt_uavgv4hi3): Likewise.
7732 (iwmmxt_tinsrb): Likewise.
7733 (iwmmxt_tinsrh): Likewise.
7734 (iwmmxt_tinsrw): Likewise.
7735 (iwmmxt_textrmub): Likewise.
7736 (iwmmxt_textrmsb): Likewise.
7737 (iwmmxt_textrmuh): Likewise.
7738 (iwmmxt_textrmsh): Likewise.
7739 (iwmmxt_textrmw): Likewise.
7740 (iwmxxt_wshufh): Likewise.
7741 (eqv8qi3): Likewise.
7742 (eqv4hi3): Likewise.
7743 (eqv2si3): Likewise.
7744 (gtuv8qi3): Likewise.
7745 (gtuv4hi3): Likewise.
7746 (gtuv2si3): Likewise.
7747 (gtv8qi3): Likewise.
7748 (gtv4hi3): Likewise.
7749 (gtv2si3): Likewise.
7750 (smax<mode>3_iwmmxt): Likewise.
7751 (umax<mode>3_iwmmxt): Likewise.
7752 (smin<mode>3_iwmmxt): Likewise.
7753 (umin<mode>3_iwmmxt): Likewise.
7754 (iwmmxt_wpackhss): Likewise.
7755 (iwmmxt_wpackwss): Likewise.
7756 (iwmmxt_wpackdss): Likewise.
7757 (iwmmxt_wpackhus): Likewise.
7758 (iwmmxt_wpackwus): Likewise.
7759 (iwmmxt_wpackdus): Likewise.
7760 (iwmmxt_wunpckihb): Likewise.
7761 (iwmmxt_wunpckihh): Likewise.
7762 (iwmmxt_wunpckihw): Likewise.
7763 (iwmmxt_wunpckilb): Likewise.
7764 (iwmmxt_wunpckilh): Likewise.
7765 (iwmmxt_wunpckilw): Likewise.
7766 (iwmmxt_wunpckehub): Likewise.
7767 (iwmmxt_wunpckehuh): Likewise.
7768 (iwmmxt_wunpckehuw): Likewise.
7769 (iwmmxt_wunpckehsb): Likewise.
7770 (iwmmxt_wunpckehsh): Likewise.
7771 (iwmmxt_wunpckehsw): Likewise.
7772 (iwmmxt_wunpckelub): Likewise.
7773 (iwmmxt_wunpckeluh): Likewise.
7774 (iwmmxt_wunpckeluw): Likewise.
7775 (iwmmxt_wunpckelsb): Likewise.
7776 (iwmmxt_wunpckelsh): Likewise.
7777 (iwmmxt_wunpckelsw): Likewise.
7778 (ror<mode>3): Likewise.
7779 (ashr<mode>3_iwmmxt): Likewise.
7780 (lshr<mode>3_iwmmxt): Likewise.
7781 (ashl<mode>3_iwmmxt): Likewise.
7782 (ror<mode>3_di): Likewise.
7783 (ashr<mode>3_di): Likewise.
7784 (lshr<mode>3_di): Likewise.
7785 (ashl<mode>3_di): Likewise.
7786 (iwmmxt_wmadds): Likewise.
7787 (iwmmxt_wmaddu): Likewise.
7788 (iwmmxt_tmia): Likewise.
7789 (iwmmxt_tmiaph): Likewise.
7790 (iwmmxt_tmiabb): Likewise.
7791 (iwmmxt_tmiatb): Likewise.
7792 (iwmmxt_tmiabt): Likewise.
7793 (iwmmxt_tmiatt): Likewise.
7794 (iwmmxt_tmovmskb): Likewise.
7795 (iwmmxt_tmovmskh): Likewise.
7796 (iwmmxt_tmovmskw): Likewise.
7797 (iwmmxt_waccb): Likewise.
7798 (iwmmxt_wacch): Likewise.
7799 (iwmmxt_waccw): Likewise.
7800 (iwmmxt_waligni): Likewise.
7801 (iwmmxt_walignr): Likewise.
7802 (iwmmxt_walignr0): Likewise.
7803 (iwmmxt_walignr1): Likewise.
7804 (iwmmxt_walignr2): Likewise.
7805 (iwmmxt_walignr3): Likewise.
7806 (iwmmxt_wsadb): Likewise.
7807 (iwmmxt_wsadh): Likewise.
7808 (iwmmxt_wsadbz): Likewise.
7809 (iwmmxt_wsadhz): Likewise.
7810 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
7811 (iwmmxt_wabsdiffb): Likewise.
7812 (iwmmxt_wabsdiffh): Likewise.
7813 (iwmmxt_wabsdiffw): Likewise.
7814 (iwmmxt_waddsubhx): Likewise
7815 (iwmmxt_wsubaddhx): Likewise.
7816 (addc<mode>3): Likewise.
7817 (iwmmxt_avg4): Likewise.
7818 (iwmmxt_avg4r): Likewise.
7819 (iwmmxt_wmaddsx): Likewise.
7820 (iwmmxt_wmaddux): Likewise.
7821 (iwmmxt_wmaddsn): Likewise.
7822 (iwmmxt_wmaddun): Likewise.
7823 (iwmmxt_wmulwsm): Likewise.
7824 (iwmmxt_wmulwum): Likewise.
7825 (iwmmxt_wmulsmr): Likewise.
7826 (iwmmxt_wmulumr): Likewise.
7827 (iwmmxt_wmulwsmr): Likewise.
7828 (iwmmxt_wmulwumr): Likewise.
7829 (iwmmxt_wmulwl): Likewise.
7830 (iwmmxt_wqmulm): Likewise.
7831 (iwmmxt_wqmulwm): Likewise.
7832 (iwmmxt_wqmulmr): Likewise.
7833 (iwmmxt_wqmulwmr): Likewise.
7834 (iwmmxt_waddbhusm): Likewise.
7835 (iwmmxt_waddbhusl): Likewise.
7836 (iwmmxt_wqmiabb): Likewise.
7837 (iwmmxt_wqmiabt): Likewise.
7838 (iwmmxt_wqmiatb): Likewise.
7839 (iwmmxt_wqmiatt): Likewise.
7840 (iwmmxt_wqmiabbn): Likewise.
7841 (iwmmxt_wqmiabtn): Likewise.
7842 (iwmmxt_wqmiatbn): Likewise.
7843 (iwmmxt_wqmiattn): Likewise.
7844 (iwmmxt_wmiabb): Likewise.
7845 (iwmmxt_wmiabt): Likewise.
7846 (iwmmxt_wmiatb): Likewise.
7847 (iwmmxt_wmiatt): Likewise.
7848 (iwmmxt_wmiabbn): Likewise.
7849 (iwmmxt_wmiabtn): Likewise.
7850 (iwmmxt_wmiatbn): Likewise.
7851 (iwmmxt_wmiattn): Likewise.
7852 (iwmmxt_wmiawbb): Likewise.
7853 (iwmmxt_wmiawbt): Likewise.
7854 (iwmmxt_wmiawtb): Likewise.
7855 (iwmmxt_wmiawtt): Likewise.
7856 (iwmmxt_wmiawbbn): Likewise.
7857 (iwmmxt_wmiawbtn): Likewise.
7858 (iwmmxt_wmiawtbn): Likewise.
7859 (iwmmxt_wmiawttn): Likewise.
7860 (iwmmxt_wmerge): Likewise.
7861 (iwmmxt_tandc<mode>3): Likewise.
7862 (iwmmxt_torc<mode>3): Likewise.
7863 (iwmmxt_torvsc<mode>3): Likewise.
7864 (iwmmxt_textrc<mode>3): Likewise.
7865 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
7866 (wmmxt_pack): Likewise.
7867 (wmmxt_mult_c1): Likewise.
7868 (wmmxt_mult_c2): Likewise.
7869 (wmmxt_alu_c1): Likewise.
7870 (wmmxt_alu_c2): Likewise.
7871 (wmmxt_alu_c3): Likewise.
7872 (wmmxt_transfer_c1): Likewise.
7873 (wmmxt_transfer_c2): Likewise.
7874 (wmmxt_transfer_c3): Likewise.
7875 (marvell_f_iwmmxt_wstr): Likewise.
7876 (marvell_f_iwmmxt_wldr): Likewise.
7878 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
7880 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
7882 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7885 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7886 * lra-constraints.c (need_for_split_p): Check call used hard regs
7887 living through calls.
7889 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
7890 call used regs for call insn.
7892 2013-06-28 Jakub Jelinek <jakub@redhat.com>
7895 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
7896 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
7898 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7900 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
7902 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
7904 * lra-constraints.c (need_for_split_p): Check call used hard regs
7905 living through calls.
7907 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
7910 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
7911 to tie with any other modes. Eliminate Altivec vector mode tests,
7912 since these are a subset of ALTIVEC or VSX vector modes. Simplify
7913 code, to return 0 if testing MODE2 for a condition, if we've
7914 already tested MODE1 for the same condition.
7916 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7918 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
7921 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7923 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7924 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
7926 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
7928 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
7930 (aarch64_symbolic_constant_p): Remove.
7931 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
7932 static. Fix line length and white space.
7933 (aarch64_symbolic_constant_p): Remove.
7934 * config/aarch64/predicates.md (aarch64_valid_symref):
7935 Use aarch64_classify_symbol_expression.
7937 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7939 * config/arm/constraints.md (Ts): New constraint.
7940 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
7942 (compare_scc): Use "Ts" constraint for operand 0.
7943 (ior_scc_scc): Likewise.
7944 (and_scc_scc): Likewise.
7945 (and_scc_scc_nodom): Likewise.
7946 (ior_scc_scc_cmp): Likewise for operand 7.
7947 (and_scc_scc_cmp): Likewise.
7948 * config/arm/thumb2.md (thumb2_movsi_insn):
7949 Add alternatives for 16-bit encodings.
7950 (thumb2_movhi_insn): Likewise.
7951 (thumb2_movsicc_insn): Likewise.
7952 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
7953 (thumb2_negscc): Use "Ts" constraint.
7954 Move mvn instruction outside cond_exec block.
7955 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
7956 for 16-bit encodings.
7958 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7960 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
7962 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
7963 (mulsi3subsi): Likewise.
7964 (mulsidi3adddi): Likewise.
7965 (mulsidi3_v6): Likewise.
7966 (umulsidi3_v6): Likewise.
7967 (umulsidi3adddi_v6): Likewise.
7968 (smulsi3_highpart_v6): Likewise.
7969 (umulsi3_highpart_v6): Likewise.
7970 (mulhisi3tb): Likewise.
7971 (mulhisi3bt): Likewise.
7972 (mulhisi3tt): Likewise.
7973 (maddhisi4): Likewise.
7974 (maddhisi4tb): Likewise.
7975 (maddhisi4tt): Likewise.
7976 (maddhidi4): Likewise.
7977 (maddhidi4tb): Likewise.
7978 (maddhidi4tt): Likewise.
7979 (zeroextractsi_compare0_scratch): Likewise.
7980 (insv_zero): Likewise.
7981 (insv_t2): Likewise.
7982 (anddi_notzesidi_di): Likewise.
7983 (anddi_notsesidi_di): Likewise.
7984 (andsi_notsi_si): Likewise.
7985 (iordi_zesidi_di): Likewise.
7986 (xordi_zesidi_di): Likewise.
7987 (andsi_iorsi3_notsi): Likewise.
7989 (smax_m1): Likewise.
7991 (not_shiftsi): Likewise.
7992 (unaligned_loadsi): Likewise.
7993 (unaligned_loadhis): Likewise.
7994 (unaligned_loadhiu): Likewise.
7995 (unaligned_storesi): Likewise.
7996 (unaligned_storehi): Likewise.
7997 (extv_reg): Likewise.
7998 (extzv_t2): Likewise.
8000 (udivsi3): Likewise.
8001 (arm_zero_extendhisi2addsi): Likewise.
8002 (arm_zero_extendqisi2addsi): Likewise.
8003 (compareqi_eq0): Likewise.
8004 (arm_extendhisi2_v6): Likewise.
8005 (arm_extendqisi2addsi): Likewise.
8006 (arm_movt): Likewise.
8007 (thumb2_ldrd): Likewise.
8008 (thumb2_ldrd_base): Likewise.
8009 (thumb2_ldrd_base_neg): Likewise.
8010 (thumb2_strd): Likewise.
8011 (thumb2_strd_base): Likewise.
8012 (thumb2_strd_base_neg): Likewise.
8013 (arm_negsi2): Add alternative for 16-bit encoding.
8014 (arm_one_cmplsi2): Likewise.
8016 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8018 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8019 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8020 (movdfcc): Likewise.
8021 * config/arm/vfp.md (*thumb2_movsf_vfp):
8022 Disable predication for arm_restrict_it.
8023 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8024 (*thumb2_movdfcc_vfp): Likewise.
8025 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8026 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8027 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8028 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8029 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8030 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8031 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8032 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8033 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8034 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8035 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8036 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8037 Disable predication for arm_restrict_it.
8039 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8041 * config/i386/bmiintrin.h (_bextr_u32): New.
8042 (_bextr_u64): Ditto.
8044 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8046 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8047 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8048 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8049 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8050 * config/mips/n32-elf.h: ...this new file.
8052 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
8055 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8056 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8058 2013-06-27 Catherine Moore <clm@codesourcery.com>
8060 * config/mips/mips-tables.opt: Regenerate.
8061 * config/mips/mips-cpus.def: Add m14ke and m14kec.
8062 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8063 * doc/invoke.texi: Add -m14kc.
8065 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8068 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8069 constraints of operand 1 and 2.
8072 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8073 to match RTL canonicalization. Swap predicates and
8074 constraints of operand 1 and 2.
8076 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8078 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8079 Process OP_INOUT regs for splitting too.
8081 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8083 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8084 decl before the loop, initialize to NULL.
8085 (vectorizable_load): Initialize ptr_incr to NULL.
8087 2013-06-27 Martin Jambor <mjambor@suse.cz>
8090 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8091 * ipa-ref.c (ipa_maybe_record_reference): New function.
8092 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8093 * ipa-cp.c (create_specialized_node): Record potential references from
8095 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8097 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8099 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8100 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8101 'mode' to force_reg.
8102 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8103 (aarch64_expand_mov_immediate): Likewise.
8105 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8107 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8108 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8110 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8112 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8113 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8114 (struct machine_function): Add tbegin_p.
8115 (s390_canonicalize_comparison): Fold CC mode compares to
8116 conditional jump if possible.
8117 (s390_emit_jump): Return the emitted jump.
8118 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8119 Handle CCRAWmode compares.
8120 (s390_option_override): Default to -mhtm if available.
8121 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8122 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8123 FPRs instead of df_regs_ever_live_p.
8124 (s390_optimize_nonescaping_tx): New function.
8125 (s390_init_frame_layout): Extend clobbered_regs array to cover
8127 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8128 (s390_expand_tbegin): New function.
8129 (enum s390_builtin): New enum definition.
8130 (code_for_builtin): New array definition.
8131 (s390_init_builtins): New function.
8132 (s390_expand_builtin): New function.
8133 (TARGET_INIT_BUILTINS): Define.
8134 (TARGET_EXPAND_BUILTIN): Define.
8135 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
8136 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8137 (s390_alc_comparison): Likewise.
8138 * config/s390/s390-modes.def: Add CCRAWmode.
8139 * config/s390/s390.h (processor_flags): Add PF_TX.
8140 (TARGET_CPU_HTM): Define macro.
8141 (TARGET_HTM): Define macro.
8142 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8143 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8144 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8145 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
8146 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
8147 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8148 ("*cc_to_int"): Move up.
8149 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8150 constants other than 0.
8151 ("*ccraw_to_int"): New insn and splitter definition.
8152 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8153 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8154 ("tx_assist"): New expander.
8155 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8156 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8157 * config/s390/s390.opt: Add -mhtm option.
8158 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8159 * config/s390/htmxlintrin.h: New file.
8160 * config/s390/htmintrin.h: New file.
8161 * config/s390/s390intrin.h: New file.
8162 * doc/extend.texi: Document htm builtins.
8163 * config.gcc: Add the new header files to extra_headers.
8165 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8167 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
8168 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
8170 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
8171 Pat Haugen <pthaugen@us.ibm.com>
8172 Peter Bergner <bergner@vnet.ibm.com>
8174 * config/rs6000/power8.md: New.
8175 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8176 setting for power8 entry.
8177 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8178 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8179 test for Power4/Power5 only.
8180 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8182 (force_new_group): Adjust comment.
8183 * config/rs6000/rs6000.md: Include power8.md.
8185 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8187 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8188 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8190 (tune_params): Update comment.
8191 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8192 (arm_max_conditional_execute): New function.
8193 (thumb2_final_prescan_insn): Use max_insn_skipped and
8194 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8196 2013-06-25 Jakub Jelinek <jakub@redhat.com>
8198 PR tree-optimization/57705
8199 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8200 SSA_NAME step, provided that it is not defined inside the loop.
8201 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
8202 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8204 2013-06-25 Martin Jambor <mjambor@suse.cz>
8207 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
8208 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
8210 (ipa_note_param_call): Initialize member_ptr flag.
8211 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
8212 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
8213 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8214 (ipa_read_indirect_edge_info): Likewise.
8216 2013-06-25 Richard Biener <rguenther@suse.de>
8219 * passes.c (init_optimization_passes): Move pass_fold_builtins
8220 and pass_dce earlier with -Og.
8222 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8224 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8225 <BIT_FIELD_REF>: Remove trailing TAB.
8226 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8229 2013-06-24 Martin Jambor <mjambor@suse.cz>
8231 PR tree-optimization/57358
8232 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8233 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8234 (ipa_analyze_params_uses): Generate pessimistic info when true.
8236 2013-06-24 Martin Jambor <mjambor@suse.cz>
8238 PR tree-optimization/57539
8239 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8240 global.inlined_to of the new node to it. All callers changed.
8241 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8242 inlining_into, pass it to cgraph_clone_node.
8243 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8244 ipa_free_edge_args_substructures.
8245 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8246 rdesc linked list. Do not assert rdesc edges have inlined caller.
8247 Assert we have found an rdesc in the rdesc list.
8249 2013-06-24 Richard Biener <rguenther@suse.de>
8251 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
8252 (pointer_set_lookup): Declare.
8253 (class pointer_map): New template class implementing a
8254 generic pointer to T map.
8255 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8256 pointer_map<T>::contains, pointer_map<T>::insert,
8257 pointer_map<T>::traverse): New functions.
8258 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8259 (pointer_set_lookup): New function.
8260 (pointer_set_contains): Use pointer_set_lookup.
8261 (pointer_set_insert): Likewise.
8262 (insert_aux): Remove.
8263 (struct pointer_map_t): Embed a pointer_set_t.
8264 (pointer_map_create): Adjust.
8265 (pointer_map_destroy): Likewise.
8266 (pointer_map_contains): Likewise.
8267 (pointer_map_insert): Likewise.
8268 (pointer_map_traverse): Likewise.
8269 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8270 pointer_map<unsigned> instead of a pointer_map_t.
8271 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8272 (streamer_tree_cache_lookup): Likewise.
8273 (streamer_tree_cache_create): Likewise.
8274 (streamer_tree_cache_delete): Likewise.
8275 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8276 pointer_map<unsigned> instead of a pointer_map_t.
8277 (lto_init_tree_ref_encoder): Adjust.
8278 (lto_destroy_tree_ref_encoder): Likewise.
8279 * lto-section-out.c (lto_output_decl_index): Likewise.
8280 (lto_record_function_out_decl_state): Likewise.
8281 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8283 2013-06-24 Richard Biener <rguenther@suse.de>
8285 PR tree-optimization/57488
8286 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8288 2013-06-24 Alan Modra <amodra@gmail.com>
8290 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8291 (gen_easy_altivec_constant): Likewise.
8292 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8293 easy_vector_constant_msb): Likewise.
8295 2013-06-23 Jakub Jelinek <jakub@redhat.com>
8298 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8299 add missing return true.
8301 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
8304 * config/sh/predicates.md (general_extend_operand): Invoke
8305 general_movsrc_operand for memory operands.
8306 (general_movsrc_operand): Allow reg+reg addressing, do not use
8307 general_operand for memory operands.
8309 2013-06-23 Sriraman Tallam <tmsriram@google.com>
8311 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8312 when current target options does not apply.
8313 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8314 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
8315 * config/i386/bmiintrin.h: Pass appropriate target
8316 attributes to header.
8317 * config/i386/mmintrin.h: Ditto.
8318 * config/i386/nmmintrin.h: Ditto.
8319 * config/i386/avx2intrin.h: Ditto.
8320 * config/i386/fxsrintrin.h: Ditto.
8321 * config/i386/tbmintrin.h: Ditto.
8322 * config/i386/xsaveintrin.h: Ditto.
8323 * config/i386/f16cintrin.h: Ditto.
8324 * config/i386/xtestintrin.h: Ditto.
8325 * config/i386/xsaveoptintrin.h: Ditto.
8326 * config/i386/bmi2intrin.h: Ditto.
8327 * config/i386/lzcntintrin.h: Ditto.
8328 * config/i386/smmintrin.h: Ditto.
8329 * config/i386/wmmintrin.h: Ditto.
8330 * config/i386/x86intrin.h: Remove all header include guards.
8331 * config/i386/prfchwintrin.h: Ditto.
8332 * config/i386/pmmintrin.h: Ditto.
8333 * config/i386/tmmintrin.h: Ditto.
8334 * config/i386/xmmintrin.h: Ditto.
8335 * config/i386/popcntintrin.h: Ditto.
8336 * config/i386/rdseedintrin.h: Ditto.
8337 * config/i386/ammintrin.h: Ditto.
8338 * config/i386/emmintrin.h: Ditto.
8339 * config/i386/immintrin.h: Remove all header include guards.
8340 * config/i386/fma4intrin.h: Ditto.
8341 * config/i386/lwpintrin.h: Ditto.
8342 * config/i386/xopintrin.h: Ditto.
8343 * config/i386/ia32intrin.h: Ditto.
8344 * config/i386/avxintrin.h: Ditto.
8345 * config/i386/rtmintrin.h: Ditto.
8346 * config/i386/fmaintrin.h: Ditto.
8347 * config/i386/mm3dnow.h: Ditto.
8349 2013-06-22 Sriraman Tallam <tmsriram@google.com>
8351 * common/config/i386/i386-common.c: Handle LZCNT.
8353 2013-06-22 Andi Kleen <ak@linux.intel.com>
8355 * doc/extend.texi: Use __atomic_store_n instead of
8356 __atomic_store in HLE example.
8358 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
8360 * config/sh/sh.c: Remove <cstdlib> workaround.
8362 2013-06-21 Andi Kleen <ak@linux.intel.com>
8364 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
8366 2013-06-21 Andi Kleen <ak@linux.intel.com>
8368 * doc/extend.texi: Document that __atomic_clear and
8369 __atomic_test_and_set should only be used with bool.
8371 2013-06-20 Jan Hubicka <jh@suse.cz>
8373 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8375 * tree.c (decls_same_for_odr): New function.
8376 (same_for_edr): New function.
8377 (types_same_for_odr): New function.
8378 (get_binfo_at_offset): Use it.
8379 * tree.h (types_same_for_odr): Declare.
8381 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
8382 Jason Merrill <jason@redhat.com>
8384 * system.h: Include <cstdlib> as well as <stdlib.h>.
8386 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
8389 * config/i386/i386.c (construct_container): Report error if
8390 long double is used with disabled x87 float returns.
8392 2013-06-20 Jan Hubicka <jh@suse.cz>
8394 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8396 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
8398 PR rtl-optimization/57425
8399 PR rtl-optimization/57569
8400 * alias.c (write_dependence_p): Remove parameters mem_mode and
8401 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8402 Changed all callers.
8403 (canon_anti_dependence): Get comments and semantics in sync.
8404 Add parameter mem_canonicalized. Changed all callers.
8405 * rtl.h (canon_anti_dependence): Update prototype.
8407 2013-06-20 Richard Biener <rguenther@suse.de>
8409 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8410 byte case, inline streamer_read_uchar and defer section
8413 2013-06-20 Richard Biener <rguenther@suse.de>
8415 PR tree-optimization/57584
8416 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8417 SSA names into the expanded expression that take part in
8418 abnormal coalescing.
8420 2013-06-19 Sharad Singhai <singhai@google.com>
8422 * gcov.c (print_usage): Handle new option.
8423 (process_args): Ditto.
8424 (get_gcov_intermediate_filename): New function.
8425 (output_intermediate_file): New function.
8426 (output_gcov_file): New function
8427 (generate_results): Handle new option.
8428 (release_function): Relase demangled name.
8429 (read_graph_file): Handle demangled name.
8430 (output_lines): Ditto.
8431 * doc/gcov.texi: Document gcov intermediate format.
8433 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8436 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8437 (lra_emit_add): Use the functions. Add comment about Y as an
8440 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
8443 * collect2.c (collect_atexit): New.
8444 (collect_exit): Delete.
8445 (main): Register collect_atexit with atexit.
8446 (collect_wait): Change collect_exit to exit.
8448 * collect2.h (collect_exit): Delete.
8449 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8451 2013-06-19 Wei Mi <wmi@google.com>
8453 PR rtl-optimization/57518
8454 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8455 if regno is used in paradoxical subreg.
8456 (update_equiv_regs): Check pdx_subregs[regno] before
8457 set a reg to be equivalent with a mem.
8459 2013-06-19 Matthias Klose <doko@ubuntu.com>
8462 * file-find.h (find_a_file): Add a mode parameter.
8463 * file-find.c (find_a_file): Likewise.
8464 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8465 with X_OK for the executables.
8466 * collect2.c (main): Call find_a_file with X_OK.
8468 2013-06-19 Steve Ellcey <sellcey@mips.com>
8471 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8472 Use NEXT_INSN instead of next_real_insn.
8474 2013-06-19 Jan Hubicka <jh@suse.cz>
8476 * cgraph.h (const_value_known_p): Replace by ...
8477 (ctor_for_folding): .. this one.
8478 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8479 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8480 * expr.c (expand_expr_real_1): Likewise.
8481 (string_constant): Likewise.
8482 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8483 * ipa.c (process_references): Likewise.
8484 (symtab_remove_unreachable_nodes): Likewise.
8485 * ipa-inline-analysis.c (param_change_prob): Likewise.
8486 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8487 (get_base_constructor): Likwise.
8488 * varpool.c (varpool_remove_node): Likewise.
8489 (varpool_remove_initializer): LIkewise.
8490 (dump_varpool_node): LIkwise.
8491 (const_value_known_p): Rewrite to ...
8492 (ctor_for_folding): ... this one.
8494 2013-06-19 Jakub Jelinek <jakub@redhat.com>
8497 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8498 PERSONALITY in $PATH derived prefixes.
8500 2013-06-19 Jeff Law <law@redhat.com>
8502 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8505 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8506 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8509 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8511 * config/arm/vfp.md: Move VFP instruction classification documentation
8513 * config/arm/arm.md: ... here. Update instruction classification
8516 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
8518 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8519 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8520 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8522 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8525 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8526 with NEXT_INSN. Use tablejump_p to check for jump table data
8529 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8532 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8533 that now in C++ the value is correct per the C++ standards.
8535 2013-06-19 Richard Biener <rguenther@suse.de>
8537 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8540 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8543 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8546 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8547 with next_active_insn.
8549 2013-06-18 Sriraman Tallam <tmsriram@google.com>
8551 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8552 functions are inlined during failures to flag an error.
8553 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8554 in early inline pass.
8556 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8558 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8561 2013-06-18 Julian Brown <julian@codesourcery.com>
8563 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8564 Permit virtual register pre-reload if !strict.
8565 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8567 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8569 * config/arm/neon.md (movmisalign<mode>): Use
8570 neon_perm_struct_or_reg_operand instead of
8571 neon_struct_or_register_operand.
8572 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8573 neon_permissive_struct_operand instead of neon_struct_operand.
8574 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8575 neon_vector_mem_operand.
8576 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8577 neon_vector_mem_operand.
8578 (neon_permissive_struct_operand): New.
8579 (neon_struct_or_register_operand): Rename to...
8580 (neon_perm_struct_or_reg_operand): This. Adjust call to
8581 neon_vector_mem_operand.
8583 2013-06-18 Richard Biener <rguenther@suse.de>
8585 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8586 * lto-streamer.h: Include pointer-set.h.
8587 (struct lto_decl_slot): Remove.
8588 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8589 Remove next_index entry.
8590 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8591 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8592 (lto_init_tree_ref_encoder): Adjust.
8593 (lto_destroy_tree_ref_encoder): Likewise.
8594 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8595 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8596 (lto_output_decl_index): Adjust.
8597 (lto_new_out_decl_state): Likewise.
8598 (lto_record_function_out_decl_state): Likewise.
8599 * lto-streamer-out.c (copy_function): Likewise.
8601 2013-06-18 Richard Biener <rguenther@suse.de>
8603 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8604 * cgraphunit.c: Include cfgloop.h.
8605 (init_lowered_empty_function): Initialize the loop tree.
8606 (assemble_thunk): Insert new BBs into loops.
8608 2013-06-18 Richard Biener <rguenther@suse.de>
8610 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8611 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8612 the map from cache entry to cache index optional.
8613 (streamer_tree_cache_replace_tree): Adjust accordingly.
8614 (streamer_tree_cache_append): Likewise.
8615 (streamer_tree_cache_delete): Likewise.
8616 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8617 streamer cache map from cache entry to cache index.
8618 * lto-streamer-out.c (create_output_block): Adjust.
8620 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8622 * config/arm/arm.md (attribute "insn"): Move multiplication and
8623 division attributes to...
8624 (attribute "type"): ... here. Remove mult.
8625 (attribute "mul32"): New attribute.
8626 (attribute "mul64"): Add umaal.
8627 (*arm_mulsi3): Update attributes.
8628 (*arm_mulsi3_v6): Likewise.
8629 (*thumb_mulsi3): Likewise.
8630 (*thumb_mulsi3_v6): Likewise.
8631 (*mulsi3_compare0): Likewise.
8632 (*mulsi3_compare0_v6): Likewise.
8633 (*mulsi_compare0_scratch): Likewise.
8634 (*mulsi_compare0_scratch_v6): Likewise.
8635 (*mulsi3addsi): Likewise.
8636 (*mulsi3addsi_v6): Likewise.
8637 (*mulsi3addsi_compare0): Likewise.
8638 (*mulsi3addsi_compare0_v6): Likewise.
8639 (*mulsi3addsi_compare0_scratch): Likewise.
8640 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8641 (*mulsi3subsi): Likewise.
8642 (*mulsidi3adddi): Likewise.
8643 (*mulsi3addsi_v6): Likewise.
8644 (*mulsidi3adddi_v6): Likewise.
8645 (*mulsidi3_nov6): Likewise.
8646 (*mulsidi3_v6): Likewise.
8647 (*umulsidi3_nov6): Likewise.
8648 (*umulsidi3_v6): Likewise.
8649 (*umulsidi3adddi): Likewise.
8650 (*umulsidi3adddi_v6): Likewise.
8651 (*smulsi3_highpart_nov6): Likewise.
8652 (*smulsi3_highpart_v6): Likewise.
8653 (*umulsi3_highpart_nov6): Likewise.
8654 (*umulsi3_highpart_v6): Likewise.
8655 (mulhisi3): Likewise.
8656 (*mulhisi3tb): Likewise.
8657 (*mulhisi3bt): Likewise.
8658 (*mulhisi3tt): Likewise.
8659 (maddhisi4): Likewise.
8660 (*maddhisi4tb): Likewise.
8661 (*maddhisi4tt): Likewise.
8662 (maddhidi4): Likewise.
8663 (*maddhidi4tb): Likewise.
8664 (*maddhidi4tt): Likewise.
8666 (udivsi3): Likewise.
8667 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8668 (thumb2_mulsi_short_compare0): Likewise.
8669 (thumb2_mulsi_short_compare0_scratch): Likewise.
8670 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8671 (1020mult2): Likewise.
8672 (1020mult3): Likewise.
8673 (1020mult4): Likewise.
8674 (1020mult5): Likewise.
8675 (1020mult6): Likewise.
8676 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8678 (cortex_a15_mult64): Likewise.
8679 (cortex_a15_sdiv): Likewise.
8680 (cortex_a15_udiv): Likewise.
8681 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8687 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8688 (pj4_ir_div): Likewise.
8689 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8690 (11_mult2): Likewise.
8691 (11_mult3): Likewise.
8692 (11_mult4): Likewise.
8693 (11_mult5): Likewise.
8694 (11_mult6): Likewise.
8695 (11_mult7): Likewise.
8696 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8697 (cortex_a8_mla): Likewise.
8698 (cortex_a8_mull): Likewise.
8699 (cortex_a8_smulwy): Likewise.
8700 (cortex_a8_smlald): Likewise.
8701 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8702 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8703 (cortex_r4_mul_3): Likewise.
8704 (cortex_r4_mla_4): Likewise.
8705 (cortex_r4_mla_3): Likewise.
8706 (cortex_r4_smlald): Likewise.
8707 (cortex_r4_mull): Likewise.
8708 (cortex_r4_sdiv): Likewise.
8709 (cortex_r4_udiv): Likewise.
8710 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8711 (cortex_a7_idiv): Likewise.
8712 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8713 (9_mult2): Likewise.
8714 (9_mult3): Likewise.
8715 (9_mult4): Likewise.
8716 (9_mult5): Likewise.
8717 (9_mult6): Likewise.
8718 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8719 (cortex_a53_sdiv): Likewise.
8720 (cortex_a53_udiv): Likewise.
8721 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8722 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8723 (mp626_mult2): Likewise.
8724 (mp626_mult3): Likewise.
8725 (mp626_mult4): Likewise.
8726 * config/arm/fa526.md (526_mult1): Update attribute change.
8727 (526_mult2): Likewise.
8728 * config/arm/arm-generic.md (mult): Update attribute change.
8729 (mult_ldsched_strongarm): Likewise.
8730 (mult_ldsched): Likewise.
8731 (multi_cycle): Likewise.
8732 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8733 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8734 (606te_mult2): Likewise.
8735 (606te_mult3): Likewise.
8736 (606te_mult4): Likewise.
8737 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8738 (cortex_a9_mac16): Likewise.
8739 (cortex_a9_multiply): Likewise.
8740 (cortex_a9_mac): Likewise.
8741 (cortex_a9_multiply_long): Likewise.
8742 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8743 (626te_mult2): Likewise.
8744 (626te_mult3): Likewise.
8745 (626te_mult4): Likewise.
8747 2013-06-18 Richard Biener <rguenther@suse.de>
8750 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8752 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8755 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8756 with next_active_insn.
8758 2013-06-18 Alan Modra <amodra@gmail.com>
8760 * config/rs6000/rs6000.h (enum data_align): New.
8761 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8762 (DATA_ABI_ALIGNMENT): Define.
8763 (CONSTANT_ALIGNMENT): Correct comment.
8764 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8765 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8767 2013-06-17 David Malcolm <dmalcolm@redhat.com>
8769 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8770 ATTRIBUTE_UNUSED marking.
8772 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8774 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8775 alternative and update.
8776 (aarch64_dup_lanedi): Delete.
8777 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8778 * config/aarch64/aarch64-simd-builtins.def: Update.
8780 2013-06-17 Richard Biener <rguenther@suse.de>
8782 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8783 (lto_input_scc): Declare.
8784 (lto_input_tree_1): Likewise.
8785 (struct lto_stats_d): Add num_tree_bodies_output and
8786 num_pickle_refs_output.
8787 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8788 (lto_read_tree_1): Split out from ...
8789 (lto_read_tree): ... this.
8790 (lto_input_scc): New function.
8791 (lto_input_tree_1): Split out from ...
8792 (lto_input_tree): ... this. Handle LTO_tree_scc.
8793 (lto_data_in_create): Create the streamer cache without hashes.
8794 * lto-streamer-out.c (create_output_block): Create the streamer
8795 cache with hashes when not doing WPA.
8796 (lto_write_tree_1): Split out from ...
8797 (lto_write_tree): ... this.
8798 (get_symbol_initial_value): New function.
8799 (lto_output_tree_1): Split out from ...
8800 (lto_output_tree): ... this. Write trees as series of SCCs
8801 using a DFS walk via DFS_write_tree.
8802 (struct sccs, struct scc_entry): New types.
8803 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
8804 (DFS_write_tree_body): New function.
8805 (DFS_write_tree): Likewise.
8806 (hash_tree): Likewise.
8807 (scc_entry_compare): Likewise.
8808 (hash_scc): Likewise.
8809 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
8810 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
8811 TREE_CHAIN as regular reference.
8812 (streamer_read_integer_cst): Remove.
8813 (streamer_get_pickled_tree): Adjust.
8814 * tree-streamer-out.c (streamer_write_chain): Disable streaming
8815 of DECL_EXTERNALs in BLOCK_VARS for now.
8816 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
8818 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
8819 Add hash value argument and record that if hashes are recorded
8821 (streamer_tree_cache_insert_1): Adjust.
8822 (streamer_tree_cache_insert): Likewise.
8823 (streamer_tree_cache_insert_at): Rename to ...
8824 (streamer_tree_cache_replace_tree): ... this and adjust.
8825 (streamer_tree_cache_append): Adjust.
8826 (record_common_node): Likewise.
8827 (streamer_tree_cache_create): Add argument whether to
8828 record hash values together with trees.
8829 (streamer_tree_cache_delete): Adjust.
8830 * tree-streamer.h (struct streamer_tree_cache_d): Add
8832 (streamer_read_integer_cst): Remove.
8833 (streamer_tree_cache_insert): Adjust.
8834 (streamer_tree_cache_append): Likewise.
8835 (streamer_tree_cache_insert_at): Rename to ...
8836 (streamer_tree_cache_replace_tree): ... this and adjust.
8837 (streamer_tree_cache_create): Add argument whether to record hashes.
8838 (streamer_tree_cache_get): Rename to ...
8839 (streamer_tree_cache_get_tree): ... this.
8840 (streamer_tree_cache_get_hash): New function.
8841 * tree.c (cache_integer_cst): New function.
8842 * tree.h (cache_integer_cst): Declare.
8843 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
8844 * lto-symtab.c (lto_varpool_replace_node): Only release
8845 DECL_INITIAL of non-prevailing decls.
8846 * varpool.c (varpool_remove_initializer): Do not release
8847 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
8849 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8851 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
8852 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
8853 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
8854 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
8855 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
8856 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
8857 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
8858 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
8859 instead of TARGET_64BIT.
8860 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
8861 Require ISA_HAS_<D>DIV.
8863 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
8865 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
8866 (mips*-*-linux*): Move default with_llsc setting to where other
8868 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
8870 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
8871 Likewise. Remove default with_tune setting. Move default float
8872 setting to its own block. Handle with_llsc in the same block as above.
8874 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
8876 PR rtl-optimization/57425
8877 PR rtl-optimization/57569
8878 * alias.c (write_dependence_p): Add new parameters mem_mode,
8879 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
8880 Changed all callers.
8881 (canon_anti_dependence): New function.
8882 * cse.c (check_dependence): Use canon_anti_dependence.
8883 * cselib.c (cselib_invalidate_mem): Likewise.
8884 * rtl.h (canon_anti_dependence): Declare.
8886 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
8888 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
8889 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
8890 ".set mips3" for 64-bit targets.
8892 2013-06-15 Dehao Chen <dehao@google.com>
8894 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
8895 * gimple-low.c (gimple_check_call_matching_types): Likewise.
8896 (gimple_check_call_args): Likewise.
8897 * value-prof.c (check_ic_target): Likewise.
8898 * ipa-inline.c (early_inliner): Likewise.
8899 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
8900 * cgraph.c (cgraph_create_edge_1): Likewise.
8901 (cgraph_make_edge_direct): Likewise.
8903 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
8906 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
8907 rs6000_output_move_128bit to handle emitting quad memory
8908 operations. Set attribute length to 8 bytes.
8910 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
8912 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
8914 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
8915 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
8916 (aarch64_<su>mlsl<mode>): Likewise.
8918 2013-06-14 Mike Stump <mikestump@comcast.net>
8920 * Makefile.in (TARGET_H): Add insn-codes.h.
8922 2013-06-14 Alan Modra <amodra@gmail.com>
8926 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
8927 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
8928 bitfield expansion when EXPAND_MEMORY.
8929 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
8931 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
8933 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
8934 test for clearing quad memory on 32-bit later.
8936 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8938 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
8939 (fold_negate_expr): Likewise.
8940 (fold_real_zero_addition_p): Handle vectors.
8941 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
8943 2013-06-14 Alan Modra <amodra@gmail.com>
8945 * varasm.c (force_const_mem): Revert 2013-06-07 change.
8947 2013-06-13 Jan Hubicka <jh@suse.cz>
8949 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
8950 Local comdats are not externally visible.
8951 * symtab.c (dump_symtab_base): Dump externally visible.
8952 (verify_symtab_base): Verify back links in the symtab hash.
8954 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8956 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
8957 CONVERT_EXPR as equal nodes.
8959 2013-06-13 Bin Cheng <bin.cheng@arm.com>
8961 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8963 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
8965 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
8966 Generalize to complex and vector.
8967 * tree.c (build_all_ones_cst): New function.
8968 * tree.h (build_all_ones_cst): Declare it.
8970 2013-06-13 Alan Modra <amodra@gmail.com>
8972 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
8973 * config/rs6000/rs6000.md (signbittf2): New insn.
8974 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
8975 (abstf2_internal, cmptf_internal2): Likewise.
8976 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
8978 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8979 Pat Haugen <pthaugen@us.ibm.com>
8980 Peter Bergner <bergner@vnet.ibm.com>
8982 * config/rs6000/rs6000.c (emit_load_locked): Add support for
8983 power8 byte, half-word, and quad-word atomic instructions.
8984 (emit_store_conditional): Likewise.
8985 (rs6000_expand_atomic_compare_and_swap): Likewise.
8986 (rs6000_expand_atomic_op): Likewise.
8988 * config/rs6000/sync.md (larx): Add new modes for power8.
8990 (AINT): New mode iterator to include TImode as well as normal
8991 integer modes on power8.
8992 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
8993 that VSX registers are not considered. Use AINT mode iterator
8994 instead of INT1 to allow inclusion of quad word atomic operations
8996 (load_locked<mode>): Likewise.
8997 (store_conditional<mode>): Likewise.
8998 (atomic_compare_and_swap<mode>): Likewise.
8999 (atomic_exchange<mode>): Likewise.
9000 (atomic_nand<mode>): Likewise.
9001 (atomic_fetch_<fetchop_name><mode>): Likewise.
9002 (atomic_nand_fetch<mode>): Likewise.
9003 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
9005 (ATOMIC): On power8, add QImode, HImode modes.
9006 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
9007 modes that promote to SImode.
9008 (load_lockedti): Convert TImode arguments to PTImode, so that we
9009 get a guaranteed even/odd register pair.
9010 (load_lockedpti): Likewise.
9011 (store_conditionalti): Likewise.
9012 (store_conditionalpti): Likewise.
9014 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9015 atomic load/store instructions.
9018 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9020 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9022 (insn_count): New attribute, with most cases extracted from...
9023 (length): ...here. Redefine most cases in terms of insn_count.
9024 (single_insn): Delete.
9025 (can_delay): Use insn_count to check for single instructions.
9026 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9027 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9028 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9029 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9030 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9031 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9032 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9033 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9034 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9035 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9036 rather than "length".
9037 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9038 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9039 Use "insn_count" rather than "length".
9040 * config/mips/mips-dsp.md
9041 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9042 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9045 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
9047 PR tree-optimization/57361
9048 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9050 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9052 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9054 (aarch64_simd_combine<mode>): New instruction expansion.
9055 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9057 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9058 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9060 2013-06-12 Jan Hubicka <jh@suse.cz>
9062 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9063 decl has when in streaming stage.
9064 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9065 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9067 2013-06-12 Roland Stigge <stigge@antcom.de>
9070 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9072 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9074 PR tree-optimization/57537
9075 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9076 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9078 2013-06-12 Richard Biener <rguenther@suse.de>
9080 * data-streamer.h (streamer_write_char_stream): CSE
9081 obs->current_pointer.
9082 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9083 streamer_write_char_stream manually and optimize the resulting loop.
9084 (streamer_write_hwi_stream): Likewise.
9086 2013-06-12 Jan Hubicka <jh@suse.cz>
9088 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9089 * cgraph.h (varpool_create_empty_node): Declare.
9090 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9092 * symtab.c (symtab_unregister_node): Be lax about missin entries
9094 (symtab_get_node): Update comment.
9095 * varpool.c (varpool_create_empty_node): Break out from ...
9096 (varpool_node_for_decl): ... here.
9097 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9099 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9101 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9102 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9103 part. Use straight-line flow at the end.
9104 <COMPONENT_REF>: Remove superfluous else.
9105 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9107 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9110 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9111 target hook even for !TREE_PUBLIC decls. If no resolution info
9112 is available, return false for common and external decls.
9114 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9116 * config/rl78/constraints.md (U): New constraint.
9117 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9120 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9123 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9124 to allow returning address to AT_PLATFORM name.
9126 2013-06-11 Jan Hubicka <jh@suse.cz>
9128 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9129 * cgraph.h (symtab_node_base): Add weakref flag.
9130 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9131 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9132 (output_weakrefs): Use weakref flag.
9133 * fold-const.c (simple_operand_p): Handle WEAK.
9134 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9135 * ipa.c (varpool_externally_visible_p): Drop weakref.
9136 (function_and_variable_visibility): Update comment; fix weakref
9137 sanity checks; do not clear DECL_WEAK on them.
9138 * lto-cgraph.c (lto_output_node): update.
9139 (lto_output_varpool_node): Update.
9140 (input_overwrite_node): Update.
9141 (input_node): Update.
9142 (input_varpool_node): Update.
9143 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9144 (lto_symtab_merge_symbols): Add sanity check.
9145 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9146 * passes.c (rest_of_decl_compilation): Set static flag, too.
9147 * symtab.c (dump_symtab_base): Dump weakref.
9148 (verify_symtab_base): Sanity check weakrefs.
9149 (symtab_make_decl_local): Remove duplicated code.
9150 (symtab_alias_ultimate_target): Simplify.
9151 * varpool.c (varpool_create_variable_alias): Set weakref flag.
9153 2013-06-11 Tom de Vries <tom@codesourcery.com>
9155 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9156 sequence_vect == NULL.
9158 2013-06-11 DJ Delorie <dj@redhat.com>
9160 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9161 (rl78_unwind_word_mode): New.
9163 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9165 * final.c (debug_prefix_maps): Make static.
9167 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9169 * function.c (initial_trampoline): Remove stray copy.
9171 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9173 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9175 2013-06-11 Martin Jambor <mjambor@suse.cz>
9177 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9178 within bounds at the beginning of the function.
9180 2013-06-11 Alan Modra <amodra@gmail.com>
9182 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9184 (get_named_section): Don't NULL !DECL_P decl.
9186 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9188 * doc/invoke.texi (core-avx2): Document.
9190 (atom): Updated with MOVBE.
9192 2013-06-11 Richard Biener <rguenther@suse.de>
9194 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
9196 2013-06-11 Anton Blanchard <anton@samba.org>
9198 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9199 correct shift value in little-endian mode.
9201 2013-06-11 Jakub Jelinek <jakub@redhat.com>
9204 * varasm.c (get_variable_align): Move #endif to the right place.
9206 2013-06-10 Cary Coutant <ccoutant@google.com>
9208 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
9209 for hash so that hash table traversal order is deterministic.
9211 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
9212 Pat Haugen <pthaugen@us.ibm.com>
9213 Peter Bergner <bergner@vnet.ibm.com>
9215 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9216 of vectors in GPRS if they are direct moves or quad word load or
9219 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9221 (direct_move_p): Likewise.
9222 (quad_load_store_p): Likewise.
9224 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9225 classes into bins based on the physical register type.
9226 (reg_class_to_reg_type): Likewise.
9227 (IS_STD_REG_TYPE): Likewise.
9228 (IS_FP_VECT_REG_TYPE): Likewise.
9229 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9230 use direct move instructions.
9231 (reload_gpr_vsx): Likewise.
9232 (reload_vsx_gpr): Likewise.
9233 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9234 information that is a simplification of register classes. Also
9235 precalculate direct move reload helpers.
9236 (direct_move_p): New function to return true if the operation can
9237 be done as a direct move instruciton.
9238 (quad_load_store_p): New function to return true if the operation
9239 is a quad memory operation.
9240 (rs6000_legitimize_address): If quad memory, only allow register
9241 indirect for TImode addresses.
9242 (rs6000_legitimate_address_p): Likewise.
9243 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9244 (rs6000_reload_register_type): Likewise.
9245 (register_to_reg_type): Return register type.
9246 (rs6000_secondary_reload_simple_move): New helper function for
9247 secondary reload and secondary memory needed to identify anything
9248 that is a simple move, and does not need reloading.
9249 (rs6000_secondary_reload_direct_move): New helper function for
9250 secondary reload to identify cases that can be done with several
9251 instructions via the direct move instructions.
9252 (rs6000_secondary_reload_move): New helper function for secondary
9253 reload to identify moves between register types that can be done.
9254 (rs6000_secondary_reload): Add support for quad memory operations
9255 and for direct move.
9256 (rs6000_secondary_memory_needed): Likewise.
9257 (rs6000_debug_secondary_memory_needed): Change argument names.
9258 (rs6000_output_move_128bit): New function to return the move to
9259 use for 128-bit moves, including knowing about the various
9260 limitations of quad memory operations.
9262 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9263 memory operations. call rs6000_output_move_128bit for the actual
9264 instruciton(s) to generate.
9265 (vsx_movti_64bit): Likewise.
9267 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9268 (UNSPEC_P8V_MTVSRWZ): Likewise.
9269 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9270 (UNSPEC_P8V_MTVSRD): Likewise.
9271 (UNSPEC_P8V_XXPERMDI): Likewise.
9272 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9273 (UNSPEC_FUSION_GPR): Likewise.
9274 (FMOVE128_GPR): New iterator for direct move.
9275 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
9278 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9279 loads and direct move instructions.
9280 (zero_extendsidi2_lfiwzx): Likewise.
9281 (extendsidi2_lfiwax): Likewise.
9282 (extendsidi2_nocell): Likewise.
9283 (floatsi<mode>2_lfiwax): Likewise.
9285 (floatunssi<mode>2_lfiwzx): Likewise.
9287 (fix_trunc<mode>_stfiwx): Likewise.
9288 (fixuns_trunc<mode>_stfiwx): Likewise.
9289 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9290 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9291 (parity<mode>2_cmpb): Set length/type attr.
9292 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9293 for 'mr.' to fast_compare.
9294 (bpermd_<mode>): Change type attr to popcnt.
9295 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9296 (p8_mtvsrwz_1): Likewise.
9297 (p8_mtvsrwz_2): Likewise.
9298 (reload_fpr_from_gpr<mode>): Likewise.
9299 (p8_mtvsrd_1): Likewise.
9300 (p8_mtvsrd_2): Likewise.
9301 (p8_xxpermdi_<mode>): Likewise.
9302 (reload_vsx_from_gpr<mode>): Likewise.
9303 (reload_vsx_from_gprsf): Likewise.
9304 (p8_mfvsrd_3_<mode>): LIkewise.
9305 (reload_gpr_from_vsx<mode>): Likewise.
9306 (reload_gpr_from_vsxsf): Likewise.
9307 (p8_mfvsrd_4_disf): Likewise.
9308 (multi-word GPR splits): Do not split direct moves or quad memory
9311 2013-06-10 David Malcolm <dmalcolm@redhat.com>
9313 * tree-into-ssa.c (interesting_blocks): Make static.
9315 2013-06-10 Jakub Jelinek <jakub@redhat.com>
9318 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9319 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9320 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9321 (get_variable_align): New function.
9322 (get_variable_section, emit_bss, emit_common,
9323 assemble_variable_contents, place_block_symbol): Use
9324 get_variable_align instead of DECL_ALIGN.
9325 (assemble_noswitch_variable): Add align argument, use it
9326 instead of DECL_ALIGN.
9327 (assemble_variable): Adjust caller. Use get_variable_align
9328 instead of DECL_ALIGN.
9329 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9331 (DATA_ABI_ALIGNMENT): Define.
9332 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9333 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9334 opt is false, only return the psABI mandated alignment increase.
9335 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9336 (DATA_ABI_ALIGNMENT): ... this.
9337 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9338 (DATA_ABI_ALIGNMENT): ... this.
9339 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9340 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9341 (DATA_ABI_ALIGNMENT): ... this.
9342 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9343 * doc/tm.texi: Regenerated.
9345 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
9347 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9348 cmp_code to construct REG_EQUAL note.
9350 2013-06-09 Jakub Jelinek <jakub@redhat.com>
9353 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9354 that operands[2] doesn't overlap with operands[0].
9356 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
9357 Jan Hubicka <jh@suse.cz>
9359 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9360 hack to mark symbols as used.
9362 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9364 PR rtl-optimization/57559
9365 * lra-constraints.c (process_alt_operands): Don't discourage
9366 memory with known offset for offsetable memory constraint.
9367 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9369 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9371 * varasm.c (struct oc_local_state): Reorder fields.
9372 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9373 and adjust accordingly.
9374 (output_constructor): Reorder initialization code and adjust call to
9375 output_constructor_bitfield.
9377 2013-06-07 Jan Hubicka <jh@suse.cz>
9379 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9381 2013-06-07 David Malcolm <dmalcolm@redhat.com>
9383 * tree-object-size.c (unknown): Make const.
9385 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9387 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9388 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9389 for last alternative in the cpu_facility attribute.
9391 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9394 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9395 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9396 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9397 * config/arm/constraints.md (Dg): New constraint.
9398 * config/arm/neon.md (xordi3_neon): Remove.
9399 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9400 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9402 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9404 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9405 Clean up alternatives.
9407 2013-06-07 Alan Modra <amodra@gmail.com>
9409 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9412 2013-06-07 Alan Modra <amodra@gmail.com>
9414 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9416 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9418 * config/arm/constraints.md (Df): New constraint.
9419 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9420 Correct length attribute for last two alternatives.
9422 2013-06-07 Alan Modra <amodra@gmail.com>
9424 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9425 override user -mfp-in-toc.
9426 (offsettable_ok_by_alignment): Consider just the current access
9427 rather than the whole object, unless BLKmode. Handle
9428 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9429 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9430 for -mcmodel=medium.
9431 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9432 override user -mfp-in-toc or -msum-in-toc. Default to
9433 -mno-fp-in-toc for -mcmodel=medium.
9435 2013-06-06 DJ Delorie <dj@redhat.com>
9437 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9438 TARGET_VALID_POINTER_MODE.
9440 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
9441 Pat Haugen <pthaugen@us.ibm.com>
9442 Peter Bergner <bergner@vnet.ibm.com>
9444 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9445 Document new power8 builtins.
9447 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9448 condition code register, to allow 128-bit logical operations to be
9449 done in the VSX or GPR registers.
9450 (nor<mode>3): Use the canonical form for nor.
9451 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9452 vclz*, and vpopcnt* vector instructions.
9453 (nand<mode>3): Likewise.
9454 (orc<mode>3): Likewise.
9455 (clz<mode>2): LIkewise.
9456 (popcount<mode>2): Likewise.
9458 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9459 that only the GPRs are recognized.
9461 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9462 support for new power8 builtins.
9464 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9466 (xscvdpspn): Likewise.
9472 (vpopcnt): Likewise.
9473 (vpopcntb): Likewise.
9474 (vpopcnth): Likewise.
9475 (vpopcntw): Likewise.
9476 (vpopcntd): Likewise.
9481 (eqv_v16qi3): Likewise.
9482 (eqv_v8hi3): Likewise.
9483 (eqv_v4si3): Likewise.
9484 (eqv_v2di3): Likewise.
9485 (eqv_v4sf3): Likewise.
9486 (eqv_v2df3): Likewise.
9488 (nand_v16qi3): Likewise.
9489 (nand_v8hi3): Likewise.
9490 (nand_v4si3): Likewise.
9491 (nand_v2di3): Likewise.
9492 (nand_v4sf3): Likewise.
9493 (nand_v2df3): Likewise.
9495 (orc_v16qi3): Likewise.
9496 (orc_v8hi3): Likewise.
9497 (orc_v4si3): Likewise.
9498 (orc_v2di3): Likewise.
9499 (orc_v4sf3): Likewise.
9500 (orc_v2df3): Likewise.
9502 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9503 allow power8 quad mode in 64-bit.
9504 (rs6000_builtin_vectorized_function): Add support to vectorize
9505 ISA 2.07 count leading zeros, population count builtins.
9506 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9507 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9508 (builtin_function_type): Add vgbbd builtin function which takes an
9510 (altivec_expand_vec_perm_const): Add support for new power8 merge
9513 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9514 that does not include TImdoe for use with 32-bit.
9515 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9517 (UNSPEC_VSX_CVDPSPN): Likewise.
9518 (vsx_xscvdpspn): Likewise.
9519 (vsx_xscvspdpn): Likewise.
9520 (vsx_xscvdpspn_scalar): Likewise.
9521 (vsx_xscvspdpn_directmove): Likewise.
9522 (vsx_and<mode>3): Split logical operations into 32-bit and
9523 64-bit. Add support to do logical operations on TImode as well as
9524 VSX vector types. Allow logical operations to be done in either
9525 VSX registers or in general purpose registers in 64-bit mode. Add
9526 splitters if GPRs were used. For AND, add clobber of CCmode to
9527 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9529 (vsx_and<mode>3_32bit): Likewise.
9530 (vsx_and<mode>3_64bit): Likewise.
9531 (vsx_ior<mode>3): Likewise.
9532 (vsx_ior<mode>3_32bit): Likewise.
9533 (vsx_ior<mode>3_64bit): Likewise.
9534 (vsx_xor<mode>3): Likewise.
9535 (vsx_xor<mode>3_32bit): Likewise.
9536 (vsx_xor<mode>3_64bit): Likewise.
9537 (vsx_one_cmpl<mode>2): Likewise.
9538 (vsx_one_cmpl<mode>2_32bit): Likewise.
9539 (vsx_one_cmpl<mode>2_64bit): Likewise.
9540 (vsx_nor<mode>3): Likewise.
9541 (vsx_nor<mode>3_32bit): Likewise.
9542 (vsx_nor<mode>3_64bit): Likewise.
9543 (vsx_andc<mode>3): Likewise.
9544 (vsx_andc<mode>3_32bit): Likewise.
9545 (vsx_andc<mode>3_64bit): Likewise.
9546 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9547 and xxlorc instructions.
9548 (vsx_eqv<mode>3_64bit): Likewise.
9549 (vsx_nand<mode>3_32bit): Likewise.
9550 (vsx_nand<mode>3_64bit): Likewise.
9551 (vsx_orc<mode>3_32bit): Likewise.
9552 (vsx_orc<mode>3_64bit): Likewise.
9554 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9556 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9558 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9559 (p8_vmrgow): Likewise.
9560 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9561 GPRs to be split under VSX.
9562 (p8v_clz<mode>2): Add power8 count leading zero support.
9563 (p8v_popcount<mode>2): Add power8 population count support.
9564 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9567 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9570 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9572 (vec_nand): Likewise.
9573 (vec_vclz): Likewise.
9574 (vec_vclzb): Likewise.
9575 (vec_vclzd): Likewise.
9576 (vec_vclzh): Likewise.
9577 (vec_vclzw): Likewise.
9578 (vec_vgbbd): Likewise.
9579 (vec_vmrgew): Likewise.
9580 (vec_vmrgow): Likewise.
9581 (vec_vpopcnt): Likewise.
9582 (vec_vpopcntb): Likewise.
9583 (vec_vpopcntd): Likewise.
9584 (vec_vpopcnth): Likewise.
9585 (vec_vpopcntw): Likewise.
9587 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9589 PR rtl-optimization/57468
9590 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9593 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9595 PR rtl-optimization/57459
9596 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9597 type when setting live regs.
9599 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9601 * config/s390/s390.opt (mlra): New option.
9602 * config/s390/s390.c (s390_decompose_address): Check displacement
9603 for all registers for LRA.
9604 (s390_secondary_reload): Don't used secondary reloads for LRA.
9605 (s390_lra_p): New function.
9606 (TARGET_LRA_P): Define.
9607 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9608 of attribute cpu_facility to zarch for the last alternative.
9609 (*cmpmem_short): Ditto.
9611 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9613 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9614 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9615 (arm_expand_prologue): Likewise.
9617 2013-06-06 Teresa Johnson <tejohnson@google.com>
9620 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9621 as this is now done by redirect_edge_and_branch_force.
9622 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9623 barriers, and fix interaction with splitting.
9624 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9625 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9626 reflect changes made in the routine.
9627 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9628 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9629 since this is called in cfglayout mode, and replace partition fixup
9630 with assert as that is now done by force_nonfallthru_and_redirect.
9631 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9632 already be marked with region crossing note.
9633 (insert_section_boundary_note): Make non-static, gate on flag
9634 has_bb_partition, rewrite to also check for multiple partitions.
9635 (rest_of_handle_reorder_blocks): Remove call to
9636 insert_section_boundary_note, now done later during free_cfg.
9637 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9638 * bb-reorder.h (insert_section_boundary_note): Declare.
9639 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9640 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9641 invoke insert_section_boundary_note.
9642 (try_redirect_by_replacing_jump): Remove unnecessary
9643 check for region crossing note.
9644 (fixup_partition_crossing): New function.
9645 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9646 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9647 in non-cfglayout mode.
9648 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9649 remove old code that tried to do this. Emit barrier correctly
9650 when we are in cfglayout mode.
9651 (last_bb_in_partition): New function.
9652 (rtl_split_edge): Correctly fixup partition boundaries.
9653 (commit_one_edge_insertion): Remove old code that tried to
9654 fixup region crossing edge since this is now handled in
9655 split_block, and set up insertion point correctly since
9656 block may now end in a jump.
9657 (verify_hot_cold_block_grouping): Guard against checking when not in
9658 linearized RTL mode.
9659 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9661 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9662 rtl_verify_flow_info, so not called in cfglayout mode.
9663 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9664 (fixup_reorder_chain): Remove old code that attempted to fixup region
9665 crossing note as this is now handled in force_nonfallthru_and_redirect.
9666 (duplicate_insn_chain): Don't duplicate switch section notes.
9667 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9669 * basic-block.h (emit_barrier_after_bb): Declare.
9671 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9673 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9674 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9675 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9677 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9679 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9681 * config/arm/ldmstm.md: Regenerate.
9683 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9685 * config/arm/sync.md (atomic_loaddi_1):
9686 Disable predication for arm_restrict_it.
9687 (arm_load_exclusive<mode>): Likewise.
9688 (arm_load_exclusivesi): Likewise.
9689 (arm_load_exclusivedi): Likewise.
9690 (arm_load_acquire_exclusive<mode>): Likewise.
9691 (arm_load_acquire_exclusivesi): Likewise.
9692 (arm_load_acquire_exclusivedi): Likewise.
9693 (arm_store_exclusive<mode>): Likewise.
9694 (arm_store_exclusive<mode>): Likewise.
9695 (arm_store_release_exclusivedi): Likewise.
9696 (arm_store_release_exclusive<mode>): Likewise.
9698 2013-06-06 Richard Biener <rguenther@suse.de>
9700 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9702 (lto_tag_is_tree_code_p): Adjust.
9703 (lto_tag_is_gimple_code_p): Likewise.
9704 (lto_gimple_code_to_tag): Likewise.
9705 (lto_tag_to_gimple_code): Likewise.
9706 (lto_tree_code_to_tag): Likewise.
9707 (lto_tag_to_tree_code): Likewise.
9708 * data-streamer.h (streamer_write_hwi_in_range): Use
9709 uhwi streaming to stream the normalized range.
9710 (streamer_read_hwi_in_range): Likewise.
9712 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9714 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9715 (predicable_short_it): Likewise.
9716 (predicated): Likewise.
9717 (enabled): Handle above.
9718 (define_cond_exec): Set predicated attribute to yes.
9720 2013-06-05 Mike Stump <mikestump@comcast.net>
9722 * gdbinit.in (__FUNCTION__): Add.
9724 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
9726 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9727 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9729 2013-06-05 Jan Hubicka <jh@suse.cz>
9731 * varasm.c (mark_decl_referenced): Revert the removal until targets
9734 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
9736 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9737 instead of mark_decl_referenced.
9739 2013-06-05 Jan Hubicka <jh@suse.cz>
9741 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9742 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9743 and symtab_used_from_object_file_p.
9744 (cgraph_make_node_local_1): Clear forced_by_abi.
9745 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9746 * cgraph.h (symtab_node_base): Add forced_by_abi.
9747 (decide_is_variable_needed): Remove.
9748 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9749 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9750 (decide_is_symbol_needed): ... this one; handle symbols in general;
9751 always analyze virtuals; honnor forced_by_abi.
9752 (cgraph_finalize_function): Update.
9753 (varpool_finalize_decl): Update.
9754 (symbol_defined_and_needed): Remove.
9755 (analyze_functions): Update.
9756 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9757 output_refs, input_overwrite_node): Handle forced_by_abi.
9758 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9759 (address_taken_from_non_vtable_p): ... this one.
9760 (comdat_can_be_unshared_p_1): New function.
9761 (cgraph_comdat_can_be_unshared_p): Rename to ...
9762 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9763 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9764 (function_and_variable_visibility): Clear forced_by_abi as needed.
9765 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9766 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
9767 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9768 * varpool.c (decide_is_variable_needed): Remove.
9770 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9772 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9773 (arm_option_override): Override arm_restrict_it where appropriate.
9774 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9775 * config/arm/arm.opt (mrestrict-it): New command-line option.
9776 * doc/invoke.texi: Document -mrestrict-it.
9778 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9780 * tsan.c (tsan_atomic_table): Make const.
9782 2013-06-05 Richard Biener <rguenther@suse.de>
9784 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9785 index associated with the tree we are supposed to replace.
9786 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9787 TREE_ASM_WRITTEN as zero for everything but SSA names.
9789 2013-06-05 David Malcolm <dmalcolm@redhat.com>
9791 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9793 2013-06-04 Jan Hubicka <jh@suse.cz>
9795 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9796 (input_node, input_varpool_node): Handle correctly external same
9798 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9799 nodes at ltrans stage.
9801 2013-06-04 Jan Hubicka <jh@suse.cz>
9803 * ipa-inline.c (update_caller_keys): Fix availability test.
9804 (update_callee_keys): Likewise.
9805 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
9806 to follow ELF standard.
9808 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
9810 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
9811 (mips64r5900el-*-elf*): New configurations.
9812 * config/mips/mips-cpus.def (r5900): New processor.
9813 * config/mips/mips-tables.opt: Regenerate.
9814 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
9815 (mips_issue_rate): Handle PROCESSOR_R5900.
9816 (mips_reorg_process_insns): Force reorder mode for the R5900.
9817 * config/mips/mips.h (TARGET_MIPS5900): Define.
9818 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
9820 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
9822 * config/mips/mips.md (processor): Add r5900.
9823 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
9825 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9827 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
9828 into function to generate MOVI instruction.
9829 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
9830 (aarch64_preferred_simd_mode): Turn into wrapper.
9831 (aarch64_output_scalar_simd_mov_immediate): New function.
9832 * config/aarch64/aarch64-protos.h: Add prototype for above.
9834 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9836 * config/aarch64/aarch64.c (simd_immediate_info): Remove
9837 element_char member.
9838 (sizetochar): Return signed char.
9839 (aarch64_simd_valid_immediate): Remove elchar and other
9840 unnecessary variables.
9841 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
9842 Calculate element_char as required.
9843 * config/aarch64/aarch64-protos.h: Update and move prototype
9844 for aarch64_output_simd_mov_immediate.
9845 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
9848 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9850 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
9851 information completed by aarch64_simd_valid_immediate.
9852 (aarch64_legitimate_constant_p): Update arguments.
9853 (aarch64_simd_valid_immediate): Work with struct rather than many
9855 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
9856 (aarch64_simd_make_constant): Update arguments.
9857 (aarch64_output_simd_mov_immediate): Work with struct rather than
9858 many pointers. Output immediate directly rather than as operand.
9859 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9861 * config/aarch64/constraints.md (Dn): Update arguments.
9863 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9865 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
9867 (aarch64_simd_immediate_valid_for_move): Remove.
9868 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
9869 (aarch64_simd_make_constant): Update call.
9870 (aarch64_output_simd_mov_immediate): Update call.
9871 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
9873 * config/aarch64/constraints.md (Dn): Update call.
9875 2013-06-04 Ian Bolton <ian.bolton@arm.com>
9877 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
9878 return type to bool for prototype.
9879 (aarch64_legitimate_constant_p): Check for true instead of not -1.
9880 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
9881 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
9883 2013-06-04 Catherine Moore <clm@codesourcery.com>
9885 * config/mips/mips.opt (meva): New.
9886 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
9887 (ASM_SPEC): Handle -meva.
9888 * doc/invoke.texi (meva): Document.
9890 2013-06-04 Alan Modra <amodra@gmail.com>
9892 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
9895 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9897 * rtl.def: Add extra fourth optional field to define_cond_exec.
9898 * gensupport.c (process_one_cond_exec): Process attributes from
9900 * doc/md.texi: Document fourth field in define_cond_exec.
9902 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
9904 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
9905 out the processing order as in store_bit_field_1.
9907 2013-06-04 Jan Hubicka <jh@suse.cz>
9910 * cgraphunit.c (cgraph_process_same_body_aliases): Create
9911 non-VAR_DECL node if it does not exist yet.
9913 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
9915 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
9916 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
9917 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
9918 target_cpu_default setting.
9920 2013-06-03 Teresa Johnson <tejohnson@google.com>
9922 * dumpfile.c (opt_info_switch_p): Change -fopt-info
9923 default to -fopt-info=optimized instead of all.
9924 * doc/invoke.texi: Ditto.
9925 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
9926 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
9927 (execute_vect_slp): Emit BB vectorization success under
9928 MSG_OPTIMIZED_LOCATIONS.
9929 * tree-vect-slp.c (vect_slp_transform_bb): Change
9930 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
9931 * tree-vect-loop.c (vect_transform_loop): Ditto.
9933 2013-06-03 Jason Merrill <jason@redhat.com>
9936 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9937 Use TARGET_EXPR for C++.
9939 2013-06-03 Jakub Jelinek <jakub@redhat.com>
9941 PR rtl-optimization/57268
9942 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
9943 if DEBUG_INSN_P (insn).
9946 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9948 PR rtl-optimization/57268
9949 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9950 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9952 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9954 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
9955 (ix86_avoid_lea_for_addr): Likewise.
9956 (exact_dependency_1): Likewise.
9957 (ix86_adjust_cost): Likewise.
9958 (swap_top_of_ready_list): Fix formatting and !reload_completed check
9960 (do_reorder_for_imul): Fix typo, formatting and
9961 !reload_completed check removed.
9962 (ix86_sched_reorder): Fix typo and formatting.
9963 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
9966 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
9968 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
9970 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
9972 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
9973 <STRING_CST>: Likewise.
9974 <VECTOR_CST>: Likewise.
9976 2013-06-01 Janus Weil <janus@gcc.gnu.org>
9977 Mikael Morin <mikael@gcc.gnu.org>
9979 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
9980 * config.in: Regenerated.
9981 * configure: Regenerated.
9983 2013-06-01 Jan Hubicka <jh@suse.cz>
9986 * cgraphunit.c (compile): When weakref is not supported,
9987 set up transparent aliases before final output pass.
9988 * varasm.c (assemble_alias): Do not try to do it here.
9990 2013-06-01 Jan Hubicka <jh@suse.cz>
9993 * passes.c (for_per_function): Skip unanalyzed functions.
9995 2013-06-01 Jan Hubicka <jh@suse.cz>
9997 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
9998 (lto_symtab_merge_symbols_1): ... this one.
9999 (lto_symtab_merge_cgraph_nodes): Rename to ...
10000 (lto_symtab_merge_symbols): ... this one; simplify.
10001 * cgraph.c (same_body_aliases_done): Rename to ...
10002 (cpp_implicit_aliases_done): ... this one.
10003 (cgraph_create_function_alias): Update.
10004 (cgraph_same_body_alias): Update.
10005 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
10006 (verify_edge_corresponds_to_fndecl): Simplify.
10007 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
10008 (cgraph_node): Remove same_body_alias.
10009 (varpool_node): Remove alias_of and extra_name_alias.
10010 (same_body_aliases_done): Rename to ..
10011 (cpp_implicit_aliases_done): ... this one.
10012 (symtab_alias_ultimate_target): Add default parameter.
10013 (symtab_resolve_alias): New function.
10014 (fixup_same_cpp_alias_visibility): Declare.
10015 (cgraph_function_node): Add default parameter.
10016 (cgraph_node_asm_name): Likewise.
10017 (cgraph_function_or_thunk_node): Add default parameter; do
10018 not ICE when it is NULL.
10019 (varpool_variable_node): Likewise.
10020 * tree-emutls.c (create_emultls_var): Update.
10021 (ipa_lower_emutls): Update.
10022 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10023 (cgraph_reset_node): Reset alias info.
10024 (cgraph_finalize_function): Update.
10025 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10026 (analyze_function): Simplify.
10027 (cgraph_process_same_body_aliases): Simplify.
10028 (analyze_functions): Fixup same body aliases.
10029 (handle_alias_pairs): Simplify.
10030 (assemble_thunk): Update.
10031 (assemble_thunks_and_aliases): Update.
10032 (output_weakrefs): Rewrite.
10033 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10034 (lto_output_varpool_node): Likewise.
10035 (compute_ltrans_boundary): Remve assert.
10036 (get_alias_symbol): New functoin.
10037 (input_node): Rewrite alias handling.
10038 (input_varpool_node): Likewise.
10039 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10040 * ipa.c (process_references): Handle weakrefs correctly.
10041 (symtab_remove_unreachable_nodes): Likewise.
10042 * trans-mem.c (get_cg_data): Update.
10043 (ipa_tm_create_version_alias): Update.
10044 (ipa_tm_execute): Update.
10045 * symtab.c (dump_symtab_base): Dump aliases.
10046 (verify_symtab_base): Verify aliases.
10047 (symtab_node_availability): New function.
10048 (symtab_alias_ultimate_target): Simplify.
10049 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10050 handle all the fixup cases.
10051 (symtab_resolve_alias): New function.
10052 * passes.c (ipa_write_summaries): Handle weakrefs.
10053 * varpool.c (varpool_analyze_node): Simplify.
10054 (assemble_aliases): Update.
10055 (varpool_create_variable_alias): Simplify.
10056 (varpool_extra_name_alias): Simplify.
10057 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10058 (lto_symtab_merge_symbols): ... this one.
10060 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10063 PR rtl-optimization/57268
10064 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10065 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10067 2013-06-01 Tobias Burnus <burnus@net-b.de>
10069 Partially reverted:
10070 2013-05-31 Tobias Burnus <burnus@net-b.de>
10072 PR middle-end/57073
10073 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10076 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10078 PR rtl-optimization/57268
10079 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10080 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10082 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10084 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10085 unordered comparison operators when -fno-trapping-math is in effect
10087 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10088 and implement unordered comparison operators properly on the e500.
10090 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10092 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10093 for constant scalar integers.
10094 (simplify_relational_operation_1): Likewise.
10096 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10098 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10099 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10102 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10103 Igor Zamyatin <igor.zamyatin@intel.com>
10105 Silvermont (SLM) architecture performance tuning.
10106 * config/i386/i386.h (enum ix86_tune_indices): Add
10107 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10108 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10110 * config/i386/i386.c (initial_ix86_tune_features)
10111 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10112 (ix86_lea_outperforms): Handle Silvermont tuning.
10113 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10115 (ix86_use_lea_for_mov): Likewise.
10116 (ix86_avoid_lea_for_addr): Likewise.
10117 (ix86_lea_for_add_ok): Likewise.
10118 (exact_dependency_1): New function.
10119 (exact_store_load_dependency): Likewise.
10120 (ix86_adjust_cost): Handle Silvermont tuning.
10121 (do_reoder_for_imul): Likewise.
10122 (swap_top_of_ready_list): New function.
10123 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10125 * config/i386/i386.md (peepholes that split memory operand in fp
10128 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10130 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10131 Remove un-necessary braces.
10133 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10135 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10136 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10138 2013-05-31 Tobias Burnus <burnus@net-b.de>
10140 PR middle-end/57073
10141 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
10143 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10146 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10147 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10148 * config/arm/neon.md (iordi3_neon): Remove.
10149 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10150 * config/arm/predicates.md (imm_for_neon_logic_operand):
10151 Move to earlier in the file.
10152 (neon_logic_op2): Likewise.
10153 (arm_iordi_operand_neon): New predicate.
10155 2013-05-31 Richard Biener <rguenther@suse.de>
10157 PR tree-optimization/57478
10158 PR tree-optimization/57453
10159 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10162 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10164 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10165 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10167 2013-05-30 Tobias Burnus <burnus@net-b.de>
10168 Thomas Koenig <tkoenig@gcc.gnu.org>
10170 PR middle-end/57073
10171 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10172 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10174 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10176 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
10178 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10180 * target.def (register_usage_leveling_p): New hook.
10181 * targhooks.c (default_register_usage_leveling_p): New.
10182 * targhooks.h (default_register_usage_leveling_p): New prototype.
10183 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10184 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10185 * doc/tm.texi: Update.
10186 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10188 2013-05-30 Ian Bolton <ian.bolton@arm.com>
10190 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10191 (*insv_reg<mode>): New define_insn.
10193 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
10195 PR rtl-optimization/57439
10196 * postreload.c (move2add_valid_value_p): Check that we have
10197 a zero subreg_regno_offset when accessing the register in
10198 the requested mode.
10200 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10201 Igor Zamyatin <igor.zamyatin@intel.com>
10203 Silvermont (SLM) architecture pipeline model, tuning and
10205 * config.gcc: Add slm config options and target.
10207 * config/i386/slm.md: New.
10209 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
10211 * config/i386/i386-c.c (ix86_target_macros_internal): New case
10213 (ix86_target_macros_internal): Likewise.
10215 * config/i386/i386.c (slm_cost): New cost.
10216 (m_SLM): New macro flag.
10217 (initial_ix86_tune_features): Set m_SLM.
10218 (x86_accumulate_outgoing_args): Likewise.
10219 (x86_arch_always_fancy_math_387): Likewise.
10220 (processor_target_table): Add slm cost.
10221 (cpu_names): Add slm cpu name.
10222 (x86_option_override_internal): Set SLM ISA.
10223 (ix86_issue_rate): New case PROCESSOR_SLM.
10224 (ia32_multipass_dfa_lookahead): Likewise.
10225 (fold_builtin_cpu): Add slm.
10227 * config/i386/i386.h (TARGET_SLM): New target macro.
10228 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10229 (processor_type): Add PROCESSOR_SLM.
10231 * config/i386/i386.md (cpu): Add new value "slm".
10232 (slm.md): Include slm.md.
10234 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10235 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10237 * config/arm/arm-protos.h: Add and update function protos.
10238 * config/arm/arm.c (use_simple_return_p): New added.
10239 (thumb2_expand_return): Check simple_return flag.
10240 * config/arm/arm.md: Add simple_return and conditional simple_return.
10241 * config/arm/iterators.md: Add iterator for return and simple_return.
10243 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10245 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10246 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10247 (arm_emit_vfp_multi_reg_pop): Likewise.
10248 (thumb2_emit_ldrd_pop): Likewise.
10249 (arm_expand_epilogue): Add misc REG_CFA notes.
10250 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10252 2013-05-29 Lawrence Crowl <crowl@google.com>
10254 * config/arm/t-arm: Update for below.
10256 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10257 Change type to hash_table. Update dependent calls and types.
10259 * config/i386/t-cygming: Update for below.
10261 * config/i386/t-interix: Update for below.
10263 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10264 Change type to hash_table. Update dependent calls and types.
10265 (i386_find_on_wrapper_list::wrappers): Likewise.
10267 * config/ia64/t-ia64: Update for below.
10269 * config/ia64/ia64.c (bundle_state_table):
10270 Change type to hash_table. Update dependent calls and types.
10272 * config/mips/mips.c (mips_reorg_process_insns::htab):
10273 Change type to hash_table. Update dependent calls and types.
10275 * config/sol2.c (solaris_comdat_htab):
10276 Change type to hash_table. Update dependent calls and types.
10278 * config/t-sol2: Update for above.
10280 2013-05-29 Teresa Johnson <tejohnson@google.com>
10282 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10283 functions are not yet marked as defined.
10285 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
10286 Pat Haugen <pthaugen@us.ibm.com>
10287 Peter Bergner <bergner@vnet.ibm.com>
10289 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10293 (vrotl<mode>3): Likewise.
10294 (vashl<mode>3): Likewise.
10295 (vlshr<mode>3): Likewise.
10296 (vashr<mode>3): Likewise.
10298 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10299 support for power8 V2DI builtins.
10301 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10302 power8 V2DI builtins.
10303 (vupkhsw): Likewise.
10304 (vupklsw): Likewise.
10305 (vaddudm): Likewise.
10306 (vminsd): Likewise.
10307 (vmaxsd): Likewise.
10308 (vminud): Likewise.
10309 (vmaxud): Likewise.
10310 (vpkudum): Likewise.
10311 (vpksdss): Likewise.
10312 (vpkudus): Likewise.
10313 (vpksdus): Likewise.
10318 (vsubudm): Likewise.
10319 (vcmpequd): Likewise.
10320 (vcmpgtsd): Likewise.
10321 (vcmpgtud): Likewise.
10322 (vcmpequd_p): Likewise.
10323 (vcmpgtsd_p): Likewise.
10324 (vcmpgtud_p): Likewise.
10325 (vupkhsw): Likewise.
10326 (vupklsw): Likewise.
10327 (vaddudm): Likewise.
10328 (vmaxsd): Likewise.
10329 (vmaxud): Likewise.
10330 (vminsd): Likewise.
10331 (vminud): Likewise.
10332 (vpksdss): Likewise.
10333 (vpksdus): Likewise.
10334 (vpkudum): Likewise.
10335 (vpkudus): Likewise.
10340 (vsubudm): Likewise.
10342 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10343 support for power8 V2DI instructions.
10345 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10346 power8 V2DI instructions. Combine pack and unpack insns to use an
10347 iterator for each mode. Check whether a particular mode supports
10348 Altivec instructions instead of just checking TARGET_ALTIVEC.
10349 (UNSPEC_VPKUWUM): Likewise.
10350 (UNSPEC_VPKSHSS): Likewise.
10351 (UNSPEC_VPKSWSS): Likewise.
10352 (UNSPEC_VPKUHUS): Likewise.
10353 (UNSPEC_VPKSHUS): Likewise.
10354 (UNSPEC_VPKUWUS): Likewise.
10355 (UNSPEC_VPKSWUS): Likewise.
10356 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10357 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10358 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10359 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10360 (UNSPEC_VUPKHSB): Likewise.
10361 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10362 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10363 (UNSPEC_VUPKHSH): Likewise.
10364 (UNSPEC_VUPKLSB): Likewise.
10365 (UNSPEC_VUPKLSH): Likewise.
10367 (VI_char): Likewise.
10368 (VI_scalar): Likewise.
10369 (VI_unit): Likewise.
10371 (VP_small): Likewise.
10372 (VP_small_lc): Likewise.
10373 (VU_char): Likewise.
10374 (add<mode>3): Likewise.
10375 (altivec_vaddcuw): Likewise.
10376 (altivec_vaddu<VI_char>s): Likewise.
10377 (altivec_vadds<VI_char>s): Likewise.
10378 (sub<mode>3): Likewise.
10379 (altivec_vsubcuw): Likewise.
10380 (altivec_vsubu<VI_char>s): Likewise.
10381 (altivec_vsubs<VI_char>s): Likewise.
10382 (altivec_vavgs<VI_char>): Likewise.
10383 (altivec_vcmpbfp): Likewise.
10384 (altivec_eq<mode>): Likewise.
10385 (altivec_gt<mode>): Likewise.
10386 (altivec_gtu<mode>): Likewise.
10387 (umax<mode>3): Likewise.
10388 (smax<mode>3): Likewise.
10389 (umin<mode>3): Likewise.
10390 (smin<mode>3): Likewise.
10391 (altivec_vpkuhum): Likewise.
10392 (altivec_vpkuwum): Likewise.
10393 (altivec_vpkshss): Likewise.
10394 (altivec_vpkswss): Likewise.
10395 (altivec_vpkuhus): Likewise.
10396 (altivec_vpkshus): Likewise.
10397 (altivec_vpkuwus): Likewise.
10398 (altivec_vpkswus): Likewise.
10399 (altivec_vpks<VI_char>ss): Likewise.
10400 (altivec_vpks<VI_char>us): Likewise.
10401 (altivec_vpku<VI_char>us): Likewise.
10402 (altivec_vpku<VI_char>um): Likewise.
10403 (altivec_vrl<VI_char>): Likewise.
10404 (altivec_vsl<VI_char>): Likewise.
10405 (altivec_vsr<VI_char>): Likewise.
10406 (altivec_vsra<VI_char>): Likewise.
10407 (altivec_vsldoi_<mode>): Likewise.
10408 (altivec_vupkhsb): Likewise.
10409 (altivec_vupkhs<VU_char>): Likewise.
10410 (altivec_vupkls<VU_char>): Likewise.
10411 (altivec_vupkhsh): Likewise.
10412 (altivec_vupklsb): Likewise.
10413 (altivec_vupklsh): Likewise.
10414 (altivec_vcmpequ<VI_char>_p): Likewise.
10415 (altivec_vcmpgts<VI_char>_p): Likewise.
10416 (altivec_vcmpgtu<VI_char>_p): Likewise.
10417 (abs<mode>2): Likewise.
10418 (vec_unpacks_hi_v16qi): Likewise.
10419 (vec_unpacks_hi_v8hi): Likewise.
10420 (vec_unpacks_lo_v16qi): Likewise.
10421 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10422 (vec_unpacks_lo_v8hi): Likewise.
10423 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10424 (vec_pack_trunc_v8h): Likewise.
10425 (vec_pack_trunc_v4si): Likewise.
10426 (vec_pack_trunc_<mode>): Likewise.
10428 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10430 (vec_vmaxsd): Likewise.
10431 (vec_vmaxud): Likewise.
10432 (vec_vminsd): Likewise.
10433 (vec_vminud): Likewise.
10434 (vec_vpksdss): Likewise.
10435 (vec_vpksdus): Likewise.
10436 (vec_vpkudum): Likewise.
10437 (vec_vpkudus): Likewise.
10438 (vec_vrld): Likewise.
10439 (vec_vsld): Likewise.
10440 (vec_vsrad): Likewise.
10441 (vec_vsrd): Likewise.
10442 (vec_vsubudm): Likewise.
10443 (vec_vupkhsw): Likewise.
10444 (vec_vupklsw): Likewise.
10446 2013-05-29 Jan Hubicka <jh@suse.cz>
10448 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10449 flags; reorder rest of fields in more consistent way.
10450 (varpool_node): Remove analyzed, finalized and alias.
10451 (cgraph_ndoe): Likewise.
10452 (symtab_alias_ultimate_target): New function.
10453 (cgraph_function_node): Move offline.
10454 (cgraph_reset_node): Declare.
10455 (cgraph_comdat_can_be_unshared_p): Remove.
10456 (varpool_remove_initializer): Declare.
10457 (varpool_first_defined_variable, varpool_next_defined_variable
10458 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10459 (cgraph_function_with_gimple_body_p): Update.
10460 (varpool_all_refs_explicit_p): Update.
10461 (symtab_alias_target): New function.
10462 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10463 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
10464 (cgraph_function_or_thunk_node): Simplify using
10465 symtab_alias_ultimate_target.
10466 (varpool_variable_node): Likewise.
10467 * cgraph.c (cgraph_create_function_alias): Update.
10468 (cgraph_add_thunk): Update.
10469 (cgraph_remove_node): Update.
10470 (dump_cgraph_node): Do not dump removed flags.
10471 (cgraph_function_body_availability): Update.
10472 (cgraph_propagate_frequency): Update.
10473 (verify_cgraph_node): Check sanity of local flag.
10474 (cgraph_function_node): Move here from cgraph.h; revamp for
10475 cgraph_function_or_thunk_node.
10476 * lto-symtab.c (lto_varpool_replace_node): Update.
10477 (lto_symtab_resolve_can_prevail_p): Update.
10478 (lto_symtab_merge_cgraph_nodes): Update.
10479 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10480 propagate_constants_accross_call, devirtualization_time_bonus,
10481 ipcp_propagate_stage): Update.
10482 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
10483 * ipa-inline-transform.c (clone_inlined_nodes,
10484 preserve_function_body_p): Update.
10485 * ipa-reference.c (propagate): Update.
10486 (write_node_summary_p): Update.
10487 * toplev.c (wrapup_global_declaration_2): Update.
10488 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10489 (analyze_function) ... this one.
10490 (cgraph_process_new_functions): Update.
10491 (cgraph_reset_node): Export.
10492 (cgraph_finalize_function): Update.
10493 (cgraph_add_new_function): Update.
10494 (process_function_and_variable_attributes): Update.
10495 (varpool_finalize_decl): Update.
10496 (symbol_finalized): Remove.
10497 (symbol_finalized_and_needed): Rename to ...
10498 (symbol_defined_and_needed): ... update.
10499 (cgraph_analyze_functions): Update.
10500 (handle_alias_pairs): Update.
10501 (mark_functions_to_output): Update.
10502 (assemble_thunk): Update.
10503 (output_in_order): Update.
10504 (output_weakrefs): Update.
10505 (finalize_compilation_unit): Update.
10506 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10507 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10508 input_node, input_varpool_node): Update.
10509 * dbxout.c (dbxout_expand_expr): Update.
10510 * cgraphclones.c (cgraph_clone_node): Update.
10511 (cgraph_copy_node_for_versioning): Update.
10512 (cgraph_materialize_clone): Update.
10513 (cgraph_materialize_all_clones): Update.
10514 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10515 propagate_pure_const, propagate_nothrow): Update.
10516 * lto-streamer-out.c (lto_output, write_symbol): Update.
10517 * ipa-utils.c (ipa_reverse_postorder): Update.
10518 * ipa-inline.c (can_inline_edge_p): Update.
10519 (update_caller_keys, ipa_inline): Update.
10520 * dwarf2out.c (reference_to_unused,
10521 premark_types_used_by_global_vars_helper): Update.
10522 * tree-eh.c (tree_could_trap_p): Update.
10523 * ipa-split.c (consider_split, execute_split_functions): Update.
10524 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
10525 has_addr_references_p): Update; move ahead in file for better
10527 (process_references): Simplify.
10528 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10529 bodies are removed.
10530 (cgraph_comdat_can_be_unshared_p): Make static.
10531 (cgraph_externally_visible_p): Update.
10532 (varpool_externally_visible_p): Update.
10533 (function_and_variable_visibility): Update.
10534 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10535 ipa_tm_mark_force_output_node): Update.
10536 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10537 estimate_edge_devirt_benefit, inline_generate_summary,
10538 inline_write_summary): Update.
10539 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10540 * ipa-prop.c (ipa_compute_jump_functions): Update.
10541 (ipa_print_node_params, ipa_prop_read_section,
10542 ipa_update_after_lto_read, read_replacements_section): Update.
10543 * varasm.c (mark_decl_referenced): Update.
10544 (assemble_alias, dump_tm_clone_pairs): Update.
10545 * tree-inline.c (copy_bb): Update.
10546 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10548 * symtab.c (dump_symtab_base): Print new flags.
10549 (verify_symtab_base): Verify new flags.
10550 (symtab_alias_ultimate_target): New function.
10551 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10552 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10554 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10556 * i386.c (ix86_get_function_versions_dispatcher,
10557 ix86_generate_version_dispatcher_body): Update.
10558 (fold_builtin_cpu): Use varpool_add_new_variable.
10559 * varpool.c (varpool_remove_initializer): Break out from ...
10560 (varpool_remove_node): ... this one.
10561 (dump_varpool_node, varpool_node_for_asm,
10562 cgraph_variable_initializer_availability, varpool_analyze_node,
10563 varpool_assemble_decl, varpool_remove_unreferenced_decls,
10564 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10567 2013-05-29 Jan Hubicka <jh@suse.cz>
10569 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10571 2013-05-29 Easwaran Raman <eraman@google.com>
10573 PR tree-optimization/57442
10574 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10575 when control exits the main loop.
10577 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10579 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10581 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
10582 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10583 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10586 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10588 PR tree-optimization/57441
10589 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10590 Don't limit size of incr_vec to number of candidates.
10592 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
10594 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10595 and mips16 directories.
10596 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
10597 (MULTILIB_DIRNAMES): Ditto.
10598 (MULTILIB_EXCEPTIONS): Add new exceptions.
10599 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10600 (MULTILIB_DIRNAMES): Ditto.
10601 (MULTILIB_EXCEPTIONS): Add new exceptions.
10603 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10604 Marcus Shawcroft <marcus.shawcroft@arm.com>
10606 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10607 SYMBOL_TINY_ABSOLUTE.
10608 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10609 SYMBOL_TINY_ABSOLUTE.
10610 (aarch64_expand_mov_immediate): Likewise.
10611 (aarch64_classify_symbol): Likewise.
10612 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10613 Permit SYMBOL_TINY_ABSOLUTE.
10614 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10616 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10617 Marcus Shawcroft <marcus.shawcroft@arm.com>
10619 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10620 Refactor if/switch. Replace gcc_assert with if.
10622 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10624 * config/i386/i386.c (initial_ix86_tune_features): Enable
10625 FP Reassociation for AMD bdver1 and bdver2.
10627 2013-05-29 Martin Jambor <mjambor@suse.cz>
10629 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10630 and IMAGPART_EXPR do not occur within other handled_components.
10632 2013-05-29 Richard Biener <rguenther@suse.de>
10634 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10635 access on whether the use is in the BB we currently try to
10637 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10638 vectorize to vect_bb_slp_scalar_cost.
10640 2013-05-29 Richard Biener <rguenther@suse.de>
10642 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10643 computing scalar cost offsetted by stmts that are kept live
10645 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10646 for computation of scalar cost.
10648 2013-05-28 Steve Ellcey <sellcey@mips.com>
10650 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10652 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10654 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10655 array notation built-in reduction functions.
10656 * doc/passes.texi (Passes): Added documentation about changes done
10658 * doc/invoke.texi (C Dialect Options): Added documentation about
10659 the -fcilkplus flag.
10660 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10661 (BUILTINS_DEF): Depend on cilkplus.def.
10662 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10663 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10664 * cilkplus.def: New file.
10666 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
10668 PR rtl-optimization/57439
10669 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10671 2013-05-28 Easwaran Raman <eraman@google.com>
10673 PR tree-optimization/57337
10674 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10675 (find_insert_point): Correctly identify the insertion point
10676 when two statements with the same UID is compared.
10678 2013-05-28 Richard Biener <rguenther@suse.de>
10680 PR tree-optimization/56787
10681 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10682 from the list of data references.
10683 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10685 (vect_analyze_loop_operations): Likewise.
10686 (vect_transform_loop): Remove clobbers.
10688 2013-05-28 Martin Jambor <mjambor@suse.cz>
10690 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10691 and REALPART_EXPRs have scalar type.
10693 2013-05-28 Richard Biener <rguenther@suse.de>
10695 PR tree-optimization/57411
10696 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10698 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10699 virtual operand propagation.
10701 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10703 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10704 destination register for bmasksi_vis.
10705 (vector_init_bshuffle): Likewise.
10706 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10708 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10710 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10711 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10712 mode if the instruction isn't available in the original mode.
10713 * config/sparc/sparc.opt (mfix-ut699): New option.
10714 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10715 (divdf3): Turn into expander.
10716 (divdf3_nofix): New insn.
10717 (divdf3_fix): Likewise.
10718 (divsf3): Disable if -mfix-ut699.
10719 (sqrtdf2): Turn into expander.
10720 (sqrtdf2_nofix): New insn.
10721 (sqrtdf2_fix): Likewise.
10722 (sqrtsf2): Disable if -mfix-ut699.
10724 2013-05-27 Richard Biener <rguenther@suse.de>
10726 PR middle-end/57412
10727 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10728 block for the new loop.
10730 2013-05-27 Richard Biener <rguenther@suse.de>
10732 PR tree-optimization/57343
10733 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10734 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10735 (number_of_iterations_cond): Do not build the folded tree.
10737 2013-05-27 Richard Biener <rguenther@suse.de>
10740 PR middle-end/57381
10741 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10742 OEP_CONSTANT_ADDRESS_OF retained.
10744 PR tree-optimization/57417
10745 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10746 for unchanged base.
10747 (set_ssa_val_to): Compare addresses using
10748 get_addr_base_and_unit_offset.
10750 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
10752 PR rtl-optimization/56833
10753 * postreload.c (move2add_record_mode): New function.
10754 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10755 (move2add_use_add2_insn): Use move2add_record_sym_value.
10756 (move2add_use_add3_insn): Likewise.
10757 (reload_cse_move2add): Use move2add_valid_value_p and
10758 move2add_record_mode. Invalidate call-clobbered and REG_INC
10759 affected regs by setting reg_mode to VOIDmode.
10760 (move2add_note_store): Don't pretend the inside of a SUBREG is
10761 the actual destination. Invalidate single/leading registers by
10762 setting reg_mode to VOIDmode.
10763 Use move2add_record_sym_value, move2add_valid_value_p and
10764 move2add_record_mode.
10766 2013-05-27 Richard Biener <rguenther@suse.de>
10768 PR tree-optimization/57396
10769 * tree-affine.c (double_int_constant_multiple_p): Properly
10770 return false for val == 0 and div != 0.
10772 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10774 * config/mips/mips.h: Use #elif in preprocessor conditions.
10776 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10779 * config/mips/constraints.md (kl): New constraint.
10780 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10781 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10782 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10783 from LO for MIPS16.
10784 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10786 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10789 * config/mips/mips.c (mips_can_inline_p): New function.
10790 (TARGET_CAN_INLINE_P): Define.
10792 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10794 * sched-int.h (ds_t, dw_t): Make unsigned int.
10795 Fix documentation that describes how all the ds_t bits are used.
10796 Reserve the last bit for delayed-branch scheduling.
10797 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10798 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10799 (gen_dep_weak_1): Remove prototype.
10800 * sched-deps.c (get_dep_weak_1): Make static.
10801 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10802 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10803 * doc/tm.texi: Regenerate.
10804 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
10806 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
10809 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
10811 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
10812 Sandra Loosemore <sandra@codesourcery.com>
10814 * config.gcc (powerpc-*): Allow native for with-cpu.
10816 2013-05-24 Jeff Law <law@redhat.com>
10818 PR tree-optimization/57124
10819 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
10820 conversion feeding a condition if the range has an overflow
10821 if -fstrict-overflow. Add warnings for when we do make the
10824 2013-05-24 Dehao Chen <dehao@google.com>
10826 * tree-cfg.c (locus_discrim_map): Fix the typo.
10827 (locus_discrim_hasher): Likewise.
10828 (locus_discrim_hasher::hash): Likewise.
10829 (locus_discrim_hasher::equal): Likewise.
10831 2013-05-24 Martin Jambor <mjambor@suse.cz>
10833 PR tree-optimization/57294
10834 * cgraph.h (ipa_record_stmt_references): Declare.
10835 * cgraphbuild.c (ipa_record_stmt_references): New function.
10836 (build_cgraph_edges): Use ipa_record_stmt_references.
10837 (rebuild_cgraph_edges): Likewise.
10838 (cgraph_rebuild_references): Likewise.
10839 * ipa-prop.c (ipa_modify_call_arguments): Discard references
10840 associated with the old statement and build references from the
10841 newly built statements.
10842 * ipa-ref.c (ipa_remove_stmt_references): New function.
10843 * ipa-ref.h (ipa_remove_stmt_references): Declare.
10845 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
10847 * lra-constraints.c (emit_spill_move): Use smaller mode for
10849 (check_and_process_move): Consider mem-reg moves for secondary
10851 (curr_insn_transform): Don't lose insns emitted before for
10852 secondary memory moves.
10853 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
10854 reg set up in the current insn.
10856 2013-05-24 Dehao Chen <dehao@google.com>
10858 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
10860 (locus_descrim_hasher::equal): Likewise.
10861 (build_gimple_cfg): New discriminator assignment algorithm.
10862 (make_edges): Likewise.
10863 (next_discriminator_for_locus): Likewise.
10864 (same_line_p): Likewise.
10865 (assign_discriminators): Likewise.
10866 (make_cond_expr_edges): Likewise.
10867 (make_gimple_switch_edges): Likewise.
10868 (make_goto_expr_edges): Likewise.
10869 (make_gimple_asm_edges): Likewise.
10871 2013-05-24 Ian Bolton <ian.bolton@arm.com>
10873 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
10874 X format specifier to only display bottom 16 bits.
10875 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
10876 immediate to match for operand 2, since it will be masked.
10878 2013-05-24 Richard Biener <rguenther@suse.de>
10880 PR tree-optimization/57287
10881 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
10882 all SSA names that occur in abnormal PHIs.
10884 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
10887 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
10888 that index is not negative.
10890 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
10892 PR rtl-optimization/55177
10893 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
10894 (simplify_byte_swapping_operation): New.
10895 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
10896 (simplify_relational_operation_1): Deal with BSWAP.
10898 2013-05-23 Richard Henderson <rth@redhat.com>
10901 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
10902 (ix86_reorg): Call it.
10904 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
10907 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
10908 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
10909 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
10911 2013-05-23 Christian Bruel <christian.bruel@st.com>
10914 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
10916 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10917 Marcus Shawcroft <marcus.shawcroft@arm.com>
10919 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
10920 * config/aarch64/constraints.md (Usa): Remove.
10921 * doc/md.texi (AArch64 Usa): Remove.
10923 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
10924 Marcus Shawcroft <marcus.shawcroft@arm.com>
10926 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
10927 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
10928 * config/aarch64/predicates.md (aarch64_const_address): Remove.
10929 (aarch64_mov_operand): Use aarch64_mov_operand_p.
10931 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
10933 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
10934 instruction (AdvSIMD).
10935 * config/aarch64/aarch64-builtins.c
10936 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
10937 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
10939 2013-05-23 Martin Jambor <mjambor@suse.cz>
10941 PR middle-end/57347
10942 * tree.h (contains_bitfld_component_ref_p): Declare.
10943 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
10944 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
10946 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
10947 not access a bit-field. Assert all final offsets are byte-aligned.
10949 2013-05-23 Richard Biener <rguenther@suse.de>
10951 PR tree-optimization/57380
10952 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
10953 least one invariant or re-used load.
10954 * passes.c (init_optimization_passes): Move pass_phiprop before
10957 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
10959 * config/aarch64/aarch64-simd.md
10960 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
10962 2013-05-23 Richard Biener <rguenther@suse.de>
10964 PR middle-end/57381
10965 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10966 OEP_CONSTANT_ADDRESS_OF retained.
10968 2013-05-23 Jakub Jelinek <jakub@redhat.com>
10970 PR middle-end/57344
10971 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
10972 don't lower unit. Handle unit not being always BITS_PER_WORD.
10974 2013-05-23 Richard Biener <rguenther@suse.de>
10976 PR rtl-optimization/57341
10977 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
10978 instead of true_dependence.
10980 2013-05-22 David Malcolm <dmalcolm@redhat.com>
10982 * bb-reorder.c (branch_threshold): Make const.
10983 (exec_threshold): Ditto.
10985 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
10986 Pat Haugen <pthaugen@us.ibm.com>
10987 Peter Bergner <bergner@vnet.ibm.com>
10989 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
10990 documentation for the power8 crypto builtins.
10992 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
10994 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
10995 macros for defining power8 builtin functions.
10996 (BU_P8V_AV_2): Likewise.
10997 (BU_P8V_AV_P): Likewise.
10998 (BU_P8V_VSX_1): Likewise.
10999 (BU_P8V_OVERLOAD_1): Likewise.
11000 (BU_P8V_OVERLOAD_2): Likewise.
11001 (BU_CRYPTO_1): Likewise.
11002 (BU_CRYPTO_2): Likewise.
11003 (BU_CRYPTO_3): Likewise.
11004 (BU_CRYPTO_OVERLOAD_1): Likewise.
11005 (BU_CRYPTO_OVERLOAD_2): Likewise.
11006 (XSCVSPDP): Fix typo, point to the correct instruction.
11007 (VCIPHER): Add power8 crypto builtins.
11008 (VCIPHERLAST): Likewise.
11009 (VNCIPHER): Likewise.
11010 (VNCIPHERLAST): Likewise.
11011 (VPMSUMB): Likewise.
11012 (VPMSUMH): Likewise.
11013 (VPMSUMW): Likewise.
11014 (VPERMXOR_V2DI): Likewise.
11015 (VPERMXOR_V4SI: Likewise.
11016 (VPERMXOR_V8HI: Likewise.
11017 (VPERMXOR_V16QI: Likewise.
11018 (VSHASIGMAW): Likewise.
11019 (VSHASIGMAD): Likewise.
11020 (VPMSUM): Likewise.
11021 (VPERMXOR): Likewise.
11022 (VSHASIGMA): Likewise.
11024 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11025 __CRYPTO__ if the crypto instructions are available.
11026 (altivec_overloaded_builtins): Add support for overloaded power8
11029 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11030 support for power8 crypto builtins.
11031 (builtin_function_type): Likewise.
11032 (altivec_init_builtins): Add support for builtins that take vector
11033 long long (V2DI) arguments.
11035 * config/rs6000/crypto.md: New file, define power8 crypto
11038 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11039 Pat Haugen <pthaugen@us.ibm.com>
11040 Peter Bergner <bergner@vnet.ibm.com>
11042 * doc/invoke.texi (Option Summary): Add power8 options.
11043 (RS/6000 and PowerPC Options): Likewise.
11045 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11046 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11047 wm, wn, wr documentation.
11049 * config/rs6000/constraints.md (wm): New constraint for VSX
11050 registers if direct move instructions are enabled.
11051 (wn): New constraint for no registers.
11052 (wq): New constraint for quad word even GPR registers.
11053 (wr): New constraint if 64-bit instructions are enabled.
11054 (wv): New constraint if power8 vector instructions are enabled.
11055 (wQ): New constraint for quad word memory locations.
11057 * config/rs6000/predicates.md (const_0_to_15_operand): New
11058 constraint for 0..15 for crypto instructions.
11059 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11060 as GPR and floating point registers.
11061 (int_reg_operand): New predicate to match only GPR registers.
11062 (base_reg_operand): New predicate to match base registers.
11063 (quad_int_reg_operand): New predicate to match even GPR registers
11064 for quad memory operations.
11065 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11066 operations in both GPR and VSX registers.
11067 (quad_memory_operand): New predicate for quad memory operations.
11068 (reg_or_indexed_operand): New predicate for direct move support.
11070 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
11071 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11072 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11073 (POWERPC_MASKS): Add power8 options.
11074 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11077 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11078 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11080 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
11081 (-mpower8-fusion): New power8 options.
11082 (-mpower8-fusion-sign): Likewise.
11083 (-mpower8-vector): Likewise.
11084 (-mcrypto): Likewise.
11085 (-mdirect-move): Likewise.
11086 (-mquad-memory): Likewise.
11088 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
11089 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11091 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
11092 (rs6000_debug_vector_unit): Add p8_vector.
11093 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11094 definitions. Also print fusion state.
11095 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11096 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11097 (rs6000_option_override_internal): Add support for power8.
11098 (rs6000_common_init_builtins): Add debugging for skipped builtins
11099 if -mdebug=builtin.
11100 (rs6000_adjust_cost): Add power8 support.
11101 (rs6000_issue_rate): Likewise.
11102 (insn_must_be_first_in_group): Likewise.
11103 (insn_must_be_last_in_group): Likewise.
11104 (force_new_group): Likewise.
11105 (rs6000_register_move_cost): Likewise.
11106 (rs6000_opt_masks): Likewise.
11108 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11109 power8 capable assembler, default to power7 options.
11110 (TARGET_DIRECT_MOVE): Likewise.
11111 (TARGET_CRYPTO): Likewise.
11112 (TARGET_P8_VECTOR): Likewise.
11113 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11114 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11115 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11116 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11117 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11118 (TARGET_XSCVDPSPN): Likewise.
11119 (TARGET_XSCVSPDPN): Likewsie.
11120 (TARGET_SYNC_HI_QI): Likewise.
11121 (TARGET_SYNC_TI): Likewise.
11122 (MASK_CRYPTO): Likewise.
11123 (MASK_DIRECT_MOVE): Likewise.
11124 (MASK_P8_FUSION): Likewise.
11125 (MASK_P8_VECTOR): Likewise.
11126 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11127 temporary used by some of the direct move instructions to get two FP
11128 temporary registers does not force creation of a stack frame.
11129 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11130 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11131 that any VSX registers are tieable, even if they are also an
11132 Altivec vector mode.
11133 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11134 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11135 (RS6000_BTM_CRYPTO): Likewise.
11136 (RS6000_BTM_COMMON): Likewise.
11138 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11139 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11140 (enum rs6000_vector): Add power8 vector support.
11142 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11146 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11147 (any_sibcall_could_use_r3): this and handle indirect calls.
11148 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11150 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11152 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11154 2013-05-22 Richard Biener <rguenther@suse.de>
11156 PR middle-end/57349
11157 * profile.c (branch_prob): Do not split blocks that are
11158 abnormally receiving from ECF_RETURNS_TWICE functions.
11160 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11162 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11163 address mode. Move pointer mode initialization to the same place.
11165 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11167 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11168 while it has any effect.
11170 2013-05-21 Easwaran Raman <eraman@google.com>
11172 PR tree-optimization/57322
11173 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11174 UID of the statement added to the BB to be 1.
11176 2013-05-21 Jakub Jelinek <jakub@redhat.com>
11178 PR tree-optimization/57331
11179 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11180 of conversion from pointer type to integral type with integer.
11182 2013-05-21 Martin Jambor <mjambor@suse.cz>
11185 * ipa-prop.c (ipa_read_node_info): Process param_used and
11186 controlled_uses in the same order as when writing.
11188 2013-05-21 Magnus Granberg <baldrick@free.fr>
11191 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
11193 2013-05-21 Richard Biener <rguenther@suse.de>
11195 PR tree-optimization/57318
11196 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11197 estimate stmts with side-effects as likely eliminated.
11199 2013-05-21 Richard Biener <rguenther@suse.de>
11201 PR tree-optimization/57330
11202 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11203 preserve the call stmts fntype.
11205 2013-05-21 Richard Biener <rguenther@suse.de>
11207 PR tree-optimization/57303
11208 * tree-ssa-sink.c (statement_sink_location): Improve killing
11209 stmt detection and properly handle self-assignments.
11211 2013-05-21 Christian Bruel <christian.bruel@st.com>
11213 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11214 spanning registers. LEAF_REG_REMAP is supported only for contiguous
11215 registers. Set register size out of the PARALLEL loop.
11217 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11220 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11221 (*fmasf4, *fmasf4_media): New insns.
11223 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11225 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11226 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11227 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11228 (mips_idiv_insns): Update the comments to say that the returned
11229 instruction counts are in units of BASE_INSN_LENGTH.
11230 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11231 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11232 using 2 rather than 4 as the length of indirect MIPS16 and
11233 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11234 length of a NOP. Don't divide MIPS16 lengths by 2.
11235 (mips16_split_long_branches): Assume a branch is long if the
11236 length is greater than 4 rather than 8.
11237 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11238 rather than multiplying them by 2. Multiply instruction counts
11239 by BASE_INSN_LENGTH rather than 4.
11240 (*jump_mips16, tls_get_tp_mips16_<mode>)
11241 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11243 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11245 * config/mips/mips.md (extended_mips16): Remove branch case.
11246 (length): Remove duplicated extended_mips16 test.
11248 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11250 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11252 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11254 * recog.h (Recog_data): Rename to...
11255 (recog_data_d): ...this.
11256 (recog_data): Update accordingly.
11257 * recog.c (recog_data): Likewise.
11258 * reload.c (save_recog_data): Likewise.
11259 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11260 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11262 2013-05-17 Julian Brown <julian@codesourcery.com>
11264 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11265 found in a REG_EQUAL note, invalidate it.
11267 2013-05-17 Easwaran Raman <eraman@google.com>
11269 * tree-ssa-reassoc.c (find_insert_point): New function.
11270 (insert_stmt_after): Likewise.
11271 (get_def_stmt): Likewise.
11272 (ensure_ops_are_available): Likewise.
11273 (not_dominated_by): Likewise.
11274 (rewrite_expr_tree): Do not move statements beyond what is
11275 necessary. Remove call to swap_ops_for_binary_stmt...
11276 (reassociate_bb): ... and move it here.
11277 (build_and_add_sum): Assign UIDs for new statements.
11278 (linearize_expr): Likewise.
11279 (do_reassoc): Renumber gimple statement UIDs.
11281 2013-05-17 Jan Hubicka <jh@suse.cz>
11283 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11285 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
11286 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11288 (output_weakrefs): Update.
11290 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11291 Martin Jambor <mjambor@suse.cz>
11293 PR middle-end/57276
11294 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11295 value that corresponds to the given aggval is found in values vector.
11297 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
11299 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11300 sse, sse2, sse3, ssse3 and sse4a flags to options.
11302 2013-05-17 David Malcolm <dmalcolm@redhat.com>
11304 * gengtype-state.c: (s_expr_writer): New class, to handle
11305 prettifying of output layout of s-expressions.
11306 (state_writer): New class, to write out gtype.state.
11307 (state_written_type_count): Move this variable into member data of
11309 (s_expr_writer::s_expr_writer): New code: constructor for new class
11310 (state_writer::state_writer(): ditto
11311 (s_expr_writer::write_new_line): New function
11312 (s_expr_writer::write_any_indent): ditto
11313 (s_expr_writer::begin_s_expr): ditto
11314 (s_expr_writer::end_s_expr): ditto
11315 (write_state_fileloc): convert to method of state_writer...
11316 (state_writer:: write_state_fileloc): ...and use methods of
11317 s_expr_writer to write indentation into the gtype.state output file
11318 to visually represent the hierarchical structure of the list
11320 (write_state_fields): ditto, renaming to...
11321 (state_writer::write_state_fields)
11322 (write_state_a_string): ditto, renaming to...
11323 (state_writer::write_state_a_string)
11324 (write_state_string_option): ditto, renaming to...
11325 (state_writer::write_state_string_option)
11326 (write_state_type_option): ditto, renaming to...
11327 (state_writer::write_state_type_option)
11328 (write_state_nested_option): ditto, renaming to...
11329 (state_writer::write_state_nested_option)
11330 (write_state_option): ditto, renaming to...
11331 (state_writer::write_state_option)
11332 (write_state_options): ditto, renaming to...
11333 (state_writer::write_state_options)
11334 (write_state_lang_bitmap): ditto, renaming to...
11335 (state_writer::write_state_lang_bitmap)
11336 (write_state_version): ditto, renaming to...
11337 (state_writer::write_state_version)
11338 (write_state_scalar_type): ditto, renaming to...
11339 (state_writer::write_state_scalar_type)
11340 (write_state_string_type): ditto, renaming to...
11341 (state_writer::write_state_string_type)
11342 (write_state_undefined_type): ditto, renaming to...
11343 (state_writer::write_state_undefined_type)
11344 (write_state_struct_union_type): ditto, renaming to...
11345 (state_writer::write_state_struct_union_type)
11346 (write_state_struct_type): ditto, renaming to...
11347 (state_writer::write_state_struct_type)
11348 (write_state_user_struct_type): ditto, renaming to...
11349 (state_writer::write_state_user_struct_type)
11350 (write_state_lang_struct_type): ditto, renaming to...
11351 (state_writer::write_state_lang_struct_type)
11352 (write_state_param_struct_type): ditto, renaming to...
11353 (state_writer::write_state_param_struct_type)
11354 (write_state_pointer_type): ditto, renaming to...
11355 (state_writer::write_state_pointer_type)
11356 (write_state_array_type): ditto, renaming to...
11357 (state_writer::write_state_array_type)
11358 (write_state_gc_used): ditto, renaming to...
11359 (state_writer::write_state_gc_used)
11360 (write_state_common_type_content): ditto, renaming to...
11361 (state_writer::write_state_common_type_content)
11362 (write_state_type): ditto, renaming to...
11363 (state_writer::write_state_type)
11364 (write_state_pair_list): ditto, renaming to...
11365 (state_writer::write_state_pair_list)
11366 (write_state_pair): ditto, renaming to...
11367 (state_writer::write_state_pair)
11368 (write_state_typedefs): ditto, renaming to...
11369 (state_writer::write_state_typedefs)
11370 (write_state_structures): ditto, renaming to...
11371 (state_writer::write_state_structures)
11372 (write_state_param_structs): ditto, renaming to...
11373 (state_writer::write_state_param_structs)
11374 (write_state_variables): ditto, renaming to...
11375 (state_writer::write_state_variables)
11376 (write_state_srcdir): ditto, renaming to...
11377 (state_writer::write_state_srcdir)
11378 (write_state_files_list): ditto, renaming to...
11379 (state_writer::write_state_files_list)
11380 (write_state_languages): ditto, renaming to...
11381 (state_writer::write_state_languages)
11382 (write_state): create a state_writer instance and use it when
11383 writing out the state file
11385 2013-05-17 Mike Stump <mikestump@comcast.net>
11387 PR rtl-optimization/57304
11388 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11389 accessing DF_REF_REAL_LOC.
11391 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11393 PR rtl-optimization/57281
11394 PR rtl-optimization/57300
11395 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11396 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11397 what the other splitter did if the registers are dead.
11399 2013-05-17 Richard Biener <rguenther@suse.de>
11401 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11404 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11406 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11409 2013-05-17 Marek Polacek <polacek@redhat.com>
11411 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11412 length when doing non-zero store of storing '\0' to '\0'.
11414 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11416 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11417 vect_external_def oprnd1 with loop_vinfo, try to emit
11418 optional cast, negation and and stmts on the loop preheader
11419 edge instead of into the pattern def seq.
11421 PR tree-optimization/57051
11422 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11423 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11425 2013-05-16 Nick Clifton <nickc@redhat.com>
11427 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11428 (rl78_is_naked_func): New function.
11429 (rl78_expand_prologue): Skip prologue generation for naked functions.
11430 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
11431 * doc/extend.texi (naked): Add RL78 to the list of processors
11432 that supports this attribute.
11434 2013-05-16 Jeff Law <law@redhat.com>
11436 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11438 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11440 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11441 cache parameters using detect_caches_amd also for CYRIX,
11442 NSC and TM2 signatures.
11444 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11445 Dzianis Kahanovich <mahatma@eu.by>
11449 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11450 VIA/Centaur processors and determine their cache parameters
11451 using detect_caches_amd.
11453 2013-05-16 Teresa Johnson <tejohnson@google.com>
11455 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11456 (rtl_verify_edges): New function.
11457 (rtl_verify_bb_insns): Ditto.
11458 (rtl_verify_bb_pointers): Ditto.
11459 (rtl_verify_bb_insn_chain): Ditto.
11460 (rtl_verify_fallthru): Ditto.
11461 (rtl_verify_bb_layout): Ditto.
11462 (rtl_verify_flow_info_1): Outline checks into new functions.
11463 (rtl_verify_flow_info): Ditto.
11465 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
11467 * cfghooks.c (copy_bbs): Add update_dominance argument.
11468 * cfghooks.h (copy_bbs): Update prototype.
11469 * tree-cfg.c (gimple_duplicate_sese_region):
11470 Add update_dominance argument.
11471 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11472 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11473 gimple_duplicate_sese_region call.
11474 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11475 Update copy_bbs call.
11476 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11477 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11479 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11481 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11482 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11483 vect_recog_rotate_pattern.
11484 (vect_recog_rotate_pattern): New function.
11486 2013-05-16 Jason Merrill <jason@redhat.com>
11488 * Makefile.in (LLINKER): New variable.
11489 (mostlyclean): Remove link mutex.
11490 * configure.ac: Handle --enable-link-mutex.
11491 * lock-and-run.sh: New script.
11493 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11496 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11499 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11501 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11503 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11505 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11506 * config/arm/arm.c (next_consecutive_mem): New function.
11507 (gen_movmem_ldrd_strd): Likewise.
11508 * config/arm/arm.md (movmemqi): Update condition and code.
11509 (unaligned_loaddi, unaligned_storedi): New patterns.
11511 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11513 * config.gcc: Obsolete *-*-solaris2.9*.
11514 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11516 2013-05-16 Richard Biener <rguenther@suse.de>
11518 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11519 earlier, after GRAPHITE transforms and IV canonicalization.
11521 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11523 * omp-low.c (extract_omp_for_data): For collapsed loops,
11524 if at least one of the loops is known at compile time to
11525 iterate zero times, set count to 0.
11526 (expand_omp_regimplify_p): New function.
11527 (expand_omp_for_generic): For collapsed loops, if at least
11528 one of the loops isn't known to iterate at least once,
11529 add runtime check with setting count to 0.
11530 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11531 For unsigned types if it isn't known at compile time that
11532 the loop will iterate at least once, add runtime check to bypass
11533 the whole loop if initial condition isn't true.
11535 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11537 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11539 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
11541 PR middle-end/57286
11542 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11543 transformations to avoid an infinite loop.
11545 2013-05-16 Marek Polacek <polacek@redhat.com>
11547 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11549 2013-05-15 Leif Ekblad <leif@rdos.net>
11551 * config/i386/i386.c (ix86_decompose_address): Use
11552 DEFAULT_TLS_SEG_REG to access TLS segment register.
11553 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11554 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11555 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11557 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11560 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11561 sibling calls to functions that would normally be lazily bound,
11562 unless $gp is call-clobbered.
11564 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
11566 * config/i386/i386.c (ix86_option_override_internal): Update
11567 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
11568 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11569 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
11570 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11572 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11574 2013-05-15 Andreas Schwab <schwab@suse.de>
11576 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11577 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11580 2013-05-15 Teresa Johnson <tejohnson@google.com>
11582 * loop-unroll.c (report_unroll_peel): Check decision before
11583 emitting unroll/peel message.
11585 2013-05-15 Teresa Johnson <tejohnson@google.com>
11587 * function.h (has_bb_partition): New rtl_data flag.
11588 (bb_reorder_complete): Ditto.
11589 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11590 instead of flag_reorder_blocks_and_partition.
11591 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11592 with some enhancements.
11593 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11594 * bb-reorder.c (connect_traces): Check for has_bb_partition
11595 instead of flag_reorder_blocks_and_partition.
11596 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11597 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11598 verify_hot_cold_block_grouping.
11599 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11601 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11604 * config/arm/predicates.md (call_insn_operand): New predicate.
11605 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11606 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11607 if insn is not a tail call.
11608 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11610 * config/arm/arm.h (enum reg_class): New caller save register class.
11611 (REG_CLASS_NAMES): Likewise.
11612 (REG_CLASS_CONTENTS): Likewise.
11613 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11616 2013-05-15 Richard Biener <rguenther@suse.de>
11618 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11619 of MSG_OPTIMIZED_LOCATIONS.
11620 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11621 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11623 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11624 of MSG_OPTIMIZED_LOCATIONS.
11625 (execute_vect_slp): Likewise.
11626 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11627 (vect_create_cond_for_alias_checks): Likewise.
11628 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11629 (vect_recog_widen_mult_pattern): Likewise.
11630 (vect_recog_widen_sum_pattern): Likewise.
11631 (vect_recog_over_widening_pattern): Likewise.
11632 (vect_recog_widen_shift_pattern): Likewise.
11633 (vect_recog_vector_vector_shift_pattern): Likewise.
11634 (vect_recog_divmod_pattern): Likewise.
11635 (vect_recog_mixed_size_cond_pattern): Likewise.
11636 (vect_recog_bool_pattern): Likewise.
11637 (vect_pattern_recog_1): Likewise.
11639 2013-05-15 Martin Jambor <mjambor@suse.cz>
11641 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11642 non-functions to builtin_unreachable.
11643 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11644 correct when new direct edges were discovered.
11646 2013-05-15 Martin Jambor <mjambor@suse.cz>
11648 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11649 header, print symbol order instead of node uid, print more information
11650 about indirect edge targets.
11651 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11653 (ipa_make_edge_direct_to_target): Likewise.
11654 (remove_described_reference): Likewise.
11655 (propagate_controlled_uses): Likewise.
11656 (ipa_print_node_params): Also print symbol order.
11657 (ipcp_transform_function): Print symbol order instead of node uids.
11658 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11659 (cgraph_get_create_real_symbol_node): Likewise.
11660 * ipa-cp.c (print_lattice): Likewise.
11661 (print_all_lattices): Likewise.
11662 (determine_versionability): Likewise.
11663 (initialize_node_lattices): Likewise.
11664 (estimate_local_effects): Likewise.
11665 (update_profiling_info): Likewise.
11666 (create_specialized_node): Likewise.
11667 (perhaps_add_new_callers): Likewise.
11668 (decide_about_value): Likewise.
11669 (decide_whether_version_node): Likewise.
11670 (identify_dead_nodes): Likewise.
11671 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11672 (dump_inline_summary): Likewise.
11673 (estimate_node_size_and_time): Likewise.
11674 (inline_analyze_function): Likewise.
11675 * ipa-inline.c (report_inline_failed_reason): Likewise.
11676 (want_early_inline_function_p): Likewise.
11677 (edge_badness): Likewise.
11678 (update_edge_key): Likewise.
11679 (inline_small_functions): Likewise. Add dumping of order to two other
11681 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11682 instead of node uids.
11683 (propagate_pure_const): Likewise.
11684 (propagate_pure_const): Likewise.
11685 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11686 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11687 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11689 * tree-pretty-print.c (dump_function_header): Likewise.
11690 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11691 Print symbol order instead of node uids.
11693 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11695 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11696 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11698 2013-05-15 Richard Biener <rguenther@suse.de>
11700 PR tree-optimization/57275
11701 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11702 return value for fail to do runtime alias checks for gather loads.
11704 2013-05-15 Jan Hubicka <jh@suse.cz>
11708 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11709 weakrefs are not external.
11710 (lto_symtab_merge_decls): Fix thinko when dealing with
11711 non-lto_symtab decls.
11712 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11713 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11714 * varpool.c (dump_varpool_node): Dump more flags.
11716 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11718 * config/i386/i386.c (processor_alias_table): Add instruction
11719 FSGSBASE for AMD bdver3 architecture.
11721 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11723 * tree.c (warn_deprecated_use): Print file:line using locus color.
11724 * diagnostic.c (diagnostic_report_current_module): Print file:line
11725 and file:line:column using locus color.
11727 2013-05-14 Mike Stump <mikestump@comcast.net>
11729 * gdbinit.in: Add __null.
11731 2013-05-14 Mike Stump <mikestump@comcast.net>
11733 * recog.h: Rename struct recog_data to Recog_data.
11734 * recog.c: Likewise.
11735 * reload.c (can_reload_into): Likewise.
11736 * config/picochip/picochip.c: Likewise.
11738 2013-05-14 Mike Stump <mikestump@comcast.net>
11740 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11742 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11744 * resource.h (struct resources): Remove unch_memory member.
11745 (CLEAR_RESOURCE): Don't clear unch_memory.
11746 * resource.c (mark_referenced_resources): Don't set it.
11747 (mark_set_resources): Likewise.
11748 (mark_target_live_regs): Don't clear it.
11749 (init_resource_info): Likewise.
11750 * reorg.c (resource_conflicts_p): Don't compare it.
11751 (redundant_insn): Don't set it.
11753 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11755 * emit-rtl.c (next_label): Remove unused function.
11756 (skip_consecutive_labels, link_cc0_insns): Move to ...
11757 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11758 only place where these functions are used, and make them static.
11760 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11762 * fold-const.c (fold_negate_expr): Handle vectors.
11763 (fold_truth_not_expr): Make it static.
11764 (fold_invert_truthvalue): New static function.
11765 (invert_truthvalue_loc): Handle vectors. Do not call
11766 fold_truth_not_expr directly.
11767 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11768 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11769 (fold_comparison): Handle vector constants.
11770 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11771 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11772 * tree.h (fold_truth_not_expr): Remove declaration.
11774 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11776 * config/aarch64/aarch64-simd.md
11777 (aarch64_vcond_internal<mode>): Rename to...
11778 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11779 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11780 float modes. Clarify all iterator modes.
11781 (vcond<mode><mode>): Use new name for vcond expanders.
11782 (vcond<v_cmp_result><mode>): Likewise.
11783 (vcondu<mode><mode>: Likewise.
11784 * config/aarch64/iterators.md (VDQF_COND): New.
11786 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
11789 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11790 variable for the shift amount. Check that we shift by non-negative
11793 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11796 * config/arm/arm.h (EPILOGUE_USES): Only return true
11797 for LR_REGNUM after epilogue_completed.
11799 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
11801 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
11802 is error_mark_node.
11804 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11807 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
11808 and Solaris 11+/x86 with gld.
11809 * configure: Regenerate.
11811 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11813 * expmed.c (expand_shift_1): Canonicalize rotates by
11814 constant bitsize / 2 to bitsize - 1.
11815 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
11816 case ROTATERT>: Likewise.
11819 2013-05-10 Jakub Jelinek <jakub@redhat.com>
11821 * config/i386/i386.md (rotateinv): New code attr.
11822 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
11823 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
11824 roll $31, %eax, etc.
11826 2013-05-14 Richard Biener <rguenther@suse.de>
11828 PR middle-end/57235
11829 * tree-eh.c (sink_clobbers): Give up for successors with
11830 multiple predecessors and no virtual uses.
11832 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
11834 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
11835 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
11837 2013-05-14 Jakub Jelinek <jakub@redhat.com>
11839 PR middle-end/57251
11840 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
11841 the case when both op0 and op1 have VOIDmode.
11843 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11845 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
11846 in multiply-accumulate mode.
11848 2013-05-13 Guozhi Wei <carrot@google.com>
11850 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
11852 2013-05-13 Kai Tietz <ktietz@redhat.com>
11855 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
11856 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
11857 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
11858 * config/i386/i386.c (ix86_option_override_internal): Likewise.
11859 (ix86_expand_prologue): Likewise.
11860 (ix86_expand_split_stack_prologue): Likewise.
11861 (legitimate_pic_address_disp_p): Likewise.
11862 (legitimize_pic_address): Likewise.
11863 (legitimize_tls_address): Likewise.
11864 (legitimize_pe_coff_symbol): Likewise.
11865 (output_pic_addr_const): Likewise.
11866 (construct_plt_address): Likewise.
11867 (ix86_expand_call): Likewise.
11868 (x86_output_mi_thunk): Likewise.
11869 (x86_function_profiler): Likewise.
11871 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
11873 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
11874 similar switch cases.
11875 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
11876 (aarch64_simd_mov_to_<mode>low): Delete.
11877 (aarch64_simd_mov_to_<mode>high): Delete.
11878 (move_lo_quad_<mode>): Add w<-r alternative.
11879 (aarch64_simd_move_hi_quad_<mode>): Likewise.
11880 (aarch64_simd_mov_from_*): Update type attribute.
11881 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
11884 2013-05-13 Jan Hubicka <jh@suse.cz>
11886 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
11887 * config/i386/i386.c (ix86_compute_frame_layout,
11888 ix86_expand_epilogue, emit_i387_cw_initialization,
11889 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
11890 ix86_local_alignment): Fix use of size/speed predicates.
11892 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11894 PR tree-optimization/45216
11895 PR tree-optimization/57157
11896 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
11897 the (-Y) & (B - 1) variant if OP is |.
11898 * expmed.c (expand_shift_1): For rotations by const0_rtx just
11899 return shifted. Use (-op1) & (prec - 1) as other_amount
11900 instead of prec - op1.
11902 2013-05-13 Martin Jambor <mjambor@suse.cz>
11904 PR middle-end/42371
11905 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
11906 (ipa_constant_data): New type.
11907 (ipa_jump_func): Use ipa_constant_data to hold information about
11908 constant jump functions.
11909 (ipa_get_jf_constant): Adjust to jump function type changes.
11910 (ipa_get_jf_constant_rdesc): New function.
11911 (ipa_param_descriptor): New field controlled_uses.
11912 (ipa_get_controlled_uses): New function.
11913 (ipa_set_controlled_uses): Likewise.
11914 * ipa-ref.h (ipa_find_reference): Declare.
11915 * ipa-prop.c (ipa_cst_ref_desc): New type.
11916 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
11918 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
11919 New parameter cs. Adjust all callers.
11920 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
11921 (remove_described_reference): New function.
11922 (jfunc_rdesc_usable): Likewise.
11923 (try_make_edge_direct_simple_call): Decrement controlled use count,
11924 attempt to remove reference if it hits zero.
11925 (combine_controlled_uses_counters): New function.
11926 (propagate_controlled_uses): Likewise.
11927 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
11928 (ipa_edge_duplication_hook): Duplicate reference descriptions.
11929 (ipa_print_node_params): Print described use counter.
11930 (ipa_write_jump_function): Adjust to jump function type changes.
11931 (ipa_read_jump_function): New parameter CS, pass it to
11932 ipa_set_jf_constant. Adjust caller.
11933 (ipa_write_node_info): Stream controlled use count
11934 (ipa_read_node_info): Likewise.
11935 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
11937 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
11938 count. Remove cloning-added reference if it reaches zero.
11939 * ipa-ref.c (ipa_find_reference): New function.
11941 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11943 * config/i386/i386.c (processor_target_table): Modified default
11944 alignment values for AMD BD and BT architectures.
11946 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
11948 * tree-vect-generic.c (uniform_vector_p): Move ...
11949 * tree.c (uniform_vector_p): ... here.
11950 * tree.h (uniform_vector_p): Declare it.
11951 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
11954 2013-05-13 Jakub Jelinek <jakub@redhat.com>
11956 PR tree-optimization/57230
11957 * tree-ssa-strlen.c (handle_char_store): Record length for
11958 array store from STRING_CST.
11960 PR tree-optimization/57230
11961 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
11964 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
11966 * config/epiphany/epiphany.c (epiphany_init): Check size of
11967 NUM_MODES_FOR_MODE_SWITCHING.
11968 (epiphany_expand_prologue):
11969 Remove CONFIG_REGNUM initial value handling code.
11970 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
11971 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
11972 (emit_set_fp_mode, epiphany_mode_after): Likewise.
11973 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
11974 Don't return 1 for FP_MODE_NONE.
11975 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
11976 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
11977 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
11978 * config/epiphany/epiphany.md (save_config): New pattern.
11980 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
11982 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
11984 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
11986 * config/i386/i386.md (memory): Handle sseishft1.
11987 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
11988 (*vec_extractv2di_1): Ditto.
11990 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
11992 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
11995 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
11997 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
11998 Add mthumb/march=armv7-a multilib.
11999 Add mthumb/march=armv7-r multilib.
12000 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
12002 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
12004 * config/v850/t-rtems: Add more multilibs.
12006 2013-05-10 Richard Biener <rguenther@suse.de>
12008 PR tree-optimization/57214
12009 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
12010 not propagate from SSA names that occur in abnormal PHI nodes.
12012 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
12014 * stor-layout.c (element_precision): New function.
12015 * machmode.h (element_precision): Declare it.
12016 * tree.c (build_minus_one_cst): New function.
12017 (element_precision): Likewise.
12018 * tree.h (build_minus_one_cst): Declare new function.
12019 (element_precision): Likewise.
12020 * fold-const.c (operand_equal_p): Use element_precision.
12021 (fold_binary_loc): Handle vector types.
12022 * convert.c (convert_to_integer): Use element_precision.
12023 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12026 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12028 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12029 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12030 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12031 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12032 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12033 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12034 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12035 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12036 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12037 (Uuw8): New constraints.
12038 (Usb4): Move into alphabetical order.
12039 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12040 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12041 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12042 previously unnamed patterns.
12043 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12044 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12045 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12046 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12047 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12048 of specific lengths.
12050 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12052 * config/i386/i386.md (rotateinv): New code attr.
12053 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12054 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12055 roll $31, %eax, etc.
12057 PR tree-optimization/45216
12058 PR tree-optimization/57157
12059 * tree-ssa-forwprop.c (simplify_rotate): New function.
12060 (ssa_forward_propagate_and_combine): Call it.
12062 2013-05-10 Richard Biener <rguenther@suse.de>
12064 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12065 disable peeling when we version for aliasing.
12066 (vector_alignment_reachable_p): Honor explicit user alignment.
12067 (vect_supportable_dr_alignment): Likewise.
12068 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12069 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12070 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12071 then peeling to arrange for the cost-model check to come first.
12073 2013-05-10 Alan Modra <amodra@gmail.com>
12075 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12076 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12077 * configure: Regenerate.
12079 2013-05-10 Alan Modra <amodra@gmail.com>
12082 * varasm.c (default_elf_select_section): Move !DECL_P check..
12083 (get_named_section): ..to here before calling get_section_name.
12085 (default_section_type_flags): Add DECL_P check.
12086 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12087 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12089 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12091 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12092 When using gen_stack_adjust_str with a register offset, add a
12093 REG_FRAME_RELATED_EXPR note.
12095 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
12097 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12098 (*vec_extractv4si_zext_mem): Ditto.
12099 (*vec_extractv2di): Add 0->x and x->x alternatives.
12100 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12101 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12103 2013-05-09 Jason Merrill <jason@redhat.com>
12105 N3639 C++1y VLA support
12106 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12109 * tree.c (build_constructor_va): New.
12110 * tree.h: Declare it.
12112 2013-05-09 Martin Jambor <mjambor@suse.cz>
12115 * gimple-fold.c (canonicalize_constructor_val): Call
12116 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12118 2013-05-09 Jan Hubicka <jh@suse.cz>
12119 Richard Biener <rguenther@suse.de>
12122 * symtab.c (symtab_make_decl_local): Do not add private names.
12124 2013-05-09 Jan Hubicka <jh@suse.cz>
12127 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12128 (unlink_from_assembler_name_hash): Likewise.
12129 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12130 symtab_unregister_node, symtab_initialize_asm_name_hash,
12131 change_decl_assembler_name): Update.
12133 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12135 * config/aarch64/aarch64.md: New movtf split.
12136 (*movtf_aarch64): Update.
12137 (aarch64_movdi_tilow): Handle TF modes and rename to
12138 aarch64_movdi_<mode>low.
12139 (aarch64_movdi_tihigh): Handle TF modes and rename to
12140 aarch64_movdi_<mode>high
12141 (aarch64_movtihigh_di): Handle TF modes and rename to
12142 aarch64_mov<mode>high_di
12143 (aarch64_movtilow_di): Handle TF modes and rename to
12144 aarch64_mov<mode>low_di
12145 (aarch64_movtilow_tilow): Remove spurious whitespace.
12146 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12148 (aarch64_print_operand): Update.
12150 2013-05-09 Alan Modra <amodra@gmail.com>
12152 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12154 * configure: Regenerate.
12156 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12158 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12159 splitter preparation statements.
12160 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12161 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12164 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12166 * gimple-ssa-strength-reduction.c (count_candidates): Change
12167 return value to int.
12168 (analyze_candidates_and_replace): Change type of length to int.
12170 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12172 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12173 (*vec_extract<mode>): Use VI12_128 mode iterator.
12174 (*vec_extract<mode>_mem): Ditto.
12175 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12178 2013-05-08 Diego Novillo <dnovillo@google.com>
12183 2012-08-17 Diego Novillo <dnovillo@google.com>
12186 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12187 * config.in: Regenerate.
12188 * configure: Regenerate.
12189 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
12191 2013-05-08 Jan Hubicka <jh@suse.cz>
12194 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12195 * cgraph.h (symtab_node_base): Add unique_name.
12196 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12197 input_overwrite_node, input_varpool_node): Stream unique_name.
12198 * cgraphclones.c (cgraph_create_virtual_clone,
12199 cgraph_function_versioning): Set unique_name.
12200 * ipa.c (function_and_variable_visibility): Set unique_name.
12202 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12204 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
12205 (alloc_cand_and_find_basis): Restrict conditional candidate
12206 processing to CAND_MULTs.
12208 2013-05-08 Jan Hubicka <jh@suse.cz>
12211 lto-symtab.c (lto_symtab_symbol_p): New function.
12212 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
12213 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12214 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12215 Skip static symbols.
12217 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12219 PR tree-optimization/57200
12220 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12221 Only call inform if the preceding warning_at returns true.
12223 2013-05-07 Han Shen <shenhan@google.com>
12225 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12226 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12227 * common.opt (fstack-protector-strong): New option.
12228 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12229 * doc/invoke.texi (Optimization Options): Document
12230 "-fstack-protector-strong".
12231 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12233 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12235 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12237 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12239 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12240 it is not lhs of the insn.
12241 (setup_reg_equiv): Remove insn having equiv memory which it is not
12243 * lra-constraints.c (process_address): Try to improve generation
12244 code for address base + disp.
12245 (lra_constraints): Make correct the code for checking insn setting
12246 up backward equivalence. Remove insn only if it is in the init
12248 * lra-eliminations.c (update_reg_eliminate): Change return value.
12249 (lra_eliminate): Use the result.
12251 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
12253 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12254 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12255 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12256 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12257 register target operands.
12258 (*vec_extractv8hi_sse2): New pattern.
12259 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12260 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12261 (*vec_extract<mode>_mem): New insn and split pattern.
12263 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12265 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12266 (TARGET_ASAN_SHADOW_OFFSET): Define.
12267 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12268 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12270 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12272 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12273 (incr_vec_index): Return -1 if increment not found.
12274 (create_add_on_incoming_edge): Assert if increment not found.
12275 (record_increment): Limit number of increments recorded.
12276 (all_phi_incrs_profitable): Return false if an increment not found.
12277 (replace_profitable_candidates): Don't process increments that were
12279 (analyze_candidates_and_replace): Limit size of incr_vec.
12281 2013-05-07 Richard Biener <rguenther@suse.de>
12283 * calls.c (special_function_p): setjmp-like functions are leaf.
12284 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12285 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12287 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12289 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12290 (aarch64_simd_mov<mode>): New expander.
12291 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12292 (aarch64_simd_mov_to_<mode>high): Likewise.
12293 (aarch64_simd_mov_from_<mode>low): Likewise.
12294 (aarch64_simd_mov_from_<mode>high): Likewise.
12295 (aarch64_dup_lane<mode>): Update.
12296 (aarch64_dup_lanedi): New instruction pattern.
12297 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12298 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12300 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12302 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12303 (replace_mult_candidate): Remove unnecessary argument; remove
12304 unnecessary parameter from call to introduce_cast_before_cand.
12305 (replace_unconditional_candidate): Remove unnecessary parameter
12306 from call to replace_mult_candidate.
12307 (replace_conditional_candidate): Likewise.
12308 (insert_initializers): Use make_temp_ssa_name.
12309 (introduce_cast_before_cand): Remove unnecessary argument; use
12310 make_temp_ssa_name.
12311 (replace_one_candidate): Remove unnecessary argument; remove
12312 unnecessary parameter from calls to introduce_cast_before_cand.
12313 (replace_profitable_candidates): Remove unnecessary parameters
12314 from calls to replace_one_candidate.
12316 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12318 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12319 phi def as possibly hiding a basis for a CAND_ADD whose operands
12320 have been commuted in the analysis.
12321 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12323 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12325 * config/aarch64/aarch64.md
12326 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12327 shift value between 0-4.
12329 2013-05-07 Richard Biener <rguenther@suse.de>
12331 * double-int.h (rshift): New overload.
12332 * double-int.c (rshift): New function.
12333 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12334 (create_reference_ops_from_ref): Remove.
12335 (vn_reference_insert): Use shared ops for constructing the
12336 reference and copy it.
12338 2013-05-07 Richard Biener <rguenther@suse.de>
12340 PR middle-end/57190
12341 * tree-eh.c (sink_clobbers): Properly propagate
12342 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12344 2013-05-07 Jakub Jelinek <jakub@redhat.com>
12346 PR tree-optimization/57149
12347 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12348 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12349 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12350 uninit_undefined_value_p instead of ssa_undefined_value_p.
12353 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12354 for modifier == EXPAND_INITIALIZER.
12356 2013-05-07 Anton Blanchard <anton@samba.org>
12358 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12359 for powerpc64 little endian.
12360 * configure: Regenerate.
12362 2013-05-06 Graham Stott <grahams@btinternet.com>
12364 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12365 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12366 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12367 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12369 2013-05-06 Graham Stott <grahams@btinternet.com>
12371 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12372 codes which allow non-lvalues.
12374 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
12376 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12377 components are all 1s.
12378 (integer_minus_onep): New function.
12379 * tree.h (integer_minus_onep): Declare it.
12380 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12381 integer_minus_onep instead of integer_all_onesp.
12383 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12386 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12387 variations of these patterns.
12389 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12391 * config/i386/i386.md (isa): Add x64_sse4 member.
12392 (enabled): Handle x64_sse4.
12393 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12394 instruction for 64bit SSE4_1 targets. Update insn attributes.
12395 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12396 instruction for SSE4_1 targets. Update insn attributes.
12397 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12398 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12399 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12401 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12402 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12403 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12405 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12408 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12410 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12412 * final.c (do_assembler_dialects): Don't handle curly braces and
12413 vertical bar escaped by % as dialect delimiters.
12414 (output_asm_insn): Print curly braces and vertical bar if escaped
12415 by % and ASSEMBLER_DIALECT defined.
12416 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12417 * doc/tm.texi: Regenerated.
12419 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12421 * config/mips/mips.c: Include tree-pass.h.
12422 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12423 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12424 (pass_mips_machine_reorg2): New machine specific pass.
12425 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12426 (mips_option_override): Register the new pass.
12427 * rtl.h (cleanup_barriers): Remove prototype.
12428 (dbr_schedule): Likewise.
12429 * jump.c (cleanup_barriers): Make static.
12430 * reorg.c (dbr_schedule): Likewise.
12432 2013-05-06 Richard Biener <rguenther@suse.de>
12434 PR tree-optimization/57185
12435 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12436 SSA names properly.
12438 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12441 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12442 between "first_arg" and "insn", not "last" and "insn".
12444 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12446 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12447 (find_candidates_in_block): Re-enable slsr_process_phi.
12448 (create_phi_basis): Fix double counting of candidate adjustment.
12450 2013-05-06 Richard Biener <rguenther@suse.de>
12452 PR middle-end/57147
12453 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12454 the edge is also fallthru, preserve it and just clear the
12456 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12457 also complex, preserve that and just clear the fallthru flag.
12458 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12459 update virtual operands.
12461 2013-05-06 Alan Modra <amodra@gmail.com>
12463 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12464 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12465 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12466 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12467 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12468 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12470 2013-05-06 Alan Modra <amodra@gmail.com>
12472 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12473 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12474 (DEFAULT_ASM_ENDIAN): Define.
12475 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12476 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12477 Update -K PIC clause from sysv4.h.
12478 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12479 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12481 2013-05-06 Alan Modra <amodra@gmail.com>
12483 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12484 twice for little-endian.
12485 (ashrdi3_no_power, ashrdi3): Support little-endian.
12487 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12490 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12491 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12493 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12494 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12495 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12498 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12499 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12501 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12502 * config/pa/pa.opt: Make mbig-switch a no-op.
12503 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12504 (CASE_VECTOR_MODE): Always return SImode.
12505 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12506 for the !TARGET_BIG_SWITCH case.
12507 * config/pa/pa-linux.h: Likewise.
12508 * config/pa/pa-openbsd.h: Likewise.
12509 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12510 * config/pa/pa.md (short_jump): Remove define_insn.
12511 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12512 (casesi0): Remove define_insn.
12513 (type): Remove btable_branch.
12514 (pa_combine_type): Likewise.
12515 (in_nullified_branch_delay): Likewise.
12516 (in_call_delay): Likewise.
12517 (define_delay): Likewise.
12518 (define_insn_reservation "Z3"): Likewise.
12519 (define_insn_reservation "Z4"): Likewise.
12520 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12521 (pa_adjust_insn_length): Remove adjustment for btable branches.
12522 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12525 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
12527 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12528 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12529 Add m->r,x alternatives.
12530 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12531 splitters using SWI48x mode iterator.
12532 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12533 TARGET_64BIT. Add m->x alternative.
12534 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12535 Add o->x alternative. Enable for TARGET_SSE.
12536 (sse_storeq): Remove expander.
12537 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12538 with memory input operand.
12539 (*vec_extractv2di_1 splitter): New.
12540 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12541 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12543 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12545 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12546 (INT_LOWPART): Delete.
12547 (extract_MB): Adjust.
12548 (extract_ME): Adjust.
12549 (print_operand): Adjust.
12551 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12553 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12554 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12555 (reg_or_logical_cint_operand, easy_fp_constant,
12556 logical_const_operand): Delete "CONST_DOUBLE" case.
12557 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12558 "HOST_BITS_PER_WIDE_INT == 64" test.
12559 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
12560 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12561 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12562 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12564 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12565 CONST_DOUBLE DImode/VOIDmode case.
12566 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12567 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12568 CONST_DOUBLE VOIDmode case.
12569 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12570 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12571 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12572 Delete CONST_DOUBLE case.
12573 (splitters for mov FMOVE64 const_double): Delete
12574 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12575 "HOST_BITS_PER_WIDE_INT >= 64" test.
12576 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12578 (mov DI const_double): Delete.
12580 2013-05-04 Jakub Jelinek <jakub@redhat.com>
12582 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12583 on op shows all bits zero in mode of a lowpart subreg, return zero.
12585 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12588 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12589 to save TFmode registers and DImode to save TImode registers for
12590 caller save operations.
12591 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12592 mark being partially clobbered since they only use the first
12595 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12596 and TDmode only use the upper 64-bits of each VSX register.
12598 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12600 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12601 (find_candidates_in_block): Disable slsr_process_phi.
12603 2013-05-03 Guozhi Wei <carrot@google.com>
12605 * coverage.c (coverage_obj_init): Move the construction of gcov
12607 (build_init_ctor): ... here.
12609 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12611 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12612 (slsr_cand_d): Redefine def_phi.
12613 (stride_status, phi_adjust_status, count_phis_status): New enums.
12614 (find_phi_def): New.
12615 (find_basis_for_base_expr): New.
12616 (find_basis_for_candidate): Handle hidden bases.
12617 (alloc_cand_and_find_basis): Handle phi candidates.
12618 (slsr_process_phi): New.
12619 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12620 (create_mul_imm_cand): Likewise.
12621 (create_add_ssa_cand): Exclude phi base candidates.
12622 (create_add_imm_cand): Likewise.
12623 (slsr_process_cast): Likewise.
12624 (slsr_process_copy): Likewise.
12625 (find_candidates_in_block): Handle phi candidates.
12626 (dump_candidate): Likewise.
12627 (unconditional_cands): Delete.
12628 (unconditional_cands_with_known_stride_p): Delete.
12629 (phi_dependent_cand_p): New.
12630 (cand_increment): Handle phi-dependent candidates.
12631 (replace_dependent): Delete.
12632 (replace_mult_candidate): New.
12633 (replace_unconditional_candidate): New.
12634 (incr_vec_index): Move to avoid forward reference.
12635 (create_add_on_incoming_edge): New.
12636 (create_phi_basis): New.
12637 (replace_dependents): Delete.
12638 (replace_conditional_candidate): New.
12639 (phi_add_costs): New.
12640 (replace_uncond_cands_and_profitable_phis): New.
12641 (record_increment): Handle phi adjustments.
12642 (record_phi_increments): New.
12643 (record_increments): Handle phi adjustments.
12644 (phi_incr_cost): New.
12645 (lowest_cost_path): Handle phis.
12646 (total_savings): Likewise.
12647 (analyze_increments): Likewise.
12648 (ncd_with_phi): New.
12649 (ncd_of_cand_and_phis): New.
12650 (nearest_common_dominator_for_cands): Handle phi increments.
12651 (all_phi_incrs_profitable): New.
12652 (replace_profitable_candidates): Handle phi-dependent candidates.
12653 (analyze_candidates_and_replace): Likewise.
12655 2013-05-03 Teresa Johnson <tejohnson@google.com>
12658 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12659 do not exceed REG_BR_PROB_BASE.
12661 2013-05-03 Jeff Law <law@redhat.com>
12663 PR tree-optimization/57144
12664 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12665 operand of the condition will bit into the new type when eliminating
12666 a cast feeding a condition.
12668 2013-05-03 Jakub Jelinek <jakub@redhat.com>
12670 PR rtl-optimization/57130
12671 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12672 of COMPARE as in_code to the recursive call if needed.
12674 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
12676 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12677 (enabled): Handle new members.
12678 * config/i386/sse.md (*vec_concatv2si): Merge from
12679 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12680 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12682 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
12684 PR tree-optimization/57027
12685 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12686 for fnms opportunity, check we got the prerequisite kind
12687 of tree / gimple before using accessor functions.
12689 2013-05-03 Richard Biener <rguenther@suse.de>
12691 * double-int.h (lshift): New overload without precision
12692 and arith argument.
12693 (operator *=, operator +=, operator -=): Move ...
12694 * double-int.c (operator *=, operator +=, operator -=): ... here
12695 and implement more efficiently.
12696 (mul_double_with_sign): Remove.
12697 (lshift_double): Adjust to take unsinged shift argument, push
12698 dispatching code to callers.
12699 (mul_double_wide_with_sign): Add early out for callers that
12700 are not interested in high parts or overflow.
12701 (lshift): New function.
12702 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12703 dispatch code here.
12704 (lrotate, rrotate): Use logical shifts.
12705 * expr.c (get_inner_reference): Use lshift.
12706 * fixed-value.c (do_fixed_divide): Likewise.
12707 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12708 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12709 (indirect_refs_may_alias_p): Likewise.
12710 (stmt_kills_ref_p_1): Likewise.
12712 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12714 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12716 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12718 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12719 scalar form of FABD instruction.
12721 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12723 * lra-constraints.c (process_alt_operands): Add checking alt
12724 number to choose the best alternative.
12726 2013-05-02 Richard Biener <rguenther@suse.de>
12728 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12729 bitmap and its handling.
12730 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12732 2013-05-02 Richard Biener <rguenther@suse.de>
12734 PR middle-end/57140
12735 * tree-inline.c (copy_loops): Properly handle removed loops.
12736 (copy_cfg_body): Mark destination loops for fixup if source
12737 loops needed fixup.
12739 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12742 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12743 generating simple_return for naked functions.
12745 2013-05-02 Martin Jambor <mjambor@suse.cz>
12747 PR middle-end/56988
12748 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12749 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12751 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12752 ipa_agg_replacement_value structures.
12753 (known_aggs_to_agg_replacement_list): Likewise.
12754 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12755 (read_agg_replacement_chain): Likewise.
12756 (ipcp_transform_function): Also check that by_ref flags match.
12758 2013-05-02 Richard Biener <rguenther@suse.de>
12760 * graphds.h (struct graph): Add obstack member.
12761 * graphds.c (new_graph): Initialize obstack and allocate
12763 (add_edge): Allocate edge from the obstack.
12764 (free_graph): Free the obstack instead of all edges and vertices.
12766 2013-05-02 Teresa Johnson <tejohnson@google.com>
12768 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12770 * cfg.c (update_bb_profile_for_threading): Ditto.
12771 * tree-inline.c (copy_bb): Ditto.
12772 (copy_edges_for_bb): Ditto.
12773 (initialize_cfun): Ditto.
12774 (copy_cfg_body): Ditto.
12775 (expand_call_inline): Ditto.
12776 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12777 (estimate_node_size_and_time): Ditto.
12778 (inline_merge_summary): Ditto.
12779 * cgraphclones.c (cgraph_clone_edge): Ditto.
12780 (cgraph_clone_node): Ditto.
12781 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12782 (compute_trg_info): Ditto.
12784 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12786 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12787 S reg when fp attribute set.
12788 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12790 2013-05-02 Ian Bolton <ian.bolton@arm.com>
12792 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12794 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12795 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12796 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12798 2013-05-02 Richard Biener <rguenther@suse.de>
12800 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12801 (struct instantiate_cache_entry): New type.
12802 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12803 (struct instantiate_cache_type): New type.
12804 (set_instantiated_value, get_instantiated_value): Remove.
12805 (get_instantiated_value_entry): New function.
12806 (instantiate_scev_name): Use the new cache and adjust.
12807 (instantiate_scev_poly): Adjust.
12808 (instantiate_scev_binary): Likewise.
12809 (instantiate_array_ref): Likewise.
12810 (instantiate_scev_convert): Likewise.
12811 (instantiate_scev_not): Likewise.
12812 (instantiate_scev_3): Likewise.
12813 (instantiate_scev_2): Likewise.
12814 (instantiate_scev_r): Likewise.
12815 (instantiate_scev): Likewise.
12816 (resolve_mixers): Likewise.
12818 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
12821 * lra-constraints.c (best_small_class_operands_num): Remove.
12822 (process_alt_operands): Remove small_class_operands_num. Take
12823 small classes operands into losers and only if the operand is not
12824 matched. Modify debugging output.
12825 (curr_insn_transform): Remove best_small_class_operands_num.
12828 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12830 * config/aarch64/aarch64-builtins.c
12831 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
12832 * config/aarch64/aarch64-simd-builtins.def
12833 (reduc_splus_): Add new modes.
12834 (reduc_uplus_): New.
12835 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
12836 (reduc_uplus_v4sf): Likewise.
12837 (reduc_splus_v4sf): Likewise.
12838 (aarch64_addv<mode>): Likewise.
12839 (reduc_uplus_<mode>): Likewise.
12840 (reduc_splus_<mode>): Likewise.
12841 (aarch64_addvv2di): Likewise.
12842 (reduc_uplus_v2di): Likewise.
12843 (reduc_splus_v2di): Likewise.
12844 (aarch64_addvv2si): Likewise.
12845 (reduc_uplus_v2si): Likewise.
12846 (reduc_splus_v2si): Likewise.
12847 (reduc_<sur>plus_<mode>): New.
12848 (reduc_<sur>plus_v2di): Likewise.
12849 (reduc_<sur>plus_v2si): Likewise.
12850 (reduc_<sur>plus_v4sf): Likewise.
12851 (aarch64_addpv4sf): Likewise.
12852 * config/aarch64/arm_neon.h
12853 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
12854 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
12855 add UNSPEC_SADDV, UNSPEC_UADDV.
12857 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
12859 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12861 * config/aarch64/arm_neon.h
12862 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
12864 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12866 * config/aarch64/aarch64-builtins
12867 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
12869 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12871 * config/aarch64/aarch64-simd-builtins.def
12872 (reduc_smax_): New.
12873 (reduc_smin_): Likewise.
12874 (reduc_umax_): Likewise.
12875 (reduc_umin_): Likewise.
12876 (reduc_smax_nan_): Likewise.
12877 (reduc_smin_nan_): Likewise.
12880 (smax): Update for V2SF, V4SF and V2DF modes.
12883 (smin_nan): Likewise.
12884 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
12885 (<su><maxmin><mode>3): ...This, refactor.
12886 (s<maxmin><mode>3): New.
12887 (<maxmin_uns><mode>3): Likewise.
12888 (reduc_<maxmin_uns>_<mode>): Refactor.
12889 (reduc_<maxmin_uns>_v4sf): Likewise.
12890 (reduc_<maxmin_uns>_v2si): Likewise.
12891 (aarch64_<fmaxmin><mode>: Remove.
12892 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
12894 (vmin<q>_f<32,64>): Likewise.
12895 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12897 (su): Add mappings for smax, smin, umax, umin.
12899 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
12900 (FMAXMIN): Rename as...
12901 (FMAXMIN_UNS): ...This.
12903 (fmaxminv): Likewise.
12904 (fmaxmin): Likewise.
12906 (maxmin_uns_op): Likewise.
12908 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12910 * config/aarch64/arm_neon.h
12911 (vac<ge, gt><sd>_f<32, 64>): Rename to...
12912 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
12913 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
12915 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12917 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
12918 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
12920 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12922 * config/aarch64/aarch64-simd.md
12923 (vcond<mode>_internal): Handle special cases for constant masks.
12924 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
12925 (vcondu<mode><mode>): Likewise.
12926 (vcond<v_cmp_result><mode>): New.
12928 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12930 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
12931 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
12932 * config/aarch64/aarch64-simd-builtins.def
12933 (cmeq): Update to BUILTIN_VALLDI.
12938 * config/aarch64/arm_neon.h
12939 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
12940 to builtins or C as appropriate.
12942 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
12944 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
12946 (cmhi): Rename to...
12948 * config/aarch64/aarch64-simd.md
12949 (simd_mode): Add SF.
12950 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
12951 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
12952 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
12953 (cstore<mode>_neg): ...This.
12954 * config/aarch64/iterators.md
12956 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
12957 (COMPARISONS): New.
12958 (UCOMPARISONS): Likewise.
12959 (optab): Add missing comparisons.
12965 (VCMP_S): Likewise.
12966 (VCMP_U): Likewise.
12967 (V_cmp_result): Add DF, SF modes.
12968 (v_cmp_result): Likewise.
12970 (vmtype): Likewise.
12971 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
12973 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
12975 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
12976 define_insn to define_insn_and_split.
12977 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
12978 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
12979 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
12980 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
12981 (thumb2_negscc): Likewise.
12983 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12985 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
12987 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
12989 * config/arm/thumb2.md: Remove trailing whitespaces.
12991 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12993 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
12994 Use gen_int_mode rather than GEN_INT.
12996 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
12998 * value-prof.c (stream_in_histogram_value): Remove the strayed
13001 2013-04-30 Richard Biener <rguenther@suse.de>
13003 PR middle-end/57122
13004 * cfghooks.c (split_edge): Properly check for the loop latch edge.
13006 2013-04-30 Richard Biener <rguenther@suse.de>
13008 PR middle-end/57107
13009 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
13011 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
13013 PR rtl-optimization/56957
13014 PR rtl-optimization/57105
13015 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13016 variable. Use just INSN_UID for determining whether an insn
13017 should be only disconnected from the insn stream.
13018 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13020 2013-04-30 Jakub Jelinek <jakub@redhat.com>
13022 PR tree-optimization/57104
13023 * tsan.c (instrument_expr): Don't instrument accesses to
13024 DECL_HARD_REGISTER VAR_DECLs.
13026 2013-04-30 Richard Biener <rguenther@suse.de>
13028 * function.h (loops_for_fn): New inline function.
13029 (set_loops_for_fn): Likewise.
13030 * cfgloop.h (place_new_loop): Add struct function parameter.
13031 (get_loop): Likewise.
13032 (get_loops): Likewise.
13033 (number_of_loops): Likewise.
13034 (fel_next): Adjust.
13035 (fel_init): Likewise.
13036 * cfg.c (get_loop_copy): Adjust.
13037 * cfgloop.c (flow_loops_dump): Likewise.
13038 (record_loop_exits): Likewise.
13039 (verify_loop_structure): Likewise.
13040 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13041 (estimate_reg_pressure_cost): Likewise.
13042 (mark_loop_exit_edges): Likewise.
13043 * cfgloopmanip.c (place_new_loop): Likewise.
13044 (add_loop): Likewise.
13045 (duplicate_loop): Likewise.
13046 * graph.c (draw_cfg_nodes): Likewise.
13047 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13048 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13049 (extract_affine_chrec): Likewise.
13050 (build_scop_iteration_domain): Likewise.
13051 * graphite.c (graphite_initialize): Likewise.
13052 * ira-build.c (create_loop_tree_nodes): Likewise.
13053 (more_one_region_p): Likewise.
13054 (rebuild_regno_allocno_maps): Likewise.
13055 (mark_loops_for_removal): Likewise.
13056 (mark_all_loops_for_removal): Likewise.
13057 (remove_unnecessary_regions): Likewise.
13058 (ira_build): Likewise.
13059 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13060 * loop-init.c (fix_loop_structure): Likewise.
13061 (gate_rtl_move_loop_invariants): Likewise.
13062 (gate_rtl_unswitch): Likewise.
13063 (gate_rtl_unroll_and_peel_loops): Likewise.
13064 (rtl_doloop): Likewise.
13065 * lto-streamer-in.c (input_cfg): Likewise.
13066 * lto-streamer-out.c (output_cfg): Likewise.
13067 * modulo-sched.c (sms_schedule): Likewise.
13068 * predict.c (tree_estimate_probability): Likewise.
13069 (tree_estimate_probability_driver): Likewise.
13070 (estimate_loops): Likewise.
13071 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13072 (move_sese_region_to_fn): Likewise.
13073 (debug_loop_num): Likewise.
13074 * tree-chrec.c (chrec_evaluate): Likewise.
13075 (hide_evolution_in_other_loops_than_loop): Likewise.
13076 (chrec_component_in_loop_num): Likewise.
13077 (reset_evolution_in_loop): Likewise.
13078 (evolution_function_is_invariant_rec_p): Likewise.
13079 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13080 * tree-inline.c (copy_loops): Likewise.
13081 (copy_cfg_body): Likewise.
13082 (tree_function_versioning): Likewise.
13083 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13084 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13086 (add_to_evolution_1): Likewise.
13087 (scev_const_prop): Likewise.
13088 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13089 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13090 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13091 (tree_ssa_lim_initialize): Likewise.
13092 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13093 (verify_loop_closed_ssa): Likewise.
13094 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13095 (tree_ssa_loop_im): Likewise.
13096 (tree_ssa_loop_unswitch): Likewise.
13097 (tree_vectorize): Likewise.
13098 (check_data_deps): Likewise.
13099 (tree_ssa_loop_ivcanon): Likewise.
13100 (tree_ssa_loop_bounds): Likewise.
13101 (tree_complete_unroll): Likewise.
13102 (tree_complete_unroll_inner): Likewise.
13103 (tree_parallelize_loops): Likewise.
13104 (tree_ssa_loop_prefetch): Likewise.
13105 (tree_ssa_loop_ivopts): Likewise.
13106 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13107 * tree-vectorizer.c (vectorize_loops): Likewise.
13109 2013-04-29 Mike Frysinger <vapier@gentoo.org>
13111 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13112 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13113 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13114 with EABI_LINK_SPEC.
13116 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13119 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13122 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13125 * lra-constraints.c (process_alt_operands): Discourage a bit more
13126 using memory for pseudos. Print cost dump for alternatives.
13127 Modify cost values for conflicts with early clobbers.
13128 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13130 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13133 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13135 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13137 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13138 from/to S register.
13139 (movdi_aarch64): Support LDR/STR from/to D register.
13141 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13143 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13144 or higher by default.
13146 2013-04-29 Richard Biener <rguenther@suse.de>
13148 PR middle-end/57075
13149 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13150 even if not adding abnormal edges for calls that can make
13153 2013-04-29 Richard Biener <rguenther@suse.de>
13155 PR middle-end/57103
13156 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13158 (move_stmt_r): Remove redundant checking.
13160 2013-04-29 Teresa Johnson <tejohnson@google.com>
13163 * basic-block.h (apply_scale): New function.
13164 (apply_probability): Use apply_scale.
13165 * gimple-streamer-in.c (input_bb): Ditto.
13166 * lto-streamer-in.c (input_cfg): Ditto.
13167 * lto-cgraph.c (merge_profile_summaries): Ditto.
13168 * tree-optimize.c (execute_fixup_cfg): Ditto.
13169 * tree-inline.c (copy_bb): Update comment to use apply_scale.
13170 (copy_edges_for_bb): Ditto.
13171 (copy_cfg_body): Ditto.
13173 2013-04-29 Tom de Vries <tom@codesourcery.com>
13175 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13176 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13177 (tail_merge_optimize): Handle current_loops == NULL.
13179 2013-04-26 Jeff Law <law@redhat.com>
13181 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13182 (simplify_cond_using_ranges): Generalize code to simplify
13183 COND_EXPRs where one argument is a constant and the other
13184 is an SSA_NAME created by an integral type conversion.
13186 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13188 * config/arm/arm.md (store_minmaxsi): Use only when
13189 optimize_insn_for_size_p.
13191 2013-04-29 Christian Bruel <christian.bruel@st.com>
13194 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13196 2013-04-29 Richard Biener <rguenther@suse.de>
13198 PR middle-end/57089
13199 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13200 loop tree make sure to schedule a fixup for the child as well.
13201 (expand_omp_for_generic): Properly add loops.
13202 (expand_omp_for_static_nochunk): Likewise.
13203 (expand_omp_for_static_chunk): Likewise.
13204 (expand_omp_for): For the degenerate case fixup loops.
13205 (expand_omp_sections): Fix default bb placement in loops.
13206 (expand_omp_atomic_pipeline): Properly add loops.
13208 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13210 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
13212 2013-04-29 Tom de Vries <tom@codesourcery.com>
13214 * tree-ssa-tail-merge.c: Update header comment.
13216 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13218 * config/aarch64/arm_neon.h
13219 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13220 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13221 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13222 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13223 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13224 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13225 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13226 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13228 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13230 * config/aarch64/aarch64-simd.md
13231 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13232 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13233 fix_trunc, fixuns_trunc.
13234 (ftrunc<VDQF:mode>2): New.
13235 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13236 (fix_trunc_optab): New.
13238 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13240 * config/aarch64/aarch64-builtins.c
13241 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13242 iceilf, lround, iroundf.
13244 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13247 * config/i386/i386.h (enum ix86_tune_indices)
13248 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13249 New, split from X86_TUNE_INTER_UNIT_MOVES.
13250 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13251 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13252 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13253 (TARGET_INTER_UNIT_MOVES): Remove.
13254 * config/i386/i386.c (initial_ix86_tune_features): Update.
13255 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13256 (ix86_expand_convert_uns_didf_sse): Use
13257 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13258 (ix86_expand_vector_init_one_nonzero): Ditto.
13259 (ix86_expand_vector_init_interleave): Ditto.
13260 (inline_secondary_memory_needed): Return true for moves from SSE class
13261 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13262 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13263 * config/i386/constraints.md (Yi, Ym): Depend on
13264 TARGET_INTER_UNIT_MOVES_TO_VEC.
13265 (Yj, Yn): New constraints.
13266 * config/i386/i386.md (*movdi_internal): Change constraints of
13267 operand 1 from Yi to Yj and from Ym to Yn.
13268 (*movsi_internal): Ditto.
13269 (*movdf_internal): Ditto.
13270 (*movsf_internal): Ditto.
13271 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13272 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13273 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13274 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13275 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13276 * config/i386/sse.md (movdi_to_sse): Ditto.
13277 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13278 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13279 TARGET_INTER_UNIT_MOVES.
13280 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13281 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13282 instead of TARGET_INTER_UNIT_MOVES.
13283 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13284 operand 1 from Yi to Yj and from Ym to Yn.
13286 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13288 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13289 (float_truncate_hi_): Likewise.
13290 (float_extend_lo_): Likewise.
13291 (float_truncate_lo_): Likewise.
13292 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13293 (aarch64_float_extend_lo_v2df): Likewise.
13294 (vec_unpacks_hi_v4sf): Likewise.
13295 (aarch64_float_truncate_lo_v2sf): Likewise.
13296 (aarch64_float_truncate_hi_v4sf): Likewise.
13297 (vec_pack_trunc_v2df): Likewise.
13298 (vec_pack_trunc_df): Likewise.
13300 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13302 * config/aarch64/aarch64-builtins.c
13303 (aarch64_fold_builtin): Fold float conversions.
13304 * config/aarch64/aarch64-simd-builtins.def
13305 (floatv2si, floatv4si, floatv2di): New.
13306 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13307 * config/aarch64/aarch64-simd.md
13308 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13309 * config/aarch64/iterators.md (FLOATUORS): New.
13310 (optab): Add float, floatuns.
13311 (su_optab): Likewise.
13313 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13315 * config/aarch64/aarch64-builtins.c
13316 (aarch64_builtin_vectorized_function): Use new names for
13318 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13319 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13320 (fcvtzu): Split as...
13321 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13322 (fcvtas): Split as...
13323 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13324 (fcvtau): Split as...
13325 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13326 (fcvtps): Split as...
13327 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13328 (fcvtpu): Split as...
13329 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13330 (fcvtms): Split as...
13331 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13332 (fcvtmu): Split as...
13333 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13334 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13335 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13336 (lfrintnusf, lfrintnudf): Likewise.
13337 * config/aarch64/aarch64-simd.md
13338 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13340 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13341 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13342 (fcvt_pattern): Likewise.
13344 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13346 * config/aarch64/aarch64-simd.md
13347 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13348 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13350 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13352 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13353 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13354 (vrnd<a,m,n,p>_f32): Implement using builtins.
13355 (vrnd<i,x><q>_f<32, 64>): New.
13357 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13359 * config/aarch64/aarch64-builtins.c
13360 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13361 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13362 (frintz): Rename to...
13364 (frintp): Rename to...
13366 (frintm): Rename to...
13368 (frinti): Rename to...
13369 (nearbyint): ...this.
13370 (frintx): Rename to...
13372 (frinta): Rename to...
13374 * config/aarch64/aarch64-simd.md
13375 (aarch64_frint<frint_suffix><mode>): Delete.
13376 (<frint_pattern><mode>2): Convert to insn.
13377 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13378 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13379 (frint_pattern): Likewise.
13380 (frint_suffix): Likewise.
13382 2013-04-29 Richard Biener <rguenther@suse.de>
13384 PR tree-optimization/57081
13385 * loop-init.c: Include tree-flow.h.
13386 (loop_optimizer_finalize): Free number of iteration estimates.
13387 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13389 2013-04-29 Jakub Jelinek <jakub@redhat.com>
13391 PR tree-optimization/57083
13392 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13393 non-singleton shift count range, zero extend low_bound for uns case.
13395 * config/i386/predicates.md (general_vector_operand): New predicate.
13396 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13397 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13398 if they aren't nonimmediate operands. If their original values
13399 satisfy const_vector_equal_evenodd_p, don't shift them.
13400 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13401 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13402 if not nonimmediate_operand.
13403 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13404 instead of register_operand.
13405 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13407 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
13409 * stor-layout.c (finalize_size_functions): Allocate a structure and
13410 reset cfun before dumping the functions.
13412 2013-04-27 Jakub Jelinek <jakub@redhat.com>
13414 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13417 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13418 use xop_pmacsdqh if uns_p.
13419 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13420 the immediate rotate count.
13422 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13424 * rtl.h (struct rtx_def): Add comment for field jump.
13425 (LRA_SUBREG_P): New macro.
13426 * recog.c (register_operand): Check LRA_SUBREG_P.
13427 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13429 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13430 (lra_final_code_change): Skip subreg change for operators.
13431 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13432 if there are no operand changes.
13433 * lra-constraints.c (curr_insn_set): New.
13434 (match_reload): Set LRA_SUBREG_P.
13435 (emit_spill_move): Ditto.
13436 (check_and_process_move): Use curr_insn_set. Process only single
13437 set insns. Don't initialize sec_mem_p and change_p.
13438 (simplify_operand_subreg): Use LRA_SUBREG_P.
13439 (reg_in_class_p): New function.
13440 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13441 of #ifdef. Add code to remove cycling.
13442 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13443 non-null disp. Reload inner instead of disp when base and index
13444 are null. Try to put lo_sum into register.
13445 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13446 (check_and_process_move): Move code for move cost check to
13447 simple_move_p. Remove equiv_substitution.
13448 (simple_move_p): New function.
13449 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13450 curr_insn_set. Call check_and_process_move only for single set
13451 insns. Use the new function. Move call of check_and_process_move
13452 after operand equiv substitution and address process.
13454 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13457 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13458 with nonlocal goto receivers or returns twice calls, ignore
13459 unininitialized values from abnormal edges to nl goto receiver
13460 or returns twice call.
13462 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13464 PR tree-optimization/57051
13465 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13466 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13469 2013-04-26 Richard Biener <rguenther@suse.de>
13471 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13472 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13473 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13474 (fixup_loop_arrays_after_move): New function.
13475 (move_sese_region_to_fn): Properly outline the loop tree parts
13476 of the SESE region.
13478 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
13480 * config/i386/i386.md (type, unit): Fix long lines.
13482 2013-04-26 Richard Biener <rguenther@suse.de>
13484 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13485 (lto-streamer-out.o): Likewise.
13486 * cfgloop.c (init_loops_structure): Export, add struct function
13487 argument and adjust.
13488 (flow_loops_find): Adjust.
13489 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13490 (init_loops_structure): Declare.
13491 * lto-streamer-in.c: Include cfgloop.h.
13492 (input_cfg): Input the loop tree.
13493 * lto-streamer-out.c: Include cfgloop.h.
13494 (output_cfg): Output the loop tree.
13495 (output_struct_function_base): Do not drop PROP_loops.
13497 2013-03-26 Richard Biener <rguenther@suse.de>
13499 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13500 (pass_build_cfg): Provide PROP_loops.
13501 (move_sese_region_to_fn): Remove loops that are outlined into fn
13503 * tree-inline.c: Include cfgloop.h.
13504 (initialize_cfun): Do not drop PROP_loops.
13505 (copy_loops): New function.
13506 (copy_cfg_body): Copy loop structure.
13507 (tree_function_versioning): Initialize destination loop tree.
13508 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13509 (pass_parallelize_loops): Do IL verification.
13510 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13511 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13512 the CFG make sure we fixup loops as well.
13513 * tree-ssa-tail-merge.c: Include cfgloop.h.
13514 (replace_block_by): When merging loop latches mark loops for fixup.
13515 * lto-streamer-out.c (output_struct_function_base): Drop
13516 PROP_loops for now.
13517 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13518 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13519 * ipa-split.c: Include cfgloop.h.
13520 (split_function): Add the new return block to the loop tree root.
13521 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13522 whether we have removed the forwarder block.
13523 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13524 * cfgloop.h (place_new_loop): Declare.
13525 * cfgloopmanip.c (place_new_loop): Export.
13526 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13527 (tree-switch-conversion.o): Likewise.
13528 (tree-complex.o): Likewise.
13529 (tree-inline.o): Likewise.
13530 (tree-ssa-tailmerge.o): Likewise.
13531 (ipa-split.o): Likewise.
13532 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13533 (tree-ssa-copy.o): Likewise.
13534 * tree-switch-conversion.c: Include cfgloop.h
13535 (process_switch): If we emit a bit-test cascade, schedule loops
13537 * tree-complex.c: Include cfgloop.h.
13538 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13539 * asan.c: Include cfgloop.h.
13540 (create_cond_insert_point): Properly add new basic-blocks to
13541 loops, schedule loop fixup.
13542 * cfgloop.c (verify_loop_structure): Check that looks are not
13544 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13546 (expand_omp_for_generic): Likewise.
13547 (expand_omp_sections): Likewise.
13548 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13549 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13550 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13551 is initialized, not when loops are present.
13552 * tree-parloops.c (parallelize_loops): Remove checking here.
13553 * passes.c (init_optimization_passes): Schedule a copy-propagation
13554 pass before complete unrolling of inner loops.
13556 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13558 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13559 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13560 (colorize_init): Add argument to _WIN32 version.
13561 * toplev.c: Include diagnostic-color.h.
13562 (process_options): Default to -fdiagnostics-color=auto if
13563 GCC_COLORS env var is in the environment.
13564 * common.opt (fdiagnostics-color=): Add Var and Init.
13565 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13566 env var is in the environment, the default is auto rather than never.
13568 * diagnostic.h (file_name_as_prefix): Add context argument.
13569 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13570 the string as locus.
13571 * langhooks.c (lhd_print_error_function): Adjust caller.
13573 2013-04-25 Lawrence Crowl <crowl@google.com>
13575 * var-tracking.c (shared_hash_def::htab):
13576 Change type to hash_table. Update dependent calls and types.
13578 2013-04-25 Lawrence Crowl <crowl@google.com>
13580 * Makefile.in: Update as needed below.
13582 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13583 Move declaration to after the type's method definitons.
13585 * attribs.c (htab_t scoped_attributes::attribute_hash):
13586 Change type to hash_table. Update dependent calls and types.
13588 * bitmap.c (htab_t bitmap_desc_hash):
13589 Change type to hash_table. Update dependent calls and types.
13591 * cselib.c (htab_t cselib_hash_table):
13592 Change type to hash_table. Update dependent calls and types.
13594 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13595 (hash_string_slot_node): Move implementation into lto-streamer.h
13596 struct string_slot_hasher.
13597 (eq_string_slot_node): Likewise.
13599 * data-streamer-out.c: Update output_block::string_hash_table
13600 dependent calls and types.
13602 * dwarf2cfi.c (htab_t trace_index):
13603 Change type to hash_table. Update dependent calls and types.
13605 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13606 Change type to hash_table. Update dependent calls and types.
13607 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13608 (htab_t optimize_external_refs::map): Likewise.
13609 (htab_t output_comp_unit::extern_map): Likewise.
13610 (htab_t output_comdat_type_unit::extern_map): Likewise.
13611 (htab_t output_macinfo::macinfo_htab): Likewise.
13612 (htab_t optimize_location_lists::htab): Likewise.
13613 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13615 * except.c (htab_t ehspec_hash_type):
13616 Change type to hash_table. Update dependent calls and types.
13617 (assign_filter_values::ttypes): Likewise.
13618 (assign_filter_values::ehspec): Likewise.
13619 (sjlj_assign_call_site_values::ar_hash): Likewise.
13620 (convert_to_eh_region_ranges::ar_hash): Likewise.
13622 * gcse.c (htab_t pre_ldst_table):
13623 Change type to hash_table. Update dependent calls and types.
13625 * ggc-common.c (htab_t saving_htab):
13626 Change type to hash_table. Update dependent calls and types.
13627 (htab_t loc_hash): Likewise.
13628 (htab_t ptr_hash): Likewise.
13629 (call_count): Rename ggc_call_count.
13630 (call_alloc): Rename ggc_call_alloc.
13631 (loc_descriptor): Rename make_loc_descriptor.
13632 (add_statistics): Rename ggc_add_statistics.
13634 * ggc-common.c (saving_htab):
13635 Change type to hash_table. Update dependent calls and types.
13637 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13638 (push_gimplify_context): Likewise.
13639 (pop_gimplify_context): Likewise.
13640 (struct gimple_temp_hash_elt): Added.
13641 (struct gimplify_hasher): Likewise.
13642 (struct gimplify_ctx.temp_htab):
13643 Change type to hash_table. Update dependent calls and types.
13645 * gimple-fold.c: Include gimplify-ctx.h.
13647 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13648 Change type to hash_table. Update dependent calls and types.
13649 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13650 avoid potential global name collision.
13652 * gimplify.c: Include gimplify-ctx.h.
13653 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13654 (htab_t gimplify_ctx::temp_htab):
13655 Update dependent calls and types for new type hash_table.
13656 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13657 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13659 * gimplify-ctx.h: New.
13660 (struct gimple_temp_hash_elt): Move from gimplify.c.
13661 (class gimplify_hasher): New.
13662 (struct gimplify_ctx): Move from gimple.h.
13663 (htab_t gimplify_ctx::temp_htab):
13664 Change type to hash_table. Update dependent calls and types.
13666 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13667 (htab_t ivs_params::newivs_index):
13668 Change type to hash_table. Update dependent calls and types.
13669 (htab_t ivs_params::params_index): Likewise.
13670 (htab_t print_generated_program::params_index): Likewise.
13671 (htab_t gloog::newivs_index): Likewise.
13672 (htab_t gloog::params_index): Likewise.
13674 * graphite.c: Include graphite-htab.h.
13675 4htab_t graphite_transform_loops::bb_pbb_mapping):
13676 Change type to hash_table. Update dependent calls and types.
13678 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13679 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13680 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13682 * graphite-dependences.c: Include graphite-htab.h.
13683 (loop_is_parallel_p): Change hash table type of parameter.
13685 * graphite-htab.h: New.
13686 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13687 (extern find_pbb_via_hash): Move from graphite-poly.h.
13688 (extern loop_is_parallel_p): Move from graphite-poly.h.
13689 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13691 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13692 (extern loop_is_parallel_p): Move to graphite-htab.h.
13693 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13695 * haifa-sched.c (htab_t delay_htab):
13696 Change type to hash_table. Update dependent calls and types.
13697 (htab_t delay_htab_i2): Likewise.
13699 * ira-color.c (htab_t allocno_hard_regs_htab):
13700 Change type to hash_table. Update dependent calls and types.
13702 * ira-costs.c (htab_t cost_classes_htab):
13703 Change type to hash_table. Update dependent calls and types.
13705 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13706 Change type to hash_table. Update dependent calls and types.
13708 * loop-iv.c (htab_t bivs):
13709 Change type to hash_table. Update dependent calls and types.
13711 * loop-unroll.c (htab_t opt_info::insns_to_split):
13712 Change type to hash_table. Update dependent calls and types.
13713 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13715 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13716 (struct string_slot_hasher): New.
13717 (htab_t output_block::string_hash_table):
13718 Change type to hash_table. Update dependent calls and types.
13720 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13721 (htab_t file_name_hash_table):
13722 Change type to hash_table. Update dependent calls and types.
13724 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13727 * lto-streamer.c (htab_t tree_htab):
13728 Change type to hash_table. Update dependent calls and types.
13730 * omp-low.c: Include gimplify-ctx.h.
13732 * passes.c (htab_t name_to_pass_map):
13733 Change type to hash_table. Update dependent calls and types.
13734 (pass_traverse): Rename to passes_pass_traverse.
13736 * plugin.c (htab_t event_tab):
13737 Change type to hash_table. Update dependent calls and types.
13739 * postreload-gcse.c (htab_t expr_table):
13740 Change type to hash_table. Update dependent calls and types.
13741 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13743 * sese.c (debug_rename_map_1): Make extern.
13744 (htab_t copy_bb_and_scalar_dependences::rename_map):
13745 Change type to hash_table. Update dependent calls and types.
13747 * sese.h (extern debug_rename_map): Move to .c file.
13749 * store-motion.c (htab_t store_motion_mems_table):
13750 Change type to hash_table. Update dependent calls and types.
13752 * trans-mem.c (htab_t tm_new_mem_hash):
13753 Change type to hash_table. Update dependent calls and types.
13755 * tree-browser.c (htab_t TB_up_ht):
13756 Change type to hash_table. Update dependent calls and types.
13758 * tree-cfg.c (htab_t discriminator_per_locus):
13759 Change type to hash_table. Update dependent calls and types.
13761 * tree-complex.c: Include tree-hasher.h
13762 (htab_t complex_variable_components):
13763 Change type to hash_table. Update dependent calls and types.
13765 * tree-eh.c (htab_t finally_tree):
13766 Change type to hash_table. Update dependent calls and types.
13768 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13769 struct int_tree_hasher.
13770 (extern int_tree_map_eq): Likewise.
13771 (uid_decl_map_hash): Removed.
13772 (extern decl_tree_map_eq): Likewise.
13774 * tree-hasher.h: New.
13775 (struct int_tree_hasher): New.
13776 (typedef int_tree_htab_type): New.
13778 * tree-inline.c: Include gimplify-ctx.h.
13780 * tree-mudflap.c: Include gimplify-ctx.h.
13782 * tree-parloops.c: Include tree-hasher.h.
13783 (htab_t eliminate_local_variables_stmt::decl_address):
13784 Change type to hash_table. Update dependent calls and types.
13785 (htab_t separate_decls_in_region::decl_copies): Likewise.
13787 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13788 Change type to hash_table. Update dependent calls and types.
13790 * tree-sra.c (candidates):
13791 Change type to hash_table. Update dependent calls and types.
13793 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13795 (int_tree_map_hash): Likewise.
13797 * tree-ssa-dom.c (htab_t avail_exprs):
13798 Change type to hash_table. Update dependent calls and types.
13800 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13801 Change type to hash_table. Update dependent calls and types.
13803 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
13804 Change type to hash_table. Update dependent calls and types.
13806 * tree-ssa-phiopt.c (seen_ssa_names):
13807 Change type to hash_table. Update dependent calls and types.
13809 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
13810 Change type to hash_table. Update dependent calls and types.
13812 * tree-ssa-uncprop.c (equiv):
13813 Change type to hash_table. Update dependent calls and types.
13815 2013-04-25 Jakub Jelinek <jakub@redhat.com>
13817 PR rtl-optimization/57003
13818 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
13819 call note_stores with kill_clobbered_value callback again after
13820 killing regs_invalidated_by_call.
13822 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13824 * config/aarch64/aarch64-simd.md
13825 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
13826 (aarch64_simd_bsl<mode>): Likewise.
13827 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
13829 2013-04-25 Marek Polacek <polacek@redhat.com>
13831 PR tree-optimization/57066
13832 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
13834 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
13836 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
13838 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13840 * config/aarch64/aarch64-builtins.c
13841 (aarch64_fold_builtin): New.
13842 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
13843 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
13844 * config/aarch64/aarch64-simd-builtins.def (abs): New.
13845 * config/aarch64/arm_neon.h
13846 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
13848 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13849 Tejas Belagod <tejas.belagod@arm.com>
13851 * config/aarch64/aarch64-builtins.c
13852 (aarch64_gimple_fold_builtin): New.
13853 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
13854 * config/aarch64/aarch64-simd-builtins.def (addv): New.
13855 * config/aarch64/aarch64-simd.md (addpv4sf): New.
13856 (addvv4sf): Update.
13857 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
13859 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13861 * config/aarch64/aarch64.md
13862 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
13864 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13866 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
13867 (*ngcsi_uxtw): New pattern.
13869 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13870 Julian Brown <julian@codesourcery.com>
13872 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
13873 (TB_DREG): Add T_V4HF.
13874 (v4hf_UP): New macro.
13875 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13876 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13877 Handle initialisation of V4HF. Adjust initialisation of reinterpret
13879 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
13880 (arm_vector_mode_supported_p): Handle V4HF.
13881 (arm_mangle_map): Handle V4HFmode.
13882 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
13883 * config/arm/arm_neon_builtins.def: Add entries for
13884 vcvtv4hfv4sf, vcvtv4sfv4hf.
13885 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
13886 (neon_vcvtv4hfv4sf): Likewise.
13887 * config/arm/neon-gen.ml: Handle half-precision floating point
13889 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
13890 * config/arm/arm_neon.h: Regenerate.
13891 * config/arm/neon.ml (type elts): Add F16.
13892 (type vectype): Add T_float16x4, T_floatHF.
13893 (type vecmode): Add V4HF.
13894 (type features): Add Requires_FP_bit feature.
13895 (elt_width): Handle F16.
13896 (elt_class): Likewise.
13897 (elt_of_class_width): Likewise.
13898 (mode_of_elt): Refactor.
13899 (type_for_elt): Handle F16, fix error messages.
13900 (vectype_size): Handle T_float16x4.
13901 (vcvt_sh): New function.
13902 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
13903 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
13904 (string_of_mode): Handle V4HF.
13905 * doc/arm-neon-intrinsics.texi: Regenerate.
13907 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
13909 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
13910 format specifier in 'X' case.
13912 2013-04-25 Alan Modra <amodra@gmail.com>
13915 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
13916 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
13917 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
13918 Repeat for many other rotate/shift and mask patterns using subregs.
13919 Name lshiftrt insns.
13920 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
13921 on WORDS_BIG_ENDIAN.
13923 2013-04-25 Alan Modra <amodra@gmail.com>
13925 * config.gcc: Support little-endian powerpc-linux targets.
13926 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
13927 (LINK_OS_LINUX_SPEC): Define.
13928 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
13929 Preserve MASK_LITTLE_ENDIAN.
13930 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
13931 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
13932 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
13933 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
13934 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
13935 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
13936 Correct fp word order for little-endian. Don't shift toc entries
13937 smaller than a word for little-endian.
13938 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
13939 (bswapdi2 splits): Correct low-part subreg for little-endian.
13940 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
13941 low/high where such is correct only for be.
13942 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
13943 little-endian for -mcall-aixdesc.
13945 2013-04-25 Alan Modra <amodra@gmail.com>
13947 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
13948 replace_equiv_address_nv.
13950 2013-04-25 Alan Modra <amodra@gmail.com>
13952 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
13954 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13957 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13958 * rtl.h (struct rtx_def): ...
13960 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13962 PR rtl-optimizations/57046
13963 * lra-constraints (split_reg): Set up lra_risky_transformations_p
13964 for multi-reg splits.
13966 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
13968 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
13970 2013-04-24 Sterling Augustine <saugustine@google.com>
13972 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
13973 (comp_dir_string, debug_str_dwo_section): New.
13974 (DEBUG_STR_DWO_SECTION): Rename to ...
13975 (DEBUG_DWO_STR_SECTION): ... this.
13976 (DEBUG_NORM_STR_SECTION): Delete.
13977 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
13978 (DEBUG_STR_DWO_SECTION_FLAGS): New.
13979 (find_AT_string): Move most logic to ...
13980 (find_AT_string_in_table): ... here. New.
13981 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
13982 add_skeleton_AT_string. Delete logic.
13983 (output_skeleton_debug_sections): Remove call to
13984 add_top_level_skeleton_die_attrs.
13985 (add_comp_dir_attribute): Move logic to comp_dir_string.
13986 (dwarf2out_init): Initialize debug_str_dwo_section.
13987 (output_indirect_string): Call find_string_form.
13988 (output_indirect_strings): Rewrite.
13989 (prune_unused_types): Empty skeleton_debug_str_hash.
13990 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
13991 (dwarf2out_finish): Call output_indirect_strings.
13993 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
13995 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
13997 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
13999 * rtl.h (struct rtx_def): Add comment for field jump.
14000 (LRA_SUBREG_P): New macro.
14001 * recog.c (register_operand): Check LRA_SUBREG_P.
14002 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14004 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14005 (lra_final_code_change): Skip subreg change for operators.
14006 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14007 if there are no operand changes.
14008 * lra-constraints.c (curr_insn_set): New.
14009 (match_reload): Set LRA_SUBREG_P.
14010 (emit_spill_move): Ditto.
14011 (check_and_process_move): Use curr_insn_set. Process only single
14012 set insns. Don't initialize sec_mem_p and change_p.
14013 (simplify_operand_subreg): Use LRA_SUBREG_P.
14014 (reg_in_class_p): New function.
14015 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14016 of #ifdef. Add code to remove cycling.
14017 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14018 non-null disp. Reload inner instead of disp when base and index
14019 are null. Try to put lo_sum into register.
14020 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14021 (check_and_process_move): Move code for move cost check to
14022 simple_move_p. Remove equiv_substitution.
14023 (simple_move_p): New function.
14024 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14025 curr_insn_set. Call check_and_process_move only for single set
14026 insns. Use the new function. Move call of check_and_process_move
14027 after operand equiv substitution and address process.
14029 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14031 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14032 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14033 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14035 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14037 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14039 2013-04-24 Marek Polacek <polacek@redhat.com>
14041 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14042 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14043 (select_loops_exit_conditions): Likewise.
14044 (number_of_iterations_for_all_loops): Likewise.
14045 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14046 (scev_analysis): Likewise.
14048 2013-04-02 Catherine Moore <clm@codesourcery.com>
14049 Chao-ying Fu <fu@mips.com>
14051 * config/mips/micromips.md (jraddiusp): New pattern.
14052 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14053 instruction if possible.
14055 2013-04-24 Alan Modra <amodra@gmail.com>
14057 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14059 2013-04-24 Julian Brown <julian@codesourcery.com>
14060 Chung-Lin Tang <cltang@codesourcery.com>
14062 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
14063 dependency behavior in enumeration type DIE generation. Add TODO note
14064 to comments about future DW_FORM_sdata/udata re-work of related code.
14066 2013-04-23 Lawrence Crowl <crowl@google.com>
14068 * Makefile.in: Update as needed below.
14070 * hash-table.h (class hash_table):
14071 Correct many methods with parameter types compare_type to the correct
14072 value_type. (Correct code was unlikely to notice the change.)
14073 (hash_table::elements_with_deleted) New.
14074 (class hashtable::iterator): New.
14075 (hashtable::begin()): New.
14076 (hashtable::end()): New.
14077 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14079 * statistics.c (statistics_hashes):
14080 Change type to hash_table. Update dependent calls and types.
14082 * tree-into-ssa.c (var_infos):
14083 Change type to hash_table. Update dependent calls and types.
14085 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14086 Change type to hash_table. Update dependent calls and types.
14088 * tree-ssa-loop-im.c (struct mem_ref.refs):
14089 Change type to hash_table. Update dependent calls and types.
14091 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14092 Change type to hash_table. Update dependent calls and types.
14094 * tree-ssa-sccvn.c (vn_tables_s::nary):
14095 Change type to hash_table. Update dependent calls and types.
14096 (vn_tables_s::phis): Likewise.
14097 (vn_tables_s::references): Likewise.
14099 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14100 (vn_reference_eq): Update parameter and return types.
14102 * tree-ssa-structalias.c (pointer_equiv_class_table):
14103 Change type to hash_table. Update dependent calls and types.
14104 (location_equiv_class_table): Likewise.
14106 * tree-vect-data-refs.c: Consequential changes for making
14107 peeling a hash_table.
14109 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14110 (destroy_loop_vec_info): Dependent hash_table update.
14112 * tree-vectorizer.h (peeling_htab):
14113 Change type to hash_table. Update dependent calls and types.
14115 2013-04-23 Shiva Chen <shiva0217@gmail.com>
14117 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14118 to check the register content is equal or not.
14119 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14120 to assign register content record.
14121 * lra-eliminations.c (update_reg_eliminate): Use
14122 lra_update_reg_val_offset to update register content offset.
14123 * lra-int.h (struct lra_reg): Add offset member.
14124 (lra_reg_val_equal_p): New static inline function.
14125 (lra_update_reg_val_offset): New static inline function.
14126 (lra_assign_reg_val): New static inline function.
14127 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14128 to assign register content record.
14129 (initialize_lra_reg_info_element): Initial offset to zero.
14131 2013-04-23 Catherine Moore <clm@codesourcery.com>
14133 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14134 operands. Record compression.
14136 2013-04-23 Xinliang David Li <davidxl@google.com>
14138 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14140 2013-04-23 Richard Biener <rguenther@suse.de>
14142 PR middle-end/57036
14143 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14144 parameter, only add abnormal goto edges from the copied body
14145 if the call could perform abnormal gotos.
14146 (copy_cfg_body): Adjust.
14148 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14150 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14152 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14154 * coretypes.h (gimple_stmt_iterator): Add struct to make
14157 2013-04-23 Richard Biener <rguenther@suse.de>
14159 PR tree-optimization/57026
14160 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14161 from SSA names occuring in abnormal PHI nodes.
14163 2013-04-22 Andi Kleen <ak@linux.intel.com>
14165 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14167 2013-04-22 Andi Kleen <ak@linux.intel.com>
14169 * lto/lto.c (print_lto_report_1): Declare early.
14170 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14172 2013-04-22 Andi Kleen <ak@linux.intel.com>
14174 * common.opt (-flto-report-wpa): Add.
14175 * doc/invoke.texi (-flto-report-wpa): Add.
14176 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14179 2013-04-22 Xinliang David Li <davidxl@google.com>
14181 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14182 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14183 * Makefile.in: New dependency
14185 David Daney <ddaney.cavm@gmail.com>
14187 * configure.ac (gcc_cv_as_micromips_support): Use the
14188 --fatal-warnings option.
14189 * configure: Regenerate.
14191 2013-04-22 Marek Polacek <polacek@redhat.com>
14194 * tsan.c (instrument_expr): Don't instrument expression
14195 in case its size is zero.
14197 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
14201 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14203 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14205 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14207 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
14208 (gimple_stmt_iterator): New typedef.
14209 * gimple.h (gimple_stmt_iterator): Rename to...
14210 (gimple_stmt_iterator_d): ... This.
14211 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
14212 trees be valid for GIMPLE and GENERIC.
14213 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14214 * gimple-fold.c (gimple_fold_call): Call target hook
14215 gimple_fold_builtin.
14216 * hooks.c (hook_bool_gsiptr_false): New.
14217 * hooks.h (hook_bool_gsiptr_false): New.
14218 * target.def (fold_stmt): New.
14219 * doc/tm.texi: Regenerate.
14221 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14224 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14225 a set sp if no stack realignment.
14227 2013-04-22 Nick Clifton <nickc@redhat.com>
14229 * config.gcc (tilegx-linux): Extend extra_objs rather than
14231 (tilepro-linux): Likewise.
14233 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14235 * config/aarch64/aarch64-builtins.c
14237 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14238 (VAR<1-12>): Add MAP parameter.
14239 (BUILTIN_*): Likewise.
14240 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14241 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14242 (aarch64_ushl_n<mode>): Likewise.
14243 (aarch64_sshr_n<mode>): Likewise.
14244 (aarch64_ushr_n<mode>): Likewise.
14245 (aarch64_<maxmin><mode>): Likewise.
14246 (aarch64_sqrt<mode>): Likewise.
14247 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14248 (vshr<q>_n_*): Likewise.
14250 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14252 * config/aarch64/aarch64-builtins.c
14253 (aarch64_simd_builtin_type_mode): Handle SF types.
14255 (BUILTIN_GPF): Define.
14256 (aarch64_init_simd_builtins): Handle SF types.
14257 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14258 (frecps): Likewise.
14259 (frecpx): Likewise.
14260 * config/aarch64/aarch64-simd.md
14261 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14262 (aarch64_frecpe<mode>): New.
14263 (aarch64_frecps<mode>): Likewise.
14264 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14265 (v8type): Add frecp<esx>.
14266 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14267 (aarch64_frecps<mode>): Likewise.
14268 * config/aarch64/iterators.md (FRECP): New.
14269 (frecp_suffix): Likewise.
14270 * config/aarch64/arm_neon.h
14271 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14273 2013-04-22 Christian Bruel <christian.bruel@st.com>
14276 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14277 (REG_CLASS_NAMES): Idem.
14278 (REG_CLASS_CONTENTS): Idem.
14279 (REGCLASS_HAS_FP_REG): Idem.
14280 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14281 (sh_conditional_register_usage): Idem.
14283 2013-04-21 Jeff Law <law@redhat.com>
14285 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14286 (ssa_forward_propagate_and_combine): Use it.
14288 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14290 * lra.c: Update the flow chart diagram.
14292 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14294 PR rtl-optimization/56847
14295 * lra-constraints.c (process_alt_operands): Discourage alternative
14296 with non-matche doffsettable memory constraint fro memory with
14299 2013-04-19 Richard Biener <rguenther@suse.de>
14301 PR tree-optimization/56982
14302 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14304 * gimplify.c (gimplify_call_expr): Notice special calls.
14305 (gimplify_modify_expr): Likewise.
14306 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14307 abnormal control flow receivers.
14308 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14309 in the same way as cfun->has_nonlocal_labels.
14310 (gimple_purge_dead_abnormal_call_edges): Likewise.
14311 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14312 receivers start a basic-block.
14314 2013-04-19 Richard Biener <rguenther@suse.de>
14316 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14318 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14319 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14320 (SLP_TREE_LOAD_PERMUTATION): Add.
14321 (vect_transform_slp_perm_load): Adjust prototype.
14322 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14323 (vect_free_slp_instance): Likewise.
14324 (vect_create_new_slp_node): Likewise.
14325 (vect_supported_slp_permutation_p): Remove.
14326 (vect_slp_rearrange_stmts): Adjust.
14327 (vect_supported_load_permutation_p): Likewise. Inline
14328 vect_supported_slp_permutation_p here.
14329 (vect_analyze_slp_instance): Compute load permutations per
14330 slp node instead of per instance.
14331 (vect_get_slp_defs): Adjust.
14332 (vect_transform_slp_perm_load): Likewise.
14333 (vect_schedule_slp_instance): Remove redundant code.
14334 (vect_schedule_slp): Remove hack for PR56270, add it ...
14335 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14336 CSE loads for SLP. Adjust.
14338 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14340 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14341 spelling in two comments.
14343 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14346 * config/arm/arm.c (load_multiple_sequence): Require SP
14347 as base register for loads if SP is in the register list.
14349 2013-04-19 Martin Jambor <mjambor@suse.cz>
14351 PR tree-optimization/56718
14352 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14353 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14354 and made public. Adjusted all callers.
14355 (ipa_intraprocedural_devirtualization): New function.
14356 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14357 (ipa_intraprocedural_devirtualization): Likewise.
14358 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14360 2013-04-19 Richard Biener <rguenther@suse.de>
14362 PR tree-optimization/57000
14363 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14365 2013-04-19 Terry Guo <terry.guo@arm.com>
14367 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14369 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14370 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14371 (cortex_m4_fmacs): Use new reservations.
14372 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14374 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14376 PR rtl-optimization/56999
14377 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14379 (lra_coalesce): Remove split_origin_bitmap and related code.
14380 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14381 ranges if necessary.
14383 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
14385 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14387 (ix86_expand_call): Remove clobbered_registers array and use
14388 x86_64_ms_sysv_extra_clobbered_registers instead.
14389 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14391 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14393 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14394 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14395 (*call_value_rex64_ms_sysv): Ditto.
14397 2013-04-18 Cary Coutant <ccoutant@google.com>
14399 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14400 parent instead of die_mark.
14402 2013-04-18 Diego Novillo <dnovillo@google.com>
14404 * gimple.c (create_gimple_tmp): New.
14405 (get_expr_type): New.
14406 (build_assign): New.
14407 (build_type_cast): New.
14408 * gimple.h (enum ssa_mode): Define.
14409 (gimple_seq_set_location): New.
14410 * asan.c (build_check_stmt): Change some gimple_build_* calls
14411 to use build_assign and build_type_cast.
14413 2013-04-18 Richard Biener <rguenther@suse.de>
14415 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14416 handle negative step. Remove redundant checks.
14417 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14418 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14419 for negative step and grouped loads fail to vectorize.
14421 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14423 * emit-rtl.c (reset_insn_used_flags): New function.
14424 (reset_all_used_flags): Use it.
14425 (verify_insn_sharing): New function.
14426 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14428 2013-04-18 Jakub Jelinek <jakub@redhat.com>
14430 PR tree-optimization/56984
14431 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14432 and (x >> M) >= N don't register any assertion if N << M is the
14435 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14437 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14438 do not delete_insn insns that have not yet been emitted, only
14439 unlink them with remove_insn.
14440 * df-scan.c (df_insn_delete): Revert r197492.
14442 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14444 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14445 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14447 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14449 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14450 define_insn_and_split.
14451 (and_scc,ior_scc,negscc): Likewise.
14452 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14454 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14456 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14457 can benefit from using a sequence of LDRD instructions in epilogue
14458 instead of a single LDM instruction.
14460 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14463 * doc/extend.texi: Fix typo.
14465 2013-04-17 Richard Biener <rguenther@suse.de>
14467 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14468 (vect_build_slp_tree): ... here.
14469 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14470 match. Remove special-casing of mismatched complex loads.
14471 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14472 re-try the match with swapped commutative operands.
14473 (vect_supported_load_permutation_p): Remove special-casing of
14474 mismatched complex loads.
14475 (vect_analyze_slp_instance): Adjust.
14477 2013-04-17 Richard Biener <rguenther@suse.de>
14479 PR rtl-optimization/56921
14480 * cfgloop.h (struct loop): Add simple_loop_desc member.
14481 (struct niter_desc): Mark with GTY(()).
14482 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14483 * loop-iv.c (get_simple_loop_desc): Likewise.
14484 (free_simple_loop_desc): Likewise.
14487 2013-04-16 Richard Biener <rguenther@suse.de>
14489 PR rtl-optimization/56921
14490 * loop-init.c (pass_rtl_move_loop_invariants): Add
14491 TODO_do_not_ggc_collect to todo_flags_finish.
14492 (pass_rtl_unswitch): Same.
14493 (pass_rtl_unroll_and_peel_loops): Same.
14494 (pass_rtl_doloop): Same.
14496 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14498 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14499 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14500 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14502 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14503 * tree-streamer.c (record_common_node): Adjust reference in comment.
14505 2013-04-17 Terry Guo <terry.guo@arm.com>
14507 * config/arm/cortex-m4.md: Add a new bypass.
14509 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14511 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14513 (*subs_<optab><mode>_multp2): New pattern.
14514 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14515 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14517 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14519 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14520 (*subs_mul_imm_<mode>): New pattern.
14522 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
14525 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14526 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14527 (vsx_movti_32bit): Change j->wa to O->wa.
14529 2013-04-16 Richard Biener <rguenther@suse.de>
14531 PR rtl-optimization/56921
14532 * loop-init.c (pass_rtl_move_loop_invariants): Add
14533 TODO_do_not_ggc_collect to todo_flags_finish.
14534 (pass_rtl_unswitch): Same.
14535 (pass_rtl_unroll_and_peel_loops): Same.
14536 (pass_rtl_doloop): Same.
14538 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
14540 * config/arm/arm.c (emit_multi_reg_push): New declaration
14541 for an existing function.
14542 (arm_emit_strd_push): New function.
14543 (arm_expand_prologue): Used here.
14544 (arm_emit_ldrd_pop): New function.
14545 (arm_expand_epilogue): Used here.
14546 (arm_get_frame_offsets): Update condition.
14547 (arm_emit_multi_reg_pop): Add a special case for load of a single
14548 register with writeback.
14550 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
14552 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14555 2013-04-16 Richard Biener <rguenther@suse.de>
14557 PR tree-optimization/56756
14558 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14559 (first_mem_ref_loc): New.
14560 (execute_sm): Place the load temporarily before a previous
14561 access instead of in the latch edge to ensure its SSA dependencies
14562 are defined at points dominating the load.
14564 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14566 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14567 correct fix by moving header and footer insn to the footer of
14568 the merged basic block. Clear BB_END of the merged-away block.
14570 PR middle-end/43631
14571 * emit-rtl.c (make_note_raw): New function.
14572 (link_insn_into_chain): New static inline function.
14573 (add_insn): Use it.
14574 (add_insn_before, add_insn_after): Factor insn chain linking code...
14575 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14576 using link_insn_into_chain.
14577 (note_outside_basic_block_p): New helper function for emit_note_after
14578 and emit_note_before.
14579 (emit_note_after): Use nobb variant of add_insn_after if the note
14580 should not be contained in a basic block.
14581 (emit_note_before): Use nobb variant of add_insn_before if the note
14582 should not be contained in a basic block.
14583 (emit_note_copy): Use make_note_raw.
14584 (emit_note): Likewise.
14585 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14586 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14587 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14588 the moved barrier the tail of the basic block it follows.
14589 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14591 2013-04-15 Jakub Jelinek <jakub@redhat.com>
14593 PR tree-optimization/56962
14594 * gimple-ssa-strength-reduction.c (record_increment): Only set
14595 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14596 either rhs1 or rhs2 is equal to c->base_expr.
14598 2013-04-15 Richard Biener <rguenther@suse.de>
14600 PR tree-optimization/56933
14601 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14603 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14604 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14605 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14606 dependence check ...
14607 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14609 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14610 GROUP_READ_WRITE_DEPENDENCE.
14612 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14614 * emit-rtl.c (reset_all_used_flags): New function.
14615 (verify_rtl_sharing): Call reset_all_used_flags before and after
14616 performing the checks.
14618 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14620 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14621 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14622 * config/arm/constraints.md (De): New constraint.
14623 * config/arm/neon.md (anddi3_neon): Delete.
14624 (neon_vand<mode>): Expand to standard anddi3 pattern.
14625 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14626 Move earlier in the file.
14627 (neon_inv_logic_op2): Likewise.
14628 (arm_anddi_operand_neon): New predicate.
14630 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14632 * configure.ac (gcc_cv_ld_as_needed): Set
14633 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14634 Use -z ignore, -z record on *-*-solaris2*.
14635 (HAVE_LD_AS_NEEDED): Update comment.
14636 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14637 * configure: Regenerate.
14638 * config.in: Regenerate.
14639 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14640 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14641 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14642 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14643 equivalents. Fix markup.
14644 * doc/tm.texi: Regenerate.
14646 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14648 * config/i386/i386.opt: New option mstack-protector-guard=.
14649 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14650 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14651 TARGET_SSP_TLS_GUARD.
14652 * config/i386/i386.c (ix86_option_override_internal): Set
14653 ix86_stack_protector_guard.
14654 * config/i386/i386.md (stack_protect_set): Enable for
14655 TARGET_SSP_TLS_GUARD only.
14656 (stack_protect_set_<mode>): Ditto.
14657 (stack_protect_test): Ditto.
14658 (stack_protect_test_<mode>): Ditto.
14659 * doc/invoke.texi (i386 Option): Document.
14661 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14664 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14665 (S_MODES): Set H_MODE bit.
14666 (SF_MODES): Set only S_MODE and SF_MODE bits.
14667 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14668 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14669 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14670 <MODE_FLOAT>: Likewise.
14672 2013-04-15 Joey Ye <joey.ye@arm.com>
14674 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14676 2013-04-15 Joey Ye <joey.ye@arm.com>
14678 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14680 (thumb_far_jump_used_p): Count instruction size and set
14683 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14685 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14686 * resource.c (mark_target_live_regs): Reformat conditional block.
14688 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14690 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14691 notes, they are emitted only just before final.
14692 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14694 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14696 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14697 * cfgrtl.c (delete_insn): Call it here instead.
14698 * lra-spills.c (lra_final_code_change): Use delete_insn.
14699 * haifa-sched.c (sched_remove_insn): Likewise.
14700 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14701 returning to the nop pool.
14702 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14703 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14705 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14707 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14708 * doc/tm.texi: Regenerated.
14710 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
14712 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14715 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14717 * df-core.c (df_find_def): Compare register numbers.
14718 (df_find_use): Likewise.
14720 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14723 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14724 lra_in_progress for return.
14726 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14728 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14729 define_insn into define_insn_and_split and emit movsicc patterns.
14731 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14733 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14735 2013-04-12 Richard Biener <rguenther@suse.de>
14737 * tree-pass.h (TODO_do_not_ggc_collect): New.
14738 * passes.c (execute_one_ipa_transform_pass): Honor
14739 TODO_do_not_ggc_collect.
14740 (execute_one_pass): Likewise.
14743 2013-04-10 Richard Biener <rguenther@suse.de>
14745 * passes.c (init_optimization_passes): Remove reload pass.
14746 * ira.c (do_reload): Merge into ...
14748 (rest_of_handle_reload): Remove.
14749 (pass_reload): Likewise.
14750 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14751 of reload for vzeroupper pass placement.
14753 2013-04-12 Jakub Jelinek <jakub@redhat.com>
14755 PR tree-optimization/56918
14756 PR tree-optimization/56920
14757 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14758 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14759 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14760 use wide_mul_with_sign method.
14762 2013-04-12 Richard Biener <rguenther@suse.de>
14764 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14765 not be considered a gimple constant.
14767 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
14769 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14770 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14773 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14774 Jakub Jelinek <jakub@redhat.com>
14776 * opts.c: Include diagnostic-color.h.
14777 (common_handle_option): Handle OPT_fdiagnostics_color_.
14778 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14779 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14780 (diagnostic-color.o): New.
14781 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14782 (diagnostic_color_rule): New enum.
14783 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14784 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14785 the location string.
14786 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14787 either NULL, or color kind.
14788 * diagnostic-color.c: New file.
14789 * diagnostic-color.h: New file.
14790 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14792 * doc/invoke.texi (-fdiagnostics-color): Document.
14793 * pretty-print.h (pp_show_color): Define.
14794 (struct pretty_print_info): Add show_color field.
14795 * diagnostic.c: Include diagnostic-color.h.
14796 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14797 macros. Colorize error:, warning: etc. strings and also the location
14799 (diagnostic_show_locus): Colorize the caret line.
14800 * pretty-print.c: Include diagnostic-color.h.
14801 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14802 inside of %< %> quotes or quoted through q format modifier.
14804 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14806 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
14808 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14810 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
14811 code in CC_NZ mode.
14812 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
14815 2013-04-11 Marek Polacek <polacek@redhat.com>
14817 PR tree-optimization/48184
14818 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
14820 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
14822 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
14823 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
14824 (skip_simple_arithmetic): Tidy up.
14825 * tree.h (skip_simple_constant_arithmetic): Declare.
14827 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14829 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
14831 2013-04-11 Richard Biener <rguenther@suse.de>
14833 * tree-vect-loop.c (get_initial_def_for_induction): Properly
14834 generate vector constants.
14836 2013-04-11 Richard Biener <rguenther@suse.de>
14838 PR tree-optimization/56878
14839 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
14840 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
14842 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
14843 Prefer to align the DR with the most invariant base address.
14845 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14847 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
14850 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
14852 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
14853 floating-point vector comparisons against 0.
14855 2013-04-11 Jakub Jelinek <jakub@redhat.com>
14857 PR tree-optimization/56899
14858 * fold-const.c (extract_muldiv_1): Apply distributive law
14859 only if TYPE_OVERFLOW_WRAPS (ctype).
14861 2013-04-11 Bin Cheng <bin.cheng@arm.com>
14864 * ira-costs.c (scan_one_insn): Check whether the source rtx of
14865 loading has side effect.
14867 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
14869 * config/sparc/sparc.c: Include tree-pass.h.
14870 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
14871 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
14872 head of file. Change return type. Split off gate function.
14873 (sparc_gate_work_around_errata): New function.
14874 (pass_work_around_errata): New pass definition.
14875 (insert_pass_work_around_errata) New pass insert definition to
14876 insert pass_work_around_errata just after delayed-branch scheduling.
14877 (sparc_option_override): Insert the pass.
14878 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
14880 2013-04-10 David S. Miller <davem@davemloft.net>
14882 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
14883 or -mcpu=hypersparc.
14885 * target.def (cstore_mode): New hook.
14886 * target.h: Include insn-codes.h
14887 * targhooks.c: Likewise.
14888 (default_cstore_mode): New function.
14889 * targhooks.h: Declare it.
14890 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
14891 * doc/tm.texi: Rebuild.
14892 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
14893 target hook, rather than inspecting the insn_data.
14894 * config/sparc/sparc.c (sparc_cstore_mode): New function.
14895 (TARGET_CSTORE_MODE): Redefine.
14896 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
14898 * config/sparc/predicates.md (cstore_result_operand): New special
14900 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
14901 Use it for operand 0.
14902 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
14903 (*snesi_special): Likewise.
14904 (*snesi_zero): Likewise.
14905 (*seqsi_zero): Likewise.
14906 (*sltu_insn): Likewise.
14907 (*sgeu_insn): Likewise.
14908 (*seqdi_special): Make operand 0 and comparison operation be of
14910 (*snedi_special): Likewise.
14911 (*snedi_special_vis3): Likewise.
14912 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
14913 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
14914 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
14915 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
14916 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
14917 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
14918 (*sltu_extend_sp64): Likewise.
14919 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
14920 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
14921 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
14922 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
14923 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
14925 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
14927 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
14928 (aarch64_start_file): Use the new function.
14930 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
14931 Jason Merrill <jason@redhat.com>
14933 * common.opt: Add -gdwarf.
14934 * opts.c (common_handle_option): Handle it.
14935 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
14937 2013-04-10 Richard Biener <rguenther@suse.de>
14939 * passes.c (execute_todo): Do not call ggc_collect conditional here.
14940 (execute_one_ipa_transform_pass): But unconditionally here.
14941 (execute_one_pass): And here.
14942 (init_optimization_passes): Remove reload pass.
14943 * tree-pass.h (TODO_ggc_collect): Remove.
14944 (pass_reload): Likewise.
14945 * ira.c (do_reload): Merge into ...
14947 (rest_of_handle_reload): Remove.
14948 (pass_reload): Likewise.
14949 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14950 of reload for vzeroupper pass placement.
14951 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
14952 and todo_flags_finish of all passes.
14954 2013-04-10 Richard Biener <rguenther@suse.de>
14956 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
14957 first_const_oprnd field, rename first_def_type to first_op_type.
14958 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
14959 (vect_get_and_check_slp_defs): Always use the type of the
14960 operand. Allow mixed vect_external_def, vect_constant_def types.
14961 (vect_get_constant_vectors): Handle mixed vect_external_def,
14962 vect_constant_def types.
14964 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
14966 PR tree-optimization/55524
14967 * tree-ssa-math-opts.c
14968 (convert_mult_to_fma): Don't use an fms construct
14969 when we don't have an fms operation, but fnma, and it looks
14970 likely that we'll be able to use the latter.
14972 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
14974 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
14976 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
14977 inline fail caused by overwritable functions.
14979 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
14981 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
14982 unnecessary bits in the constant power of two case.
14984 2013-04-10 Richard Biener <rguenther@suse.de>
14986 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
14987 broken code swapping operands.
14988 (vect_build_slp_tree): Do not compute load permutations here.
14989 (vect_analyze_slp_instance): Compute load permutations here,
14990 after building the SLP tree.
14992 2013-04-09 Christian Bruel <christian.bruel@st.com>
14994 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
14995 of next/prev_real_insn.
14997 2013-04-09 Jan Hubicka <jh@suse.cz>
14999 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15000 Drop aliased parameter.
15001 (function_and_variable_visibility): Do not handle alias pairs.
15002 * cgraph.c (varpool_externally_visible_p): Update prototype.
15003 * varpool.c (varpool_add_new_variable): Update.
15005 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15007 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
15009 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
15011 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15013 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15015 2013-04-09 Marek Polacek <polacek@redhat.com>
15017 PR tree-optimization/48762
15018 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
15020 2013-04-09 Richard Biener <rguenther@suse.de>
15022 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15024 (vect_build_slp_tree): Likewise.
15025 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15026 calculating the cost of a SLP instance.
15027 (vect_analyze_slp_instance): Use it from here, after building
15030 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15032 PR middle-end/56883
15033 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15034 expand_omp_for_static_chunk): Use simple_p = true in
15035 force_gimple_operand_gsi calls when assigning to addressable decls.
15037 2013-04-09 Jeff Law <law@redhat.com>
15039 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15040 when the boolean was created by converting a wider object which
15041 had a boolean range.
15043 2013-04-09 Richard Biener <rguenther@suse.de>
15045 * tree-vectorizer.h (slp_void_p): Remove.
15046 (slp_tree): Typedef before _slp_tree declaration.
15047 (struct _slp_tree): Use a vector of slp_tree as children.
15048 (vect_get_place_in_interleaving_chain): Remove.
15049 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15051 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15053 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15054 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15055 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15056 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15057 Use slp_node instead of slp_void_p and adjust.
15059 2013-04-09 Richard Biener <rguenther@suse.de>
15061 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15062 work that is not necessary.
15064 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15066 PR tree-optimization/56854
15067 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15068 forward into clobber stmts if it would change MEM_REF lhs into
15071 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15073 * tree.c (type_hash_lookup, type_hash_add): Make static.
15074 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15076 2013-04-09 Richard Biener <rguenther@suse.de>
15078 * tree.h (unsave_expr_now): Remove.
15079 * tree-inline.c (mark_local_for_remap_r): Remove.
15080 (unsave_expr_1): Likewise.
15081 (unsave_r): Likewise.
15082 (unsave_expr_now): Likewise.
15083 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15084 (propagate_tree_value): Likewise.
15086 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15088 * doc/rtl.texi (sequence): Rewrite documentation to match the
15089 current use of SEQUENCE rtl objects.
15090 * rtl.def (SEQUENCE): Likewise.
15092 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15093 Update documentation.
15094 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15095 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15097 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15099 2013-04-08 Teresa Johnson <tejohnson@google.com>
15101 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15102 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
15104 (estimate_edge_size_and_time): Add comment to suggest using rounding
15106 (estimate_node_size_and_time): Ditto.
15107 (remap_edge_change_prob): Use helper rounding divide methods.
15108 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15109 (gimple_mod_pow2_value_transform): Ditto.
15110 (gimple_mod_subtract_transform): Ditto.
15111 (gimple_ic_transform): Ditto.
15112 (gimple_stringops_transform): Ditto.
15113 * stmt.c (conditional_probability): Ditto.
15114 (emit_case_dispatch_table): Ditto.
15115 * lto-cgraph.c (merge_profile_summaries): Ditto.
15116 * tree-optimize.c (execute_fixup_cfg): Ditto.
15117 * cfgcleanup.c (try_forward_edges): Ditto.
15118 * cfgloopmanip.c (scale_loop_profile): Ditto.
15120 (duplicate_loop_to_header_edge): Ditto.
15121 (lv_adjust_loop_entry_edge): Ditto.
15122 * tree-vect-loop.c (vect_transform_loop): Ditto.
15123 * profile.c (compute_branch_probabilities): Ditto.
15124 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15125 * lto-streamer-in.c (input_cfg): Ditto.
15126 * gimple-streamer-in.c (input_bb): Ditto.
15127 * ipa-cp.c (update_profiling_info): Ditto.
15128 (update_specialized_profile): Ditto.
15129 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15130 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
15132 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15133 (compute_trg_info): Ditto.
15134 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15135 (purge_dead_edges): Ditto.
15136 * loop-unswitch.c (unswitch_loop): Ditto.
15137 * cgraphclones.c (cgraph_clone_edge): Ditto.
15138 (cgraph_clone_node): Ditto.
15139 * tree-inline.c (copy_bb): Ditto.
15140 (copy_edges_for_bb): Ditto.
15141 (initialize_cfun): Ditto.
15142 (copy_cfg_body): Ditto.
15143 (expand_call_inline): Ditto.
15145 2013-04-08 Kai Tietz <ktietz@redhat.com>
15147 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15148 TARGET_CYGWIN64 by TARGET_64BIT.
15150 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15152 * config/epiphany/epiphany.md (GPR_1): New constant.
15153 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15154 * config/epiphany/epiphany.c (gen_compare_reg):
15155 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15156 is already in place.
15157 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15158 Don't require being called during rtl expansion; If y operlaps r0,
15160 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15161 (epiphany_expand_epilogue): Likewise.
15163 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15164 Don't use CC_FPmode for ORDERED / UNORDERED.
15165 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
15167 * config/epiphany/constraints.md (CnL): New constraint.
15168 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15169 * config/epiphany/predicates.md (add_operand): Allow 1024.
15171 * config/epiphany/epiphany.md (logical_op): New code iterator.
15172 (op_mnc): New code attribute.
15173 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15174 (mov_f+1, mov_f+2): New peephole2 patterns.
15176 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15177 (cstoresi4): Also allow re-use of zero result when doing a NE
15178 comparison to a non-zero operand.
15179 Use (clobber (scratch)) for first insn if the gpr output is not needed.
15181 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15182 Use gen_addsi3_i / gen_subsi3_i.
15184 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15188 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15189 contain anything but clobbers, at most one __builtin_stack_restore,
15190 optionally debug stmts and final resx, and if it has at least one
15191 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15192 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15193 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15194 which isn't defaut definition, remove them.
15195 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15196 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15197 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15198 with MEM_REF LHS with SSA_NAME address.
15200 2013-04-08 Jeff Law <law@redhat.com>
15202 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
15204 2013-04-08 Richard Biener <rguenther@suse.de>
15206 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
15208 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
15209 determined vector type.
15210 (vect_analyze_data_refs): Likewise.
15211 (vect_get_new_vect_var): Adjust.
15212 (vect_create_destination_var): Preserve SSA name versions.
15213 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15214 not dump anything here.
15216 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15218 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15219 Add member lr_slot_known.
15220 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15222 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15223 Remove code that sets lr_slot_offset according to what a previous
15224 version of epiphany_emit_save_restore used to do.
15225 (epiphany_emit_save_restore): When doing an lr save or restore,
15226 set/verify lr_slot_known and lr_slot_offset.
15228 2013-04-08 Xinyu Qi <xyqi@marvell.com>
15231 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
15234 2013-04-08 Richard Biener <rguenther@suse.de>
15236 * alias.c (find_base_term): Fix thinko in previous change.
15238 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15240 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15241 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15242 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15243 if possible to compute val.
15244 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15245 For QImode integers don't require anything about precision. Use
15246 const_with_all_bytes_same to find out if the constant doesn't have
15247 repeated bytes in it.
15249 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15251 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15254 2013-04-08 Marek Polacek <polacek@redhat.com>
15256 PR rtl-optimization/48182
15257 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15260 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15263 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15264 nuses, make sure we have a label.
15266 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15269 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15270 (rs6000_emit_swdiv_low_precision): Remove.
15271 (rs6000_emit_swdiv): Rewrite to handle between one and four
15272 iterations of Newton-Raphson generally; modify required number of
15273 iterations for some cases.
15274 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15276 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15278 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15279 set-but-unused variable.
15281 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15282 basic blocks of released function bodies garbage-collectable.
15284 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15285 (struct rtl_opt_pass): Add TODO_df_finish.
15287 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15289 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15291 * config/arm/constraints.md (q): New constraint.
15292 * config/arm/ldrdstrd.md: New file.
15293 * config/arm/arm.md (ldrdstrd.md) New include.
15294 (arm_movdi): Use "q" instead of "r" constraint
15295 for double-word memory access.
15296 (movdf_soft_insn): Likewise.
15297 * config/arm/vfp.md (movdi_vfp): Likewise.
15298 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
15299 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
15300 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15301 (mem_ok_for_ldrd_strd): Likewise.
15302 (output_move_double): Update assertion.
15304 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15306 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15308 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15310 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15311 define_insn_and_split.
15312 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15314 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15316 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15317 define_insn_and_split.
15318 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
15319 (shiftsi3_compare): New pattern.
15320 (rrx): New pattern.
15321 * config/arm/unspecs.md (UNSPEC_RRX): New.
15323 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15325 * config/arm/arm.md (negdi_extendsidi): New pattern.
15326 (negdi_zero_extendsidi): Likewise.
15328 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15330 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15331 define_insn_and_split.
15332 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15333 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15335 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15337 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15338 define_insn_and_split.
15339 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15340 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15342 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15344 * config/arm/arm.md (subsi3_carryin): New pattern.
15345 (subsi3_carryin_const): Likewise.
15346 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15347 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15349 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15351 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15353 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15355 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
15356 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
15358 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15360 * config/arm/arm.c (arm_expand_builtin): Change fcode
15361 type to unsigned int.
15363 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15365 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15367 2013-04-04 Ian Lance Taylor <iant@google.com>
15369 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15372 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15374 PR middle-end/56729
15375 * df-scan.c (df_insn_delete): Disable failing assert.
15377 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15379 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15380 New function prototype.
15381 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15382 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15383 (arm_builtin_vectorized_function): New function.
15385 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15387 * config/arm/arm_neon_builtins.def: New file.
15388 * config/arm/arm.c (neon_builtin_data): Move contents to
15389 arm_neon_builtins.def.
15390 (enum arm_builtins): Include neon builtin definitions.
15391 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
15392 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
15394 2013-04-04 Marek Polacek <polacek@redhat.com>
15396 PR tree-optimization/48186
15397 * predict.c (maybe_hot_frequency_p): Return false if
15398 HOT_BB_FREQUENCY_FRACTION is 0.
15399 (cgraph_maybe_hot_edge_p): Likewise.
15401 2013-04-04 Richard Biener <rguenther@suse.de>
15403 PR tree-optimization/56826
15404 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15407 2013-04-04 Richard Biener <rguenther@suse.de>
15409 PR tree-optimization/56213
15410 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
15411 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
15413 2013-04-04 Richard Biener <rguenther@suse.de>
15415 PR tree-optimization/56837
15416 * tree-loop-distribution.c (classify_partition): For non-zero
15417 values require that the value has the same precision as its
15418 mode to be useful as memset value.
15420 2013-04-03 Nick Clifton <nickc@redhat.com>
15422 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
15423 (fmssf4): Use fmsf.s on E3V5 architectures.
15424 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15425 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15427 2013-04-03 Jeff Law <law@redhat.com>
15429 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15430 (lra-eliminations.o): Likewise.
15432 2013-04-03 Teresa Johnson <tejohnson@google.com>
15434 * gcov-io.c (compute_working_sets): Moved most of body of old
15435 compute_working_sets here from profile.c.
15436 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
15437 (gcov_working_set_t): Moved typedef here from basic-block.h
15438 (compute_working_set): Declare.
15439 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15440 (get_working_sets): Renamed from compute_working_set,
15441 replace most of body with call to new compute_working_sets.
15442 (get_exec_counts): Replace call to compute_working_sets
15443 to get_working_sets.
15444 * profile.h (get_working_sets): Renamed from compute_working_set.
15445 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
15446 to get_working_sets.
15447 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15448 * gcov-dump.c (dump_working_sets): New function.
15450 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
15452 * hwint.c (sext_hwi, zext_hwi): New functions.
15453 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15454 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15455 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15456 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15457 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15458 (sext_hwi, zext_hwi): New functions.
15460 2013-04-03 Jeff Law <law@redhat.com>
15462 PR tree-optimization/56799
15463 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15464 back test for widening conversion erroneously dropped in prior change.
15466 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15469 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15470 instead of next_real_insn.
15472 2013-04-03 Marek Polacek <polacek@redhat.com>
15475 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
15477 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15480 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15482 (thumb1_output_casesi): Likewise.
15483 (thumb2_output_casesi): Likewise.
15485 2013-04-03 Richard Biener <rguenther@suse.de>
15487 PR tree-optimization/56817
15488 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15490 (tree_unroll_loops_completely_1): ... new function to manually
15491 walk the loop tree, properly defering outer loops of unrolled
15492 loops to later iterations.
15494 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15496 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15497 (vectorizable_load): Likewise.
15498 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15499 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15501 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15503 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15506 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15508 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15510 2013-04-03 Bin Cheng <bin.cheng@arm.com>
15512 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15514 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15516 PR tree-optimization/56790
15517 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15520 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15522 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15524 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15525 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15528 2013-04-03 Jakub Jelinek <jakub@redhat.com>
15531 * tree.h (force_folding_builtin_constant_p): New decl.
15532 * builtins.c (force_folding_builtin_constant_p): New variable.
15533 (fold_builtin_constant_p): Fold immediately also if
15534 force_folding_builtin_constant_p.
15536 2013-04-03 Richard Biener <rguenther@suse.de>
15538 PR tree-optimization/56812
15539 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15540 DRs of the same interleaving chain are independent.
15542 2013-04-02 Jason Merrill <jason@redhat.com>
15544 * gdbinit.in (pbb): Use debug fn.
15546 2013-04-02 Lawrence Crowl <crowl@google.com>
15548 * sese.h (struct ivtype_map_elt_s): Remove unused.
15549 (extern debug_ivtype_map): Remove unused.
15550 (extern eq_ivtype_map_elts): Remove unused.
15551 * sese.c (debug_ivtype_map): Removed unused.
15552 (debug_ivtype_map_1): Removed unused.
15553 (debug_ivtype_elt): Remove unused.
15554 (eq_ivtype_map_elts): Remove unused.
15557 2013-04-02 Kai Tietz <ktietz@redhat.com>
15560 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15561 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
15562 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15564 (legitimize_pe_coff_symbol): Likewise.
15565 (is_imported_p): New helper-function.
15566 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15567 for Windows x64 targets.
15568 (ix86_expand_prologue): Optimize for pe-coff targets.
15569 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15570 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15571 medium/large code-model.
15572 (legitimize_pic_address): Likewise.
15573 (legitimize_tls_address): Likewise.
15574 (ix86_expand_call): Likewise.
15575 (x86_output_mi_thunk): Likewise.
15576 (get_dllimport_decl): Add new beimport argument.
15577 (construct_plt_address): Don't assert for x64 pe-coff targets.
15578 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15580 (SYMBOL_FLAG_STUBVAR): New macro.
15581 (SYMBOL_REF_STUBVAR_P): Likewise.
15582 * config/i386/winnt.c (stub_list): New structure.
15583 (stub_head): New local variable.
15584 (i386_pe_record_stub): New function.
15585 (i386_pe_file_end): Emit refptr-stubs.
15587 2013-04-02 Jakub Jelinek <jakub@redhat.com>
15589 PR rtl-optimization/56745
15590 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15591 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15594 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15595 and both of them are MEM_REFs, just compare first argument for
15596 equality and attempt to deal even with differing offsets.
15599 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15600 of gimple_clobber_p to be MEM_REF.
15601 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15602 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15603 after gimplification.
15604 * asan.c (get_mem_ref_of_assignment): Don't instrument
15605 gimple_clobber_p stmts.
15606 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15607 gimple_clobber_p stmt if they have MEM_REF lhs and
15608 are dead because of another gimple_clobber_p stmt.
15609 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15610 gimple_clobber_p stmts like debug stmts.
15611 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15612 that refer to unused VAR_DECLs or uninitialized values.
15613 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15614 gimple_clobber_p stmts if they refer to removed parameters.
15615 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15618 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
15620 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15621 using SWI48 mode attribute.
15623 2013-04-02 Wei Mi <wmi@google.com>
15625 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15626 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15627 *<rotate_insn><mode>3_mask in i386.md.
15629 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15631 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15633 2013-04-02 Richard Biener <rguenther@suse.de>
15635 PR tree-optimization/56778
15636 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15637 Runtime alias tests are not supported for gather loads.
15638 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15639 stmts referenced from SSA operands before updating SSA form.
15641 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15642 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15644 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
15645 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15646 * config/arm/cortex-a53.md: New file.
15647 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15648 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15649 * config/arm/arm.c (arm_issue_rate): Likewise.
15650 * config/arm/arm-tune.md: Regenerate
15651 * config/arm/arm-tables.opt: Regenerate.
15652 * config/arm/arm-cores.def: Add cortex-a53.
15654 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15656 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15659 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15661 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15662 scalar load/store operations using B/H registers.
15663 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15665 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15667 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15669 * config/aarch64/aarch64.c
15670 (aarch64_simd_scalar_immediate_valid_for_move): New.
15671 * config/aarch64/aarch64-protos.h
15672 (aarch64_simd_scalar_immediate_valid_for_move): New.
15673 * config/aarch64/constraints.md (Dh, Dq): New.
15674 * config/aarch64/iterators.md (hq): New.
15676 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15678 * reorg.c (get_branch_condition): Deal with conditional returns.
15679 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15681 2013-04-01 Wei Mi <wmi@google.com>
15683 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15684 Truncate operand 2 using %b asm operand modifier.
15685 (*<shift_insn><mode>3_mask): Ditto.
15686 (*<rotate_insn><mode>3_mask): Ditto.
15688 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15690 PR middle-end/56798
15691 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15693 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15695 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15697 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15699 2013-03-30 Lawrence Crowl <crowl@google.com>
15701 * dse.c (clear_alias_sets): Remove never set.
15702 (disqualified_clear_alias_sets): Remove never set.
15703 (clear_alias_mode_pool): Remove never set.
15704 (dse_step0): Remove condition that is never true.
15705 (canon_address): Remove condition that is never true.
15706 (dse_step7): Remove condition that is never true.
15707 (rest_of_handle_dse): Remove condition that is never true.
15708 (rest_of_handle_dse::did_global): Remove never read from above.
15709 (dse_step2_spill): Remove never called from above.
15710 (dse_step5_spill): Remove never called from above.
15712 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15714 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15715 JUMP_TABLE_DATA changes.
15716 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15717 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15718 (Insns) <jump_table_data>: New entry.
15719 * doc/tm.texi: Regenerate.
15721 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15723 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15724 for table jump at the end of a basic block using tablejump_p.
15725 * targhooks.c (default_invalid_within_doloop): Likewise.
15726 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15727 target hook implementation that is identical to the default hook.
15728 (rs6000_invalid_within_doloop): Remove.
15730 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15731 unused variable from tablejump_p call.
15733 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15734 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15735 (INSN_DELETED_P): Likewise.
15736 (emit_jump_table_data): New prototype.
15737 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15738 after 4th as unused.
15739 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15740 * sched-vis.c (print_insn): Likewise.
15741 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15742 insn for compatibility with back ends that use next_active_insn to
15743 identify jump table data.
15744 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15745 (remove_insn): Likewise.
15746 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15748 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15749 (emit_jump_table_data): New function.
15751 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15752 basic block, a JUMP_TABLE_DATA never is.
15753 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15754 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15755 off from code handling real insns.
15756 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15757 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15758 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15759 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15760 is not a NONDEBUG_INSN_P.
15761 * ira-costs.c (scan_one_insn): Likewise.
15762 * jump.c (mark_all_labels): Likewise.
15763 (mark_jump_label_1): Likewise.
15764 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15765 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15767 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15768 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15769 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15770 JUMP_TABLE_DATA_P insns.
15771 (calculate_elim_costs_all_insns): Likewise.
15772 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15773 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15774 (delete_output_reload): Code style fixups.
15775 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15776 insn flags on this non-insn.
15777 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15778 as scheduling barriers, for pre-change compatibility.
15779 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15780 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15782 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15783 redundant JUMP_TABLE_DATA_P test.
15784 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15785 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15786 (frv_for_each_packet): Likewise.
15787 * config/i386/i386.c (min_insn_size): Likewise.
15788 (ix86_avoid_jump_mispredicts): Likewise.
15789 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15790 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15791 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15792 (mips16_insn_length): Robustify.
15793 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15794 (mips16_split_long_branches): Likewise.
15795 * config/pa/pa.c (pa_combine_instructions): Likewise.
15796 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15797 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15798 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15799 as contributing to pool range lengths.
15800 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15801 Remove redundant JUMP_TABLE_DATA_P test.
15802 (sh_loop_align): Likewise.
15803 (split_branches): Likewise.
15804 (sh_insn_length_adjustment): Likewise.
15805 * config/spu/spu.c (get_branch_target): Likewise.
15807 2013-03-29 Jan Hubicka <jh@suse.cz>
15809 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
15810 gcov streaming; stream hot bb threshold to ltrans.
15811 * predict.c (get_hot_bb_threshold): Break out from ....
15812 (maybe_hot_count_p): ... here.
15813 (set_hot_bb_threshold): New function.
15814 * lto-section-in.c (lto_section_name): Add profile.
15815 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
15816 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
15817 and data-streamer.h
15818 (histogram_entry): New structure.
15819 (histogram, histogram_pool): New global vars.
15820 (histogram_hash): New structure.
15821 (histogram_hash::hash): New method.
15822 (histogram_hash::equal): Likewise.
15823 (account_time_size): New function.
15824 (cmp_counts): New function.
15825 (dump_histogram): New function.
15826 (ipa_profile_generate_summary): New function.
15827 (ipa_profile_write_summary): New function.
15828 (ipa_profile_read_summary): New function.
15829 (ipa_profile): Decide on threshold.
15830 (pass_ipa_profile): Add ipa_profile_write_summary and
15831 ipa_profile_read_summary.
15832 * Makefile.in (ipa.o): Update dependencies.
15833 * lto-streamer.h (LTO_section_ipa_profile): New section.
15835 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
15837 * tree.h (VAR_P): New.
15839 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
15842 * doc/invoke.texi ([-fwhole-program]): Fix typo.
15844 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
15846 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
15847 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15848 (control_flow_insn_p): Likewise.
15849 * cfgrtl.c (duplicate_insn_chain): Likewise.
15850 * final.c (get_attr_length_1): Likewise.
15851 (shorten_branches): Likewise.
15852 (final_scan_insn): Likewise.
15853 * function.c (instantiate_virtual_regs): Likewise.
15854 * gcse.c (insert_insn_end_basic_block): Likewise.
15855 * ira-costs.c (scan_one_insn): Likewise.
15856 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15857 * lra.c (check_rtl): Likewise.
15858 * reload1.c (elimination_costs_in_insn): Likewise.
15859 * reorg.c (follow_jumps): Likewise.
15861 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
15862 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
15863 (thumb_far_jump_used_p): Likewise.
15864 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
15865 (workaround_speculation): Likewise.
15866 (add_sched_insns_for_speculation): Likewise.
15867 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
15868 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15869 (frv_for_each_packet): Likewise.
15870 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
15871 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
15872 (final_emit_insn_group_barriers): Likewise.
15873 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15874 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15875 (mips16_insn_length): Likewise.
15876 * config/pa/pa.c (pa_reorg): Likewise.
15877 (pa_combine_instructions): Likewise.
15878 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
15879 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
15880 (sh_reorg): Likewise.
15881 (split_branches): Likewise.
15882 * config/spu/spu.c (get_branch_target): Likewise.
15884 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
15887 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
15889 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
15890 Fix declaration name.
15892 2013-03-28 Lawrence Crowl <crowl@google.com>
15894 * graphds.h (struct graph.indicies): Remove unused.
15895 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
15896 (SCOP_ORIGINAL_PDDRS): Remove unused.
15897 * sese.h (extern insert_loop_close_phis): Removed unused.
15898 (extern insert_guard_phis): Removed unused.
15899 (extern ivtype_map_elt_info): Removed unused.
15900 (new_ivtype_map_elt): Removed unused.
15901 * sese.c (ivtype_map_elt_info): Removed unused.
15903 2013-03-28 Lawrence Crowl <crowl@google.com>
15905 * Makefile.in: Add several missing include dependences.
15907 (test-dump.o): New. This object is not added to any executable,
15908 but is present for ad-hoc testing.
15910 (debug (const bitmap_head_def &)): New.
15911 (debug (const bitmap_head_def *)): New.
15913 (extern debug (const bitmap_head_def &)): New.
15914 (extern debug (const bitmap_head_def *)): New.
15916 (debug (edge_def &)): New.
15917 (debug (edge_def *)): New.
15919 (debug (basic_block_def &)): New.
15920 (debug (basic_block_def *)): New.
15922 (dump_node (const_tree, int, FILE *)): Correct source file.
15924 (debug (die_struct &)): New.
15925 (debug (die_struct *)): New.
15927 (extern debug (die_struct &)): New.
15928 (extern debug (die_struct *)): New.
15929 * gimple-pretty-print.c
15930 (debug (gimple_statement_d &)): New.
15931 (debug (gimple_statement_d *)): New.
15932 * gimple-pretty-print.h
15933 (extern debug (gimple_statement_d &)): New.
15934 (extern debug (gimple_statement_d *)): New.
15936 (debug (ira_allocno_copy &)): New.
15937 (debug (ira_allocno_copy *)): New.
15938 (debug (ira_allocno &)): New.
15939 (debug (ira_allocno *)): New.
15941 (extern debug (ira_allocno_copy &)): New.
15942 (extern debug (ira_allocno_copy *)): New.
15943 (extern debug (ira_allocno &)): New.
15944 (extern debug (ira_allocno *)): New.
15946 (debug (live_range &)): New.
15947 (debug (live_range *)): New.
15949 (debug (lra_live_range &)): New.
15950 (debug (lra_live_range *)): New.
15952 (debug (lra_live_range &)): New.
15953 (debug (lra_live_range *)): New.
15955 (debug (omega_pb_d &)): New.
15956 (debug (omega_pb_d *)): New.
15958 (extern debug (omega_pb_d &)): New.
15959 (extern debug (omega_pb_d *)): New.
15961 (debug (const rtx_def &)): New.
15962 (debug (const rtx_def *)): New.
15964 (debug_tree (tree): Move within file.
15965 (debug_raw (const tree_node &)): New.
15966 (debug_raw (const tree_node *)): New.
15967 (dump_tree_via_hooks (const tree_node *, int)): New.
15968 (debug (const tree_node &)): New.
15969 (debug (const tree_node *)): New.
15970 (debug_verbose (const tree_node &)): New.
15971 (debug_verbose (const tree_node *)): New.
15972 (debug_head (const tree_node &)): New.
15973 (debug_head (const tree_node *)): New.
15974 (debug_body (const tree_node &)): New.
15975 (debug_body (const tree_node *)): New.
15976 (debug_vec_tree (tree): Move and reimplement in terms of dump.
15977 (debug (vec<tree, va_gc> &)): New.
15978 (debug (vec<tree, va_gc> *)): New.
15980 (extern debug (const rtx_def &)): New.
15981 (extern debug (const rtx_def *)): New.
15983 (debug_raw (simple_bitmap_def &)): New.
15984 (debug_raw (simple_bitmap_def *)): New.
15985 (debug (simple_bitmap_def &)): New.
15986 (debug (simple_bitmap_def *)): New.
15988 (extern debug (simple_bitmap_def &)): New.
15989 (extern debug (simple_bitmap_def *)): New.
15990 (extern debug_raw (simple_bitmap_def &)): New.
15991 (extern debug_raw (simple_bitmap_def *)): New.
15993 (debug (vinsn_def &)): New.
15994 (debug (vinsn_def *)): New.
15995 (debug_verbose (vinsn_def &)): New.
15996 (debug_verbose (vinsn_def *)): New.
15997 (debug (expr_def &)): New.
15998 (debug (expr_def *)): New.
15999 (debug_verbose (expr_def &)): New.
16000 (debug_verbose (expr_def *)): New.
16001 (debug (vec<rtx> &)): New.
16002 (debug (vec<rtx> *)): New.
16004 (extern debug (vinsn_def &)): New.
16005 (extern debug (vinsn_def *)): New.
16006 (extern debug_verbose (vinsn_def &)): New.
16007 (extern debug_verbose (vinsn_def *)): New.
16008 (extern debug (expr_def &)): New.
16009 (extern debug (expr_def *)): New.
16010 (extern debug_verbose (expr_def &)): New.
16011 (extern debug_verbose (expr_def *)): New.
16012 (extern debug (vec<rtx> &)): New.
16013 (extern debug (vec<rtx> *)): New.
16015 (_list_iter_cond_expr): Make inline instead of static.
16017 (debug (sreal &)): New.
16018 (debug (sreal *)): New.
16020 (extern debug (sreal &)): New.
16021 (extern debug (sreal *)): New.
16023 (extern debug_raw (const tree_node &)): New.
16024 (extern debug_raw (const tree_node *)): New.
16025 (extern debug (const tree_node &)): New.
16026 (extern debug (const tree_node *)): New.
16027 (extern debug_verbose (const tree_node &)): New.
16028 (extern debug_verbose (const tree_node *)): New.
16029 (extern debug_head (const tree_node &)): New.
16030 (extern debug_head (const tree_node *)): New.
16031 (extern debug_body (const tree_node &)): New.
16032 (extern debug_body (const tree_node *)): New.
16033 (extern debug (vec<tree, va_gc> &)): New.
16034 (extern debug (vec<tree, va_gc> *)): New.
16036 (debug (struct loop &)): New.
16037 (debug (struct loop *)): New.
16038 (debug_verbose (struct loop &)): New.
16039 (debug_verbose (struct loop *)): New.
16040 * tree-dump.c: Add header dependence.
16042 (extern debug (struct loop &)): New.
16043 (extern debug (struct loop *)): New.
16044 (extern debug_verbose (struct loop &)): New.
16045 (extern debug_verbose (struct loop *)): New.
16047 (debug (data_reference &)): New.
16048 (debug (data_reference *)): New.
16049 (debug (vec<data_reference_p> &)): New.
16050 (debug (vec<data_reference_p> *)): New.
16051 (debug (vec<ddr_p> &)): New.
16052 (debug (vec<ddr_p> *)): New.
16054 (extern debug (data_reference &)): New.
16055 (extern debug (data_reference *)): New.
16056 (extern debug (vec<data_reference_p> &)): New.
16057 (extern debug (vec<data_reference_p> *)): New.
16058 (extern debug (vec<ddr_p> &)): New.
16059 (extern debug (vec<ddr_p> *)): New.
16061 (debug (pt_solution &)): New.
16062 (debug (pt_solution *)): New.
16064 (extern debug (pt_solution &)): New.
16065 (extern debug (pt_solution *)): New.
16067 (debug (_var_map &)): New.
16068 (debug (_var_map *)): New.
16069 (debug (tree_live_info_d &)): New.
16070 (debug (tree_live_info_d *)): New.
16072 (extern debug (_var_map &)): New.
16073 (extern debug (_var_map *)): New.
16074 (extern debug (tree_live_info_d &)): New.
16075 (extern debug (tree_live_info_d *)): New.
16077 2013-03-28 Jan Hubicka <jh@suse.cz>
16079 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16081 2013-03-28 Ian Bolton <ian.bolton@arm.com>
16083 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16084 record only when desired or required.
16086 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
16088 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16089 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16091 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16093 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16094 (*andsi3_compare0_uxtw): New pattern.
16095 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16096 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16098 2013-03-28 Jan Hubicka <jh@suse.cz>
16100 * data-streamer-in.c (streamer_read_gcov_count): New function.
16101 * gimple-streamer-out.c: Include value-prof.h.
16102 (output_gimple_stmt): Output histogram.
16103 (output_bb): Use streamer_write_gcov_count.
16104 * value-prof.c: Include data-streamer.h
16105 (dump_histogram_value): Add HIST_TYPE_MAX.
16106 (stream_out_histogram_value): New function.
16107 (stream_in_histogram_value): New function.
16108 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16109 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16110 * data-streamer-out.c (streamer_write_gcov_count): New function.
16111 (streamer_write_gcov_count_stream): New function.
16112 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16113 (lto_output_node): Likewise.
16114 (input_node, input_edge): Likewise.
16115 * lto-streamer-out.c (output_cfg): Update streaming.
16116 * lto-streamer-in.c (input_cfg): Likewise.
16117 * data-streamer.h (streamer_write_gcov_count,
16118 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16119 * gimple-streamer-in.c: Include value-prof.h
16120 (input_gimple_stmt): Input histograms.
16121 (input_bb): Update profile streaming.
16123 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16125 * genmodes.c (emit_max_int): New function.
16126 (emit_insn_modes_h): Added call to emit_max_function.
16127 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16129 * machmode.def: Fixed comment.
16131 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16133 * combine.c (try_combine): Removed useless assert.
16134 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16136 2013-03-28 Marek Polacek <polacek@redhat.com>
16137 Richard Biener <rguenther@suse.de>
16139 PR tree-optimization/56695
16140 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16141 build signed result of a vector comparison.
16142 * tree-cfg.c (verify_gimple_comparison): Check that a result
16143 of a vector comparison has signed type.
16145 2013-03-28 Richard Biener <rguenther@suse.de>
16147 PR tree-optimization/37021
16148 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16149 do not restrict gaps between groups.
16150 * tree-vect-stmts.c (vectorizable_load): Properly account for
16151 a gap between groups.
16153 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16155 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16156 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16159 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16161 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16162 * doc/extend.texi (Named Address Spaces): Ditto.
16163 (Variable Attributes): Ditto.
16165 2013-03-27 Kai Tietz <ktietz@redhat.com>
16167 * config.build: Add support for cygwin x64 target.
16168 * config.gcc: Likewise.
16169 * config.host: Likewise.
16170 * configure.ac: Likewise
16171 * configure: Regenerated.
16173 2013-03-27 Kai Tietz <ktietz@redhat.com>
16175 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16176 * config/i386/t-cygwin-w64: New file.
16177 * config/i386/cygwin-w64.h: New file.
16178 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16179 and add support for x64-cygwin target.
16180 (CPP_SPEC): Likewise.
16181 (CXX_WRAP_SPEC_LIST): Undefine before define.
16182 (LIBGCJ_SONAME): Use 15 as version.
16184 2013-03-27 Richard Biener <rguenther@suse.de>
16186 PR tree-optimization/56716
16187 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16188 dumping for ref nodes.
16190 2013-03-27 Martin Jambor <mjambor@suse.cz>
16192 PR tree-optimization/55334
16193 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16194 restricted pointers to arrays.
16196 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16198 * Makefile.in (.SUFFIXES): Add .cc.
16199 (.c.o): Apply same recipe for implicit rule .cc.o.
16201 2013-03-27 Richard Biener <rguenther@suse.de>
16203 PR tree-optimization/37021
16204 * tree-vect-data-refs.c (vect_check_strided_load): Allow
16205 REALPART/IMAGPART_EXPRs around the supported refs.
16206 * tree-ssa-structalias.c (find_func_aliases): Assume that
16207 floating-point values are not used to transfer pointers.
16209 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
16211 * target.def (TARGET_HAS_IFUNC_P): New target hook.
16212 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
16213 * doc/tm.texi: Regenerate.
16214 * targhooks.h (default_has_ifunc_p): New.
16215 * targhooks.c (default_has_ifunc_p): Ditto.
16216 * config/linux-protos.h: New file.
16217 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16218 hook for linux which disables support of indirect functions in android.
16219 * config/linux-android.c: New file.
16220 * config/t-linux-android.c: Ditto.
16221 * config.gcc: Added new object file linux-android.o.
16222 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
16223 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
16224 * varasm.c (do_assemble_alias): Likewise.
16225 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16226 doesn't support indirect functions.
16227 * configure: Regenerate.
16229 2013-03-27 Bin Cheng <bin.cheng@arm.com>
16232 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16233 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16236 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16238 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16240 2013-03-27 Terry Guo <terry.guo@arm.com>
16242 * config/arm/arm-cores.def: Added core cortex-r7.
16243 * config/arm/arm-tune.md: Regenerated.
16244 * config/arm/arm-tables.opt: Regenerated.
16245 * doc/invoke.texi: Added entry for core cortex-r7.
16247 2013-03-27 Walter Lee <walt@tilera.com>
16249 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16250 double-decrement of next_scratch_regno.
16252 2013-03-27 Walter Lee <walt@tilera.com>
16254 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
16256 (insn_v1mulus): Ditto.
16257 (insn_v2muls): Ditto.
16259 2013-03-27 Walter Lee <walt@tilera.com>
16261 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
16262 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16264 2013-03-27 Walter Lee <walt@tilera.com>
16266 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
16267 (*sibcall_value): Ditto.
16269 2013-03-27 Walter Lee <walt@tilera.com>
16271 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16272 (insn_mnz_v8qi): ... this ...
16273 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16275 (insn_v<n>mnz): Replaced by ...
16276 (insn_v1mnz): ... this ...
16277 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16279 (insn_mz_<mode>): Replaced by ...
16280 (insn_mz_v8qi): ... this ...
16281 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16283 (insn_v<n>mz): Replaced by ...
16284 (insn_v1mz): ... this ...
16285 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16288 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16290 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16292 2013-03-26 Roland McGrath <mcgrathr@google.com>
16294 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16295 than fprintf with a non-constant, non-format string.
16297 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16299 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16300 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16301 operand 0 predicate.
16302 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16303 attribute. Use general_x64nomem_operand as operand 1 predicate.
16304 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16305 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16306 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16307 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16308 (mov<mode>_insv_1): Remove expander. Merge insn with
16309 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16310 Use general_x64nomem_operand as operand 1 predicate.
16311 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16312 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16313 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16314 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16315 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16316 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16317 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16318 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16319 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16320 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16321 (general_x64nomem_operand): Ditto.
16323 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
16325 * config/rtems.opt: Add -pthread option.
16327 2013-03-26 Richard Biener <rguenther@suse.de>
16329 * alias.c (find_base_term): Avoid redundant and not used recursion.
16330 (base_alias_check): Get the initial base term from the caller.
16331 (true_dependence_1): Compute and pass base terms to base_alias_check.
16332 (write_dependence_p): Likewise.
16333 (may_alias_p): Likewise.
16335 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16337 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16338 PC-relative load in SI modes and above only.
16340 2013-03-26 Xinyu Qi <xyqi@marvell.com>
16342 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
16343 * config/arm/iwmmxt.md (WCGR0): Update.
16344 (WCGR1, WCGR2, WCGR3): Likewise.
16346 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16348 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16349 Use x64 and nox64 isa attributes.
16351 2013-03-26 Richard Biener <rguenther@suse.de>
16353 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16354 alignment computations and rely on get_object_alignment_1
16355 for the !TYPE_P case.
16356 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16358 2013-03-26 Walter Lee <walt@tilera.com>
16360 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16361 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16363 2013-03-25 Jeff Law <law@redhat.com>
16365 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16366 check for INTEGRAL_TYPE_P that was missing due to checking in
16367 wrong version of prior patch.
16369 2013-03-25 Walter Lee <walt@tilera.com>
16371 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16372 TILEGX_INSN_SHUFFLEBYTES1.
16373 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16375 (tilegx_builtins): Ditto.
16376 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16378 2013-03-25 Walter Lee <walt@tilera.com>
16380 * config/tilegx/tilegx.md (floatsisf2): New pattern.
16381 (floatunssisf2): New pattern.
16382 (floatsidf2): New pattern.
16383 (floatunssidf2): New pattern.
16385 2013-03-25 Walter Lee <walt@tilera.com>
16387 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16388 tests for constraint J, K, N, P.
16390 2013-03-25 Walter Lee <walt@tilera.com>
16392 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16393 Use indirect/pcrel encoding.
16394 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16397 2013-03-25 Steve Ellcey <sellcey@mips.com>
16399 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16400 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16401 * config/mips/mips.c (mips_option_override): Set IMADD default.
16402 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16403 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16404 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16405 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16406 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16408 2013-03-25 Jeff Law <law@redhat.com>
16410 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16411 slightly to avoid creating and folding useless trees. Simplify
16412 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16414 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
16416 * config/i386/i386.md (*zero_extendsidi2): Merge with
16417 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16418 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16419 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16420 targets. Clarify comment.
16422 2013-03-25 Martin Jambor <mjambor@suse.cz>
16424 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16425 pass-through jump functions differently.
16426 (ipa_read_jump_function): Likewise. Also use setter functions to set
16429 2013-03-25 Martin Jambor <mjambor@suse.cz>
16431 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16432 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16434 (ipa_get_indirect_edge_target): New function.
16435 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16436 ipa_get_indirect_edge_target. Update all callers.
16437 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16438 ipa_get_indirect_edge_target_1 instead of calling
16439 ipa_get_indirect_edge_target.
16440 (create_specialized_node): Pass aggvlas to
16441 ipcp_discover_new_direct_edges.
16443 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16445 * config/arm/arm.md (f_sels, f_seld): New types.
16446 (*cmov<mode>): New pattern.
16447 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16450 2013-03-25 Kai Tietz <ktietz@redhat.com>
16452 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16453 POSIX-printf for mingw-hosted builds.
16455 2013-03-25 Richard Biener <rguenther@suse.de>
16457 PR middle-end/56694
16458 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16459 must-not-throw stmt location.
16461 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16463 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16464 Emit load-acquire versions when acq is true.
16465 (arm_emit_store_exclusive): Add rel parameter.
16466 Emit store-release versions when rel is true.
16467 (arm_split_compare_and_swap): Use acquire-release instructions
16469 of barriers when appropriate.
16470 (arm_split_atomic_op): Likewise.
16471 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16472 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16473 (VUNSPEC_SLX): Likewise.
16474 (VUNSPEC_LDA): Likewise.
16475 (VUNSPEC_STL): Likewise.
16476 * config/arm/sync.md (atomic_load<mode>): New pattern.
16477 (atomic_store<mode>): Likewise.
16478 (arm_load_acquire_exclusive<mode>): Likewise.
16479 (arm_load_acquire_exclusivesi): Likewise.
16480 (arm_load_acquire_exclusivedi): Likewise.
16481 (arm_store_release_exclusive<mode>): Likewise.
16483 2013-03-25 Catherine Moore <clm@codesourcery.com>
16485 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16486 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16487 * config/mip/predicates.md (lwsp_swsp_operand,
16488 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16489 sb16_operand, db4_operand, db7_operand, ib3_operand,
16490 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16491 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16492 andi16_operand): New predicates.
16493 * config/mips/mips.md (compression): New attribute.
16494 (enabled): New attribute.
16495 (length): Consider compression in computing length.
16496 (shift_compression): New code attribute.
16497 (*add<mode>3): New operands. Record compression.
16498 (sub<mode>3): Likewise.
16499 (one_cmpl<mode>2): Likewise.
16500 (*and<mode>3): Likewise.
16501 (*ior<mode>3): Likewise.
16502 (unnamed pattern for xor): Likewise.
16503 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16504 (*<optab><mode>3): Likewise.
16505 (*mov<mode>_internal: Likewise.
16506 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16507 (mips_unsigned_immediate_p): New.
16508 (umips_lwsp_swsp_address_p): New.
16509 (m16_based_address_p): New.
16510 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16511 (mips_unsigned_immediate_p): New prototype.
16512 (lwsp_swsp_address_p): New prototype.
16513 (m16_based_address_p): New prototype.
16514 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16515 (mips_signed_immediate_p): New function.
16516 (m16_based_address_p): New function.
16517 (lwsp_swsp_address_p): New function.
16518 (mips_print_operand_punctuation): Recognize short delay slot insns
16519 for microMIPS.add<mode>3"
16521 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16524 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16525 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16527 2013-03-25 Richard Biener <rguenther@suse.de>
16529 PR tree-optimization/56689
16530 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16533 2013-03-25 Richard Biener <rguenther@suse.de>
16535 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16537 (memory_references): Likewise.
16538 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16539 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16540 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16541 (gather_mem_refs_in_loops): Fold into ...
16542 (analyze_memory_references): ... this. Move initialization
16543 to tree_ssa_lim_initialize.
16544 (fill_always_executed_in): Rename to ...
16545 (fill_always_executed_in_1): ... this.
16546 (fill_always_executed_in): Move contains_call computation to
16547 this new function from ...
16548 (tree_ssa_lim_initialize): ... here.
16549 (tree_ssa_lim): Call fill_always_executed_in.
16551 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16553 * postreload.c (reload_combine): Fix code detecting returns.
16555 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16557 * function.c (emit_use_return_register_into_block): On cc0 targets,
16558 do not emit the sequence between cc0 setter and user.
16560 2013-03-25 Kai Tietz <ktietz@redhat.com>
16562 * config/i386/predicates.md (local_symbolic_operand): Interpret
16563 dll-imported symbols as none-local.
16565 2013-03-25 Richard Biener <rguenther@suse.de>
16567 * tree-ssa-loop-im.c (struct depend): Remove.
16568 (struct lim_aux_data): Make depends a vec of gimples.
16569 (free_lim_aux_data): Adjust.
16570 (add_dependency): Likewise.
16571 (set_level): Likewise.
16573 2013-03-25 Richard Biener <rguenther@suse.de>
16575 PR middle-end/56434
16576 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16577 the pointer returned by calls with ECF_MALLOC set.
16579 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16581 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16583 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
16585 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
16586 using MMXMODE mode iterator.
16587 (*move<mode>_internal): Merge with *movv2sf_internal and
16588 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16590 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16592 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16593 (record_last_mem_set_info): Likewise.
16595 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16596 of XNEWVEC followed by memset.
16597 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16599 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16601 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16602 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16603 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16604 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16605 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16606 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16607 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16608 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16609 BARRIER_P instead of GET_CODE.
16611 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16613 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16614 inaccuracy in the probing code.
16616 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16617 (ctrapdi4): Likewise.
16619 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16621 * calls.c (expand_call): Add missing guard to code handling return
16622 of non-BLKmode structures in MSB.
16623 * function.c (expand_function_end): Likewise.
16625 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16627 * combine.c (try_combine): Adjust comment. Do not add the set of
16628 insn #0 if the destination indirectly is set or dies in insn #2.
16629 Tidy up code to distribute a new note.
16631 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16633 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16634 also for alternatives 16 and 17.
16636 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
16638 * config/i386/sse.md (*mov<mode>_internal): Merge with
16639 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16640 Emit insn template depending on type attribute. Use
16641 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16642 movd instead of movq mnemonic for interunit moves. Rewrite mode
16643 attribute calculation. Remove unit attribute calculation.
16644 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16645 Set prefix_data16 attribute for DImode ssemov types.
16646 Use Ym instead of y for SSE-MMX conversion alternatives.
16647 Reorder operand constraints.
16649 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16651 * df.h (df_insn_delete): Adjust prototype.
16652 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16653 and let it decide whether mark the basic block dirty.
16654 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16655 * df-scan.c (df_insn_info_delete): New helper function, split
16656 off from df_insn_delete.
16657 (df_scan_free_bb_info): Use it.
16658 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16660 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16661 that the insn is actually an insn and it has a non-NULL basic block.
16662 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16664 2013-03-22 Richard Biener <rguenther@suse.de>
16666 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16668 (mem_ref_alloc): Do not allocate them.
16669 (refs_independent_p): Do not query or maintain a cache.
16671 2013-03-22 Richard Biener <rguenther@suse.de>
16673 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16674 (gather_mem_refs_in_loops): Do not compute it.
16675 (analyze_memory_references): Do not allocate it.
16676 (tree_ssa_lim_finalize): Do not free it.
16677 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16679 2013-03-22 Richard Biener <rguenther@suse.de>
16681 * is-a.h (as_a): Use gcc_checking_assert.
16683 2013-03-22 Ian Bolton <ian.bolton@arm.com>
16685 * config/aarch64/aarch64.c (aarch64_print_operand): New
16686 format specifier for printing a constant in hex.
16687 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16688 format specifier for printing second operand.
16690 2013-03-22 Richard Biener <rguenther@suse.de>
16692 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16694 (gather_mem_refs_in_loops): Perform store accumulation here.
16695 (create_vop_ref_mapping_loop): Remove.
16696 (create_vop_ref_mapping): Likewise.
16697 (analyze_memory_references): Initialize refs_stored_in_loop.
16698 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
16699 (record_indep_loop): Remove.
16700 (record_dep_loop): New function.
16701 (ref_indep_loop_p_1): Adjust to only walk over references
16702 in the loop, not its subloops.
16703 (ref_indep_loop_p): Rename to ...
16704 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16705 maintaining a more fine-grained cache.
16706 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16707 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16709 2013-03-22 Richard Biener <rguenther@suse.de>
16711 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16712 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16713 aggregate mem_ref_loc.
16714 (free_mem_ref_locs): Inline into ...
16715 (memref_free): ... this and adjust.
16716 (mem_ref_alloc): Adjust.
16717 (mem_ref_locs_alloc): Remove.
16718 (record_mem_ref_loc): Adjust.
16719 (get_all_locs_in_loop): Rewrite into ...
16720 (for_all_locs_in_loop): ... this iterator.
16721 (rewrite_mem_ref_loc): New functor.
16722 (rewrite_mem_refs): Use for_all_locs_in_loop.
16723 (sm_set_flag_if_changed): New functor.
16724 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16725 (ref_always_accessed): New functor.
16726 (ref_always_accessed_p): Use for_all_locs_in_loop.
16728 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16730 * tree-pass.h (PROP_gimple_lvec): New.
16731 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16732 (init_optimization_passes): Move pass_lower_vector.
16733 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16735 (pass_lower_vector): Provide PROP_gimple_lvec.
16736 (pass_lower_vector_ssa): Likewise.
16737 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16739 2013-03-21 Mark Wielaard <mjw@redhat.com>
16741 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16743 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16745 * config/i386/i386.md (*movdi_internal): Disparage slightly
16746 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16747 conversion alternatives.
16749 2013-03-21 Jakub Jelinek <jakub@redhat.com>
16751 PR middle-end/48087
16752 * diagnostic.def (DK_WERROR): New kind.
16753 * diagnostic.h (werrorcount): Define.
16754 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16755 promoted to DK_ERROR, increment DK_WERROR counter instead of
16757 * toplev.c (toplev_main): Call print_ignored_options even if
16758 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16759 even if just werrorcount is non-zero.
16762 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16764 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16765 (string_cst_pool_decl): New function.
16766 (optimize_one_addr_into_implicit_ptr): New function.
16767 (resolve_addr_in_expr): Optimize DWARF location expression
16768 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16769 which doesn't live in memory, but has DW_AT_location or
16770 DW_AT_const_value, or refers to a string literal, into
16771 DW_OP_GNU_implicit_pointer.
16772 (optimize_location_into_implicit_ptr): New function.
16773 (resolve_addr): If removing DW_AT_location of a variable because
16774 it was DW_OP_addr of address of the variable, but the variable doesn't
16775 live in memory, try to emit const value attribute for the initializer.
16777 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16779 * tree.h (VECTOR_TYPE_P): New macro.
16780 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16781 TYPE_MODE): Use it.
16782 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16783 VEC_COND_EXPR cannot be lvalues.
16784 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16786 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
16788 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16789 Restrict the transformation to equal modes.
16791 2013-03-21 Richard Biener <rguenther@suse.de>
16793 PR tree-optimization/39326
16794 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16795 (MEM_ANALYZABLE): Adjust.
16796 (record_mem_ref_loc): Move bitmap ops ...
16797 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16798 unanalyzable refs, do not record locations for it.
16799 (analyze_memory_references): Allocate ref zero as shared
16801 (refs_independent_p): Do not test for unanalyzed mems here.
16802 (ref_indep_loop_p_1): Special-case disambiguation against
16803 the unanalyzed ref.
16804 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
16806 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
16808 * config/arm/arm-protos.h (tune_params): Add
16809 prefer_neon_for_64bits field.
16810 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
16811 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
16812 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
16813 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
16814 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
16815 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
16816 (arm_option_override): Handle -mneon-for-64bits new option.
16817 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
16818 (prefer_neon_for_64bits): Declare new variable.
16819 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
16820 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
16821 (arch_enabled): Handle new arch types. Remove support for onlya8
16823 (one_cmpldi2): Use new arch names.
16824 (zero_extend<mode>di2, extend<mode>di2): Ditto.
16825 * config/arm/arm.opt (mneon-for-64bits): Add option.
16826 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
16827 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
16828 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
16830 * doc/invoke.texi (-mneon-for-64bits): Document.
16832 2013-03-21 Richard Biener <rguenther@suse.de>
16834 PR tree-optimization/39326
16835 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
16836 (sort_bbs_in_loop_postorder_cmp): New function.
16837 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
16839 2013-03-21 Richard Biener <rguenther@suse.de>
16841 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
16842 (vect_insert_into_interleaving_chain): Likewise.
16843 (vect_drs_dependent_in_basic_block): Inline ...
16844 (vect_slp_analyze_data_ref_dependence): ... here. New function,
16846 (vect_analyze_data_ref_dependence): ... here. Simplify.
16847 (vect_check_interleaving): Simplify.
16848 (vect_analyze_data_ref_dependences): Likewise. Split out ...
16849 (vect_slp_analyze_data_ref_dependences): ... this new function.
16850 (dr_group_sort_cmp): New function.
16851 (vect_analyze_data_ref_accesses): Compute data-reference groups
16852 here instead of in vect_analyze_data_ref_dependence. Use
16853 a more efficient algorithm.
16854 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
16855 vect_slp_analyze_data_ref_dependences. Call
16856 vect_analyze_data_ref_accesses earlier.
16857 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
16858 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
16859 (vect_slp_analyze_data_ref_dependences): New prototype.
16861 2013-03-21 Richard Biener <rguenther@suse.de>
16863 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
16864 ref is stored in the loop.
16865 (find_refs_for_sm): Walk only over all stores.
16866 (store_motion_loop): Allocate from lim_bitmap_obstack.
16867 (store_motion): Likewise.
16869 2013-03-21 Richard Biener <rguenther@suse.de>
16871 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
16872 Update virtual SSA form.
16874 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16876 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
16877 * configure: Regenerate.
16878 * config.in: Regenerate.
16879 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
16880 if !HAVE_LD_EH_FRAME_CIEV3.
16882 2013-03-21 Richard Biener <rguenther@suse.de>
16884 * tree-cfg.c (verify_expr_no_block): New function.
16885 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
16886 nor DECL_VALUE_EXPR have locations with associated blocks.
16887 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
16888 (clear_unused_block_pointer): Remove code dealing with
16889 blocks in DECL_DEBUG_EXPR locations.
16891 2013-03-21 Richard Biener <rguenther@suse.de>
16893 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
16894 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
16895 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
16896 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
16897 instead of DECL_DEBUG_EXPR_IS_FROM.
16898 * gimplify.c (gimplify_modify_expr): Likewise.
16899 * tree-cfg.c (verify_expr_location_1): Likewise.
16900 * tree-complex.c (create_one_component_var): Likewise.
16901 * tree-sra.c (create_access_replacement): Likewise.
16902 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
16903 (clear_unused_block_pointer): Likewise.
16904 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
16905 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
16906 * var-tracking.c (var_debug_decl): Likewise.
16907 (track_expr_p): Likewise.
16908 * tree-inline.c (add_local_variables): Likewise. Set
16909 DECL_HAS_DEBUG_EXPR_P after copying it.
16910 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
16911 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
16913 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
16916 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
16917 * configure: Regenerate.
16918 * config.in: Regenerate.
16919 * config/i386/i386.md (*movdf_internal): Use
16920 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16921 movd instead of movq mnemonic for interunit moves.
16922 (*movdi_internal): Ditto.
16924 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16926 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
16927 (abd<mode>_3): New pattern.
16928 (aba<mode>_3): New pattern.
16929 (fabd<mode>_3): New pattern.
16931 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16933 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
16934 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
16935 occurrence of REGISTER_PREFIX as its empty string.
16937 2013-03-20 Jeff Law <law@redhat.com>
16939 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
16940 addititional equivalences for equality comparisons between an SSA_NAME
16941 and a constant where the SSA_NAME was set from a widening conversion.
16943 2013-03-20 Walter Lee <walt@tilera.com>
16945 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
16947 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16949 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
16950 depending on type attribute.
16951 (*movti_internal): Ditto.
16952 (*movtf_internal): Ditto.
16953 (*movxf_internal): Ditto.
16954 (*movdf_internal): Ditto.
16955 (*movsf_internal): Ditto.
16957 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16959 * config/i386/i386.md (*movti_internal): Set prefix attribute to
16960 maybe_vex for sselog1 and ssemov types.
16961 (*movdi_internal): Reorder operand constraints.
16962 (*movsi_internal): Ditto. Set prefix attribute to
16963 maybe_vex for sselog1 and ssemov types.
16964 (*movtf_internal): Set prefix attribute to maybe_vex
16965 for sselog1 and ssemov types.
16966 (*movdf_internal): Ditto. Set prefix_data16 attribute for
16967 DImode ssemov types. Reorder operand constraints.
16968 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
16969 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
16970 attribute for SImode ssemov types. Reorder operand constraints.
16972 2013-03-20 Martin Jambor <mjambor@suse.cz>
16974 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
16975 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
16977 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
16979 * config/rs6000/predicates.md (indexed_address, update_address_mem
16980 update_indexed_address_mem): New predicates.
16981 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
16982 attribute for load/store instructions.
16983 * config/rs6000/dfp.md (movsd_store): Likewise.
16984 (movsd_load): Likewise.
16985 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
16986 (unnamed HI->DI extend define_insn): Likewise.
16987 (unnamed SI->DI extend define_insn): Likewise.
16988 (unnamed QI->SI extend define_insn): Likewise.
16989 (unnamed QI->HI extend define_insn): Likewise.
16990 (unnamed HI->SI extend define_insn): Likewise.
16991 (unnamed HI->SI extend define_insn): Likewise.
16992 (extendsfdf2_fpr): Likewise.
16993 (movsi_internal1): Likewise.
16994 (movsi_internal1_single): Likewise.
16995 (movhi_internal): Likewise.
16996 (movqi_internal): Likewise.
16997 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
16998 attribute for load/store instructions.
16999 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
17001 (mov<mode>_softfloat): Likewise.
17002 (mov<mode>_hardfloat32): Likewise.
17003 (mov<mode>_hardfloat64): Likewise.
17004 (mov<mode>_softfloat64): Likewise.
17005 (movdi_internal32): Likewise.
17006 (movdi_internal64): Likewise.
17007 (probe_stack_<mode>): Likewise.
17009 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
17011 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
17012 floating point, and decimal floating point to reload iterator.
17014 * config/rs6000/constraints.md (wl constraint): New constraints to
17015 return FLOAT_REGS if certain options are used to reduce the number
17016 of separate patterns that exist in the file.
17017 (wx constraint): Likewise.
17018 (wz constraint): Likewise.
17020 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17021 -mdebug=reg, print wg, wl, wx, and wz constraints.
17022 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17023 Initialize the reload functions for 64-bit binary/decimal floating
17025 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17026 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17027 create the buffer on the stack to overcome not having a 32-bit
17029 (rs6000_emit_move): Likewise.
17030 (rs6000_secondary_memory_needed_rtx): Likewise.
17031 (rs6000_alloc_sdmode_stack_slot): Likewise.
17032 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17033 via xxlxor, just like DFmode 0.0.
17035 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17036 define as 1 if we are running on a power7 or newer.
17037 (enum r6000_reg_class_enum): Add new constraints.
17039 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17040 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17041 with other moves by using conditional constraits (wg). Use LFIWZX
17042 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
17043 (movsd splitter): Likewise.
17044 (movsd_hardfloat): Likewise.
17045 (movsd_softfloat): Likewise.
17047 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17048 binary and decimal floating point moves.
17049 (fmove_ok): New attributes to combine binary and decimal floating
17050 point moves, and to combine power6x (mfpgpr) moves along normal
17052 (real_value_to_target): Likewise.
17053 (f32_lr): Likewise.
17054 (f32_lm): Likewise.
17055 (f32_li): Likewise.
17056 (f32_sr): Likewise.
17057 (f32_sm): Likewise.
17058 (f32_si): Likewise.
17059 (movsf): Combine binary and decimal floating point moves. Combine
17060 power6x (mfpgpr) moves with other moves by using conditional
17061 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
17062 (mov<mode> for SFmode/SDmode); Likewise.
17063 (SFmode/SDmode splitters): Likewise.
17064 (movsf_hardfloat): Likewise.
17065 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17066 (movsf_softfloat): Likewise.
17067 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17069 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17070 wx and wz constraints.
17072 * config/rs6000/constraints.md (wg constraint): New constraint to
17073 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17075 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17078 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17079 -mdebug=reg, print wg, wl, wx, and wz constraints.
17080 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17081 Initialize the reload functions for 64-bit binary/decimal floating
17083 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17084 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17085 create the buffer on the stack to overcome not having a 32-bit
17087 (rs6000_emit_move): Likewise.
17088 (rs6000_secondary_memory_needed_rtx): Likewise.
17089 (rs6000_alloc_sdmode_stack_slot): Likewise.
17090 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17091 via xxlxor, just like DFmode 0.0.
17093 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17094 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17095 with other moves by using conditional constraits (wg). Use LFIWZX
17096 and STFIWX for loading SDmode on power7.
17097 (movdd splitters): Likewise.
17098 (movdd_hardfloat32): Likewise.
17099 (movdd_softfloat32): Likewise.
17100 (movdd_hardfloat64_mfpgpr): Likewise.
17101 (movdd_hardfloat64): Likewise.
17102 (movdd_softfloat64): Likewise.
17104 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17105 64-bit binary and decimal floating point moves.
17106 (FMOVE64X): Likewise.
17107 (movdf): Combine 64-bit binary and decimal floating point moves.
17108 Combine power6x (mfpgpr) moves with other moves by using
17109 conditional constraits (wg).
17110 (mov<mode> for DFmode/DDmode): Likewise.
17111 (DFmode/DDmode splitters): Likewise.
17112 (movdf_hardfloat32): Likewise.
17113 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17114 (movdf_softfloat32): Likewise.
17115 (movdf_hardfloat64_mfpgpr): Likewise.
17116 (movdf_hardfloat64): Likewise.
17117 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17118 (movdf_softfloat64): Likewise.
17119 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17120 (reload_<mode>_load): Move to later in the file so they aren't in
17121 the middle of the floating point move insns.
17122 (reload_<mode>_store): Likewise.
17124 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17127 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17128 constraint if -mdebug=reg.
17129 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17130 Enable using dd reload support if needed.
17132 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17133 binary and decimal floating point moves in rs6000.md.
17134 (movtd_internal): Likewise.
17136 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17137 decimal floating point moves.
17139 (movtf_internal): Likewise.
17140 (mov<mode>_internal, TDmode/TFmode): Likewise.
17141 (movtf_softfloat): Likewise.
17142 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17144 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17145 movdi_internal64, using wg constraint for move direct operations.
17146 (movdi_internal64): Likewise.
17148 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17149 MODES_TIEABLE_P for selected modes. Print the numerical value of
17150 the various virtual registers. Use GPR/FPR first/last values,
17151 instead of hard coding the register numbers. Print which modes
17152 have reload functions registered.
17153 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17154 settings before/after setting cpu, target and subtarget settings.
17155 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17156 and for secondary reload failures in rs6000_secondary_reload_inner.
17157 (rs6000_secondary_reload_fail): Likewise.
17158 (rs6000_secondary_reload_inner): Likewise.
17160 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17161 macros for first/last GPR and FPR registers.
17162 (LAST_GPR_REGNO): Likewise.
17163 (FIRST_FPR_REGNO): Likewise.
17164 (LAST_FPR_REGNO): Likewise.
17166 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17167 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17168 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17169 (vcond<mode><mode>): Likewise.
17170 (vcondu<mode><mode>): Likewise.
17171 (vector_gtu<mode>): Likewise.
17172 (vector_gte<mode>): Likewise.
17173 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
17174 to prevent the compiler from converting DImode operations to TImode.
17175 (ior<mode>3): Likewise.
17176 (and<mode>3): Likewise.
17177 (one_cmpl<mode>2): Likewise.
17178 (nor<mode>3): Likewise.
17179 (andc<mode>3): Likewise.
17181 * config/rs6000/constraints.md (wt constraint): New constraint
17182 that returns VSX_REGS if TImode is allowed in VSX registers.
17184 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17185 constant under VSX.
17187 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17188 similar to TImode, but it is restricted to being in the GPRs.
17190 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17191 TImode to occupy a single VSX register.
17193 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17194 -mvsx-timode for power7/power8.
17195 (power7 cpu): Likewise.
17196 (power8 cpu): Likewise.
17198 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17199 sure that TFmode/TDmode take up two registers if they are ever
17200 allowed in the upper VSX registers.
17201 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17203 (rs6000_init_hard_regno_mode_ok): Likewise.
17204 (rs6000_debug_reg_global): Add debugging for PTImode and wt
17205 constraint. Print if LRA is turned on.
17206 (rs6000_option_override_internal): Give an error if -mvsx-timode
17207 and VSX is not enabled.
17208 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
17209 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
17210 to reg+offset addressing. Use PTImode when checking offset
17211 addresses for validity.
17212 (reg_offset_addressing_ok_p): Likewise.
17213 (rs6000_legitimate_offset_address_p): Likewise.
17214 (rs6000_legitimize_address): Likewise.
17215 (rs6000_legitimize_reload_address): Likewise.
17216 (rs6000_legitimate_address_p): Likewise.
17217 (rs6000_eliminate_indexed_memrefs): Likewise.
17218 (rs6000_emit_move): Likewise.
17219 (rs6000_secondary_reload): Likewise.
17220 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17221 reloads to fpr registers to continue to use reg+offset addressing,
17222 but 64-bit reloads to altivec registers need reg+reg addressing.
17223 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17224 it. Treat LO_SUM like a PLUS operation.
17225 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
17226 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
17227 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17228 registers to share a register with a smaller sized type, since VSX
17229 puts scalars in the upper 64-bits.
17230 (print_operand): Add support for PTImode.
17231 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17232 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17233 registers, but don't have arithmetic support.
17234 (rs6000_memory_move_cost): Add test for VSX.
17235 (rs6000_opt_masks): Add -mvsx-timode.
17237 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17240 (VSr): Use wt constraint for TImode.
17241 (VSv): Drop TImode support.
17242 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17243 (vsx_movti_64bit): Likewise.
17244 (vsx_movti_32bit): Likewise.
17245 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17246 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17247 one '?' on the appropriate output constraint. Do not allow TImode
17248 logical operations on 32-bit systems.
17249 (vsx_ior<mode>3): Likewise.
17250 (vsx_xor<mode>3): Likewise.
17251 (vsx_one_cmpl<mode>2): Likewise.
17252 (vsx_nor<mode>3): Likewise.
17253 (vsx_andc<mode>3): Likewise.
17254 (vsx_concat_<mode>): Likewise.
17255 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17257 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17258 OPTION_MASK_VSX_TIMODE.
17259 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17260 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17262 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17263 (TI2 iterator): New iterator for TImode, PTImode.
17264 (wd mode attribute): Add values for vector types.
17265 (movti_string): Replace TI move operations with operations for TImode
17266 and PTImode. Add support for TImode being allowed in VSX registers.
17267 (mov<mode>_string, TImode/PTImode): Likewise.
17268 (movti_ppc64): Likewise.
17269 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17270 (TI mode splitters): Likewise.
17272 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17275 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
17277 PR tree-optimization/56355
17278 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17279 Also handle integers with undefined overflow.
17281 2013-03-20 Catherine Moore <clm@codesourcery.com>
17282 Maciej W. Rozycki <macro@codesourcery.com>
17283 Tom de Vries <tom@codesourcery.com>
17284 Nathan Sidwell <nathan@codesourcery.com>
17285 Iain Sandoe <iain@codesourcery.com>
17286 Nathan Froyd <froydnj@codesourcery.com>
17287 Chao-ying Fu <fu@mips.com>
17289 * doc/extend.texi: (micromips, nomicromips, nocompression):
17290 Document new function attributes.
17291 * doc/invoke.texi (minterlink-compressed, mmicromips,
17292 m14k, m14ke, m14kec): Document new options.
17293 (minterlink-mips16): Update documentation.
17294 * doc/md.texi (ZC, ZD): Document new constraints.
17295 * configure.ac (gcc_cv_as_micromips): Check if linker
17296 supports the .set micromips directive.
17297 * configure: Regenerate.
17298 * config.in: Regenerate.
17299 * config/mips/mips-tables.opt: Regenerate.
17300 * config/mips/micromips.md: New file.
17301 * constraints.md (ZC, ZD): New constraints.
17302 * config/mips/predicates.md (movep_src_register): New predicate.
17303 (movep_src_operand): New predicate.
17304 (non_volatile_mem_operand): New predicate.
17305 * config/mips/mips.md (multimem): New type.
17306 (length): Differentiate between 17-bit and 18-bit branch offsets.
17307 (MOVEP1, MOVEP2): New mode iterator.
17308 (mov_<load>l): Use ZC constraint.
17309 (mov_<load>r): Likewise.
17310 (mov_<store>l): Likewise.
17311 (mov_<store>r): Likewise.
17312 (*branch_equality<mode>_inverted): Add microMIPS support.
17313 (*branch_equality<mode>): Likewise.
17314 (*jump_absolute): Likewise.
17315 (indirect_jump_<mode>): Likewise.
17316 (tablejump_<mode>): Likewise.
17317 (<optab>_internal): Likewise.
17318 (sibcall_internal): Likewise.
17319 (sibcall_value_internal): Likewise.
17320 (prefetch): Use constraint ZD.
17321 * config/mips/mips.opt (minterlink-compressed): New option.
17322 (minterlink-mips16): Now an alias for minterlink-compressed.
17323 (mmicromips): New option.
17324 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17325 (compare_and_swap_12): Likewise.
17326 (sync_add<mode>): Likewise.
17327 (sync_<optab>_12): Likewise.
17328 (sync_old_<optab>_12): Likewise.
17329 (sync_new_<optab>_12): Likewise.
17330 (sync_nand_12): Likewise.
17331 (sync_old_nand_12): Likewise.
17332 (sync_new_nand_12): Likewise.
17333 (sync_sub<mode>): Likewise.
17334 (sync_old_add<mode>): Likewise.
17335 (sync_old_sub<mode>): Likewise.
17336 (sync_new_add<mode>): Likewise.
17337 (sync_new_sub<mode>): Likewise.
17338 (sync_<optab><mode>): Likewise.
17339 (sync_old_<optab><mode>): Likewise.
17340 (sync_new_<optab><mode>): Likewise.
17341 (sync_nand<mode>): Likewise.
17342 (sync_old_nand<mode>): Likewise.
17343 (sync_new_nand<mode>): Likewise.
17344 (sync_lock_test_and_set<mode>): Likewise.
17345 (test_and_set_12): Likewise.
17346 (atomic_compare_and_swap<mode>): Likewise.
17347 (atomic_exchange<mode>_llsc): Likewise.
17348 (atomic_fetch_add<mode>_llsc): Likewise.
17349 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17350 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17351 (umips_save_restore_pattern_p): Likewise.
17352 (umips_load_store_pair_p): Likewise.
17353 (umips_output_load_store_pair): Likewise.
17354 (umips_movep_target_p): Likewise.
17355 (umips_12bit_offset_address_p): Likewise.
17356 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17357 (mips_base_mips16): Rename this...
17358 (mips_base_compression_flags): ...to this. Update all uses.
17359 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17360 (mips_mips16_decl_p): Delete.
17361 (mips_nomips16_decl_p): Delete.
17362 (mips_get_compress_on_flags): New function.
17363 (mips_get_compress_off_flags): New function.
17364 (mips_get_compress_mode): New function.
17365 (mips_get_compress_on_name): New function.
17366 (mips_get_compress_off_name): New function.
17367 (mips_insert_attributes): Support multiple compression types.
17368 (mips_merge_decl_attributes): Likewise.
17369 (umips_12bit_offset_address_p): New function.
17370 (mips_start_function_definition): Emit .set micromips directive.
17371 (mips_call_may_need_jalx_p): New function.
17372 (mips_function_ok_for_sibcall): Add microMIPS support.
17373 (mips_print_operand_punctuation): Support short delay slots and
17375 (umips_swm_mask, umips_swm_encoding): New.
17376 (umips_build_save_restore): New function.
17377 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17378 (was_mips16_p): Remove.
17379 (old_compression_mode): New.
17380 (mips_set_compression_mode): New function.
17381 (mips_set_current_function): Add microMIPS support.
17382 (mips_option_override): Likewise.
17383 (umips_save_restore_pattern_p): New function.
17384 (umips_output_save_restore): New function.
17385 (umips_load_store_pair_p_1): New function.
17386 (umips_load_store_pair_p): New function.
17387 (umips_output_load_store_pair_1): New function.
17388 (umips_output_load_store_pair): New function.
17389 (umips_movep_target_p) New function.
17390 (mips_prepare_pch_save): Add microMIPS support.
17391 * config/mips/mips.h (TARGET_COMPRESSION): New.
17392 (TARGET_CPU_CPP_BUILTINS): Update macro
17393 to use new compression flags and to support microMIPS.
17394 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17395 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17396 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17397 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17398 (ASM_SPEC): Support mmicromips and mno-micromips.
17399 (M16STORE_REG_P): New macro.
17400 (MIPS_CALL): Support TARGET_MICROMIPS.
17401 (MICROMIPS_J): New macro.
17402 (mips_base_mips16): Rename this...
17403 (mips_base_compression_flags): ...to this.
17404 (UMIPS_12BIT_OFFSET_P): New macro.
17405 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17406 (MULTILIB_DIRNAMES): Likewise.
17407 2013-03-20 Richard Biener <rguenther@suse.de>
17409 PR tree-optimization/56661
17410 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17411 the result does not have to be distinct.
17413 2013-03-20 Richard Biener <rguenther@suse.de>
17415 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17418 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17419 Steven Bosscher <steven@gcc.gnu.org>
17421 PR rtl-optimization/56605
17422 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17424 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17427 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17428 that require movd instead of movq.
17430 2013-03-20 Richard Biener <rguenther@suse.de>
17432 * tree-ssa-structalias.c (struct variable_info): Add pointer
17433 to the first field of an aggregate with sub-vars. Make
17434 this and the pointer to the next subfield its ID.
17435 (vi_next): New function.
17436 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17437 storedanything_id, integer_id): Increment by one.
17438 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17439 get_call_clobber_vi): Adjust.
17440 (solution_set_expand): Simplify and speedup.
17441 (solution_set_add): Inline into ...
17442 (set_union_with_increment): ... this. Adjust accordingly.
17443 (do_sd_constraint): Likewise.
17444 (do_ds_constraint): Likewise.
17445 (do_complex_constraint): Simplify.
17446 (build_pred_graph): Adjust.
17447 (solve_graph): Likewise. Simplify and speedup.
17448 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17449 get_constraint_for_component_ref, get_constraint_for_1,
17450 first_vi_for_offset, first_or_preceding_vi_for_offset,
17451 create_function_info_for, create_variable_info_for_1,
17452 create_variable_info_for, intra_create_variable_infos): Adjust.
17453 (init_base_vars): Push NULL for ID zero.
17454 (compute_points_to_sets): Adjust.
17456 2013-03-20 Richard Biener <rguenther@suse.de>
17458 * cfgloop.c (verify_loop_structure): Streamline and avoid
17459 ICEing on corrupt loop tree.
17460 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17463 2013-03-20 Richard Biener <rguenther@suse.de>
17465 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17466 check whether an SSA update is needed.
17468 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17470 * config/mips/constraints.md (T): Rename to...
17474 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17475 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17477 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17479 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17480 (*subsi3_carryin_uxtw): Likewise.
17482 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17484 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17485 (*rorsi3_insn_uxtw): Likewise.
17487 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17489 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17490 (*extrsi5_insn_uxtw): Likewise.
17492 2013-03-19 Richard Biener <rguenther@suse.de>
17494 PR tree-optimization/56273
17495 * passes.c (init_optimization_passes): Move second VRP after DOM.
17497 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
17499 * config/i386/i386.md (*movti_internal): Merge from
17500 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17501 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17502 nox64 isa attributes.
17504 2013-03-18 Richard Biener <rguenther@suse.de>
17506 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17508 (merge_node_constraints): Likewise.
17509 (build_succ_graph): Likewise.
17510 (valid_graph_edge): Inline into single caller.
17511 (unify_nodes): Likewise. Use bitmap_set_bit return value
17513 (scc_visit): Fix formatting and variable use.
17514 (do_sd_constraint): Use gcc_checking_assert.
17515 (do_ds_constraint): Likewise.
17516 (do_complex_constraint): Likewise.
17517 (condense_visit): Likewise. Cleanup.
17518 (dump_pred_graph): New function.
17519 (perform_var_substitution): Dump the pred-graph before
17520 variable substitution.
17521 (find_equivalent_node): Use gcc_checking_assert.
17522 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17524 2013-03-18 Richard Biener <rguenther@suse.de>
17526 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17527 Remove cond_expr_stmt_list argument and do not gimplify the
17529 (vect_loop_versioning): Adjust.
17530 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17531 Cleanup to use less temporaries.
17532 (vect_create_data_ref_ptr): Cleanup.
17534 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17536 PR tree-optimization/56635
17537 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17538 require types_compatible_p types.
17540 2013-03-18 Nick Clifton <nickc@redhat.com>
17542 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17543 spurious backslash.
17545 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17546 Add missing line to comment describing function.
17548 2013-03-18 Richard Biener <rguenther@suse.de>
17550 PR tree-optimization/56210
17551 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17552 Handle string / character search functions.
17553 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17555 2013-03-18 Richard Biener <rguenther@suse.de>
17557 PR middle-end/56483
17558 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17559 and implement properly.
17560 * gimple.h (gimple_cond_single_var_p): Remove.
17562 2013-03-18 Richard Biener <rguenther@suse.de>
17564 * tree-data-ref.h (find_data_references_in_loop): Declare.
17565 * tree-data-ref.c (get_references_in_stmt): Use a stack
17566 vector pre-allocated in the callers.
17567 (find_data_references_in_stmt): Adjust.
17568 (graphite_find_data_references_in_stmt): Likewise.
17569 (create_rdg_vertices): Likewise.
17570 (find_data_references_in_loop): Export.
17571 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17572 Compute dependences here...
17573 (vect_analyze_data_refs): ...not here. When we encounter
17574 a non-vectorizable data reference in basic-block vectorization
17575 truncate the data reference vector. Do not bother to
17576 fixup data-dependence information for gather loads.
17577 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17578 of data references, as reported.
17580 2013-03-18 Richard Biener <rguenther@suse.de>
17582 PR tree-optimization/3713
17583 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17584 has_constants and expr.
17585 (stmt_has_constants): Properly valueize SSA names when deciding
17586 whether the stmt has constants.
17588 2013-03-18 Richard Biener <rguenther@suse.de>
17590 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17591 whole function when there is nothing to do.
17592 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17593 * tree-vectorizer.c (vectorize_loops): Update virtual and
17594 loop-closed SSA once.
17595 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17597 2013-03-18 Richard Biener <rguenther@suse.de>
17599 PR middle-end/56113
17600 * domwalk.c (bb_postorder): New global static.
17601 (cmp_bb_postorder): New function.
17602 (walk_dominator_tree): Replace scheme imposing an order for
17603 visiting dominator sons by one sorting them at the time they
17604 are pushed on the stack.
17606 2013-03-18 Richard Biener <rguenther@suse.de>
17608 PR tree-optimization/39326
17609 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17610 (struct mem_ref): Replace mem member with ao_ref typed member.
17611 (MEM_ANALYZABLE): Adjust.
17612 (memref_eq): Likewise.
17613 (mem_ref_alloc): Likewise.
17614 (gather_mem_refs_stmt): Likewise.
17615 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17616 (execute_sm_if_changed_flag_set): Adjust.
17617 (execute_sm): Likewise.
17618 (ref_always_accessed_p): Likewise.
17619 (refs_independent_p): Likewise.
17620 (can_sm_ref_p): Likewise.
17622 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17625 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17626 return 1 even for !unsignedp.
17628 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17630 * config/i386/i386.md (isa): Add x64 and nox64.
17631 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17632 (*pushtf): Enable *roF alternative for x64 isa only.
17633 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17634 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17635 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17636 (*movtf_internal): Merge from *movtf_internal_rex64 and
17637 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17638 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17639 nox64 isa attributes.
17640 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17641 nox64 isa attributes.
17642 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17644 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17646 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17648 2013-03-17 Jakub Jelinek <jakub@redhat.com>
17651 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17652 class names. Remove trailing comma after "ALL_REGS".
17654 2013-03-16 Jan Hubicka <jh@suse.cz>
17656 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17657 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17658 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17659 of cgraph_get_create_node.
17660 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17662 2013-03-16 Jason Merrill <jason@redhat.com>
17665 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17666 with DW_AT_default_value.
17668 2013-03-16 Jakub Jelinek <jakub@redhat.com>
17670 * BASE-VER: Set to 4.9.0.
17672 2013-03-14 Andi Kleen <ak@linux.intel.com>
17675 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17676 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17677 Document _x* TSX intrinsics.
17679 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17680 David Holsgrove <david.holsgrove@xilinx.com>
17682 * configure.ac: Add MicroBlaze TLS support detection.
17683 * configure: Regenerate.
17684 * config/microblaze/microblaze-protos.h
17685 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17686 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17687 * config/microblaze/microblaze.c (microblaze_address_type): Add
17688 ADDRESS_TLS and tls_reloc address types.
17689 (microblaze_address_info): Add tls_reloc.
17690 (TARGET_HAVE_TLS): Define.
17691 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17692 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
17693 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17694 load_tls_operand, microblaze_call_tls_get_addr,
17695 microblaze_legitimize_tls_address): New functions.
17696 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17697 (get_base_reg): Use microblaze_tls_symbol_p.
17698 (microblaze_classify_address): Handle TLS.
17699 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17700 label_mentioned_p and microblaze_tls_referenced_p.
17701 (microblaze_legitimize_address): Handle TLS.
17702 (microblaze_address_insns): Handle ADDRESS_TLS.
17703 (pic_address_needs_scratch): Handle TLS.
17704 (print_operand_address): Handle TLS.
17705 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17706 (microblaze_expand_move): Handle TLS.
17707 (microblaze_legitimate_constant_p): Check
17708 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
17709 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17710 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17711 (PIC_OFFSET_TABLE_REGNUM): Set.
17712 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17713 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17714 (addsi3, movsi_internal2, movdf_internal): Update constraints
17715 * config/microblaze/predicates.md (arith_plus_operand): Define
17716 (move_operand): Redefine as move_src_operand,
17717 check microblaze_tls_referenced_p.
17719 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17721 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17722 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17724 2013-03-14 Ian Bolton <ian.bolton@arm.com>
17726 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17729 2013-03-14 Jakub Jelinek <jakub@redhat.com>
17731 PR tree-optimization/53265
17732 * common.opt (Waggressive-loop-optimizations): New option.
17733 * tree-ssa-loop-niter.c: Include tree-pass.h.
17734 (do_warn_aggressive_loop_optimizations): New function.
17735 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17736 if number_of_latch_executions returned constant.
17737 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17738 early. If number_of_latch_executions returned constant, set
17739 nb_iterations_upper_bound back to it.
17740 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17742 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17743 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17745 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17746 (MULTILIB_OSDIRNAMES): Set.
17747 * genmultilib: If defaultosdirname doesn't start with :: , set
17748 defaultosdirname2 instead, clear it and emit two . multilib_raw
17749 entries instead of just one.
17751 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17753 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17754 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17755 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17756 (SUBTARGET_OVERRIDE_OPTIONS): New.
17758 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17761 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17762 (musermode): Convert to Var(TARGET_USERMODE).
17763 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17764 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17765 * config/sh/sh.c (sh_option_override): Use
17766 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17767 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17769 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17771 (udivsi3_i4_single, divsi3_i4_single): Use
17772 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17774 2013-03-13 Dave Korn <dave.korn.cygwin@....>
17776 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17779 2013-03-13 Richard Biener <rguenther@suse.de>
17781 PR tree-optimization/56608
17782 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17783 calls when vectorizing basic-blocks.
17785 2013-03-13 Jakub Jelinek <jakub@redhat.com>
17788 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17791 2013-03-12 Jakub Jelinek <jakub@redhat.com>
17793 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17795 2013-03-11 Jan Hubicka <jh@suse.cz>
17798 * lto-streamer-out.c (output_symbol_p): Skip references from
17799 constructors of external variables.
17801 2013-03-11 Jan Hubicka <jh@suse.cz>
17803 PR middle-end/56571
17804 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
17806 * emit-rtl.c (verify_rtx_sharing): Likewise.
17807 (copy_insn_1): Likewise.
17808 * rtl.c (copy_rtx): Likewise.
17810 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
17813 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
17814 output_operand_lossage message.
17816 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
17819 * arm.c (shift_op): Validate RTL pattern on the fly.
17820 (arm_print_operand, case 'S'): Don't use shift_operator to validate
17823 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17826 * config/pa/pa.md (call_value): Check for calls to powf and direct to
17827 new call patterns that clobber %fr12.
17828 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
17829 split and postreload patterns.
17830 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
17831 registers %fr12 and %fr12R as call used.
17833 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
17835 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
17836 (canon_address, record_store, replace_read, check_mem_read_rtx,
17837 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
17838 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
17839 rest_of_handle_dse): Likewise.
17841 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
17843 PR middle-end/56524
17844 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
17846 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
17847 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
17848 (save_optabs_if_changed): Replace with...
17849 (init_tree_optimization_optabs): ...this.
17850 * optabs.c (save_optabs_if_changed): Rename to...
17851 (init_tree_optimization_optabs): ...this. Take the optimization node
17852 as argument. Do nothing if the base optabs are already correct.
17853 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
17854 to recompute optabs.
17855 * function.h (function): Remove optabs field.
17856 * function.c (invoke_set_current_function_hook): Call
17857 init_tree_optimization_optabs. Use the result to initialize
17860 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
17862 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
17863 if GTMA_HAS_NO_INSTRUMENTATION.
17864 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
17865 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
17866 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
17867 * gimple-pretty-print.c (dump_gimple_transaction): Handle
17868 GTMA_HAS_NO_INSTRUMENTATION.
17870 2013-03-08 Jakub Jelinek <jakub@redhat.com>
17872 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
17875 2013-03-08 Marek Polacek <polacek@redhat.com>
17876 Jakub Jelinek <jakub@redhat.com>
17878 PR tree-optimization/56478
17879 * predict.c (is_comparison_with_loop_invariant_p): Change the
17880 type of loop_step to tree.
17881 (predict_loops): Adjust.
17882 (predict_iv_comparison): Perform the computations on double_ints.
17884 2013-03-08 Richard Biener <rguenther@suse.de>
17886 PR tree-optimization/56570
17887 * tree-cfg.c (verify_expr_location_1): Verify locations for
17889 * tree-sra.c (create_access_replacement): Strip locations
17890 from DECL_DEBUG_EXPRs.
17892 2013-03-08 Richard Biener <rguenther@suse.de>
17894 * tree-inline.c (expand_call_inline): Do not associate
17895 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
17896 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
17898 2013-03-08 Richard Biener <rguenther@suse.de>
17900 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
17901 or block changes with -Og. Fix for location / block encoding
17902 changes and PHI arguments with locations.
17904 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
17906 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
17908 (struct output_info): Likewise.
17909 (register_overhead): Remove bad gcc_assert.
17910 (bitmap_find_bit): If there is only a single bitmap element, do not
17911 count a miss as a search.
17912 (print_statistics): Update for counter type changes.
17913 (dump_bitmap_statistics): Likewise. Print headers such that they
17914 are properly lined up with the printed counters.
17916 2013-03-07 Jakub Jelinek <jakub@redhat.com>
17918 PR tree-optimization/56559
17919 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
17920 check that it has only a single use.
17922 2013-03-07 Richard Biener <rguenther@suse.de>
17924 * doc/invoke.texi (fwhole-program): Discourage use in combination
17927 2013-03-06 Jakub Jelinek <jakub@redhat.com>
17929 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
17931 PR tree-optimization/56539
17932 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
17933 instead of GSI_CONTINUE_LINKING as last argument to
17934 force_gimple_operand_gsi. Adjust function comment.
17936 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
17939 PR middle-end/56548
17940 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
17941 promoted mode, convert the result back to the original mode.
17943 2013-03-06 Richard Biener <rguenther@suse.de>
17945 PR middle-end/56294
17946 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
17947 (insert_updated_phi_nodes_compare_uids): New function.
17948 (update_ssa): Sort symbols_to_rename after UID before
17949 traversing it to insert PHI nodes.
17951 2013-03-06 Richard Biener <rguenther@suse.de>
17953 PR middle-end/50494
17954 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
17955 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
17958 2013-02-13 Richard Biener <rguenther@suse.de>
17961 * varasm.c (output_constant_def_1): Get the decl representing
17962 the constant as argument.
17963 (output_constant_def): Wrap output_constant_def_1.
17964 (make_decl_rtl): Use output_constant_def_1 with the decl
17965 representing the constant.
17966 (build_constant_desc): Optionally re-use a decl already
17967 representing the constant.
17968 (tree_output_constant_def): Adjust.
17970 2013-03-06 Joey Ye <joey.ye@arm.com>
17973 * gcc.c (convert_white_space): New function.
17974 (main): Handles white space in function name.
17976 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
17979 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
17980 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
17981 to SH_DIV_CALL_TABLE for TARGET_SH2.
17982 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
17984 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
17985 call-table options.
17987 2013-03-05 Sterling Augustine <saugustine@google.com>
17988 Cary Coutant <ccoutant@google.com>
17991 * dwarf2out.c (resolve_addr): Don't call
17992 remove_loc_list_addr_table_entries a second time for the same
17995 2013-03-05 Jakub Jelinek <jakub@redhat.com>
17998 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
17999 (avoid_complex_debug_insns): New function.
18000 (expand_debug_locations): Call it.
18002 PR rtl-optimization/56484
18003 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
18004 lifetimes of hard registers on small register class machines.
18006 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18008 * config/microblaze/microblaze-protos.h: Rename
18009 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
18010 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
18012 (microblaze_fast_interrupt_function_p): New function.
18013 (microblaze_is_interrupt_handler): Rename to
18014 microblaze_is_interrupt_variant and add fast_interrupt check.
18015 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18016 (save_restore_insns): Likewise.
18017 (compute_frame_size): Likewise.
18018 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18019 (microblaze_globalize_label): Likewise.
18020 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18021 * config/microblaze/microblaze.md: Use wrapper
18022 microblaze_is_interrupt_variant.
18024 2013-03-05 Kai Tietz <ktietz@redhat.com>
18026 * sdbout.c (sdbout_one_type): Switch to current function's section
18027 supporting cold/hot.
18029 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18031 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18034 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18036 PR middle-end/56461
18037 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18038 if VALGRIND_GET_VBITS is defined, temporarily make object
18039 memory all defined, and restore previous valgrind addressability
18040 and definability afterwards. Free this_object at the end.
18042 PR middle-end/56461
18043 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18044 right before calling lra_create_live_ranges, also call it
18045 when clearing live_p. Only call lra_clear_live_ranges
18046 at the end if live_p.
18048 PR middle-end/56461
18049 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18051 2013-03-05 Richard Biener <rguenther@suse.de>
18053 PR tree-optimization/56521
18054 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18057 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18060 * except.h (remove_unreachable_eh_regions): New prototype.
18061 * except.c (remove_eh_handler_splicer): New function, split out
18062 of remove_eh_handler.
18063 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18064 warning about running it on many EH regions one at a time.
18065 (remove_unreachable_eh_regions_worker): New function, walk the
18066 EH tree in depth-first order and remove non-marked regions.
18067 (remove_unreachable_eh_regions): New function.
18068 * tree-eh.c (mark_reachable_handlers): New function, split out
18069 from remove_unreachable_handlers.
18070 (remove_unreachable_handlers): Use mark_reachable_handlers and
18071 remove_unreachable_eh_regions.
18072 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18073 and remove_unreachable_eh_regions.
18075 2013-03-05 Richard Biener <rguenther@suse.de>
18077 PR middle-end/56525
18078 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18079 not freeing them until the end.
18081 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18083 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18085 2013-03-05 Richard Biener <rguenther@suse.de>
18087 PR tree-optimization/56270
18088 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18089 of loads after scheduling an SLP instance.
18091 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18093 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18095 (check_gcc_parallelize): Run guality.exp as a separate job from
18096 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18097 struct-layout-1.exp with stackalign.exp.
18099 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18101 PR middle-end/56461
18102 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18103 load_index sbitmap even if some bit in it isn't set.
18105 PR middle-end/56461
18106 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18107 (discover_iteration_bound_by_body_walk): Change queues to
18108 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18109 spelling in comment. Call safe_push on queues[bound_index] directly.
18110 Release queues[queue_index] in every iteration unconditionally.
18111 Release bounds vector.
18113 PR middle-end/56461
18114 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18115 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18116 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18119 2013-03-05 Richard Biener <rguenther@suse.de>
18122 * tree-inline.c (remap_blocks_to_null): New function.
18123 (expand_call_inline): When expanding a call stmt without
18124 an associated block inline remap all callee blocks to NULL.
18126 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18128 PR rtl-optimization/56494
18129 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18130 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18131 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18133 PR middle-end/56461
18134 * sel-sched-ir.c (free_sched_pools): Release
18135 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18136 if succs_info_pool.max_top isn't -1.
18139 * opts.c (opts_obstack, opts_concat): Moved to...
18140 * opts-common.c (opts_obstack, opts_concat): ... here.
18142 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18144 PR middle-end/56461
18145 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18147 2013-03-04 Martin Jambor <mjambor@suse.cz>
18149 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18150 all appropriate places.
18152 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18154 PR tree-optimization/56424
18155 * ipa-split.c (split_function): Do not set the RSO flag if result is
18156 not by reference and its type is a register type.
18158 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
18160 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
18161 (microblaze_legitimate_pic_operand): Likewise
18162 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
18163 new function microblaze_legitimate_pic_operand
18164 * config/microblaze/microblaze-protos.h
18165 (microblaze_legitimate_pic_operand): Declare.
18167 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18169 * config/microblaze/predicates.md (call_insn_simple_operand):
18170 New predicate for supported rtx code types.
18171 * config/microblaze/microblaze.md (call_internal1): Use
18172 call_insn_simple_operand predicate.
18174 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18176 PR middle-end/56461
18177 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18178 partitions.ordered_remove.
18180 PR middle-end/56461
18181 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18182 vec_oprnds0.create (1) for modifier == NONE.
18184 PR middle-end/56461
18185 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18186 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18187 vec_oprnds1 right before pushing anything to it for
18190 PR middle-end/56461
18191 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18192 set nbbs to 0 instead of having separate code path.
18193 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18194 instead of false as last argument if returning NULL.
18196 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18198 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18199 the attribute is now called "target" instead of "option".
18200 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18201 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18202 attribute/pragma name for TARGET_OPTION_VALID_P and
18203 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18204 * doc/tm.texi: Regenerated.
18206 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
18208 * config/microblaze/microblaze.c:
18209 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
18210 * config/microblaze/microblaze.h: Add -mxl-reorder to
18212 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
18213 instructions emitted if TARGET_REORDER.
18214 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18215 or 0 for -m/-mno case, but initialises as 2 to detect default use case
18218 2013-03-01 Xinliang David Li <davidxl@google.com>
18220 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18223 2013-03-01 Jakub Jelinek <jakub@redhat.com>
18225 PR middle-end/56461
18226 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18227 vector even when returning true. Fix up function comment formatting.
18229 PR middle-end/56461
18230 * ira-build.c (ira_loop_nodes_count): New variable.
18231 (create_loop_tree_nodes): Initialize it.
18232 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18234 PR middle-end/56461
18235 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18236 method on dr_chain and result_chain.
18237 * tree-vect-stmts.c (vectorizable_store): Only call
18238 result_chain.create if j == 0.
18240 PR middle-end/56461
18241 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18242 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18243 before overwriting it.
18245 2013-03-01 Tobias Burnus <burnus@net-b.de>
18247 * doc/extended.texi (C Extensions): Change order in @menu
18249 (Other MIPS Built-in Functions): Move last MIPS entry before
18250 "picoChip Built-in Functions".
18251 (SH Built-in Functions): Move after RX Built-in Functions.
18252 * doc/gcc.texi (Introduction): Change order in @menu
18254 * doc/md.texi (Constraints): Ditto.
18255 * gty.texi (Type Information): Ditto.
18256 (User-provided marking routines for template types): Make
18258 * doc/invoke.texi (AArch64 Options): Move before
18259 "Adapteva Epiphany Options".
18261 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18262 Jakub Jelinek <jakub@redhat.com>
18265 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18266 no_address_safety_analysis attribute.
18267 * doc/extend.texi (no_address_safety_attribute): Rename to
18268 no_sanitize_address attribute, mention no_address_safety_analysis
18269 attribute as deprecated alias.
18271 2013-02-28 Jakub Jelinek <jakub@redhat.com>
18273 PR middle-end/56461
18274 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18275 type to vec<vec<tree> > *.
18276 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18277 to be vec<tree> instead of vec<tree> *, set vec_defs
18278 to vNULL and call vec_defs.create (number_of_vects), adjust other
18280 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18281 vectorizable_condition): Adjust vect_get_slp_defs callers.
18283 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18285 * config/aarch64/aarch64.c
18286 (aarch64_float_const_representable): Remove unused variable.
18288 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18290 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18292 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18294 * config/aarch64/aarch64-builtins.c
18295 (aarch64_init_simd_builtins): Make static.
18297 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18299 * config/aarch64/aarch64.c
18300 (aarch64_simd_make_constant): Make static.
18302 2013-02-28 Martin Jambor <mjambor@suse.cz>
18304 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18305 with no initialization to the RHS of debug statements.
18307 2013-02-28 Martin Jambor <mjambor@suse.cz>
18309 PR tree-optimization/56294
18310 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18312 (get_access_replacement): Do not call create_access_replacement.
18313 Assert a replacement exists.
18314 (get_repl_default_def_ssa_name): Create the replacement declaration
18317 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18319 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18320 final_end_function.
18322 2013-02-28 Marek Polacek <polacek@redhat.com>
18324 PR rtl-optimization/56466
18325 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18326 if we're changing a loop.
18327 (peel_loops_completely): Likewise.
18329 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18332 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18334 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18337 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18338 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18339 INTX_FTYPE_FX, FX_FTYPE_INTX.
18340 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18342 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18344 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18345 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18346 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18347 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18348 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18349 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18350 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18351 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18352 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18353 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18354 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18355 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18356 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18357 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18358 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18359 (avrxmega6): Increase max flash segments from 5 to 6.
18360 * config/avr/t-multilib: Regenerate.
18361 * config/avr/avr-tables.opt: Regenerate.
18362 * doc/avr-mmcu.texi: Regenerate.
18364 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18366 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18367 (avr_device_to_arch): Rename to avr_device_to_ld.
18368 (avr_device_to_as): New prototype.
18369 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18370 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18371 * config/avr/driver-avr.c (avr_device_to_as): New.
18372 (avr_device_to_arch): Rename to avr_device_to_ld.
18374 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18376 PR middle-end/56461
18377 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18378 method on dr_chain and result_chain.
18380 PR middle-end/56461
18381 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18382 pointer_set_destroy on not_executed_last_iteration.
18384 PR middle-end/56461
18385 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
18387 PR middle-end/56461
18388 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18389 FOR_EACH_DEFINED_FUNCTION when freeing state.
18391 PR middle-end/56461
18392 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18394 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18397 PR middle-end/56461
18398 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18399 known_aggs[i].items and release known_aggs vector.
18401 PR middle-end/56461
18402 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18404 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18406 * config/microblaze/microblaze.c (microblaze_emit_compare):
18407 Use xor for EQ/NE comparisions.
18408 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18409 (cbranchsf4): Adjust operator to comparison_operator.
18411 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18413 PR middle-end/56461
18414 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18416 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18417 vec_safe_push, always update *slot.
18418 (redirect_edge_var_map_clear): Use vec_free.
18419 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18420 (free_var_map_entry): Use vec_free.
18421 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18422 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18424 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
18426 PR middle-end/45472
18427 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18428 when the may_trap_p bit of the exprs being merged differs.
18429 Reorder tests for speculativeness in the logical and operator.
18431 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18433 * incpath.c (add_standard_paths): Use reconcat instead of concat
18434 where appropriate and avoid leaking memory.
18436 * opts.h: Include obstack.h.
18437 (opts_concat): New prototype.
18438 (opts_obstack): New declaration.
18439 * opts.c (opts_concat): New function.
18440 (opts_obstack): New variable.
18441 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18442 (finish_options): Use opts_concat instead of concat
18443 and XOBNEWVEC instead of XNEWVEC.
18444 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18445 generate_option): Likewise.
18446 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18447 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18450 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18451 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18453 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18455 PR middle-end/56461
18456 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18458 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18460 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18461 (arm_block_move_unaligned_straight): Likewise.
18462 (arm_adjust_block_mem): Likewise.
18464 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18467 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18468 temp, cond and label.
18469 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18472 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18473 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18476 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18477 of prologue/epilogue functions.
18480 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18481 Remove unused variable cfa_offset.
18482 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18485 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18486 type promotion to unsigned.
18489 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18490 for HOST_WIDE_INT of 32 bit / same size as int.
18491 (arm_block_move_unaligned_straight): Likewise.
18492 (arm_adjust_block_mem): Likewise.
18495 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18498 2013-02-26 Marek Polacek <polacek@redhat.com>
18500 PR tree-optimization/56426
18501 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
18503 2013-02-26 Richard Biener <rguenther@suse.de>
18506 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18507 unused variable loops.
18509 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18511 PR tree-optimization/56448
18512 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18513 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18514 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18515 later operands of the references, or even first operand for
18516 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18518 PR tree-optimization/56443
18519 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18520 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18521 to type_for_mode langhook.
18523 2013-02-25 Matt Turner <mattst88@gmail.com>
18525 * doc/invoke.texi: Document r4700.
18527 2013-02-25 Richard Biener <rguenther@suse.de>
18529 PR tree-optimization/56175
18530 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18532 (simplify_bitwise_binary): ... here. Also guard the conversion
18533 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18535 2013-02-25 Catherine Moore <clm@codesourcery.com>
18538 2013-02-24 Catherine Moore <clm@codesourcery.com>
18539 Maciej W. Rozycki <macro@codesourcery.com>
18540 Tom de Vries <tom@codesourcery.com>
18541 Nathan Sidwell <nathan@codesourcery.com>
18542 Iain Sandoe <iain@codesourcery.com>
18543 Nathan Froyd <froydnj@codesourcery.com>
18544 Chao-ying Fu <fu@mips.com>
18546 * doc/extend.texi: (micromips, nomicromips, nocompression):
18547 Document new function attributes.
18548 * doc/invoke.texi (minterlink-compressed, mmicromips,
18549 m14k, m14ke, m14kec): Document new options.
18550 (minterlink-mips16): Update documentation.
18551 * doc/md.texi (ZC, ZD): Document new constraints.
18552 * configure.ac (gcc_cv_as_micromips): Check if linker
18553 supports the .set micromips directive.
18554 * configure: Regenerate.
18555 * config.in: Regenerate.
18556 * config/mips/mips-tables.opt: Regenerate.
18557 * config/mips/micromips.md: New file.
18558 * constraints.md (ZC, AD): New constraints.
18559 * config/mips/predicates.md (movep_src_register): New predicate.
18560 (movep_src_operand): New predicate.
18561 (non_volatile_mem_operand): New predicate.
18562 * config/mips/mips.md (multimem): New type.
18563 (length): Differentiate between 17-bit and 18-bit branch offsets.
18564 (MOVEP1, MOVEP2): New mode iterator.
18565 (mov_<load>l): Use ZC constraint.
18566 (mov_<load>r): Likewise.
18567 (mov_<store>l): Likewise.
18568 (mov_<store>r): Likewise.
18569 (*branch_equality<mode>_inverted): Add microMIPS support.
18570 (*branch_equality<mode>): Likewise.
18571 (*jump_absolute): Likewise.
18572 (indirect_jump_<mode>): Likewise.
18573 (tablejump_<mode>): Likewise.
18574 (<optab>_internal): Likewise.
18575 (sibcall_internal): Likewise.
18576 (sibcall_value_internal): Likewise.
18577 (prefetch): Use constraint ZD.
18578 * config/mips/mips.opt (minterlink-compressed): New option.
18579 (minterlink-mips16): Now an alias for minterlink-compressed.
18580 (mmicromips): New option.
18581 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18582 (compare_and_swap_12): Likewise.
18583 (sync_add<mode>): Likewise.
18584 (sync_<optab>_12): Likewise.
18585 (sync_old_<optab>_12): Likewise.
18586 (sync_new_<optab>_12): Likewise.
18587 (sync_nand_12): Likewise.
18588 (sync_old_nand_12): Likewise.
18589 (sync_new_nand_12): Likewise.
18590 (sync_sub<mode>): Likewise.
18591 (sync_old_add<mode>): Likewise.
18592 (sync_old_sub<mode>): Likewise.
18593 (sync_new_add<mode>): Likewise.
18594 (sync_new_sub<mode>): Likewise.
18595 (sync_<optab><mode>): Likewise.
18596 (sync_old_<optab><mode>): Likewise.
18597 (sync_new_<optab><mode>): Likewise.
18598 (sync_nand<mode>): Likewise.
18599 (sync_old_nand<mode>): Likewise.
18600 (sync_new_nand<mode>): Likewise.
18601 (sync_lock_test_and_set<mode>): Likewise.
18602 (test_and_set_12): Likewise.
18603 (atomic_compare_and_swap<mode>): Likewise.
18604 (atomic_exchange<mode>_llsc): Likewise.
18605 (atomic_fetch_add<mode>_llsc): Likewise.
18606 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18607 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18608 (umips_save_restore_pattern_p): Likewise.
18609 (umips_load_store_pair_p): Likewise.
18610 (umips_output_load_store_pair): Likewise.
18611 (umips_movep_target_p): Likewise.
18612 (umips_12bit_offset_address_p): Likewise.
18613 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18614 (mips_base_mips16): Rename this...
18615 (mips_base_compression_flags): ...to this. Update all uses.
18616 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18617 (mips_mips16_decl_p): Delete.
18618 (mips_nomips16_decl_p): Delete.
18619 (mips_get_compress_on_flags): New function.
18620 (mips_get_compress_off_flags): New function.
18621 (mips_get_compress_mode): New function.
18622 (mips_get_compress_on_name): New function.
18623 (mips_get_compress_off_name): New function.
18624 (mips_insert_attributes): Support multiple compression types.
18625 (mips_merge_decl_attributes): Likewise.
18626 (umips_12bit_offset_address_p): New function.
18627 (mips_start_function_definition): Emit .set micromips directive.
18628 (mips_call_may_need_jalx_p): New function.
18629 (mips_function_ok_for_sibcall): Add microMIPS support.
18630 (mips_print_operand_punctuation): Support short delay slots and
18632 (umips_swm_mask, umips_swm_encoding): New.
18633 (umips_build_save_restore): New function.
18634 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18635 (was_mips16_p): Remove.
18636 (old_compression_mode): New.
18637 (mips_set_compression_mode): New function.
18638 (mips_set_current_function): Add microMIPS support.
18639 (mips_option_override): Likewise.
18640 (umips_save_restore_pattern_p): New function.
18641 (umips_output_save_restore): New function.
18642 (umips_load_store_pair_p_1): New function.
18643 (umips_load_store_pair_p): New function.
18644 (umips_output_load_store_pair_1): New function.
18645 (umips_output_load_store_pair): New function.
18646 (umips_movep_target_p) New function.
18647 (mips_prepare_pch_save): Add microMIPS support.
18648 * config/mips/mips.h (TARGET_COMPRESSION): New.
18649 (TARGET_CPU_CPP_BUILTINS): Update macro
18650 to use new compression flags and to support microMIPS.
18651 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18652 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18653 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18654 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18655 (ASM_SPEC): Support mmicromips and mno-micromips.
18656 (M16STORE_REG_P): New macro.
18657 (MIPS_CALL): Support TARGET_MICROMIPS.
18658 (MICROMIPS_J): New macro.
18659 (mips_base_mips16): Rename this...
18660 (mips_base_compression_flags): ...to this.
18661 (UMIPS_12BIT_OFFSET_P): New macro.
18662 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18663 (MULTILIB_DIRNAMES): Likewise.
18665 2013-02-25 Tom de Vries <tom@codesourcery.com>
18667 PR rtl-optimization/56131
18668 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18669 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18670 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18672 2013-02-25 Tobias Burnus <burnus@net-b.de>
18674 * doc/invoke.texi (-fsanitize=): Move from optimization
18675 to debugging options.
18677 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18679 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18681 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
18682 Alexander Monakov <amonakov@ispras.ru>
18684 PR middle-end/56077
18685 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18686 flush pending lists also on non-jumps. Adjust comment.
18688 2013-02-24 Catherine Moore <clm@codesourcery.com>
18689 Maciej W. Rozycki <macro@codesourcery.com>
18690 Tom de Vries <tom@codesourcery.com>
18691 Nathan Sidwell <nathan@codesourcery.com>
18692 Iain Sandoe <iain@codesourcery.com>
18693 Nathan Froyd <froydnj@codesourcery.com>
18694 Chao-ying Fu <fu@mips.com>
18696 * doc/extend.texi: (micromips, nomicromips, nocompression):
18697 Document new function attributes.
18698 * doc/invoke.texi (minterlink-compressed, mmicromips,
18699 m14k, m14ke, m14kec): Document new options.
18700 (minterlink-mips16): Update documentation.
18701 * doc/md.texi (ZC, ZD): Document new constraints.
18702 * configure.ac (gcc_cv_as_micromips): Check if linker
18703 supports the .set micromips directive.
18704 * configure: Regenerate.
18705 * config.in: Regenerate.
18706 * config/mips/mips-tables.opt: Regenerate.
18707 * config/mips/micromips.md: New file.
18708 * constraints.md (ZC, AD): New constraints.
18709 * config/mips/predicates.md (movep_src_register): New predicate.
18710 (movep_src_operand): New predicate.
18711 (non_volatile_mem_operand): New predicate.
18712 * config/mips/mips.md (multimem): New type.
18713 (length): Differentiate between 17-bit and 18-bit branch offsets.
18714 (MOVEP1, MOVEP2): New mode iterator.
18715 (mov_<load>l): Use ZC constraint.
18716 (mov_<load>r): Likewise.
18717 (mov_<store>l): Likewise.
18718 (mov_<store>r): Likewise.
18719 (*branch_equality<mode>_inverted): Add microMIPS support.
18720 (*branch_equality<mode>): Likewise.
18721 (*jump_absolute): Likewise.
18722 (indirect_jump_<mode>): Likewise.
18723 (tablejump_<mode>): Likewise.
18724 (<optab>_internal): Likewise.
18725 (sibcall_internal): Likewise.
18726 (sibcall_value_internal): Likewise.
18727 (prefetch): Use constraint ZD.
18728 * config/mips/mips.opt (minterlink-compressed): New option.
18729 (minterlink-mips16): Now an alias for minterlink-compressed.
18730 (mmicromips): New option.
18731 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18732 (compare_and_swap_12): Likewise.
18733 (sync_add<mode>): Likewise.
18734 (sync_<optab>_12): Likewise.
18735 (sync_old_<optab>_12): Likewise.
18736 (sync_new_<optab>_12): Likewise.
18737 (sync_nand_12): Likewise.
18738 (sync_old_nand_12): Likewise.
18739 (sync_new_nand_12): Likewise.
18740 (sync_sub<mode>): Likewise.
18741 (sync_old_add<mode>): Likewise.
18742 (sync_old_sub<mode>): Likewise.
18743 (sync_new_add<mode>): Likewise.
18744 (sync_new_sub<mode>): Likewise.
18745 (sync_<optab><mode>): Likewise.
18746 (sync_old_<optab><mode>): Likewise.
18747 (sync_new_<optab><mode>): Likewise.
18748 (sync_nand<mode>): Likewise.
18749 (sync_old_nand<mode>): Likewise.
18750 (sync_new_nand<mode>): Likewise.
18751 (sync_lock_test_and_set<mode>): Likewise.
18752 (test_and_set_12): Likewise.
18753 (atomic_compare_and_swap<mode>): Likewise.
18754 (atomic_exchange<mode>_llsc): Likewise.
18755 (atomic_fetch_add<mode>_llsc): Likewise.
18756 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18757 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18758 (umips_save_restore_pattern_p): Likewise.
18759 (umips_load_store_pair_p): Likewise.
18760 (umips_output_load_store_pair): Likewise.
18761 (umips_movep_target_p): Likewise.
18762 (umips_12bit_offset_address_p): Likewise.
18763 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18764 (mips_base_mips16): Rename this...
18765 (mips_base_compression_flags): ...to this. Update all uses.
18766 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18767 (mips_mips16_decl_p): Delete.
18768 (mips_nomips16_decl_p): Delete.
18769 (mips_get_compress_on_flags): New function.
18770 (mips_get_compress_off_flags): New function.
18771 (mips_get_compress_mode): New function.
18772 (mips_get_compress_on_name): New function.
18773 (mips_get_compress_off_name): New function.
18774 (mips_insert_attributes): Support multiple compression types.
18775 (mips_merge_decl_attributes): Likewise.
18776 (umips_12bit_offset_address_p): New function.
18777 (mips_start_function_definition): Emit .set micromips directive.
18778 (mips_call_may_need_jalx_p): New function.
18779 (mips_function_ok_for_sibcall): Add microMIPS support.
18780 (mips_print_operand_punctuation): Support short delay slots and
18782 (umips_swm_mask, umips_swm_encoding): New.
18783 (umips_build_save_restore): New function.
18784 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18785 (was_mips16_p): Remove.
18786 (old_compression_mode): New.
18787 (mips_set_compression_mode): New function.
18788 (mips_set_current_function): Add microMIPS support.
18789 (mips_option_override): Likewise.
18790 (umips_save_restore_pattern_p): New function.
18791 (umips_output_save_restore): New function.
18792 (umips_load_store_pair_p_1): New function.
18793 (umips_load_store_pair_p): New function.
18794 (umips_output_load_store_pair_1): New function.
18795 (umips_output_load_store_pair): New function.
18796 (umips_movep_target_p) New function.
18797 (mips_prepare_pch_save): Add microMIPS support.
18798 * config/mips/mips.h (TARGET_COMPRESSION): New.
18799 (TARGET_CPU_CPP_BUILTINS): Update macro
18800 to use new compression flags and to support microMIPS.
18801 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18802 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18803 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18804 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18805 (ASM_SPEC): Support mmicromips and mno-micromips.
18806 (M16STORE_REG_P): New macro.
18807 (MIPS_CALL): Support TARGET_MICROMIPS.
18808 (MICROMIPS_J): New macro.
18809 (mips_base_mips16): Rename this...
18810 (mips_base_compression_flags): ...to this.
18811 (UMIPS_12BIT_OFFSET_P): New macro.
18812 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18813 (MULTILIB_DIRNAMES): Likewise.
18815 2013-02-24 Jakub Jelinek <jakub@redhat.com>
18818 * target-globals.c (save_target_globals): For init_reg_sets and
18819 target_reinit remporarily set this_fn_optabs to this_target_optabs.
18821 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
18823 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
18824 * config/aarch64/t-aarch64
18825 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
18827 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
18829 PR inline-asm/56148
18830 * lra-constraints.c (process_alt_operands): Reload operand
18831 conflicting with earlier clobber only if no more other conflicting
18834 2013-02-22 Jakub Jelinek <jakub@redhat.com>
18837 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
18838 if not linking a shared library.
18840 2013-02-22 Seth LaForge <sethml@google.com>
18842 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
18844 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
18846 * config/arm/arm.md (split for extendsidi): Update condition.
18847 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
18848 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
18849 (qhs_zextenddi_cstr): Likewise.
18851 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18853 PR middle-end/56420
18854 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
18855 avoid signed wrapping.
18856 (expand_mult): Handle properly multiplication by
18857 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
18858 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
18859 in the compiler if coeff is HOST_WIDE_INT_MIN.
18860 (expand_divmod): Don't make ext_op1 static, change it's type to
18861 uhwi. Avoid undefined behavior in -INTVAL (op1).
18863 PR rtl-optimization/50339
18864 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
18866 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
18867 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
18868 into splitting_ashiftrt field.
18869 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
18871 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
18874 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
18876 PR middle-end/56108
18877 * trans-mem.c (execute_tm_mark): Do not expand transactions that
18878 are sure to go irrevocable.
18880 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
18882 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
18883 scalars are valid operands.
18885 2013-02-21 Martin Jambor <mjambor@suse.cz>
18887 PR tree-optimization/56310
18888 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
18889 only matching indices and non-negative final offsets.
18890 (intersect_aggregates_with_edge): Pass src_idx to
18891 agg_replacements_to_vector. Pass src_idx insstead of index to
18892 intersect_with_agg_replacements.
18894 2013-02-21 Martin Jambor <mjambor@suse.cz>
18896 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
18897 instead of hard-wired defaults.
18899 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
18901 * doc/invoke.texi (MIPS Options): Update documentation of the
18902 floating-point multiply-accumulate instruction restrictions.
18904 2013-02-21 Kostya Serebryany <kcc@google.com>
18906 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
18907 asan_shadow_offset on x86_64 linux.
18909 2013-02-21 Richard Biener <rguenther@suse.de>
18911 PR tree-optimization/56415
18913 2013-02-11 Richard Biener <rguenther@suse.de>
18915 PR tree-optimization/56273
18916 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18919 2013-02-21 Jakub Jelinek <jakub@redhat.com>
18922 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
18925 PR inline-asm/56405
18926 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
18927 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
18929 2013-02-20 Jan Hubicka <jh@suse.cz>
18931 PR tree-optimization/56265
18932 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
18933 when target is referenced for first time.
18935 2013-02-20 Richard Biener <rguenther@suse.de>
18937 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
18938 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
18939 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
18940 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
18941 not return anything.
18942 (rename_ssa_copies): Do not remove unused locals.
18943 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
18944 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
18945 * passes.c (execute_function_todo): Do not schedule unused locals
18946 removal if cleanup_tree_cfg did something.
18947 * tree-ssa-live.c (remove_unused_locals): Dump statistics
18948 about the number of removed locals.
18950 2013-02-20 Richard Biener <rguenther@suse.de>
18952 PR tree-optimization/56398
18953 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
18955 2013-02-20 Martin Jambor <mjambor@suse.cz>
18957 PR tree-optimization/55334
18958 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
18959 restricted pointers to arrays.
18961 2013-02-20 Richard Biener <rguenther@suse.de>
18962 Jakub Jelinek <jakub@redhat.com>
18964 PR tree-optimization/56396
18965 * tree-ssa-ccp.c (n_const_val): New static variable.
18966 (get_value): Return NULL for SSA names we don't have a lattice
18968 (ccp_initialize): Initialize n_const_val.
18969 * tree-ssa-copy.c (n_copy_of): New static variable.
18970 (init_copy_prop): Initialize n_copy_of.
18971 (get_value): Return NULL_TREE for SSA names we don't have a
18974 2013-02-20 Martin Jambor <mjambor@suse.cz>
18976 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
18978 2013-02-20 Richard Biener <rguenther@suse.de>
18980 * genpreds.c (write_lookup_constraint): Do not compare first
18981 letter of the constraint again.
18983 2013-02-20 Richard Biener <rguenther@suse.de>
18985 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
18987 (get_use_iv_cost): Terminate hashtable walk when coming across
18990 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
18992 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
18993 reassociation for avx2 targets.
18995 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18997 * config/microblaze/microblaze.c: microblaze_has_clz = 0
18998 Add version check for v8.10.a to enable microblaze_has_clz
18999 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
19000 version and TARGET_PATTERN_COMPARE check
19001 * config/microblaze/microblaze.md: New clzsi2 instruction
19003 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19005 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
19006 function before branching.
19008 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19010 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
19012 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19014 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19016 PR middle-end/55889
19017 * sel-sched.c: Include ira.h.
19018 (implicit_clobber_conflict_p): New function.
19019 (moveup_expr): Use it.
19020 * Makefile.in (sel-sched.o): Depend on ira.h.
19022 2013-02-19 Richard Biener <rguenther@suse.de>
19024 PR tree-optimization/56384
19025 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19026 (vn_hash_type): Split out from ...
19027 (vn_hash_constant_with_type): ... here.
19028 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19029 (vn_phi_eq): Compare types from vn_phi_s structure.
19030 (vn_phi_lookup): Populate vn_phi_s type.
19031 (vn_phi_insert): Likewise.
19033 2013-02-19 Jakub Jelinek <jakub@redhat.com>
19035 PR tree-optimization/56350
19036 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19037 if haven't found reduction or nested cycle operand, rather than
19038 asserting we must find it.
19040 PR tree-optimization/56381
19041 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19044 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
19045 Jakub Jelinek <jakub@redhat.com>
19048 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19049 (swap_optab_enable): Same.
19050 (init_all_optabs): Use argument instead of global.
19051 * tree.h (struct tree_optimization_option): New field target_optabs.
19052 * expr.h (init_all_optabs): Add argument to prototype.
19053 (TREE_OPTIMIZATION_OPTABS): New.
19054 (save_optabs_if_changed): Protoize.
19055 * optabs.h: Declare this_fn_optabs.
19056 * optabs.c (save_optabs_if_changed): New.
19057 Declare this_fn_optabs.
19058 (init_optabs): Add argument to init_all_optabs() call.
19059 * function.c (invoke_set_current_function_hook): Handle per
19061 * function.h (struct function): New field optabs.
19062 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19063 optimization_current_node has changed.
19064 * target-globals.h (save_target_globals_default_opts): Protoize.
19065 * target-globals.c (save_target_globals_default_opts): New.
19067 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19070 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19071 registers %fr12 and %fr12R as call used.
19074 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
19075 and HImode, require all displacements to be an integer multiple of
19077 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19078 only allow QImode and HImode when reload is in progress and strict is
19079 true. Likewise for symbolic addresses. Use base14_operand to check
19080 displacements in REG+BASE addresses.
19082 2013-02-18 Richard Biener <rguenther@suse.de>
19084 PR tree-optimization/56366
19085 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19086 handle sign-conversion of outer-loop initial induction value.
19088 2013-02-18 Richard Biener <rguenther@suse.de>
19090 PR middle-end/56349
19091 * cfghooks.c (merge_blocks): If we merge a latch into another
19092 block adjust references to it.
19093 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19094 (verify_loop_structure): Verify that a recorded latch is in fact
19097 2013-02-18 Richard Biener <rguenther@suse.de>
19099 PR tree-optimization/56321
19100 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19101 order SSA name release and virtual operand unlinking.
19103 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19105 * config/microblaze/microblaze.md (save_stack_block): Define.
19106 (restore_stack_block): Likewise.
19108 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19110 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19111 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19112 * config/microblaze/microblaze.c (microblaze_option_override):
19113 Bail out early for PIC modes when target does not support PIC.
19115 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19117 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
19118 Replace with a microblaze version.
19119 (microblaze_trampoline_init): Adapt for microblaze.
19120 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
19123 2013-02-16 Jakub Jelinek <jakub@redhat.com>
19124 Dodji Seketeli <dodji@redhat.com>
19127 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
19128 (instrument_mem_region_access): Do not forget to always put
19129 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19130 0) statement, even for cases where either 'base' or 'base + len'
19131 are not instrumented -- because they have been previously
19132 instrumented. Simplify the logic by putting all the statements
19133 instrument 'base + len' inside a sequence, and then insert that
19134 sequence right before the current insertion point. Then, to
19135 instrument 'base + len', just get an iterator on that statement.
19136 And do not forget to update the pointer to iterator the function
19137 received as argument.
19139 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19141 PR rtl-optimization/56348
19142 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19144 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19146 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19147 (clean_graph_dump_file): Pass base to start_graph_dump.
19149 2013-02-14 Richard Henderson <rth@redhat.com>
19152 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19154 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19156 * collect2-aix.h: Define F_LOADONLY.
19158 2013-02-14 Richard Biener <rguenther@suse.de>
19161 * varasm.c (output_constant_def_1): Get the decl representing
19162 the constant as argument.
19163 (output_constant_def): Wrap output_constant_def_1.
19164 (make_decl_rtl): Use output_constant_def_1 with the decl
19165 representing the constant.
19166 (build_constant_desc): Optionally re-use a decl already
19167 representing the constant.
19168 (tree_output_constant_def): Adjust.
19170 2013-02-14 Dodji Seketeli <dodji@redhat.com>
19173 * asan.c (instrument_builtin_call): Really put the length of the
19174 second source argument into src1_len.
19176 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19178 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19179 argument. If it is false, don't create edge from then_bb to
19181 (insert_if_then_before_iter): Pass true to it.
19182 (build_check_stmt): Pass false to it.
19183 (transform_statements): Flush hash table only on extended basic
19184 block boundaries, rather than at the beginning of every bb.
19185 Don't flush hash table on nonfreeing_call_p calls.
19186 * tree-flow.h (nonfreeing_call_p): New prototype.
19187 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19189 2013-02-13 David S. Miller <davem@davemloft.net>
19191 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19193 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19196 * ira.c (max_regno_before_ira): Move from ...
19198 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19201 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19203 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
19205 2013-02-13 Richard Biener <rguenther@suse.de>
19208 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
19209 globals in MEM_REFs.
19211 2013-02-13 Richard Biener <rguenther@suse.de>
19213 * loop-init.c (loop_optimizer_init): Clear loop state when
19214 re-initializing preserved loops.
19215 * loop-unswitch.c (unswitch_single_loop): Return whether
19216 we unswitched the loop. Do not verify loop state here.
19217 (unswitch_loops): When we unswitched a loop discover new loops.
19219 2013-02-13 Kostya Serebryany <kcc@google.com>
19221 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19223 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19225 2013-02-12 Dodji Seketeli <dodji@redhat.com>
19227 Avoid instrumenting duplicated memory access in the same basic block
19228 * Makefile.in (asan.o): Add new dependency on hash-table.h
19229 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19230 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19231 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19232 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19233 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19234 (get_mem_ref_of_assignment): New functions.
19235 (get_mem_refs_of_builtin_call): Extract from
19236 instrument_builtin_call and tweak a little bit to make it fit with
19238 (instrument_builtin_call): Use the new
19239 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19240 of is_gimple_builtin_call.
19241 (instrument_derefs, instrument_mem_region_access): Insert the
19242 instrumented memory reference into the hash table.
19243 (maybe_instrument_assignment): Renamed instrument_assignment into
19244 this, and change it to advance the iterator when instrumentation
19245 actually happened and return true in that case. This makes it
19246 homogeneous with maybe_instrument_assignment, and thus give a
19247 chance to callers to be more 'regular'.
19248 (transform_statements): Clear the memory reference hash table
19249 whenever we enter a new BB, when we cross a function call, or when
19250 we are done transforming statements. Use
19251 maybe_instrument_assignment instead of instrumentation. No more
19252 need to special case maybe_instrument_assignment and advance the
19253 iterator after calling it; it's now handled just like
19254 maybe_instrument_call. Update comment.
19256 2013-02-13 Richard Biener <rguenther@suse.de>
19258 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19259 Fix loop discovery code.
19261 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19263 PR inline-asm/56148
19264 * lra-constraints.c (process_alt_operands): Match early clobber
19265 operand with itself. Check conflicts with earlyclobber only if
19266 the operand is not reloaded. Prefer to reload conflicting operand
19267 if earlyclobber and matching operands are the same.
19269 2013-02-12 Richard Biener <rguenther@suse.de>
19272 * lto-streamer-out.c (write_symbol): Do not output symbols
19273 for hard register variables.
19275 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
19278 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19279 (umulsidi3_insn, mulsidi3_insn): New insns.
19281 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
19283 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19284 (struct tune_params): Add vec_costs field.
19285 * config/arm/arm.c (arm_builtin_vectorization_cost)
19286 (arm_add_stmt_cost): New functions.
19287 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19288 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19289 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19290 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19291 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19292 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19293 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19295 2013-02-12 Richard Biener <rguenther@suse.de>
19298 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19299 decls again if possible.
19301 2013-02-12 Richard Biener <rguenther@suse.de>
19303 PR middle-end/56288
19304 * tree-ssa.c (verify_ssa_name): Fix check, move
19305 SSA_NAME_IN_FREE_LIST check up.
19307 2013-02-12 Jakub Jelinek <jakub@redhat.com>
19308 Steven Bosscher <steven@gcc.gnu.org>
19310 PR rtl-optimization/56151
19311 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19312 equal to op0 or op1, and last_insn pattern is CODE operation
19313 with MEM dest and one of the operands matches that MEM.
19315 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19317 * doc/extend.texi: Document Function Multiversioning and "default"
19318 parameter string to target attribute.
19319 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19320 target attribute parameter is "default".
19321 (ix86_compare_version_priority): Remove checks for target attribute.
19322 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19323 Remove check for target attribute equal to NULL. Add assert.
19324 (ix86_generate_version_dispatcher_body): Change error to sorry.
19326 2013-02-11 Iain Sandoe <iain@codesourcery.com>
19327 Jack Howarth <howarth@bromo.med.uc.edu>
19328 Patrick Marlier <patrick.marlier@gmail.com>
19331 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19332 define ENDFILE_SPEC as TM_DESTRUCTOR.
19333 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19335 2013-02-11 Alexander Potapenko <glider@google.com>
19336 Jack Howarth <howarth@bromo.med.uc.edu>
19337 Jakub Jelinek <jakub@redhat.com>
19340 * config/darwin.c (cdtor_record): Rename ctor_record.
19341 (sort_cdtor_records): Rename sort_ctor_records.
19342 (finalize_dtors): New routine to sort destructors by
19343 priority before use in assemble_integer.
19344 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19346 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
19348 PR rtl-optimization/56275
19349 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19350 offset is non-negative and less than cmode size before
19351 calling simplify_subreg.
19353 2013-02-11 Richard Biener <rguenther@suse.de>
19355 PR tree-optimization/56264
19356 * cfgloop.h (fix_loop_structure): Adjust prototype.
19357 * loop-init.c (fix_loop_structure): Return the number of
19358 newly discovered loops.
19359 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19360 are discovered, do a full loop-closed SSA rewrite.
19362 2013-02-11 Richard Biener <rguenther@suse.de>
19364 PR tree-optimization/56273
19365 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19367 (check_array_ref): Fix missing newline in dumps.
19368 (search_for_addr_array): Likewise.
19370 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
19372 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19374 2013-02-09 Jakub Jelinek <jakub@redhat.com>
19377 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19379 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19381 PR rtl-optimization/56246
19382 * lra-constraints.c (simplify_operand_subreg): Try to reuse
19384 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19385 constraints are satisfied.
19387 2013-02-08 Jeff Law <law@redhat.com>
19390 * emit-rtl.c (reg_is_parm_p): New function.
19391 * regs.h (reg_is_parm_p): New prototype.
19392 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19393 callee-clobbered registers.
19395 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19398 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19399 If there is no implicit builtin declaration, just return NULL.
19401 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19403 * config/i386/sse.md (FMAMODEM): New mode iterator.
19404 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19405 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19407 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19409 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19410 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19411 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19413 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19415 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19416 (microblaze*-*-elf): Likewise.
19417 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19419 * config/microblaze/microblaze-c.c: Add builtin defines for
19420 _LITTLE_ENDIAN and _BIG_ENDIAN.
19421 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19422 add to TARGET_DEFAULT flags.
19423 Expand ASM_SPEC and LINK_SPEC.
19424 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19425 * config/microblaze/microblaze.md: Update extendsidi2 and
19426 movdi_internal instructions to use low-order / high-order reg
19428 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19429 options and inversemask / mask of LITTLE_ENDIAN.
19430 * config/microblaze/t-microblaze: Expand multilib options to
19431 include mlittle-endian (le) and update exceptions patterns.
19433 2013-02-08 Jakub Jelinek <jakub@redhat.com>
19435 PR rtl-optimization/56195
19436 * lra-constraints.c (get_reload_reg): Don't reuse regs
19437 if they have smaller mode than requested, if they have
19438 wider mode than requested, try to return a SUBREG.
19440 PR tree-optimization/56250
19441 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19442 if type is unsigned and code isn't MULT_EXPR.
19444 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19446 PR tree-optimization/56064
19447 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19448 bits according to mode.
19449 * fixed-value.h (fixed_from_double_int)
19450 (const_fixed_from_double_int): Adjust comments.
19452 2013-02-08 Richard Biener <rguenther@suse.de>
19455 * lto-streamer.h (struct data_in): Remove current_file, current_line
19456 and current_col members.
19457 * lto-streamer-out.c (lto_output_location): Stream changed bits
19458 en-block for efficiency.
19459 * lto-streamer-in.c (clear_line_info): Remove.
19460 (lto_input_location): Cache current file, line and column
19461 globally via local statics. Read changed bits en-block.
19462 (input_function): Do not call clear_line_info.
19463 (lto_read_body): Likewise.
19464 (lto_input_toplevel_asms): Likewise.
19466 2013-02-08 Michael Matz <matz@suse.de>
19468 PR tree-optimization/52448
19469 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19470 (nt_call_phase): New static.
19471 (add_or_mark_expr): Only mark accesses with newer phase than any
19473 (nonfreeing_call_p): New.
19474 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19475 (nt_fini_block): Keep blocks marked as visited.
19476 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19478 2013-02-08 Richard Biener <rguenther@suse.de>
19480 * ira.c (ira): Free broken dominator information.
19482 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19484 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19486 2013-02-08 Marek Polacek <polacek@redhat.com>
19488 * cfgloop.c (verify_loop_structure): Add more checking of headers.
19490 2013-02-08 Richard Biener <rguenther@suse.de>
19492 PR middle-end/56181
19493 * cfgloop.h (flow_loops_find): Adjust.
19494 (bb_loop_header_p): Declare.
19495 * cfgloop.c (bb_loop_header_p): New function split out from ...
19496 (flow_loops_find): ... here. Adjust function signature,
19497 support incremental loop structure update.
19498 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19499 * cfgloopmanip.c (fix_loop_structure): Move ...
19500 * loop-init.c (fix_loop_structure): ... here.
19501 (apply_loop_flags): Split out from ...
19502 (loop_optimizer_init): ... here.
19503 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19504 in incremental mode, only remove dead loops here.
19506 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19509 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19510 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19511 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19512 (*round<mode>3.libgcc): New insns for fixed-modes.
19513 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19514 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19515 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19516 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19517 implementations. Define to __builtin_avr_absFX,
19518 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19519 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19520 __builtin_avr_countlsFX, respectively.
19521 * config/avr/avr-c.c (target.h): Include it.
19522 (enum avr_builtin_id): New enum.
19523 (avr_resolve_overloaded_builtin): New static function.
19524 (avr_register_target_pragmas): Use it to set
19525 targetm.resolve_overloaded_builtin.
19526 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19527 tree nodes used by DEF_BUILTIN.
19528 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19529 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19530 <AVR_BUILTIN_xxBITS>: Same.
19532 2013-02-08 Richard Biener <rguenther@suse.de>
19534 * cfgloop.c (verify_loop_structure): Properly handle
19535 a loop exiting to another loop header.
19536 * ira-int.h (ira_loops): Remove.
19537 * ira.c (ira_loops): Remove.
19538 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19539 (do_reload): Use loop_optimizer_finalize.
19540 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19541 number_of_loops to access the loop tree.
19542 (more_one_region_p): Likewise.
19543 (finish_loop_tree_nodes): Likewise.
19544 (rebuild_regno_allocno_maps): Likewise.
19545 (mark_loops_for_removal): Likewise.
19546 (mark_all_loops_for_removal): Likewise.
19547 (remove_unnecessary_regions): Likewise.
19548 (ira_build): Likewise.
19549 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19551 2013-02-08 Richard Biener <rguenther@suse.de>
19553 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19554 * ipa-pure-const.c (analyze_function): Avoid calling
19555 mark_irreducible_loops twice.
19556 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
19558 2013-02-07 David S. Miller <davem@davemloft.net>
19560 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19562 * var-tracking.c (vt_add_function_parameter): Test the presence of
19563 HAVE_window_save properly and do not remap argument registers when
19564 we have a leaf function.
19566 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
19569 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19571 * config/i386/i386.c (ix86_print_operand): Ditto.
19573 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19575 * lra-constraints.c (process_alt_operands): Fix recently added comment.
19577 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19579 PR rtl-optimization/56225
19580 * lra-constraints.c (process_alt_operands): Check that reload hard
19581 reg can hold value for strict_low_part.
19583 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19586 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19587 dwarf2out_end_function.
19588 (in_first_function_p, maybe_at_text_label_p,
19589 first_loclabel_num_not_at_text_label): New variables.
19590 (dwarf2out_var_location): In the first function find out
19591 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
19592 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19595 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19597 PR rtl-optimization/56178
19598 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19599 SUBREG of a register. Tidy up related block of code.
19600 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19601 note if the source is a register or a SUBREG of a register.
19603 2013-02-07 Jakub Jelinek <jakub@redhat.com>
19606 * config/rs6000/rs6000.md (ptrm): New mode attr.
19607 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19608 call_value_indirect_aix<pttrsize>,
19609 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19612 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19614 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19615 if -bnortl. Convert to strcmp and strncmp.
19617 2013-02-07 Alan Modra <amodra@gmail.com>
19620 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19621 addresses won't wrap when offsetting.
19622 (rs6000_secondary_reload): Provide secondary reloads needed for
19623 wrapping LO_SUM addresses.
19625 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19627 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19628 MACH, just __MACH__.
19630 2013-02-06 Richard Biener <rguenther@suse.de>
19632 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19633 instead of calling fix_loop_structure.
19635 2013-02-06 Jakub Jelinek <jakub@redhat.com>
19637 PR middle-end/56217
19638 * omp-low.c (use_pointer_for_field): Return false if
19639 lower_send_shared_vars doesn't generate any copy-out code.
19641 2013-02-06 Tom de Vries <tom@codesourcery.com>
19643 PR rtl-optimization/56131
19644 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19645 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19646 of the label is NULL. Add comment.
19648 2013-02-05 Jakub Jelinek <jakub@redhat.com>
19650 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19653 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19654 (STATIC_LIBTSAN_LIBS): Likewise.
19655 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19656 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19657 is defined, don't add anything else beyond that.
19658 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19659 (LINK_COMMAND_SPEC): Use them.
19661 PR tree-optimization/56205
19662 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19663 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19664 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19666 2013-02-05 Richard Biener <rguenther@suse.de>
19668 PR tree-optimization/53342
19669 PR tree-optimization/53185
19670 * tree-vectorizer.h (vect_check_strided_load): Remove.
19671 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19672 not disallow peeling for vectorized strided loads.
19673 (vect_check_strided_load): Make static and simplify.
19674 (vect_analyze_data_refs): Adjust.
19675 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19676 correctly when vectorizing strided loads.
19678 2013-02-05 Richard Biener <rguenther@suse.de>
19680 * doc/install.texi: Refer to ISL, not PPL.
19682 2013-02-05 Jan Hubicka <jh@suse.cz>
19684 PR tree-optimization/55789
19685 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19687 2013-02-05 Jan Hubicka <jh@suse.cz>
19689 PR tree-optimization/55789
19690 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19691 the dead call anyway.
19693 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19696 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19698 2013-02-04 Alexander Potapenko <glider@google.com>
19699 Jack Howarth <howarth@bromo.med.uc.edu>
19700 Jakub Jelinek <jakub@redhat.com>
19703 * config/darwin.c (sort_ctor_records): Stabilized qsort
19704 on constructor priority by using original position.
19705 (finalize_ctors): New routine to sort constructors by
19706 priority before use in assemble_integer.
19707 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19709 2013-02-04 Jakub Jelinek <jakub@redhat.com>
19712 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19713 about visibility on artificial decls.
19714 * config/sol2.c (solaris_assemble_visibility): Likewise.
19716 2013-02-04 Kai Tietz <ktietz@redhat.com>
19719 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19720 argument and improve checking of return-argument types for 16-byte
19722 (ix86_function_value_1): Add additional valtype argument on call
19723 of function_value_64.
19724 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19725 handling infunction_value_64 function.
19727 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19729 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
19731 2013-02-04 Richard Biener <rguenther@suse.de>
19733 PR tree-optimization/56188
19734 * tree-ssa-structalias.c (label_visit): Consider case with
19735 initially non-empty points-to set.
19736 (perform_var_substitution): Dump node mapping and clean up.
19738 2013-02-04 Richard Guenther <rguenther@suse.de>
19741 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19742 node prevail as last resort.
19743 (lto_symtab_merge_decls): Remove guard on LTRANS here.
19744 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
19746 2013-02-04 Richard Biener <rguenther@suse.de>
19748 PR tree-optimization/56113
19749 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19751 (equiv_class_lookup_or_add): ... this.
19752 (label_visit): Adjust and fix error in previous patch.
19753 (perform_var_substitution): Adjust.
19755 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19757 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19758 * config/sh/sh4-300.md: Likewise.
19759 * config/sh/sh4a.md: Likewise.
19760 * config/sh/constraints.md: Likewise.
19761 * config/sh/sh.md: Likewise.
19762 * config/sh/netbsd-elf.h: Likewise.
19763 * config/sh/predicates.md: Likewise.
19764 * config/sh/sh-protos.h: Likewise.
19765 * config/sh/ushmedia.h: Likewise.
19766 * config/sh/linux.h: Likewise.
19767 * config/sh/sh.c: Likewise.
19768 * config/sh/superh.h: Likewise.
19769 * config/sh/elf.h: Likewise.
19770 * config/sh/sh4.md: Likewise.
19771 * config/sh/sh.h: Likewise.
19773 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19775 * config/pa/constraints.md: Adjust unused letters. Change "T"
19776 constraint to match_test floating_point_store_memory_operand().
19777 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19778 (base14_operand): New.
19779 (floating_point_store_memory_operand): New.
19780 (integer_store_memory_operand): Revise to use base14_operand and
19781 reg_plus_base_memory_operand.
19782 (move_dest_operand): Allow symbolic_memory_operands.
19783 (symbolic_memory_operand): Check for LO_SOM.
19784 (symbolic_operand): Change default case to break.
19785 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19786 CONST_DOUBLE values to be reloaded by putting them into memory when
19787 the destination is a floating point register.
19788 (movdf): Remove code to handle CONST_DOUBLE.
19790 (reload_indf_r1): New.
19791 (reload_insf_r1): New.
19792 Consistently use "Q" and "T" constraints with integer and floating
19793 point move instructions, respectively.
19794 (movdi): Remove FAIL.
19795 Change predicate for source operand unamed DImode move from
19796 general_operand to move_src_operand.
19797 (umulsidi3): Change predicate for destination operand to
19799 Likewise for similar unamed patterns.
19800 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19801 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19802 (hppa_legitimize_address): Simplify mask calculation.
19803 (pa_emit_move_sequence): Revised handling of secondary reloads from
19804 REG+D addresses for floating point loads and stores. Directly handle
19805 loading CONST0_RTX (mode) to a floating point register.
19806 (pa_secondary_reload): Handle reloading DF and SFmode constant values
19807 to floating point registers. Don't restrict secondary reloads to
19808 floating point registers to integer modes. Revise some comments and
19810 (TARGET_LEGITIMATE_ADDRESS_P): Define.
19811 (pa_legitimate_address_p): New.
19812 (pa_legitimize_reload_address): New.
19813 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
19814 (STRICT_REG_OK_FOR_BASE_P): New.
19815 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
19816 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
19818 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
19819 Andrew Dixie <andrewd@gentrack.com>
19821 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
19824 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
19826 * expmed.c (extract_bit_field_1): Pass the full width of the
19827 structure to get_best_reg_extraction_insn.
19829 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19832 * configure.ac (use_cxa_atexit): Add AIX.
19833 * configure: Regenerate.
19835 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
19837 2013-02-01 Jakub Jelinek <jakub@redhat.com>
19840 * final.c (need_profile_function): New variable.
19841 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
19842 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
19843 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
19844 notes, targetm.asm_out.function_prologue doesn't emit anything,
19845 HAVE_prologue and profiler should be emitted before prologue,
19846 set need_profile_function instead of emitting it.
19847 (final_scan_insn): If need_profile_function, emit
19848 profile_function on the first NOTE_INSN_BASIC_BLOCK or
19849 NOTE_INSN_FUNCTION_BEG note.
19851 2013-02-01 Richard Henderson <rth@redhat.com>
19853 * config/rs6000/rs6000.md (smulditi3): New.
19856 * config/alpha/alpha.md (umulditi3): New.
19858 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
19860 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
19861 (ASM_OUTPUT_ALIGNED_LOCAL): New.
19863 2013-02-01 Richard Biener <rguenther@suse.de>
19865 PR tree-optimization/56113
19866 * tree-ssa-structalias.c (label_visit): Reduce work for
19867 single-predecessor nodes.
19869 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
19871 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
19872 range isn't testing for zero.
19874 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
19876 PR middle-end/56113
19877 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
19879 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
19880 Nick Clifton <nickc@redhat.com>
19882 * config/v850/constraints.md (Q): Define as a memory constraint.
19883 * config/v850/predicates.md (label_ref_operand): New predicate.
19884 (e3v5_shift_operand): New predicate.
19885 (ior_operator): New predicate.
19886 * config/v850/t-v850: Add e3v5 multilib.
19887 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
19888 (v850_gen_movdi): Prototype.
19889 * config/v850/v850.c: Add support for e3v5 architecture.
19890 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
19892 (construct_save_jarl): Add e3v5 long JARL support.
19893 (v850_adjust_insn_length): New function. Adjust length of call
19894 insns when using e3v5 instructions.
19895 (v850_gen_movdi): New function: Generate instructions to move a
19897 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
19898 (CPP_SPEC): Define __v850e3v5__ as appropriate.
19899 (TARGET_USE_FPU): Enable for e3v5.
19900 (CONST_OK_FOR_W): New macro.
19901 (ADJUST_INSN_LENGTH): Define.
19902 * config/v850/v850.md (UNSPEC_LOOP): Define.
19903 (attr cpu): Add v850e3v5.
19904 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
19905 (movdi): New pattern.
19906 (movdi_internal): New pattern.
19907 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
19908 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
19909 (cstoresf4): Likewise.
19910 (cstoredf4): Likewise.
19911 (insv): New pattern.
19912 (rotlso3_a): New pattern.
19913 (rotlsi3_b): New pattern
19914 (rotlsi3_v850e3v5): New pattern.
19915 (doloop_begin): New pattern.
19916 (fix_loop_counter): New pattern.
19917 (doloop_end): New pattern.
19918 (branch_normal): Add e3v5 long branch support.
19919 (branch_invert): Likewise.
19920 (branch_z_normal): Likewise.
19921 (branch_z_invert): Likewise.
19922 (branch_nz_normal): Likewise.
19923 (branch_nz_invert): Likewise.
19924 (call_internal_short): Add e3v5 register-indirect JARL support.
19925 (call_internal_long): Likewise.
19926 (call_value_internal_short): Likewise.
19927 (call_value_internal_long): Likewise.
19928 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
19929 (mloop): New option.
19930 * config.gcc: Add support for configuring v840e3v5 target.
19931 * doc/invoke.texi: Document new v850 specific command line options.
19933 2013-01-31 Paul Koning <ni1d@arrl.net>
19937 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
19938 children if parent is a class.
19939 (prune_unused_types_prune): Don't add DW_AT_declaration.
19941 2013-01-31 Richard Biener <rguenther@suse.de>
19943 PR tree-optimization/56157
19944 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
19945 match up operand with SLP child.
19947 2013-01-31 Jason Merrill <jason@redhat.com>
19950 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
19951 parameters the first time.
19952 (gen_scheduled_generic_parms_dies): Check completeness here.
19954 2013-01-31 Richard Biener <rguenther@suse.de>
19956 PR middle-end/53073
19957 * common.opt (faggressive-loop-optimizations): New flag,
19958 enabled by default.
19959 * doc/invoke.texi (faggressive-loop-optimizations): Document.
19960 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
19961 infer_loop_bounds_from_undefined by it.
19963 2013-01-31 Richard Biener <rguenther@suse.de>
19965 PR tree-optimization/56150
19966 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
19967 visit virtual operands.
19968 (find_uses_to_rename_bb): Likewise.
19970 2013-01-31 Richard Biener <rguenther@suse.de>
19972 PR tree-optimization/56150
19973 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
19974 mixed store non-store stmts.
19976 2013-01-30 Jakub Jelinek <jakub@redhat.com>
19979 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
19980 LIBASAN_EARLY_SPEC is defined.
19981 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
19982 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
19984 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
19987 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
19988 invalid args instead of ICEing on it.
19989 (ix86_valid_target_attribute_tree): Return error_mark_node if
19990 ix86_valid_target_attribute_inner_p failed.
19991 (ix86_valid_target_attribute_p): Return false only if
19992 ix86_valid_target_attribute_tree returned error_mark_node. Allow
19993 target("default") attribute.
19994 (sorted_attr_string): Change argument from const char * to tree,
19995 merge in all target attribute arguments rather than just one.
19996 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
19997 instead of free. Avoid using strcat.
19998 (ix86_mangle_function_version_assembler_name): Mangle
19999 target("default") as if no target attribute is present. Adjust
20000 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
20001 instead of xmalloc and XDELETEVEC instead of free.
20002 (ix86_function_versions): Don't return true if one of the decls
20003 doesn't have target attribute. If they don't and one of the decls
20004 is DECL_FUNCTION_VERSIONED, report an error. Adjust
20005 sorted_attr_string caller. Use XDELETEVEC instead of free.
20006 (ix86_supports_function_versions): Remove.
20007 (make_name): Fix up formatting.
20008 (make_dispatcher_decl): Remove resolver_name and its initialization.
20009 Avoid leaking memory.
20010 (is_function_default_version): Return true if there is
20011 target("default") attribute rather than no target attribute at all.
20012 (make_resolver_func): Avoid leaking memory.
20013 (ix86_generate_version_dispatcher_body): Likewise.
20014 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20015 * target.def (supports_function_versions): Remove.
20016 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20017 * doc/tm.texi: Regenerated.
20019 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20021 PR rtl-optimization/56144
20022 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20023 for values with side effects.
20025 2013-01-30 Richard Biener <rguenther@suse.de>
20027 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20028 (sparseset_pop): Likewise.
20029 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20030 to be able to use quick_push in the worker loop.
20032 2013-01-30 Marek Polacek <polacek@redhat.com>
20034 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20036 2013-01-30 Richard Biener <rguenther@suse.de>
20039 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
20041 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
20043 PR tree-optimization/56064
20044 * fixed-value.c (fixed_from_double_int): New function.
20045 * fixed-value.h (fixed_from_double_int): New prototype.
20046 (const_fixed_from_double_int): New static inline function.
20047 * fold-const.c (native_interpret_fixed): New static function.
20048 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20049 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20050 (native_encode_fixed): New static function.
20051 (native_encode_expr) <FIXED_CST>: Use it.
20052 (native_interpret_int): Move double_int worker code to...
20053 * double-int.c (double_int::from_buffer): ...this new static method.
20054 * double-int.h (double_int::from_buffer): Prototype it.
20056 2013-01-30 Richard Biener <rguenther@suse.de>
20058 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20059 New pointer-map and obstack.
20060 (init_alias_vars): Allocate pointer-map and obstack.
20061 (delete_points_to_sets): Free them.
20062 (find_what_var_points_to): Cache result.
20063 (find_what_p_points_to): Adjust for changed interface of
20064 find_what_var_points_to.
20065 (compute_points_to_sets): Likewise.
20066 (ipa_pta_execute): Likewise.
20068 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20070 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20071 * configure: Regenerate.
20072 * config.in: Regenerate.
20073 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20074 #nobits/#progbits if supported.
20076 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20079 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20080 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20081 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20083 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20085 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20086 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
20088 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20090 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20091 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
20093 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20095 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20097 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20098 * config/arm/cortex-a7.md: New bypasses using
20099 arm_mac_accumulator_is_result.
20101 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20103 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
20104 (cortex_a7_neon_mla): Likewise.
20105 (cortex_a7_fpfmad): New reservation.
20106 (cortex_a7_fpmacs): Use ffmas and update required units.
20107 (cortex_a7_fpmuld): Update required units and latency.
20108 (cortex_a7_fpmacd): Likewise.
20109 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20110 (cortex_a7_neon). Likewise.
20111 (bypass) Update participating units.
20113 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20115 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20116 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20118 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20119 (vfp_fmul): Use ffmad.
20120 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20121 (cortex_r4_fmacd): Use ffmad.
20122 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20123 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20124 (cortex_a9_fmacd): Use ffmad.
20125 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20126 (cortex_a8_vfp_macd): Use ffmad.
20127 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20128 (cortex_a5_fpmacd): Use ffmad.
20129 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20130 (cortex_a15_vfp_macd): Use ffmad.
20131 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
20133 2013-01-29 Jason Merrill <jason@redhat.com>
20136 * varasm.c (default_assemble_visibility): Don't warn about
20137 visibility on artificial decls.
20139 2013-01-29 Richard Biener <rguenther@suse.de>
20141 PR tree-optimization/56113
20142 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20144 (label_visit): Free duplicate bitmaps and record the leader instead.
20145 (perform_var_substitution): Adjust.
20147 2013-01-29 Richard Biener <rguenther@suse.de>
20149 PR tree-optimization/55270
20150 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20151 the CFG, schedule loops for fixup.
20153 2013-01-29 Nick Clifton <nickc@redhat.com>
20155 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20158 2013-01-28 Leif Ekblad <leif@rdos.net>
20160 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20161 * config/i386/i386.h (TARGET_RDOS): New macro.
20162 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20163 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20164 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20165 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20166 DEFAULT_LARGE_SECTION_THRESHOLD.
20167 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20168 * config/i386/rdos.h: New file.
20169 * config/i386/rdos64.h: New file.
20171 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20174 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20177 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20179 PR rtl-optimization/56117
20180 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20181 call cselib_lookup_from_insn on the MEM before calling
20182 add_insn_mem_dependence.
20184 2013-01-28 Richard Biener <rguenther@suse.de>
20186 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20187 to a stmt that didn't have one.
20188 (copy_phis_for_bb): Likewise for PHI arguments.
20189 (copy_debug_stmt): Likewise for debug stmts.
20191 2013-01-28 Richard Biener <rguenther@suse.de>
20193 PR tree-optimization/56034
20194 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
20195 (partition_builtin_p): Adjust.
20196 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20197 it is the last partition.
20198 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20199 up the vertex for the definition.
20200 (classify_partition): Classify whether a partition is a
20201 PKIND_REDUCTION, thus has uses outside of the loop.
20202 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20203 Merge all PKIND_REDUCTION partitions into the last partition.
20204 (tree_loop_distribution): Seed partitions from reductions as well.
20206 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20208 PR tree-optimization/56125
20209 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
20210 pow(x,c) into sqrt(x) * powi(x, n/2) or
20211 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
20212 optimizing for size.
20213 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20214 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20217 PR tree-optimization/56094
20218 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20219 to UNKNOWN_LOCATION while gimplifying expr.
20221 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
20224 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20225 operand 0 in movabs insn template for -masm=intel asm alternative.
20226 (*movabs<mode>_2): Ditto for operand 1.
20228 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
20231 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20234 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20236 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20239 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20241 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20242 Undef to avoid warning.
20244 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20246 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20247 * configure: Regenerate.
20249 2013-01-25 Jakub Jelinek <jakub@redhat.com>
20251 PR tree-optimization/56098
20252 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20253 for stmts with volatile ops.
20254 (cond_store_replacement): Don't optimize if assign has volatile ops.
20255 (cond_if_else_store_replacement_1): Don't optimize if either
20256 then_assign or else_assign have volatile ops.
20257 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20260 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20262 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20264 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20266 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20267 missing ':' in asm example.
20269 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20271 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20272 entries into lane and laneq entries.
20273 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20274 Remove AdvSIMD scalar modes.
20275 (aarch64_sq<r>dmulh_laneq<mode>): New.
20276 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20278 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20279 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20280 * config/aarch64/iterators.md (VCOND): New.
20283 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20286 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20287 Add NULL LIBNAME argument to existing definitions.
20288 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20289 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20290 * config/avr/avr.c (DEF_BUILTIN): Same.
20291 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20292 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20293 implementation is available (DECL_ASSEMBLER_NAME is set).
20294 (avr_fold_absfx): New static function.
20295 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20296 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20297 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20298 AVR_BUILTIN_ABSLLK.
20299 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20300 (abshk, absk, abslk, absllk): Provide as static inline functions.
20302 2013-01-25 Marek Polacek <polacek@redhat.com>
20304 PR tree-optimization/56035
20305 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20307 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
20309 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20310 (*movtf_internal_rex64): Add (!o,C) alternative
20311 (*movxf_internal_rex64): Ditto.
20312 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20314 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
20316 * doc/invoke.texi: fix typo.
20317 * doc/objc.texi: fix typo.
20319 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20321 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20322 for the first two alternatives.
20324 2013-01-24 Diego Novillo <dnovillo@google.com>
20326 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
20327 (ggc-zone.o): Remove.
20328 * configure.ac: Remove option --with-gc.
20329 * configure: Re-generate.
20330 * doc/install.texi: Remove documentation for --with-gc.
20331 * gengtype.c (write_enum_defn): Remove. Update all users.
20332 (write_Types_process_field): Remove generation of gt_e_* argument.
20333 (output_type_enum): Remove. Update all users.
20334 (write_enum_defn): Remove. Update all users.
20335 (enum alloc_zone): Remove. Update all users.
20336 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
20337 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20338 Update all callers.
20339 (struct ptr_data): Remove field TYPE. Update all users.
20340 (gt_pch_note_object): Remove argument TYPE. Update all users.
20341 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20343 * ggc-none.c (ggc_alloc_typed_stat): Remove.
20344 (struct alloc_zone): Remove.
20345 (ggc_internal_alloc_zone_stat): Remove.
20346 (ggc_internal_cleared_alloc_zone_stat): Remove.
20347 * ggc-page.c (ggc_alloc_typed_stat): Remove.
20348 (ggc_pch_count_object): Remove last argument. Update all users.
20349 (ggc_pch_alloc_object): Remove last argument. Update all users.
20350 (struct alloc_zone): Remove.
20351 * ggc-zone.c: Remove.
20352 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
20353 (struct alloc_zone): Remove.
20354 (ggc_alloc_typed_stat): Remove.
20355 (ggc_alloc_typed): Remove.
20356 (ggc_splay_alloc): Remove first argument.
20357 (rtl_zone): Remove. Update all users.
20358 (tree_zone): Remove. Update all users.
20359 (tree_id_zone): Remove. Update all users.
20360 (ggc_internal_zone_alloc_stat): Remove. Update all users.
20361 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
20362 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
20363 * tree-ssanames.c: Remove references to zone allocator in comments.
20365 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
20367 * config/avr/avr.c (avr_out_fract): Make register numbers that
20368 might be outside of source operand signed.
20370 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
20372 * config/i386/constraints.md (Yf): New constraint.
20373 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20374 of f constraint to conditionaly disable x87 register preferences.
20375 (*movdf_internal): Ditto.
20376 (*movsf_internal): Ditto.
20378 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20380 PR inline-asm/55934
20381 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20382 that have operands with impossible constraints.
20383 Add a FIXME for a speed-up opportunity.
20384 * lra-constraints.c (process_alt_operands): Verify that a class
20385 selected from constraints on asms is valid for the operand mode.
20386 (curr_insn_transform): Remove incorrect comment.
20388 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
20390 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20391 TOC operand is a valid symbol ref in the constant pool.
20393 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20395 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
20397 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
20400 * config/avr/stdfix.h: New file.
20401 * t-avr (stdfix-gcc.h): New rule to build it.
20402 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20404 2013-01-23 Kostya Serebryany <kcc@google.com>
20406 * config/darwin.h: remove dependency on
20407 CoreFoundation (asan on Mac OS).
20409 2013-01-23 Jakub Jelinek <jakub@redhat.com>
20412 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20413 instead of cmpdi_operand for first comparison operand.
20414 Don't assert that comparison operands aren't both constants.
20416 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20418 * doc/install.texi (Downloading the Source): Update references to
20419 downloading separate components.
20421 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20423 * doc/extend.texi (__int128): Improve grammar.
20425 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
20428 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20429 alternative to (o,r).
20430 (*movdi_internal_rex64): Remove (!o,n) alternative.
20431 (DImode immediate->memory splitter): Remove.
20432 (DImode immediate->memory peephole2): Remove.
20433 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20434 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20435 alternative to (!o,*r).
20436 (*movtf_internal_sse): New pattern.
20437 (*movxf_internal_rex64): New pattern.
20438 (*movxf_internal): Disable for TARGET_64BIT.
20439 (*movdf_internal_rex64): Remove (!o,F) alternative.
20441 2013-01-22 Jakub Jelinek <jakub@redhat.com>
20443 PR middle-end/56074
20444 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20445 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20446 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20447 stmt locations where LOCATION_LOCUS of the stmt location is
20448 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20451 * config/i386/i386.md (UNSPEC_STOS): New.
20452 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20453 *strsetqi_1): Add UNSPEC_STOS.
20455 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20458 * doc/invoke.texi: Remove left over -Wsynth example.
20460 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20462 PR tree-optimization/56051
20463 * fold-const.c (fold_binary_loc): Don't fold
20464 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20465 a narrowing conversion, or widening conversion from signed
20468 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20470 PR rtl-optimization/56023
20471 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20472 dependent on debug instruction.
20474 2013-01-21 Martin Jambor <mjambor@suse.cz>
20476 PR middle-end/56022
20477 * function.c (allocate_struct_function): Call
20478 invoke_set_current_function_hook earlier.
20480 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20482 * reload1.c (init_reload): Only initialize reload_obstack
20483 during the first call.
20485 2013-01-21 Marek Polacek <polacek@redhat.com>
20487 * cfgloop.c (verify_loop_structure): Fix up grammar.
20489 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20491 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20492 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20494 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20497 * config/arm/marvell-pj4.md: Update copyright year.
20498 Fix up use of alu to alu_reg and simple_alu_imm.
20500 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20502 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20504 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20507 * lra-constraints.c (curr_insn_transform): Don't reuse original
20508 insn for secondary memory move when memory mode should be different.
20510 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20512 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20513 atomic_storedi_1): New patterns.
20515 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20517 btver2 pipeline descriptions.
20518 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20520 * config/i386/i386.md (btver2_decode): New type attributes.
20521 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
20523 * config/i386/btver2.md: New file describing btver2 pipelines.
20525 2013-01-19 Andrew Pinski <apinski@cavium.com>
20527 PR tree-optimization/52631
20528 * tree-ssa-sccvn (visit_use): Before looking up the original
20529 statement, try looking up the simplified expression.
20531 2013-01-19 Anthony Green <green@moxielogic.com>
20533 * config/moxie/moxie.c (moxie_expand_prologue): Set
20534 current_function_static_stack_size.
20536 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20538 PR tree-optimization/56029
20539 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20540 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20542 2013-01-18 Sharad Singhai <singhai@google.com>
20544 PR tree-optimization/55995
20545 * dumpfile.c (dump_loc): Print location only if available.
20546 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20548 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20551 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20552 secondary memory move.
20553 (inherit_reload_reg): Use rclass instead of cl for
20554 check_secondary_memory_needed_p.
20556 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20558 PR middle-end/56015
20559 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
20560 the case where writing real complex part of target modifies op1.
20562 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20564 * config/aarch64/aarch64-simd.md
20565 (aarch64_vcond_internal<mode>): Handle unordered cases.
20566 * config/aarch64/iterators.md (v_cmp_result): New.
20568 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20569 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20571 * config/arm/marvell-pj4.md: New file.
20572 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20573 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20574 (generic_vfp): Likewise.
20575 * config/arm/arm-cores.def: Add marvell-pj4.
20576 * config/arm/arm-tune.md: Regenerate.
20577 * config/arm/arm-tables.opt: Regenerate.
20578 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20579 * doc/invoke.texi: Document marvell-pj4.
20581 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20583 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20585 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20590 * var-tracking.c (negative_power_of_two_p): New.
20591 (global_get_addr_cache, local_get_addr_cache): New.
20592 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20593 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20595 (vt_stack_offset_p): Remove.
20596 (vt_canon_true_dep): Always canonicalize loc's address.
20597 (clobber_overlapping_mems): Make sure we have a MEM.
20598 (local_get_addr_clear_given_value): New.
20599 (val_reset): Clear local cached entries.
20600 (compute_bb_dataflow): Create and release the local cache.
20601 Disable duplicate MEMs clobbering.
20602 (emit_notes_in_bb): Clobber MEMs likewise.
20603 (vt_emit_notes): Create and release the local cache.
20604 (vt_initialize, vt_finalize): Create and release the global
20605 cache, respectively.
20606 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
20608 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20610 PR libmudflap/53359
20611 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20612 not found in the symtab.
20614 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
20617 PR rtl-optimization/55547
20618 PR rtl-optimization/53827
20621 * alias.c (offset_overlap_p): New, factored out of...
20622 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20623 the conservative special case for symbolic constants. Don't
20624 adjust zero sizes on alignment.
20626 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20628 PR rtl-optimization/52573
20629 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20630 REG_UNUSED for the same register.
20632 2013-01-17 Richard Biener <rguenther@suse.de>
20633 Marek Polacek <polacek@redhat.com>
20635 PR rtl-optimization/55833
20636 * loop-unswitch.c (unswitch_loops): Move loop verification...
20637 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20638 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20639 Set it to true when we're removing a loop from hierarchy tree in
20640 an irreducible region.
20641 (fix_bb_placements): Adjust caller.
20642 (fix_loop_placements): Likewise.
20644 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
20646 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20647 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20648 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20649 Remove ID. Adjust comments.
20650 * config/avr/avr-c.c (avr_builtin_name): Remove.
20651 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20652 * config/avr/avr.c (avr_tolower): New static function.
20653 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20654 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20655 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20658 2013-01-17 Jan Hubicka <jh@suse.cz>
20660 PR tree-optimization/55273
20661 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20663 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
20666 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20667 store through atomic_store<mode>_1.
20668 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20670 2013-01-17 Martin Jambor <mjambor@suse.cz>
20672 PR tree-optimizations/55264
20673 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20674 for virtual methods.
20675 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20676 virtual methods before inlining is over.
20677 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20679 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20682 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20684 PR rtl-optimization/56005
20685 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20686 pending reads for prefetch.
20688 2013-01-16 Ian Bolton <ian.bolton@arm.com>
20690 * config/aarch64/aarch64.md
20691 (*cstoresi_neg_uxtw): New pattern.
20692 (*cmovsi_insn_uxtw): New pattern.
20693 (*<optab>si3_uxtw): New pattern.
20694 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20695 (*<optab>si3_insn_uxtw): New pattern.
20696 (*bswapsi2_uxtw): New pattern.
20698 2013-01-16 Richard Biener <rguenther@suse.de>
20700 * tree-inline.c (tree_function_versioning): Remove set but
20701 never used variable.
20703 2013-01-16 Richard Biener <rguenther@suse.de>
20705 PR tree-optimization/55964
20706 * tree-flow.h (rename_variables_in_loop): Remove.
20707 (rename_variables_in_bb): Likewise.
20708 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20709 (copy_loop_before): Adjust and delete update-ssa status.
20710 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
20711 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
20712 (rename_variables_in_loop): Remove.
20713 (slpeel_update_phis_for_duplicate_loop): Likewise.
20714 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20715 use available cfg machinery instead of duplicating it.
20716 Update PHI nodes and perform poor-mans SSA update here.
20717 (slpeel_tree_peel_loop_to_edge): Adjust.
20719 2013-01-16 Richard Biener <rguenther@suse.de>
20721 PR tree-optimization/54767
20722 PR tree-optimization/53465
20723 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20724 (vrp_visit_phi_node): For PHI arguments coming via backedges
20725 drop all symbolical range information.
20726 (execute_vrp): Compute backedges.
20728 2013-01-16 Richard Biener <rguenther@suse.de>
20730 * doc/install.texi: Update CLooG and ISL requirements to
20733 2013-01-16 Christian Bruel <christian.bruel@st.com>
20736 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20737 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20738 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20740 2013-01-16 DJ Delorie <dj@redhat.com>
20742 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20743 (UNSPECV_SP_SWITCH_E): New.
20744 (sp_switch_1): Change to an unspec.
20745 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20748 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
20750 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20751 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20752 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20753 (expand_mem_thread_fence): Ditto.
20754 (expand_mem_signal_fence): Ditto.
20755 (expand_atomic_load): Ditto.
20756 (expand_atomic_store): Ditto.
20758 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
20760 PR rtl-optimization/55547
20761 PR rtl-optimization/53827
20764 * alias.c (memrefs_conflict_p): Set sizes to negative after
20767 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20770 * function.c (thread_prologue_and_epilogue_insns): Always
20771 add crtl->drap_reg to set_up_by_prologue.set, even if
20772 stack_realign_drap is false.
20774 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20776 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20777 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20778 *call): Fix indention.
20780 2013-01-15 Tom de Vries <tom@codesourcery.com>
20783 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20786 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20788 PR rtl-optimization/55153
20789 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20791 2013-01-15 Martin Jambor <mjambor@suse.cz>
20793 PR tree-optimization/55920
20794 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20795 accesses as grp_to_be_debug_replaced.
20797 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20799 PR tree-optimization/55920
20800 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20801 there is non-useless type conversion needed from debug rhs to lhs,
20802 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20804 2013-01-15 Joseph Myers <joseph@codesourcery.com>
20805 Mikael Pettersson <mikpe@it.uu.se>
20808 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
20810 (ASM_OUTPUT_CASE_LABEL): Remove.
20811 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
20812 * final.c (shorten_branches): Update alignment of labels before
20813 jump tables if CASE_VECTOR_SHORTEN_MODE.
20815 2013-01-15 Richard Biener <rguenther@suse.de>
20818 * system.h: Do not include gmp.h for building host tools.
20820 2013-01-15 Richard Biener <rguenther@suse.de>
20822 PR middle-end/55882
20823 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
20824 account for bitpos when computing alignment.
20826 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
20828 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
20829 (ix86_target_macros_internal): Likewise.
20831 * config/i386/i386.c (m_CORE2I7): Removed.
20832 (m_CORE_HASWELL): New macro.
20833 (m_CORE_ALL): Likewise.
20834 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
20835 (initial_ix86_arch_features): Likewise.
20836 (processor_target_table): Initializations for Core avx2.
20837 (cpu_names): New names "core-avx2".
20838 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
20839 PROCESSOR_CORE_HASWELL.
20840 (ix86_issue_rate): New case.
20841 (ia32_multipass_dfa_lookahead): Likewise.
20842 (ix86_sched_init_global): Likewise.
20844 * config/i386/i386.h (TARGET_HASWELL): New macro.
20845 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
20846 (processor_type): New PROCESSOR_HASWELL.
20848 2013-01-15 Jakub Jelinek <jakub@redhat.com>
20850 PR tree-optimization/55955
20851 * tree-vect-loop.c (vectorizable_reduction): Give up early on
20852 *SHIFT_EXPR and *ROTATE_EXPR codes.
20854 PR tree-optimization/48766
20855 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
20856 -ftrapv disable -fwrapv.
20858 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20861 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
20862 etc. to 1 and not to __flash.
20863 Use LL suffix for __INT24_MAX__ with -mint8.
20864 Use ULL suffix for __UINT24_MAX__ with -mint8.
20866 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20868 * config/avr/avr-arch.h
20869 (struct base_arch_s): Use typedef avr_arch_t instead.
20870 (struct arch_info_s): Use typedef avr_arch_info_t instead.
20871 (struct mcu_type_s): Use typedef avr_mcu_t instead.
20872 * config/avr/avr.c: Same.
20873 * config/avr/avr-devices.c: Same.
20874 * config/avr/driver-avr.c: Same.
20875 * config/avr/gen-avr-mmcu-texi.c: Same.
20876 * config/avr/avr-mcus.def: Adjust comment.
20878 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
20880 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
20881 * config/aarch64/iterators.md (VALLDI): New.
20883 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
20884 Andi Kleen <ak@linux.intel.com>
20887 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
20888 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
20891 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20893 * config/avr/avr-stdint.h: Remove trailing blanks.
20894 * config/avr/avr-log.h: Same.
20895 * config/avr/avr-arch.h: Same.
20896 * config/avr/avr-devices.c: Same.
20897 * config/avr/avr-dimode.md: Same.
20898 * config/avr/predicates.md: Same.
20899 * config/avr/avr-c.c: Same. And fix typo.
20901 * config/avr/avr-protos.h: Same. And:
20902 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
20903 (init_cumulative_args): Rename to avr_init_cumulative_args.
20904 (expand_prologue): Rename to avr_expand_prologue.
20905 (expand_epilogue): Rename to avr_expand_epilogue.
20906 (adjust_insn_length): Rename to avr_adjust_insn_length.
20907 (notice_update_cc): Rename to avr_notice_update_cc.
20908 (final_prescan_insn): Rename to avr_final_prescan_insn.
20909 * config/avr/avr.c: Same.
20910 * config/avr/avr.h: Same.
20911 * config/avr/avr.md: Remove trailing blanks.
20912 (prologue): Use avr_expand_prologue.
20913 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
20915 2013-01-14 Richard Biener <rguenther@suse.de>
20917 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
20918 verify_location, collect_subblocks): New functions.
20919 (verify_gimple_in_cfg): Verify that locations only reference
20920 BLOCKs in the functions BLOCK tree.
20922 2013-01-14 Richard Biener <rguenther@suse.de>
20924 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
20926 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
20928 (insert_out_of_ssa_copy_on_edge): Likewise.
20929 (rewrite_close_phi_out_of_ssa): Likewise.
20930 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
20932 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
20933 propagated constants.
20934 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
20937 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
20939 * config/avr/avr-modes.def: Add GPL copyright notice.
20941 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
20943 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
20944 MEMMODEL_MASK to determine memory model.
20945 (atomic_store<mode>): Ditto from operands[2].
20946 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
20948 2013-01-13 Jakub Jelinek <jakub@redhat.com>
20951 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
20952 (fold_gimple_assign): Don't call unshare_expr here.
20953 (fold_ctor_reference): Call unshare_expr.
20955 2013-01-13 Terry Guo <terry.guo@arm.com>
20957 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
20958 * doc/fragments.texi: Document MULTILIB_REUSE.
20959 * gcc.c (multilib_reuse): New internal spec.
20960 (set_multilib_dir): Also search multilib from multilib_reuse.
20961 * genmultilib (tmpmultilib3): Refactor code.
20962 (tmpmultilib4): Ditto.
20963 (multilib_reuse): New multilib argument.
20965 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
20967 * Makefile.in: Update copyright.
20969 2013-01-12 Tom de Vries <tom@codesourcery.com>
20971 PR middle-end/55890
20972 * calls.c (expand_call): Check if arg_nr is valid.
20974 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
20976 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
20977 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
20978 documentation. Add missing '__' in front of
20979 __builtin_ia32_packssdw256.
20981 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20984 * config/s390/s390.c (s390_preferred_reload_class): Do not return
20985 NO_REGS for larl operands.
20986 (s390_reload_larl_operand): Use s390_load_address instead of
20989 2013-01-11 Richard Biener <rguenther@suse.de>
20991 * tree-cfg.c (verify_node_sharing_1): Split out from ...
20992 (verify_node_sharing): ... here.
20993 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
20995 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
20997 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
20998 Substitute TREECHECKING.
20999 * configure: Regenerate.
21000 * Makefile.in (TREECHECKING): New.
21002 2013-01-11 Richard Guenther <rguenther@suse.de>
21004 PR tree-optimization/44061
21005 * tree-vrp.c (extract_range_basic): Compute zero as
21006 value-range for __builtin_constant_p of function parameters.
21008 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
21010 Update copyright years.
21012 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21014 PR rtl-optimization/55672
21015 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21016 const to be eliminable.
21018 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
21020 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21021 * configure: Regenerate.
21023 2013-01-10 Richard Biener <rguenther@suse.de>
21025 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
21027 2013-01-10 Richard Biener <rguenther@suse.de>
21030 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21031 locations for virtual PHI arguments.
21032 (rewrite_update_phi_arguments): Likewise.
21034 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21036 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21039 2013-01-10 Jakub Jelinek <jakub@redhat.com>
21041 PR tree-optimization/55921
21042 * tree-complex.c (expand_complex_asm): New function.
21043 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21045 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21048 * config/s390/s390.c (s390_symref_operand_p)
21049 (s390_loadrelative_operand_p): Merge the two functions.
21050 (s390_check_qrst_address, print_operand_address): Add parameters
21051 to s390_loadrelative_operand_p invokation.
21052 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21053 (s390_reload_larl_operand, s390_secondary_reload): Use
21054 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21055 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21057 2013-01-09 Mike Stump <mikestump@comcast.net>
21059 * dse.c (record_store): Remove unnecessary assert.
21061 2013-01-09 Jan Hubicka <jh@suse.cz>
21063 PR tree-optimization/55569
21064 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21065 * cfgloop.h (scale_loop_profile): Likewise.
21067 2013-01-09 Jan Hubicka <jh@suse.cz>
21070 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21072 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21074 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21076 PR middle-end/55114
21077 * expr.h (maybe_emit_group_store): Declare.
21078 * expr.c (maybe_emit_group_store): New function.
21079 * builtins.c (expand_builtin_int_roundingfn): Call it.
21080 (expand_builtin_int_roundingfn_2): Likewise.
21082 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21084 PR rtl-optimization/55829
21085 * lra-constraints.c (match_reload): Add code for absent output.
21086 (curr_insn_transform): Add code for reloads of matched inputs
21089 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
21091 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21092 attribute of movddup insn to DF.
21093 (*vec_interleave_lowv2df): Ditto.
21094 (vec_dupv2df): Ditto.
21096 2013-01-09 Jan Hubicka <jh@suse.cz>
21098 PR tree-optimiation/55875
21099 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21100 EVERY_ITERATION parameter.
21101 (number_of_iterations_exit): Check if exit is executed every iteration.
21102 (idx_infer_loop_bounds): Similarly here.
21103 (n_of_executions_at_most): Simplify
21104 to only test for cases where statement is dominated by the
21105 particular bound; handle correctly the "postdominance" test.
21106 (scev_probably_wraps_p): Use max loop iterations info
21107 as a global bound first.
21109 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
21110 Nick Clifton <nickc@redhat.com>
21112 * config/v850/v850.md (cbranchsf4): New pattern.
21113 (cstoresf4): New pattern.
21114 (cbranchdf4): New pattern.
21115 (cstoredf4): New pattern.
21116 (movsicc): Disallow floating point comparisons.
21117 (cmpsf_le_insn): Fix order of operators.
21118 (cmpsf_lt_insn): Likewise.
21119 (cmpsf_eq_insn): Likewise.
21120 (cmpdf_le_insn): Likewise.
21121 (cmpdf_lt_insn): Likewise.
21122 (cmpdf_eq_insn): Likewise.
21123 (cmpsf_ge_insn): Use LE comparison.
21124 (cmpdf_ge_insn): Likewise.
21125 (cmpsf_gt_insn): Use LT comparison.
21126 (cmpdf_gt_insn): Likewise.
21127 (cmpsf_ne_insn): Delete pattern.
21128 (cmpdf_ne_insn): Delete pattern.
21129 * config/v850/v850.c (v850_gen_float_compare): Use
21130 gen_cmpdf_eq_insn for NE comparison.
21131 (v850_float_z_comparison_operator)
21132 (v850_float_nz_comparison_operator): Move from here ...
21133 * config/v850/predicates.md: ... to here. Move GT and GE
21134 comparisons into v850_float_z_comparison_operator.
21135 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21137 (v850_float_nz_comparison_operator): Likewise.
21139 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21141 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21142 with calls to gen_insvsi/gen_insvdi.
21144 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21146 * config/i386/i386.c (initial_ix86_tune_features): Set up
21147 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21149 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21150 Jakub Jelinek <jakub@redhat.com>
21152 PR tree-optimization/48189
21153 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21154 If nitercst is 0, don't predict the exit edge.
21156 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21158 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21159 in asm_fprintf with reg_names.
21160 (aarch64_print_operand_address): Likewise.
21161 (aarch64_return_addr): Likewise.
21162 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21164 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21166 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21167 (INT_U6_BITS): Likewise.
21168 * config/pa/predicates.md (uint6_operand): New predicate.
21169 (shift5_operand, shift6_operand): Likewise.
21170 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21172 (lshrdi3): Use shift6_operand.
21173 (shrpsi4, shrpdi4): New insn patterns.
21174 (extzv): Delete expander.
21175 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21176 predicates in unamed zero extract patterns. Tighten common constraint.
21177 (extv): Delete expander.
21178 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21179 predicates in unamed sign extract patterns. Tighten common constraint.
21180 (insv): Delete expander.
21181 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21182 predicates in unamed insert patterns. Tighten common constraint.
21183 Change uint32_operand predicate to uint6_operand predicate in unamed
21184 DImode pattern to insert constant values of type 1...1xxxx.
21186 2013-01-04 Jan Hubicka <jh@suse.cz>
21188 PR tree-optimization/55823
21189 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21192 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21193 Uros Bizjak <ubizjak@gmail.com>
21195 PR rtl-optimization/55845
21196 * df-problems.c (can_move_insns_across): Stop scanning at
21197 volatile_insn_p source instruction or give up if
21198 across_from .. across_to range contains any volatile_insn_p
21201 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21203 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21204 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
21206 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
21207 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
21209 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21212 * asan.c (asan_clear_shadow): New function.
21213 (asan_emit_stack_protection): Use it.
21215 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21217 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21218 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21219 with tab instead of space.
21221 2013-01-08 Nick Clifton <nickc@redhat.com>
21223 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21224 register bank 0 at the start of an interrupt handler.
21225 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21228 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21230 * config/aarch64/aarch64-simd.md
21231 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21232 (aarch64_simd_bsl): Likewise.
21233 (aarch64_vcond_internal<mode>): Likewise.
21234 (vcond<mode><mode>): Likewise.
21235 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21236 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21238 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21240 * config/aarch64/aarch64-builtins.c
21241 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21243 2013-01-08 Martin Jambor <mjambor@suse.cz>
21246 * tree-sra.c (analyze_access_subtree): Return true also after
21247 potentially creating a debug-only replacement.
21249 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21251 PR middle-end/55890
21252 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21254 PR tree-optimization/54120
21255 * tree-vrp.c (range_fits_type_p): Don't allow
21256 src_precision < precision from signed vr to unsigned_p
21257 if vr->min or vr->max is negative.
21258 (simplify_float_conversion_using_ranges): Test can_float_p
21259 against CODE_FOR_nothing.
21261 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21262 Richard Biener <rguenther@suse.de>
21264 PR middle-end/55851
21265 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21266 types instead of just INTEGER_TYPE types.
21268 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
21270 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21273 2013-01-07 Steve Ellcey <sellcey@mips.com>
21276 * config/mips/mips.opt: Change mad to mmad to match documentation.
21278 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21281 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21282 .progmemx.data now.
21284 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21287 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21288 (avr_addrspace_t): Add .section_name field.
21289 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21291 (avr_addrspace): Same. Initialize .section_name. Remove last
21292 NULL entry. Put __memx into .progmemx.data.
21293 (progmem_section_prefix): Remove.
21294 (avr_asm_init_sections): No need to initialize progmem_section.
21295 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21296 section name prefix.
21297 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21298 retrieve the progmem section.
21299 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21300 boundary to run over avr_addrspace[].
21301 (avr_register_target_pragmas): Ditto.
21303 2013-01-06 Jakub Jelinek <jakub@redhat.com>
21305 * varasm.c (output_constant_def_contents): For asan_protect_global
21306 protected strings, adjust DECL_ALIGN if needed, before testing for
21308 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21309 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21311 (output_object_block): For asan protected decls, emit asan padding
21312 after their contents.
21313 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21314 (asan_finish_file): Test it here instead.
21316 2013-01-07 Nick Clifton <nickc@redhat.com>
21317 Matthias Klose <doko@debian.org>
21318 Doug Kwan <dougkwan@google.com>
21319 H.J. Lu <hongjiu.lu@intel.com>
21322 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21324 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21326 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21328 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
21330 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21332 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21335 * doc/install.texi (Cross-Compiler-Specific Options): Document
21338 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21340 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21341 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21342 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21343 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21344 vqmovun_high_s64): Fix source operand number and update copyright.
21346 2013-01-07 Richard Biener <rguenther@suse.de>
21348 PR middle-end/55890
21349 * gimple.h (gimple_call_builtin_p): New overload.
21350 * gimple.c (validate_call): New function.
21351 (gimple_call_builtin_p): Likewise.
21352 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21353 Use gimple_call_builtin_p.
21354 (find_func_clobbers): Likewise.
21355 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21356 (strlen_optimize_stmt): Likewise.
21358 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21360 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21361 (vld1q_dup_*): Likewise.
21362 (vld1_*): Likewise.
21363 (vld1q_*): Likewise.
21364 (vld1_lane_*): Likewise.
21365 (vld1q_lane_*): Likewise.
21367 2013-01-07 Richard Biener <rguenther@suse.de>
21369 * lto-streamer.h (LTO_minor_version): Bump to 2.
21371 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21373 * config/aarch64/aarch64-protos.h
21374 (aarch64_const_double_zero_rtx_p): Rename to...
21375 (aarch64_float_const_zero_rtx_p): ...this.
21376 (aarch64_float_const_representable_p): New.
21377 (aarch64_output_simd_mov_immediate): Likewise.
21378 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21379 move immediate case.
21380 * config/aarch64/aarch64.c
21381 (aarch64_const_double_zero_rtx_p): Rename to...
21382 (aarch64_float_const_zero_rtx_p): ...this.
21383 (aarch64_print_operand): Allow printing of new constants.
21384 (aarch64_valid_floating_const): New.
21385 (aarch64_legitimate_constant_p): Check for valid floating-point
21387 (aarch64_simd_valid_immediate): Likewise.
21388 (aarch64_vect_float_const_representable_p): New.
21389 (aarch64_float_const_representable_p): Likewise.
21390 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21391 (aarch64_output_simd_mov_immediate): New.
21392 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21393 (*movdf_aarch64): Likewise.
21394 * config/aarch64/constraints.md (Ufc): New.
21395 (Y): call aarch64_float_const_zero_rtx.
21396 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21398 2013-01-07 Richard Biener <rguenther@suse.de>
21400 PR tree-optimization/55888
21401 PR tree-optimization/55862
21402 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21403 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21404 not if it is contained therein.
21406 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21408 * config/avr/t-avr: Typo.
21410 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21413 * config/avr/t-avr: Don't automatically rebuild
21414 $(srcdir)/config/avr/t-multilib
21415 $(srcdir)/config/avr/avr-tables.opt
21416 $(srcdir)/doc/avr-mmcu.texi
21417 (avr-mcus): New phony target to build them on request.
21418 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21419 * avr/avr-mcus.def: Adjust comments.
21421 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
21423 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21425 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21427 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21429 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21431 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21433 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
21435 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21436 to generate profiling.
21437 * config/rs6000/aix64.h (LIB_SPEC): Same.
21439 2013-01-04 Andrew Pinski <apinski@cavium.com>
21441 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21443 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21445 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
21447 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21449 (ix86_expand_move): Ditto.
21450 (ix86_zero_extend_to_Pmode): Ditto.
21451 (ix86_expand_call): Ditto.
21452 (ix86_expand_special_args_builtin): Ditto.
21453 (ix86_expand_builtin): Ditto.
21455 2013-01-04 Richard Biener <rguenther@suse.de>
21457 PR tree-optimization/55862
21458 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21459 translating them through PHI nodes.
21461 2013-01-04 Martin Jambor <mjambor@suse.cz>
21463 PR tree-optimization/55755
21464 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21465 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21467 2013-01-04 Marek Polacek <polacek@redhat.com>
21469 PR middle-end/55859
21470 * opts.c (default_options_optimization): Clarify error message.
21472 2013-01-04 Richard Biener <rguenther@suse.de>
21474 PR middle-end/55863
21475 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21478 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21481 * config/pa/pa.md (movsi): Revert previous change.
21482 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21485 2013-01-03 Richard Henderson <rth@redhat.com>
21487 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21488 after eliminating TLS symbols.
21490 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
21493 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21494 * graphite-poly.c (debug_gmp_value): Likewise.
21496 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
21499 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21500 selected code model, define __code_mode_small__, __code_model_medium__,
21501 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21502 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21503 xchg temporary register with %k. Declare temporary register as
21505 [__x86_64__]: For medium and large code models, preserve %rbx register.
21507 2013-01-03 Richard Biener <rguenther@suse.de>
21509 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
21510 (dump_subscript): Adjust.
21511 (finalize_ddr_dependent): Do not dump redundant info.
21512 (analyze_siv_subscript): Adjust.
21513 (subscript_dependence_tester): Likewise.
21514 (compute_affine_dependence): Likewise.
21516 2013-01-03 Richard Biener <rguenther@suse.de>
21519 2013-01-03 Richard Biener <rguenther@suse.de>
21521 PR tree-optimization/55857
21522 * tree-vect-stmts.c (vectorizable_load): Do not setup
21523 re-alignment for invariant loads.
21525 2013-01-02 Richard Biener <rguenther@suse.de>
21527 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21528 invariant load do not generate a vector load from the scalar location.
21530 2013-01-03 Richard Biener <rguenther@suse.de>
21532 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21533 for not vectorizing.
21534 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21535 not build INDIRECT_REFs, call get_name once only.
21536 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21537 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21539 2013-01-03 Richard Biener <rguenther@suse.de>
21541 PR tree-optimization/55857
21542 * tree-vect-stmts.c (vectorizable_load): Do not setup
21543 re-alignment for invariant loads.
21545 2013-01-03 Richard Biener <rguenther@suse.de>
21548 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21549 prefer a built-in decl.
21551 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21553 * gcc.c (process_command): Update copyright notice dates.
21554 * gcov.c (print_version): Likewise.
21555 * gcov-dump.c (print_version): Likewise.
21557 PR rtl-optimization/55838
21558 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21559 iv0.step, iv1.step and step.
21561 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21562 Marc Glisse <marc.glisse@inria.fr>
21564 PR tree-optimization/55832
21565 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21566 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21567 integer_{one,zero}_node.
21569 2013-01-03 Jakub Jelinek <jakub@redhat.com>
21572 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21573 * var-tracking.c (reverse_op): Don't add reverse ops to
21574 VALUEs that have already
21575 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
21577 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21579 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21581 2013-01-02 Teresa Johnson <tejohnson@google.com>
21583 * dumpfile.c (dump_loc): Print filename with location.
21584 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21585 new location_t parameter to emit complete unroll message with
21586 new dump framework.
21587 (canonicalize_loop_induction_variables): Compute loops location
21588 and pass to try_unroll_loop_completely.
21589 * loop-unroll.c (report_unroll_peel): New function.
21590 (peel_loops_completely): Use new dump format with location
21591 for main dumpfile message, and invoke report_unroll_peel on success.
21592 (decide_unrolling_and_peeling): Ditto.
21593 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21594 by report_unroll_peel.
21595 (decide_peel_completely): Ditto.
21596 (decide_unroll_constant_iterations): Ditto.
21597 (decide_unroll_runtime_iterations): Ditto.
21598 (decide_peel_simple): Ditto.
21599 (decide_unroll_stupid): Ditto.
21600 * cfgloop.c (get_loop_location): New function.
21601 * cfgloop.h (get_loop_location): Declare.
21603 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21605 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21608 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21610 PR middle-end/55198
21611 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21612 BLKmode objects when EXPAND_MEMORY is specified.
21614 2013-01-02 Sriraman Tallam <tmsriram@google.com>
21616 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21618 (fold_builtin_cpu): Do not share cpu model decls across statements.
21620 2013-01-02 Jason Merrill <jason@redhat.com>
21623 * tree.c (build_array_type_1): Revert earlier change.
21625 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21627 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21629 * config/aarch64/aarch64-tune.md: Re-generate.
21631 2013-01-02 Richard Biener <rguenther@suse.de>
21633 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21634 invariant load do not generate a vector load from the scalar location.
21636 2013-01-02 Richard Biener <rguenther@suse.de>
21639 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21640 * configure: Regenerate.
21642 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21644 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21645 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21646 (expand_builtin_int_roundingfn_2): Keep the original target around
21647 for the fallback case.
21649 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21651 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21652 to be clear for sign changes.
21654 2013-01-01 Jan Hubicka <jh@suse.cz>
21656 * ipa-inline-analysis.c: Fix formatting.
21658 2013-01-01 Jakub Jelinek <jakub@redhat.com>
21660 PR tree-optimization/55831
21661 * tree-vect-loop.c (get_initial_def_for_induction): Use
21662 gsi_after_labels instead of gsi_start_bb.
21664 Copyright (C) 2013 Free Software Foundation, Inc.
21666 Copying and distribution of this file, with or without modification,
21667 are permitted in any medium without royalty provided the copyright
21668 notice and this notice are preserved.