1 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
4 ceildf2, btruncdf2, instead of vsx_* name.
6 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
7 iterators to only do V2DF and V4SF here. Move the DF code to
8 rs6000.md where it is combined with SF mode. Replace <VSv> with
9 just 'v' since only vector operations are handled with these insns
10 after moving the DF support to rs6000.md.
11 (vsx_sub<mode>3): Likewise.
12 (vsx_mul<mode>3): Likewise.
13 (vsx_div<mode>3): Likewise.
14 (vsx_fre<mode>2): Likewise.
15 (vsx_neg<mode>2): Likewise.
16 (vsx_abs<mode>2): Likewise.
17 (vsx_nabs<mode>2): Likewise.
18 (vsx_smax<mode>3): Likewise.
19 (vsx_smin<mode>3): Likewise.
20 (vsx_sqrt<mode>2): Likewise.
21 (vsx_rsqrte<mode>2): Likewise.
22 (vsx_fms<mode>4): Likewise.
23 (vsx_nfma<mode>4): Likewise.
24 (vsx_copysign<mode>3): Likewise.
25 (vsx_btrunc<mode>2): Likewise.
26 (vsx_floor<mode>2): Likewise.
27 (vsx_ceil<mode>2): Likewise.
28 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
29 (vsx_sminsf3): Likewise.
30 (vsx_fmadf4): Likewise.
31 (vsx_fmsdf4): Likewise.
32 (vsx_nfmadf4): Likewise.
33 (vsx_nfmsdf4): Likewise.
34 (vsx_cmpdf_internal1): Likewise.
36 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
37 simpler to select whether a target has SPE or traditional floating
38 point support in iterators.
39 (TARGET_DF_SPE): Likewise.
40 (TARGET_SF_FPR): Likewise.
41 (TARGET_DF_FPR): Likewise.
42 (TARGET_SF_INSN): Macros to say whether floating point support
43 exists for a given operation for expanders.
44 (TARGET_DF_INSN): Likewise.
46 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
47 combining of SF/DF mode operations, using both traditional and VSX
55 (abs<mode>2): Combine SF/DF modes using traditional floating point
56 instructions. Add support for using the upper DF registers with
57 VSX support, and SF registers with power8-vector support. Update
58 expanders for operations supported by both the SPE and traditional
60 (abs<mode>2_fpr): Likewise.
61 (nabs<mode>2): Likewise.
62 (nabs<mode>2_fpr): Likewise.
63 (neg<mode>2): Likewise.
64 (neg<mode>2_fpr): Likewise.
65 (add<mode>3): Likewise.
66 (add<mode>3_fpr): Likewise.
67 (sub<mode>3): Likewise.
68 (sub<mode>3_fpr): Likewise.
69 (mul<mode>3): Likewise.
70 (mul<mode>3_fpr): Likewise.
71 (div<mode>3): Likewise.
72 (div<mode>3_fpr): Likewise.
73 (sqrt<mode>3): Likewise.
74 (sqrt<mode>3_fpr): Likewise.
76 (rsqrt<mode>2): Likewise.
77 (cmp<mode>_fpr): Likewise.
78 (smax<mode>3): Likewise.
79 (smin<mode>3): Likewise.
80 (smax<mode>3_vsx): Likewise.
81 (smin<mode>3_vsx): Likewise.
82 (negsf2): Delete SF operations that are merged with DF.
89 (fmasf4_fpr): Likewise.
90 (fmssf4_fpr): Likewise.
91 (nfmasf4_fpr): Likewise.
92 (nfmssf4_fpr): Likewise.
94 (rsqrtsf_internal1): Likewise.
97 (cmpsf_internal1): Likewise.
98 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
99 (negdf2): Delete DF operations that are merged with SF.
107 (rsqrtdf_internal1): Likewise.
108 (fmadf4_fpr): Likewise.
109 (fmsdf4_fpr): Likewise.
110 (nfmadf4_fpr): Likewise.
111 (nfmsdf4_fpr): Likewise.
115 (cmpdf_internal1): Likewise.
116 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
117 (btrunc<mode>2): Delete separate expander, and combine with the
118 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
119 (btrunc<mode>2_fpr): Likewise.
120 (ceil<mode>2): Likewise.
121 (ceil<mode>2_fpr): Likewise.
122 (floor<mode>2): Likewise.
123 (floor<mode>2_fpr): Likewise.
124 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
125 Add support for using the upper registers with VSX and
126 power8-vector. Move insns to be closer to the define_expands. On
127 VSX systems, prefer the traditional form of FMA over the VSX
128 version, since the traditional form allows the target not to
129 overlap with the inputs.
130 (fms<mode>4_fpr): Likewise.
131 (nfma<mode>4_fpr): Likewise.
132 (nfms<mode>4_fpr): Likewise.
134 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
135 Richard Earnshaw <richard.earnshaw@arm.com>
137 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
138 (struct mult_cost_table): Likewise.
139 (struct mem_cost_table): Likewise.
140 (struct fp_cost_table): Likewise.
141 (struct vector_cost_table): Likewise.
142 (cpu_cost_table): Likewise.
143 * config/arm/arm.opt (mold-rts-costs): New option.
144 (mnew-generic-costs): Likewise.
145 * config/arm/arm.c (generic_extra_costs): New table.
146 (cortexa15_extra_costs): Likewise.
147 (arm_slowmul_tune): Use NULL as new costs.
148 (arm_fastmul_tune): Likewise.
149 (arm_strongarm_tune): Likewise.
150 (arm_xscale_tune): Likewise.
151 (arm_9e_tune): Likewise.
152 (arm_v6t2_tune): Likewise.
153 (arm_cortex_a5_tune): Likewise.
154 (arm_cortex_a9_tune): Likewise.
155 (arm_v6m_tune): Likewise.
156 (arm_fa726te_tune): Likewise.
157 (arm_cortex_a15_tune): Use cortex15_extra_costs.
158 (arm_cortex_tune): Use generict_extra_costs.
159 (shifter_op_p): New function.
160 (arm_unspec_cost): Likewise.
161 (LIBCALL_COST): Define.
162 (arm_new_rtx_costs): New function.
163 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
164 table is available. Use old costs otherwise unless mnew-generic-costs
166 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
167 (cpu_cost_table): Declare.
169 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
172 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
173 (*subs_mul_imm_<mode>)
174 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
175 (*sub_<shift>_<mode>)
176 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
179 2013-10-03 Ian Bolton <ian.bolton@arm.com>
181 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
183 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
185 2013-10-02 Teresa Johnson <tejohnson@google.com>
187 * predict.c (probably_never_executed): New function.
188 (probably_never_executed_bb_p): Invoke probably_never_executed.
189 (probably_never_executed_edge_p): Ditto.
190 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
191 Treat profile insanities conservatively.
193 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
195 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
197 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
199 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
200 use it. Use smaller increase for scratch. Don't increase reject
201 for early clobber scratch.
202 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
203 setting eliminated regs except setting fp from hfp.
204 (lra_eliminate): Check lra_insn_recog_data on NULL.
206 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
209 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
210 setting -mvsx-timode by default until the underlying problem is
212 (RS6000_CPU, power7 defaults): Likewise.
214 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
216 * config/x-linux (host-linux.o): Remove header dependencies.
217 Use $(COMPILE) and $(POSTCOMPILE).
218 * config/t-linux-android (linux-android.o): Ditto.
220 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
222 * Makefile.in (expmed.o-warn): Remove.
224 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
226 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
227 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
229 2013-10-02 Teresa Johnson <tejohnson@google.com>
231 * dojump.c (do_jump_1): Divide probability between
232 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
234 2013-10-02 Tom Tromey <tromey@redhat.com>
236 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
238 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
240 * tree-flow.h: Remove some prototypes.
241 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
242 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
243 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
244 mark_virtual_phi_result_for_renaming): Relocate here.
245 * tree-into-ssa.h: Add prototypes.
246 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
247 single_pred_before_succ_order.
248 (blocks_in_phiopt_order): Rename and move to cfganal.c.
249 (nonfreeing_call_p) Move to gimple.c.
250 * cfganal.c (single_pred_before_succ_order): Move and renamed from
252 * basic-block.h (single_pred_before_succ_order): Add prototype.
253 * gimple.c (nonfreeing_call_p): Relocate here.
254 * gimple.h: Add prototype.
255 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
256 * tree-ssa-dom.h: New file. Relocate prototypes here.
257 * tree-ssa.h: Include tree-ssa-dom.h.
259 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
261 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
262 Use $(COMPILE) and $(POSTCOMPILE).
264 * config/alpha/x-alpha (driver-alpha.o): Ditto.
266 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
268 * tree-flow.h: Remove some prototypes.
269 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
270 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
271 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
272 propagate_tree_value*): Move from here to...
273 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
274 propagate_tree_value*): Relocate here.
275 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
276 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
277 * gimple-fold.c: Remove gimple-fold.h from include list.
278 * tree-vrp.c: Remove gimple-fold.h from include list.
279 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
280 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
281 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
282 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
283 * sese.c: Add tree-ssa-propagate.h to include list.
285 2013-10-02 Richard Biener <rguenther@suse.de>
287 * tree-loop-distribution.c: Include tree-vectorizer.h for
289 (enum partition_kind): Remove PKIND_REDUCTION.
290 (struct partition_s): Remove has_writes member, add reduction_p member.
291 (partition_alloc): Adjust.
292 (partition_builtin_p): Likewise.
293 (partition_has_writes): Remove.
294 (partition_reduction_p): New function.
295 (partition_merge_into): Likewise.
296 (generate_code_for_partition): Commonize builtin partition
298 (rdg_cannot_recompute_vertex_p): Remove.
299 (already_processed_vertex_p): Likewise.
300 (rdg_flag_vertex): Do not set has_writes.
301 (classify_partition): Adjust.
302 (rdg_build_partitions): Do not set has_writes, treat all
303 partitions as useful.
304 (distribute_loop): Record number of library calls generated. Adjust.
305 (tree_loop_distribution): Report number of loops and library
306 calls generated as opt-info.
308 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
310 * tree-flow.h: Include new .h files. Move prototypes.
311 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
312 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
313 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
314 * tree-pretty-print.h: Add prototypes from tree-flow.h.
315 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
316 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
317 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
318 * tree.h (get_ref_base_and_extent): Move prototype out.
319 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
321 * gimple-low.h: New File. Add prototypes from tree-flow.h.
322 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
323 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
324 * tree-scalar-evolution.c: Include tree.h.
325 * sese.c: Include tree.h.
326 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
327 * dwarf2out.c: Include tree-dfa.h.
328 * tree-chrec.c: Include tree.h.
329 * tree-data-ref.c: Include tree.h.
331 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
333 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
336 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
338 * config/t-sol2 (sol2-c.o): Remove header dependencies.
339 Use $(COMPILE) and $(POSTCOMPILE).
340 (sol2-cxx.o): Likewise.
341 (sol2-stubs.o): Likewise.
343 * config/x-solaris (host-solaris.o): Likewise.
345 * config/sparc/t-sparc (sparc.o): Remove.
346 (sparc-c.o): Remove header dependencies.
347 Use $(COMPILE) and $(POSTCOMPILE).
348 * config/sparc/x-sparc: Likewise.
350 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
352 * config/arc/arc-opts.h: Add 2013 to Copyright years.
353 * config/arc/arc700.md: Likewise.
354 * config/arc/arc-modes.def: Likewise.
355 * config/arc/arc-simd.h: Likewise.
356 * config/arc/t-arc-uClibc: Likewise.
357 * config/arc/t-arc-newlib: Likewise.
359 2013-10-02 Renlin Li <renlin.li@arm.com>
361 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
363 (aarch64_expand_epilogue): Likewise.
365 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
366 Yufeng Zhang <yufeng.zhang@arm.com>
368 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
370 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
371 'base_in' represent a conversion and legal_cast_p_1 holds; set
372 'base_in' with the returned value from get_unwidened.
374 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
376 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
377 plus_constant is not used.
379 2013-10-01 Wei Mi <wmi@google.com>
381 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
382 * config/i386/i386.md: Add define_peephole2 to
383 break partial reg stall for cvtss2sd/cvtsd2ss.
385 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
387 * config/arc/arc.c (pass_arc_ifcvt::clone):
388 Update for ctxt_ -> m_ctxt change.
390 2013-10-01 Jeff Law <law@redhat.com>
392 * tree-ssa-threadupdate.c (struct redirection_data): Delete
393 outgoing_edge and intermediate_edge fields. Instead store the path.
394 (redirection_data::hash): Hash on the last edge's destination index.
395 (redirection_data::equal): Check the entire thread path.
396 (lookup_redirectio_data): Corresponding changes.
397 (create_edge_and_update_destination_phis): Likewise.
398 (thread_single_edge): Likewise.
400 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
401 Diego Novillo <dnovillo@google.com>
403 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
404 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
405 (vld32wh_insn, vld32wl_insn): Delete commented-out old
406 versions of these patterns.
408 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
409 (__builtin_arc_aligned): Likewise.
411 * config/arc/arc.md: Expand adc_0 comment stating the intended
412 purpose and why it isn't ready.
413 Replace commented out call_value_via_label_mixed with a
414 plain comment about bl_s.
416 * config/arc/arc.c (stdio.h): Don't include directly.
417 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
418 Remove if (1) condition.
419 (arc_encode_section_info): Fix comment.
421 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
423 * config/arc/arc.c (arc_conditional_register_usage):
424 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
425 Also set reg_alloc_order for DMA config regs.
427 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
428 Jeremy Bennett <jeremy.bennett@embecosm.com>
430 * doc/install.texi (--with-cpu): Mention ARC.
431 (arc-*-elf32): New paragraph.
432 (arc-linux-uclibc): Likewise.
433 * doc/md.texi (Machine Constraints): Add ARC part.
434 * doc/invoke.texi: (menu): Add ARC Options.
435 (Machine Dependent Options) <ARC Options>: Add synopsis.
436 (node ARC Options): Add.
437 * doc/extend.texi (long_call / short_call attribute): Add ARC.
438 (ARC Built-in Functions): New section defining
439 generic ARC built-in functions.
440 (ARC SIMD Built-in Functions): New section defining SIMD specific
442 (Declaring Attributes of Functions): Extended
443 description of short_call and long_call attributes for ARC and
446 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
447 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
448 Joern Rennecke <joern.rennecke@embecosm.com>
449 Muhammad Khurram Riaz <khurram.riaz@arc.com>
450 Brendan Kehoe <brendan@zen.org>
451 Michael Eager <eager@eagercon.com>
452 Simon Cook <simon.cook@embecosm.com>
453 Jeremy Bennett <jeremy.bennett@embecosm.com>
455 * config/arc, common/config/arc: New directories.
457 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
458 Brendan Kehoe <brendan@zen.org>
459 Simon Cook <simon.cook@embecosm.com>
461 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
463 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
465 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
466 * tree-ssa-coalesce.h: New. Move prototype to here.
467 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
468 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
469 (gimple_can_coalesce_p): Move to...
470 * gimple.c (gimple_can_coalesce_p): Here.
472 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
474 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
475 (dump_decl_set): Move to gimple.c.
476 * gimple.h: Don't include tree-ssa-operands.h.
477 (dump_decl_set): Add prototype.
478 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
479 Move to gimple-ssa.h.
480 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
481 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
482 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
483 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
484 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
486 (dump_decl_set): Relocate here.
487 * gimple-ssa.h: New file.
488 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
489 Relocate from gimple.h.
490 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
491 * tree-ssa-operands.c (swap_ssa_operands): Rename from
492 swap_tree_operands and remove non-ssa path.
493 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
494 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
496 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
497 vect_is_simple_reduction_1): Use swap_ssa_operands.
498 * tree-flow.h: Move various prototypes to tree-phinodes.h.
499 (enum need_phi_state): Move to tree-into-ssa.c.
500 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
501 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
502 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
503 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
504 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
505 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
506 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
507 num_imm_uses): Move to ssa-iterators.h.
508 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
509 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
511 (op_iter_done, op_iter_next_def, op_iter_next_tree,
512 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
513 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
514 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
515 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
516 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
517 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
518 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
519 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
520 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
521 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
522 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
523 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
524 (set_phi_nodes): Move to tree-phinodes.h.
525 * tree-ssa-operands.h (enum ssa_op_iter_type,
526 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
527 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
528 (dump_decl_set): Remove prototype.
529 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
530 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
531 (set_phi_nodes): Relocate from tree-flow-inline.h.
532 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
534 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
535 include list. Temporarily add gimple.h to include list.
536 * ssa-iterators.h: New file.
537 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
538 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
539 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
540 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
541 Relocate from tree-ssa-operands.h.
542 (delink_imm_use, link_imm_use_to_list, link_imm_use,
543 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
544 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
545 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
546 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
547 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
548 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
549 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
550 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
551 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
552 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
553 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
554 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
555 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
556 Relocate from tree-flow-inline.h.
557 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
559 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
562 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
563 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
564 Insert '\t' to output template.
565 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
566 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
567 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
568 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
570 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
572 * doc/install.texi (Host/target specific installation notes for GCC):
573 Put @anchor before @heading.
574 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
575 Use @email for email addresses.
577 2013-10-01 Jeff Law <law@redhat.com>
579 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
580 a vec. Only delete the path if we create one without successfully
581 registering a jump thread.
582 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
584 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
585 (paths): New vector of jump threading paths.
586 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
587 (THREAD_PATH): New accessor macro for the entire thread path.
588 (lookup_redirection_data): Get intermediate and final outgoing edge
589 from the thread path.
590 (create_edge_and_update_destination_phis): Copy the threading path.
591 (ssa_fix_duplicate_block_edges): Get edges and block types from the
593 (ssa_redirect_edges): Get edges and block types from the jump threading
594 path. Free the path vector.
595 (thread_block): Get edges from the jump threading path. Look at the
596 entire path to see if we thread to a loop exit. If we cancel a jump
597 thread request, then free the path vector.
598 (thread_single_edge): Get edges and block types from the jump threading
599 path. Free the path vector.
600 (thread_through_loop_header): Get edges and block types from the jump
601 threading path. Free the path vector.
602 (mark_threaded_blocks): Iterate over the vector of paths and store
603 the path on the appropriate edge. Get edges and block types from the
605 (mark_threaded_blocks): Get edges and block types from the jump
606 threading path. Free the path vector.
607 (thread_through_all_blocks): Use the vector of paths rather than
608 a vector of 3-edge sets.
609 (register_jump_thread): Accept pointer to a path vector rather
610 than the path vector itself. Store the path vector for later use.
613 2013-10-01 Jakub Jelinek <jakub@redhat.com>
614 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
617 * config/s390/s390.c (s390_split_branches): Modify check for table
619 (s390_chunkify_start): Rearrange table jump insn check in order to
620 deal with compare and branch insns correctly.
622 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
626 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
627 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
628 define_insn_and_split.
629 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
630 (shiftsi3_compare): New pattern.
632 * config/arm/unspecs.md (UNSPEC_RRX): New.
634 2013-10-01 Alan Modra <amodra@gmail.com>
636 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
637 casing inout operands.
639 2013-10-01 Richard Biener <rguenther@suse.de>
641 PR tree-optimization/58553
642 * tree-loop-distribution.c (struct partition_s): Add niter member.
643 (classify_partition): Populate niter member for the partition
644 and properly identify whether the relevant store happens before
645 or after the loop exit.
646 (generate_memset_builtin): Use niter member from the partition.
647 (generate_memcpy_builtin): Likewise.
649 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
651 * vec.h (vec_prefix, vec): Prefix member names with "m_".
652 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
654 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
656 * basic-block.h (edge_list): Prefix member names with "m_".
657 * context.h (context): Likewise.
658 * domwalk.h (dom_walker): Likewise.
659 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
660 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
661 * hash-table.h (hash_table): Likewise.
662 * machmode.h (bit_field_mode_iterator): Likewise.
663 * pass_manager.h (pass_list): Likewise.
664 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
665 * tree-pass.h (pass_data): Likewise.
666 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
667 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
668 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
669 * asan.c (pass_data_asan): Update accordingly.
670 * cfganal.c (control_dependences::find_control_dependence): Likewise.
671 (control_dependences::control_dependences): Likewise.
672 (control_dependences::~control_dependences): Likewise.
673 (control_dependences::~control_dependences): Likewise.
674 (control_dependences::get_edges_dependent_on): Likewise.
675 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
676 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
677 * context.c (gcc::context::context): Likewise.
678 * cprop.c (pass_rtl_cprop::clone): Likewise.
679 * domwalk.c (dom_walker::walk): Likewise.
680 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
681 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
682 * mode-switching.c (pass_mode_switching::clone): Likewise.
683 * passes.c (opt_pass::opt_pass): Likewise.
684 (pass_manager::pass_manager): Likewise.
685 * predict.c (pass_strip_predict_hints::clone): Likewise.
686 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
687 (pass_split_all_insns::clone): Likewise.
688 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
690 (bit_field_mode_iterator::next_mode): Likewise.
691 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
692 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
693 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
694 * tree-complex.c (pass_lower_complex::clone): Likewise.
695 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
696 * tree-object-size.c (pass_object_sizes::clone): Likewise.
697 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
698 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
699 (pass_fold_builtins::clone): Likewise.
700 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
701 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
702 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
703 (pass_cd_dce::clone): Likewise.
704 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
705 (pass_phi_only_cprop::clone): Likewise.
706 * tree-ssa-dse.c (pass_dse::clone): Likewise.
707 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
708 * tree-ssa-loop.c (pass_lim::clone): Likewise.
709 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
710 * tree-ssa-pre.c (pass_fre::clone): Likewise.
711 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
712 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
713 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
714 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
715 * tree-vrp.c (pass_vrp::clone): Likewise.
716 * tsan.c (pass_tsan::clone): Likewise.
718 2013-09-30 Jakub Jelinek <jakub@redhat.com>
721 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
722 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
725 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
726 optimization, punt if sign_bit_p looked through any zero extension.
728 2013-09-30 Teresa Johnson <tejohnson@google.com>
730 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
731 Update redirected out edge count in joiner case.
732 (ssa_redirect_edges): Common the joiner and non-joiner cases
733 so that joiner case gets profile updates.
735 2013-09-30 Richard Biener <rguenther@suse.de>
737 PR tree-optimization/58554
738 * tree-loop-distribution.c (classify_partition): Require
739 unconditionally executed stores for memcpy and memset recognition.
740 (tree_loop_distribution): Calculate dominance info.
742 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
744 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
745 (NO_PROFILE_COUNTERS): Likewise.
746 (PROFILE_HOOK): Likewise.
747 (FUNCTION_PROFILER): Likewise.
748 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
750 2013-09-30 Iain Sandoe <iain@codesourcery.com>
752 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
753 calls and defines in TARGET_MACHO conditional.
754 (load_macho_picbase_di): Likewise.
755 (reload_macho_picbase): Likewise.
756 (reload_macho_picbase_si): Likewise.
757 (reload_macho_picbase_di): Likewise.
758 (nonlocal_goto_receiver): Likewise.
760 2013-09-30 Nick Clifton <nickc@redhat.com>
762 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
763 that use the MSP430X ISA.
764 (msp430_option_override): Scan -mmcu command line option for any
765 MCU name that supports the MSP430X ISA.
766 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
767 -mmcu options which enable the MSP430X ISA.
769 2013-09-30 Richard Biener <rguenther@suse.de>
772 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
773 before looking for setjmp-like calls.
775 2013-09-29 Iain Sandoe <iain@codesourcery.com>
778 * config/darwin-protos.h (machopic_get_function_picbase): New.
779 * config/darwin.c (machopic_get_function_picbase): New.
780 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
781 label for a new func. (load_macho_picbase_di): Likewise.
782 (reload_macho_picbase): New expand.
783 (reload_macho_picbase_si): New insn.
784 (reload_macho_picbase_di): New insn.
785 (nonlocal_goto_receiver): New define and split.
786 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
787 (unspecv enum): Add UNSPECV_NLGR.
789 2013-09-29 Iain Sandoe <iain@codesourcery.com>
791 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
792 that altivec registers are correctly sized on Darwin.
794 2013-09-29 Iain Sandoe <iain@codesourcery.com>
796 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
797 darwin-driver.o): Use COMPILE and POSTCOMPILE.
798 * config/x-darwin (host-darwin.o): Likewise.
799 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
800 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
801 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
803 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
805 * doc/invoke.texi: Fix usage of @tie{} command.
807 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
809 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
811 2013-09-28 Jan Hubicka <jh@suse.cz>
813 * config/i386/x86-tune.def: Add documentation for each of the options;
816 2013-09-28 Jan Hubicka <jh@suse.cz>
818 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
820 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
821 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
822 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
823 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
824 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
826 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
828 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
829 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
830 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
831 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
832 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
833 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
834 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
835 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
836 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
837 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
838 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
839 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
840 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
841 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
842 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
843 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
844 gimple.h, godump.c, graphite-clast-to-gimple.c,
845 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
846 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
847 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
848 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
849 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
850 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
851 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
852 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
853 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
854 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
855 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
856 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
857 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
858 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
859 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
860 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
861 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
862 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
863 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
864 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
865 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
866 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
867 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
868 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
869 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
870 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
871 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
872 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
873 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
874 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
875 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
876 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
877 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
878 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
879 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
880 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
881 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
882 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
883 whitespace before "(".
885 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
887 * expr.h (extract_bit_field): Remove packedp parameter.
888 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
889 from forward declaration.
890 (store_split_bit_field): Remove packedp arg from calls to
891 extract_fixed_bit_field.
892 (extract_bit_field_1): Remove packedp parameter and packedp
893 argument from recursive calls and calls to extract_fixed_bit_field.
894 (extract_bit_field): Remove packedp parameter and corresponding
895 arg to extract_bit_field_1.
896 (extract_fixed_bit_field): Remove packedp parameter. Remove code
898 (extract_split_bit_field): Remove packedp arg from call to
899 extract_fixed_bit_field.
900 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
901 (copy_blkmode_from_reg): Likewise.
902 (copy_blkmode_to_reg): Likewise.
903 (read_complex_part): Likewise.
904 (store_field): Likewise.
905 (expand_expr_real_1): Likewise.
906 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
907 to extract_bit_field.
908 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
909 call to extract_bit_field.
910 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
911 call to extract_bit_field.
912 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
913 and special packedp behavior from -fstrict-volatile-bitfields
916 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
918 * lra-eliminations.c (init_elim_table): Guard value_p.
920 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
922 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
923 DFmode, DImode, and SFmode in the upper VSX registers based on the
924 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
925 if -mpower8-vector. Combine -mvsx-timode handling with the rest
926 of the VSX register handling.
928 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
930 (zero_extendsidi2_lfiwzx): Add support for loading into the
931 Altivec registers with -mpower8-vector. Use wu/wv constraints to
932 only do VSX memory options on Altivec registers.
933 (extendsidi2_lfiwax): Likewise.
934 (extendsfdf2_fpr): Likewise.
935 (mov<mode>_hardfloat, SF/SD modes): Likewise.
936 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
937 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
938 (movdi_internal64): Likewise.
940 2013-09-27 Xinliang David Li <davidxl@google.com>
942 * opts.c (finish_options): Adjust parameters
943 according to vect cost model.
944 (common_handle_option): Set dynamic vect cost
946 targhooks.c (default_add_stmt_cost): Compute stmt cost
948 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
950 * tree-vectorizer.h (unlimited_cost_model): New function.
951 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
952 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
954 (vect_enhance_data_refs_alignment): Ditto.
955 * flag-types.h: New enum.
956 * common/config/i386/i386-common.c (ix86_option_init_struct):
957 No need to initialize vect_cost_model flag.
958 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
961 2013-09-27 Diego Novillo <dnovillo@google.com>
963 * gimple.h (enum ssa_mode): Remove.
965 2013-09-27 Paulo Matos <pmatos@broadcom.com>
968 * gcc.dg/pr58463.c: New test.
970 2013-09-27 Paulo Matos <pmatos@broadcom.com>
972 * cfgloop.h (number_of_loops): Fix typo in check for null.
974 2013-09-27 Jakub Jelinek <jakub@redhat.com>
977 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
978 are children of outermost saved_cfun's loop, and set it up to
979 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
980 if loop != loop0 and SESE region contains bbs that belong to loop0.
982 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
984 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
985 (binary_scale_code_p, get_base_term, get_index_term): New functions.
986 (set_address_segment, set_address_base, set_address_index)
987 (set_address_disp): Accept the argument unconditionally.
988 (baseness): Remove must_be_base_p and must_be_index_p checks.
989 (decompose_normal_address): Classify as much as possible in the
992 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
994 * cse.c (count_reg_usage): Handle INT_LIST.
995 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
996 * reginfo.c (reg_scan_mark_refs): Likewise.
997 * reload1.c (eliminate_regs_1): Likewise.
999 2013-09-27 Iain Sandoe <iain@codesourcery.com>
1002 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
1005 2013-09-27 Richard Biener <rguenther@suse.de>
1007 PR tree-optimization/58459
1008 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
1009 restriction not propagating into loops.
1011 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
1013 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
1014 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
1015 * doc/tree-ssa.texi (Walking use-def chains): Delete.
1017 2013-09-26 Richard Biener <rguenther@suse.de>
1019 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
1021 2013-09-26 Richard Biener <rguenther@suse.de>
1023 * alias.h (component_uses_parent_alias_set): Rename to ...
1024 (component_uses_parent_alias_set_from): ... this.
1025 * alias.c (component_uses_parent_alias_set): Rename to ...
1026 (component_uses_parent_alias_set_from): ... this and return
1028 (reference_alias_ptr_type_1): Use the result from
1029 component_uses_parent_alias_set_from instead of stripping
1030 components one at a time.
1031 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
1033 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1035 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
1036 Move prototypes to...
1037 * tree-ssa-ter.h: New File. Move prototypes here.
1038 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
1039 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
1040 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
1041 from is_replaceable_p.
1042 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
1043 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
1044 (process_replaceable): Use ter_is_replaceable_p.
1045 (find_replaceable_in_bb): Use ter_is_replaceable_p.
1046 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
1047 newly refactored ssa_is_replaceable_p.
1048 * cfgexpand.c: Include tree-outof-ssa.h.
1049 * ssaexpand.h: Delete.
1051 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1053 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
1054 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
1056 (create_gimple_tmp): Delete.
1057 (get_expr_type, build_assign, build_type_cast): Move to...
1058 * gimple-builder.c: New File.
1059 (get_expr_type): Relocate from gimple.c.
1060 (build_assign, build_type_cast): Change to only create ssanames.
1061 * gimple.h: Move prototypes to...
1062 * gimple-builder.h: New File. Here.
1063 * tree-ssa.h: And here.
1064 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
1065 count_uses_and_derefs): Relocate from gimple.c.
1066 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
1067 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
1068 * tree-ssa-math-opts (execute_cse_reciprocals): Use
1069 gimple_replace_ssa_lhs.
1070 * asan.c: Include gimple-builder.h.
1071 * Makefile.in: Add gimple-builder.o.
1073 2013-09-26 Richard Biener <rguenther@suse.de>
1075 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
1076 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
1077 (loe_visit_block): Use gcc_checking_assert.
1078 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
1079 gimple_assign_ssa_name_copy_p.
1080 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
1083 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
1085 * config/rs6000/t-rs6000 (rs6000.o): Remove.
1086 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
1088 2013-09-26 Richard Biener <rguenther@suse.de>
1090 PR tree-optimization/58539
1091 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
1092 the fact that debug statements are not taking part in loop-closed
1095 2013-09-26 Nick Clifton <nickc@redhat.com>
1097 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
1098 time warning message.
1099 (msp430_print_operand_raw): Delete unused letter parameter.
1100 (TARGET_PRINT_OPERAND_ADDRESS): Define.
1101 (msp430_print_operand_address): New function.
1102 (msp430_print_operand): Move address printing code from here to
1104 * config/msp430/msp430.md (movsipsi2): Add comment in generated
1106 (zero_extendpsisi2): Likewise.
1107 (extendpsisi2): New pattern.
1108 (andneghi3): New pattern.
1110 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
1112 * config/aarch64/aarch64.opt (mlra): New option.
1113 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
1114 (TARGET_LRA_P): Define.
1116 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
1118 * expr.c (expand_assignment): Remove obsolete comment.
1120 2013-09-25 Jeff Law <law@redhat.com>
1122 * tree-flow.h (thread_through_all_blocks): Prototype moved into
1123 tree-ssa-threadupdate.h.
1124 (register_jump_thread): Similarly.
1125 * tree-ssa-threadupdate.h: New header file.
1126 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
1127 * tree-vrp.c: Likewise.
1128 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
1129 (thread_around_empty_blocks): Change type of path vector argument to
1130 an edge,type pair from just an edge. Initialize both elements when
1131 appending to a jump threading path. Tweak references to elements
1133 (thread_across_edge): Similarly. Release memory for the elements
1135 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
1136 (dump_jump_thread_path): New function broken out from
1137 register_jump_thread.
1138 (register_jump_thread): Use dump_jump_thread_path. Change type of
1139 path vector entries. Search the path for NULL edges and dump
1140 the path if one is found. Tweak the conversion of path to 3-edge
1141 form to use the block copy type information embedded in the path.
1143 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1145 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
1147 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1148 Vladimir Makarov <vmakarov@redhat.com>
1150 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
1151 from the least significant bit.
1152 (strip_address_mutations): Add bitfield operations handling.
1153 (must_be_index_p): Add shifting and rotate operations handling.
1154 (set_address_base): Use must_be_base_p predicate.
1155 (set_address_index): Use must_be_index_p predicate.
1157 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
1158 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1159 Sergey Lega <sergey.s.lega@intel.com>
1160 Anna Tikhonova <anna.tikhonova@intel.com>
1161 Ilya Tocar <ilya.tocar@intel.com>
1162 Andrey Turetskiy <andrey.turetskiy@intel.com>
1163 Ilya Verbin <ilya.verbin@intel.com>
1164 Kirill Yukhin <kirill.yukhin@intel.com>
1165 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1167 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
1169 (ix86_expand_vector_move_misalign): Support new unaligned load and
1170 stores and use new names.
1171 (CODE_FOR_sse2_storedqu): Rename to ...
1172 (CODE_FOR_sse2_storedquv16qi): ... this.
1173 (CODE_FOR_sse2_loaddqu): Rename to ...
1174 (CODE_FOR_sse2_loaddquv16qi): ... this.
1175 (CODE_FOR_avx_loaddqu256): Rename to ...
1176 (CODE_FOR_avx_loaddquv32qi): ... this.
1177 (CODE_FOR_avx_storedqu256): Rename to ...
1178 (CODE_FOR_avx_storedquv32qi): ... this.
1179 * config/i386/i386.md (fpint_logic): New.
1180 * config/i386/sse.md (VMOVE): Extend for AVX512.
1184 (VI_UNALIGNED_LOADSTORE): Ditto.
1185 (sse2_avx_avx512f): Ditto.
1186 (sse2_avx2): Extend for AVX512.
1187 (sse4_1_avx2): Ditto.
1188 (avx2_avx512f): New.
1189 (sse): Extend for AVX512.
1192 (avxsizesuffix): Ditto.
1193 (sseintvecmode): Ditto.
1195 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
1196 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
1197 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1198 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
1199 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1200 (<sse2_avx_avx512f>_storedqu<mode): ... this.
1201 (<sse>_movnt<mode>): Replace constraint "x" with "v".
1202 (STORENT_MODE): Extend for AVX512.
1203 (*absneg<mode>2): Replace constraint "x" with "v".
1204 (*mul<mode>3): Ditto.
1205 (*ieee_smin<mode>3): Ditto.
1206 (*ieee_smax<mode>3): Ditto.
1207 (avx_cmp<mode>3): Replace VF with VF_128_256.
1208 (*<sse>_maskcmp<mode>3_comm): Ditto.
1209 (<sse>_maskcmp<mode>3): Ditto.
1210 (<sse>_andnot<mode>3): Extend for AVX512.
1211 (<code><mode>3, anylogic): Replace VF with VF_128_256.
1212 (<code><mode>3, fpint_logic): New.
1213 (*<code><mode>3): Extend for AVX512.
1214 (avx512flogicsuff): New.
1215 (avx512f_<logic><mode>): Ditto.
1216 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
1218 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
1219 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
1220 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
1221 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
1222 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
1223 (xop_vpermil2<mode>3): Ditto.
1224 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
1225 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
1226 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
1227 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
1229 2013-09-25 Tom Tromey <tromey@redhat.com>
1231 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
1232 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
1233 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
1234 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
1235 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
1236 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
1237 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
1238 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
1239 (GRAPHITE_HTAB_H): Remove.
1241 2013-09-25 Tom Tromey <tromey@redhat.com>
1243 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
1245 2013-09-25 Tom Tromey <tromey@redhat.com>
1247 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
1249 2013-09-25 Tom Tromey <tromey@redhat.com>
1251 * config/i386/t-i386 (i386.o): Remove.
1252 (i386-c.o): Use COMPILE and POSTCOMPILE.
1254 2013-09-25 Tom Tromey <tromey@redhat.com>
1256 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
1258 2013-09-25 Tom Tromey <tromey@redhat.com>
1260 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
1261 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
1262 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
1263 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
1264 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
1265 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
1266 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
1267 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
1268 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
1269 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
1270 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
1271 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
1272 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
1273 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
1274 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
1275 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
1276 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
1277 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
1278 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
1279 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
1280 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
1281 (tree-ssa-pre.o, tree-ssa-sccvn.o)
1282 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
1283 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
1284 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
1285 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
1286 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
1287 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
1288 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
1289 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
1290 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
1291 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
1292 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
1293 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
1294 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
1295 (graphite-clast-to-gimple.o, graphite-dependences.o)
1296 (graphite-interchange.o, graphite-poly.o)
1297 (graphite-scop-detection.o, graphite-sese-to-poly.o)
1298 (graphite-optimize-isl.o, tree-vect-loop.o)
1299 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
1300 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
1301 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
1302 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
1303 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
1304 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
1305 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
1306 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
1307 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
1308 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
1309 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
1310 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
1311 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
1312 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
1313 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
1314 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
1315 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
1316 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
1317 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
1318 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
1319 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
1320 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
1321 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
1322 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
1323 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
1324 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
1325 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
1326 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
1327 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
1328 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
1329 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
1330 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
1331 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
1332 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
1333 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
1334 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
1335 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
1336 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
1337 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
1338 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
1339 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
1340 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
1341 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
1342 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
1343 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
1344 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
1345 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
1346 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
1347 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
1348 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
1349 (gcov-dump.o): Remove.
1350 (default-c.o): Use COMPILE and POSTCOMPILE.
1351 (CFLAGS-gcc.o): New variable.
1352 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
1354 2013-09-25 Tom Tromey <tromey@redhat.com>
1356 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
1357 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
1358 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
1359 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
1360 (c-family/c-pragma.o, c-family/c-pretty-print.o)
1361 (c-family/c-semantics.o, c-family/c-ada-spec.o)
1362 (c-family/array-notation-common.o, c-family/stub-objc.o)
1363 (c-family/c-ubsan.o): Remove.
1365 2013-09-25 Tom Tromey <tromey@redhat.com>
1367 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
1369 2013-09-25 Tom Tromey <tromey@redhat.com>
1371 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
1372 to add -DENABLE_SHARED_LIBGCC.
1373 (gcc.o): Don't use subshell.
1375 2013-09-25 Tom Tromey <tromey@redhat.com>
1377 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
1378 * configure.ac: Don't invoke AM_PROG_CC_C_O.
1379 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
1380 * configure, config.in: Rebuild.
1382 2013-09-25 Tom Tromey <tromey@redhat.com>
1384 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
1385 (COMPILE, POSTCOMPILE): New variables.
1386 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
1387 (DEPFILES): New variable.
1388 Include ".Po" files.
1389 * configure.ac: Add checks for dependency checking.
1390 * configure, aclocal.m4: Regenerate.
1392 2013-09-25 Tom Tromey <tromey@redhat.com>
1394 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
1395 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
1397 2013-09-25 Tom Tromey <tromey@redhat.com>
1399 * Makefile.in (generated_files): Add options.h,
1400 target-hooks-def.h, insn-opinit.h,
1401 common/common-target-hooks-def.h, pass-instances.def,
1402 c-family/c-target-hooks-def.h.
1404 2013-09-25 Jeff Law <law@redhat.com>
1406 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
1407 than foo[foo.length () - 1] to access last member in a vec.
1408 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
1410 2013-09-25 Richard Biener <rguenther@suse.de>
1413 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
1415 2013-09-25 Jan Hubicka <jh@suse.cz>
1417 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
1420 2013-09-25 Marek Polacek <polacek@redhat.com>
1423 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
1424 when determining the type name.
1426 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
1428 * config/sh/sh.md: Fix formatting.
1430 2013-09-24 Xinliang David Li <davidxl@google.com>
1432 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
1433 max peel iterations parameter.
1434 * param.def: New parameter.
1435 * doc/invoke.texi: Document New parameter.
1437 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
1439 * gimple-pretty-print.c: Various whitespace tweaks.
1440 * tree-core.h: Likewise.
1441 * tree-pretty-print.c: Likewise.
1442 * tree-ssa-alias.c: Likewise.
1443 * tree-ssa-copy.c: Likewise.
1444 * tree-ssanames.c: Likewise.
1445 * tree-ssanames.h: Likewise.
1446 * tree-vrp.c: Likewise.
1448 2013-09-24 Alan Modra <amodra@gmail.com>
1452 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
1453 for output operands that disallow regs. Don't use EXPAND_WRITE on
1456 2013-09-24 Richard Biener <rguenther@suse.de>
1459 * tree.c (reference_alias_ptr_type): Move ...
1460 * alias.c (reference_alias_ptr_type): ... here and implement
1461 in terms of the new reference_alias_ptr_type_1.
1462 (ref_all_alias_ptr_type_p): New helper.
1463 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
1464 use ref_all_alias_ptr_type_p.
1465 (get_deref_alias_set): Add flag_strict_aliasing check here.
1466 (reference_alias_ptr_type_1): New function, split out from ...
1467 (get_alias_set): ... here.
1468 (alias_ptr_types_compatible_p): New function.
1469 * alias.h (reference_alias_ptr_type): Declare.
1470 (alias_ptr_types_compatible_p): Likewise.
1471 * tree.h (reference_alias_ptr_type): Remove.
1472 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
1473 to compare MEM_REF alias types.
1475 2013-09-24 Richard Biener <rguenther@suse.de>
1477 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
1479 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1481 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
1482 reload helper function arrays into a single array reg_addr.
1483 (reload_fpr_gpr): Likewise.
1484 (reload_gpr_vsx): Likewise.
1485 (reload_vsx_gpr): Likewise.
1486 (struct rs6000_reg_addr): Likewise.
1487 (reg_addr): Likewise.
1488 (rs6000_debug_reg_global): Change rs6000_vector_reload,
1489 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
1490 (rs6000_init_hard_regno_mode_ok): Likewise.
1491 (rs6000_secondary_reload_direct_move): Likewise.
1492 (rs6000_secondary_reload): Likewise.
1494 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
1495 constraints: wu, ww, and wy. Repurpose wv constraint added during
1496 power8 changes. Put wg constraint in alphabetical order.
1498 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
1499 for future work to add ISA 2.07 VSX single precision support.
1500 (-mvsx-scalar-double): Change default from -1 to 1, update
1501 documentation comment.
1502 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
1503 (-mupper-regs-df): New debug switch to control whether DF values
1504 can go in the traditional Altivec registers.
1505 (-mupper-regs-sf): New debug switch to control whether SF values
1506 can go in the traditional Altivec registers.
1508 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
1510 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
1511 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
1512 Add new constraints, wu/ww/wy. Repurpose wv constraint.
1513 (rs6000_debug_legitimate_address_p): Print if we are running
1514 before, during, or after reload.
1515 (rs6000_secondary_reload): Add a comment.
1516 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
1518 * config/rs6000/constraints.md (wa constraint): Sort w<x>
1519 constraints. Update documentation string.
1520 (wd constraint): Likewise.
1521 (wf constraint): Likewise.
1522 (wg constraint): Likewise.
1523 (wn constraint): Likewise.
1524 (ws constraint): Likewise.
1525 (wt constraint): Likewise.
1526 (wx constraint): Likewise.
1527 (wz constraint): Likewise.
1528 (wu constraint): New constraint for ISA 2.07 SFmode scalar
1530 (ww constraint): Likewise.
1531 (wy constraint): Likewise.
1532 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
1533 the previous submissions.
1534 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
1536 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
1538 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
1539 (REG_BR_PROB): Say that the probability is stored in an int_list.
1540 * reg-notes.def: Update commentary to mention INT_LIST.
1541 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
1542 (INT_LIST): New rtx.
1543 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
1544 * rtlanal.c (int_reg_note_p): New function.
1545 (alloc_reg_note): Assert that the note does not have an int argument.
1546 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
1547 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
1548 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
1549 rather than an INSN_LIST. Handle INT_LIST.
1550 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
1551 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
1552 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
1553 Manipulate them as ints rather than rtxes.
1554 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
1555 * regmove.c (copy_src_to_dest): Likewise.
1556 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
1558 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
1559 into the cases that need it.
1560 * config/arm/arm.c (arm_unwind_emit): Likewise.
1562 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
1563 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
1564 * loop-doloop.c (add_test, doloop_modify): Likewise.
1565 * loop-unswitch.c (compare_and_jump_seq): Likewise.
1566 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
1567 * predict.c (combine_predictions_for_insn): Likewise.
1568 * print-rtl.c (print_rtx): Handle INT_LIST.
1569 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
1570 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
1571 * config/arm/arm.c (emit_unlikely_jump): Likewise.
1572 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
1573 (ix86_split_fp_branch, predict_jump): Likewise.
1574 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
1575 * config/sh/sh.c (expand_cbranchsi4): Likewise.
1576 * config/spu/spu.c (ea_load_store_inline): Likewise.
1578 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
1579 value of a REG_BR_PROB note.
1580 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
1581 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
1582 * emit-rtl.c (try_split): Likewise.
1583 * predict.c (br_prob_note_reliable_p): Likewise.
1584 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
1585 * reorg.c (mostly_true_jump): Likewise.
1586 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
1587 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
1588 * config/i386/i386.c (ix86_print_operand): Likewise.
1589 * config/ia64/ia64.c (ia64_print_operand): Likewise.
1590 * config/mmix/mmix.c (mmix_print_operand): Likewise.
1591 * config/rs6000/rs6000.c (output_cbranch): Likewise.
1592 * config/s390/s390.c (s390_expand_tbegin): Likewise.
1593 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
1594 * config/sparc/sparc.c (output_cbranch): Likewise.
1595 * config/spu/spu.c (get_branch_target): Likewise.
1596 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
1597 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
1599 2013-09-23 Jan Hubicka <jh@suse.cz>
1601 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
1603 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
1604 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
1606 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
1607 (gimple_fold_call): Dump inconsistent devirtualizations; add
1608 sanity check for type based devirtualizations.
1609 * ipa-prop.c: Include ipa-utils.h
1610 (ipa_intraprocedural_devirtualization): Add sanity check.
1611 (try_make_edge_direct_virtual_call): Likewise.
1613 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
1615 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
1616 assignment statements.
1618 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1620 * gimple-pretty-print.c (dump_ssaname_info): New function.
1621 (dump_gimple_phi): Call it.
1622 (pp_gimple_stmt_1): Likewise.
1623 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
1624 (range_info_def): Declare.
1625 * tree-pretty-print.c (pp_double_int): New function.
1626 (dump_generic_node): Call it.
1627 * tree-pretty-print.h (pp_double_int): Declare.
1628 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
1629 * tree-ssanames.h (range_info_def): New structure.
1630 (value_range_type): Move definition here.
1631 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
1633 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
1635 (set_range_info): New function.
1636 (get_range_info): Likewise.
1637 (duplicate_ssa_name_range_info): Likewise.
1638 (duplicate_ssa_name_fn): Check pointer type and call
1639 duplicate_ssa_name_range_info.
1640 * tree-ssa-copy.c (fini_copy_prop): Likewise.
1641 * tree-vrp.c (value_range_type): Remove definition, now in
1643 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
1644 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
1645 (SSA_NAME_RANGE_INFO): New macro.
1647 2013-09-23 Richard Biener <rguenther@suse.de>
1649 PR tree-optimization/58464
1650 * tree-ssa-pre.c (phi_trans_lookup): Remove.
1651 (phi_trans_add): Change to add conditionally on being not present.
1652 (phi_translate_1): Remove recursion detection here.
1653 (phi_translate): Pre-seed the cache with NULL to catch
1654 recursion here in a more generic way.
1655 (bitmap_find_leader): Adjust comment.
1656 (get_representative_for): Dump value-numbers.
1657 (create_expression_by_pieces): Likewise.
1658 (insert_into_preds_of_block): Likewise.
1660 2013-09-23 Christian Bruel <christian.bruel@st.com>
1663 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
1664 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
1666 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
1669 2013-09-20 Renlin Li <renlin.li@arm.com>
1671 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1673 (aarch64_expand_epilogue): Likewise.
1674 (aarch64_legitimize_reload_address): Likewise.
1676 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
1678 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
1679 NULL_TREE before pushing them onto the vector. Likewise for labels.
1681 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
1683 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
1684 comparison operators when -fno-trapping-math is in effect.
1685 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
1686 comparison operators in TFmode and assert that unsupported operators
1688 (ia64_print_operand): Likewise.
1690 2013-09-21 Jan Hubicka <jh@suse.cz>
1692 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
1693 (sse_typeless_stores): Enable for core
1694 (sse_load0_by_pxor): Likewise.
1695 (four_jump_limit): Disable for core.
1696 (pad_returns): Likewise.
1697 (avoid_vector_decode): Likewise.
1698 (fuse_cmp_and_branch): Enable for cores.
1699 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
1701 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1704 * config/pa/pa.c (pa_option_override): Disable auto increment and
1705 decrement instructions until reload is completed.
1707 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
1708 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
1709 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
1711 2013-09-20 DJ Delorie <dj@redhat.com>
1712 Nick Clifton <nickc@redhat.com>
1714 * config/rl78/rl78.c: Various whitespace and comment tweaks.
1715 (need_to_save): Save bank 0 on interrupts.
1716 (characterize_address): Strip far address wrappers.
1717 (rl78_as_legitimate_address): Likewise.
1718 (transcode_memory_rtx): Likewise.
1719 (rl78_peep_movhi_p): Disable this peephole after devirt.
1720 (rl78_propogate_register_origins): Forget all origins when a
1722 * config/rl78/rl78-virt.md: Various whitespace tweaks.
1723 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
1725 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
1726 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
1727 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
1729 * config/rl78/constraints.md: Various whitespace and paren tweaks.
1731 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1733 * config/pa/pa.md: In "scc" insn patterns, change output template to
1734 handle const0_rtx in reg_or_0_operand operands.
1736 2013-09-20 Martin Husemann <martin@NetBSD.org>
1739 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
1740 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
1742 2013-09-20 Richard Biener <rguenther@suse.de>
1745 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
1746 remembering SSA name version and block index.
1747 (new_scev_info_str): Adjust.
1748 (hash_scev_info): Likewise. Also hash the block index.
1749 (eq_scev_info): Adjust.
1750 (find_var_scev_info): Likewise.
1751 (struct instantiate_cache_entry): Remove.
1752 (struct instantiate_cache_type): Use a htab to map name, block
1754 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1755 (get_instantiated_value_entry): Likewise.
1756 (hash_idx_scev_info, eq_idx_scev_info): New functions.
1757 (instantiate_scev_name): Adjust.
1759 2013-09-20 Jeff Law <law@redhat.com>
1761 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
1763 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
1765 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
1766 Call aarch64_simd_expand_args to update op[argc].
1768 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1770 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
1773 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1775 * gengtype.c (file_rules): Added rule for *.cc files.
1776 (get_output_file_with_visibility): Give fatal message when no
1779 2013-09-20 Renlin Li <renlin.li@arm.com>
1781 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
1782 (aarch64_expand_epilogue): Likewise.
1783 (aarch64_legitimize_reload_address): Likewise.
1785 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
1788 * expr.c (expand_assignment): Remove misalignp code path.
1790 2013-09-20 Marek Polacek <polacek@redhat.com>
1793 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
1794 TYPE_PRECISION. Add asserts.
1796 2013-09-20 Richard Biener <rguenther@suse.de>
1798 PR tree-optimization/58453
1799 * tree-loop-distribution.c (distribute_loop): Apply the cost
1800 model for -ftree-loop-distribute-patterns, too.
1802 2013-09-20 Richard Biener <rguenther@suse.de>
1805 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
1806 make type comparison less strict.
1808 2013-09-20 Alan Modra <amodra@gmail.com>
1810 * configure: Regenerate.
1811 * aclocal.m4: Regenerate.
1813 2013-09-20 Marek Polacek <polacek@redhat.com>
1816 * doc/extend.texi: Document that attribute used is meant to be used
1817 on variables with static storage duration.
1819 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1821 PR tree-optimization/58472
1822 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
1823 simd_lane_access set inv_p = false.
1824 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
1825 the simduid magic VAR_DECL.
1827 2013-09-19 Jan Hubicka <jh@suse.cz>
1829 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
1831 2013-09-17 Jeff Law <law@redhat.com>
1833 * tree-ssa-dom.c (record_temporary_equivalences): New function
1834 split out of dom_opt_dom_walker::after_dom_children.
1835 (dom_opt_dom_walker::thread_across_edge): Move common code
1836 in here from dom_opt_dom_walker::after_dom_children.
1837 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
1839 2013-09-19 Jan Hubicka <jh@suse.cz>
1841 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
1842 (TARGET_GENERIC): Use PROCESOR_GENERIC
1843 (enum processor_type): Unify generic32 and 64.
1844 * i386.md (cpu): Likewise.
1845 * x86-tune.def (use_leave): Enable for generic32.
1846 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
1847 * athlon.md: Change generic64 to generic in all occurences.
1848 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
1849 (ix86_target_macros_internal): Likewise.
1850 * driver-i386.c (host_detect_local_cpu): Likewise.
1851 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
1853 (generic_memcpy, generic_memset, generic_cost): This one.
1854 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
1855 (m_GENERIC32, m_GENERIC64): Remove.
1856 (m_GENERIC): Turn into one flag.
1857 (processor_target): Unify generic tunnings.
1858 (ix86_option_override_internal): Replace generic32/64 by generic.
1859 (ix86_issue_rate): Likewise.
1860 (ix86_adjust_cost): Likewise.
1862 2013-09-19 Jan Hubicka <jh@suse.cz>
1864 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
1865 of speculative flag.
1867 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1869 * omp-low.c (expand_omp_sections): Always pass len - 1 to
1870 GOMP_sections_start, even if !exit_reachable.
1872 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1874 * lra-constraints.c (need_for_all_save_p): Use macro
1875 HARD_REGNO_CALL_PART_CLOBBERED.
1876 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
1877 set up pseudo conflict hard regs.
1879 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1882 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
1885 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1887 PR rtl-optimization/58438
1888 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
1889 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
1890 from previous subpasses.
1892 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
1894 * arm.c (arm_get_frame_offsets): Validate architecture supports
1895 LDRD/STRD before accepting the tuning preference.
1896 (arm_expand_prologue): Likewise.
1897 (arm_expand_epilogue): Likewise.
1899 2013-09-18 Richard Biener <rguenther@suse.de>
1901 PR tree-optimization/58417
1902 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
1903 have chrecs with symbols defined in the loop as operands.
1904 (chrec_fold_multiply): Likewise.
1905 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
1906 parameters before folding binary operations.
1907 (struct instantiate_cache_entry_hasher): Remove.
1908 (struct instantiate_cache_type): Use a pointer-map.
1909 (instantiate_cache_type::instantiate_cache_type): New function.
1910 (instantiate_cache_type::get): Likewise.
1911 (instantiate_cache_type::set): Likewise.
1912 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1913 (get_instantiated_value_entry): Likewise.
1914 (global_cache): New global.
1915 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
1916 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
1917 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
1918 (instantiate_scev_name): Adjust.
1919 (instantiate_scev): Construct global instead of local cache.
1920 (resolve_mixers): Likewise.
1922 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
1923 Paolo Carlini <paolo.carlini@oracle.com>
1926 * doc/implement-cxx.texi: Fix references to the C++ standards.
1928 2013-09-18 Jakub Jelinek <jakub@redhat.com>
1930 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
1931 * tree-vect-data-refs.c (vect_analyze_data_refs): For
1932 simd_lane_access drs, update also DR_ALIGNED_TO.
1934 2013-09-18 Marek Polacek <polacek@redhat.com>
1937 * doc/extend.texi: Document no_sanitize_undefined attribute.
1938 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
1939 no_sanitize_undefined attribute.
1941 2013-09-18 Nick Clifton <nickc@redhat.com>
1943 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
1944 (ASM_DECLARE_FUNCTION_NAME): Define.
1946 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
1948 * compare-elim.c (find_comparison_dom_walker): New class
1949 (find_comparisons_in_bb): Rename to
1950 find_comparison_dom_walker::before_dom_children
1951 (find_comparisons): Adjust
1952 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
1954 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
1955 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
1956 (init_walk_dominator_tree): Remove declaration.
1957 (fini_walk_dominator_tree): Remove declaration.
1958 * fwprop.c (single_def_use_dom_walker): New class
1959 (single_def_use_enter_block): Convert to
1960 single_def_use_dom_walker::before_dom_children.
1961 (single_def_use_leave_block): Convert to
1962 single_def_use_dom_walker::after_dom_children.
1963 (build_single_def_use_links): Adjust.
1964 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1966 (find_candidates_in_block): Convert to
1967 find_candidates_dom_walker::before_dom_children.
1968 (execute_strength_reduction): Adjust.
1969 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1970 (sese_dom_walker): New class.
1971 (sese_dom_walker::sese_dom_walker): New constructor.
1972 (sese_dom_walker::~sese_dom_walker): New destructor.
1973 (build_sese_conditions_before): Convert to
1974 sese_dom_walker::before_dom_children.
1975 (build_sese_conditions_after): Convert to
1976 sese_dom_walker::after_dom_children.
1977 (build_poly_scop): Adjust
1978 * tree-into-ssa.c (rewrite_dom_walker): New class
1979 (rewrite_enter_block): Convert to
1980 rewrite_dom_walker::before_dom_children.
1981 (rewrite_leave_block): Convert to
1982 rewrite_dom_walker::after_dom_children.
1983 (rewrite_update_dom_walker): New class.
1984 (rewrite_update_enter_block): Convert to
1985 rewrite_update_dom_walker::before_dom_children.
1986 (rewrite_update_leave_block): Convert to
1987 rewrite_update_dom_walker::after_dom_children.
1988 (rewrite_blocks, rewrite_into_ssa): Adjust.
1989 (mark_def_dom_walker): New class.
1990 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1991 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1992 (mark_def_sites_blocks): Convert to
1993 mark_def_dom_walker::before_dom_children.
1994 (mark_def_site_blocks): Remove.
1995 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1996 (tree_ssa_dominator_optimize): Adjust.
1997 (dom_thread_across_edge): Convert to method
1998 dom_opt_dom_walker::thread_across_edge.
1999 (dom_opt_enter_block): Convert to member function
2000 dom_opt_dom_walker::before_dom_children.
2001 (dom_opt_leave_block): Convert to member function
2002 dom_opt_dom_walker::after_dom_children.
2003 * tree-ssa-dse.c (dse_dom_walker): New class.
2004 (dse_enter_block): Convert to member function
2005 dse_dom_walker::before_dom_children.
2006 (tree_ssa_dse): Adjust.
2007 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
2008 (determine_invariantness_stmt): Convert to method
2009 invariantness_dom_walker::before_dom_children.
2010 (determine_invariantness): Remove
2011 (move_computations_dom_walker): New class.
2012 (move_computations_stmt): Convert to method
2013 move_computations_dom_walker::before_dom_children.
2014 (move_computations, tree_ssa_lim): Adjust.
2015 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
2016 (nt_init_block): Make method
2017 notrappping_dom_walker::before_dom_children.
2018 (nt_fini_block): Make
2019 method nontrapping_dom_walker::after_dom_children.
2020 (get_non_trapping): Adjust.
2021 * tree-ssa-pre.c (eliminate_dom_walker): New class.
2022 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
2023 (eliminate_leave_block): Make method.
2024 eliminate_dom_walker::after_dom_children.
2026 * tree-ssa-strlen.c (strlen_dom_walker): New class.
2027 (strlen_enter_block): Make method
2028 strlen_dom_walker::before_dom_children.
2029 (strlen_leave_block): Make
2030 method strlen_dom_walker::after_dom_children.
2031 (tree_ssa_strlen): Adjust.
2032 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
2033 (tree_ssa_uncprop): Adjust.
2034 (uncprop_leave_block): Make method
2035 uncprop_dom_walker::after_dom_children.
2036 (uncprop_leave_block): Make method
2037 uncprop_dom_walker::before_dom_children.
2039 2013-09-18 Bin Cheng <bin.cheng@arm.com>
2041 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
2042 branch in same basic block. Check both src and dest of the move insn.
2044 2013-09-17 Nick Clifton <nickc@redhat.com>
2046 * config/rl78/rl78-real.md (bf): New pattern.
2048 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
2049 (rl78_print_operand): Do not put a # before a %B.
2050 * config/rl78/rl78.opt: Tweak doc strings.
2052 2013-09-17 DJ Delorie <dj@redhat.com>
2054 * config/rl78/constraints.md (Wcv): Allow up to $r31.
2055 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
2056 (rl78_option_override): Likewise, if -mallregs.
2057 (is_virtual_register): Likewise.
2058 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
2059 (REGNO_OK_FOR_BASE_P): Likewise.
2060 * config/rl78/rl78.opt (-mallregs): New.
2062 2013-09-17 Nick Clifton <nickc@redhat.com>
2064 * config/rl78/rl78.c (need_to_save): Change return type to bool.
2065 For interrupt functions: save all call clobbered registers if the
2066 interrupt handler is not a leaf function.
2067 (rl78_expand_prologue): Always recompute the frame information.
2068 For interrupt functions: only select bank 0 if one of the bank 0
2069 registers is going to be psuhed.
2071 2013-09-17 DJ Delorie <dj@redhat.com>
2073 * config/rl78/constraints.md: For each W* constraint, rename to C*
2074 and create a W* constraint that checks for an optional ES: prefix
2076 * config/rl78/rl78.md (UNS_ES_ADDR): New.
2077 (es_addr): New. Used to wrap far addresses.
2078 * config/rl78/rl78-protos.h (rl78_es_addr): New.
2079 (rl78_es_base): New.
2080 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
2081 wrapped far addresses.
2082 (rl78_print_operand_1): Unwrap far addresses before processing.
2083 (rl78_lo16): Wrap far addresses in unspecs.
2084 (rl78_es_addr): New.
2085 (rl78_es_base): New.
2086 (insn_ok_now): Check for not-yet-wrapped far addresses.
2087 (transcode_memory_rtx): Properly re-wrap far addresses.
2089 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
2091 * config/sparc/t-rtems: Add leon3 multilibs.
2093 2013-09-17 Cong Hou <congh@google.com>
2095 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
2096 when checking the dot production pattern. The type of rhs operand
2097 of multiply is now checked correctly.
2099 2013-09-17 Jeff Law <law@redhat.com>
2101 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2102 edge implied equivalences into successor phis.
2103 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
2104 here from tree-ssa-threadedge.c.
2105 (mark_threaded_blocks): When threading through a joiner, if both
2106 successors of the joiner's clone reach the same block, verify the
2107 PHI arguments are equal. If not, cancel the jump threading request.
2108 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
2109 tree-ssa-threadupdate.c
2110 (thread_across_edge): Don't check PHI argument equality when
2111 threading through joiner block here.
2113 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
2115 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
2116 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
2117 (warn_uninit, warn_uninitialized_vars,
2118 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2119 Move to tree-ssa-uninit.c.
2120 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
2121 (has_undefined_value_p): New. Pass dependant parts of
2122 ssa_undefined_value_p.
2123 (uninit_undefined_value_p): Use has_undefined_value_p.
2124 (warn_uninit, warn_uninitialized_vars,
2125 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2126 Move from tree-ssa.c.
2127 * tree-ssa.h: Adjust prototypes.
2129 2013-09-17 Jan Hubicka <jh@suse.cz>
2132 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
2133 * ipa-inline.c (can_inline_edge_p): Do not downgrade
2134 FUNCTION_NOT_OPTIMIZED.
2135 * ipa-inline-analysis.c (compute_inline_parameters): Function
2136 not optimized is not inlinable unless it is alwaysinline.
2137 (inline_analyze_function): Force calls in not optimized
2138 function not inlinable.
2140 2013-09-17 Jan Hubicka <jh@suse.cz>
2143 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
2145 * ipa.c (function_and_variable_visibility): Likewise.
2146 * ipa-profile.c (ipa_profile): Likewise.
2148 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
2151 * cgraph.c (cgraph_function_body_availability): Check for ifunc
2152 attribute, and don't inline the resolver in this case.
2154 2013-09-17 Teresa Johnson <tejohnson@google.com>
2156 * coverage.c (get_coverage_counts): Add missing newline.
2158 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2160 PR tree-optimization/58088
2161 * fold-const.c (mask_with_trailing_zeros): New function.
2162 (fold_binary_loc): Make sure we don't recurse infinitely
2163 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
2164 Use mask_with_trailing_zeros where appropriate.
2166 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
2168 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
2169 of 'prev' var to get better distance estimation.
2171 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
2173 * tree-inline.h (struct copy_body_data): Add transform_parameter.
2174 * tree-inline.c (is_parameter_of): New predicate.
2175 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
2176 a parameter has been remapped.
2177 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
2178 (optimize_inline_calls): Initialize transform_parameter.
2179 (copy_gimple_seq_and_replace_locals): Likewise.
2180 (tree_function_versioning): Likewise.
2181 (maybe_inline_call_in_expr): Likewise.
2183 2013-09-17 Nick Clifton <nickc@redhat.com>
2185 * config/msp430/msp430-protos.h: Add prototypes for new functions.
2186 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
2188 (is_attr_func): New function.
2189 (msp430_is_interrupt_func): New function.
2190 (is_naked_func): New function.
2191 (is_reentrant_func): New function.
2192 (is_critical_func): New function.
2193 (msp430_start_function): Add annotations for function attributes.
2194 (msp430_attr): New function.
2195 (msp430_attribute_table): New.
2196 (msp430_function_section): New function.
2197 (TARGET_ASM_FUNCTION_SECTION): Define.
2198 (msp430_builtin): New enum.
2199 (msp430_init_builtins): New function.
2200 (msp430_builtin_devl): New function.
2201 (msp430_expand_builtin): New function.
2202 (TARGET_INIT_BUILTINS): Define.
2203 (TARGET_EXPAND_BUILTINS): Define.
2204 (TARGET_BUILTIN_DECL): Define.
2205 (msp430_expand_prologue): Add support for naked, interrupt,
2206 critical and reentrant functions.
2207 (msp430_expand_epilogue): Likewise.
2208 (msp430_print_operand): Handle 'O' character.
2209 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
2211 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
2212 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
2213 (pushm): Use a 'n' rather than an 'i' constraint.
2214 (msp_return): Add generation of the interrupt return instruction.
2215 (disable_interrupts): New pattern.
2216 (enable_interrupts): New pattern.
2217 (push_intr_state): New pattern.
2218 (pop_intr_state): New pattern.
2219 (bic_SR): New pattern.
2220 (bis_SR): New pattern.
2221 * doc/extend.texi: Document MSP430 function attributes and builtin
2224 2013-09-17 Richard Biener <rguenther@suse.de>
2226 PR tree-optimization/58432
2227 * tree-loop-distribution.c (tree_loop_distribution): Also
2228 scan PHIs for outside loop uses and seed a partition from them.
2230 2013-09-17 Bin Cheng <bin.cheng@arm.com>
2232 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
2233 (restructure_reference): Call backtrace_base_for_ref.
2235 2013-09-17 Alan Modra <amodra@gmail.com>
2238 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
2241 2013-09-16 DJ Delorie <dj@redhat.com>
2243 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
2244 vregs location for RL78/G10.
2245 (rl78_expand_prologue): Avoid SEL on G10.
2246 (rl78_expand_epilogue): Likewise.
2247 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
2248 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
2249 __RL78_G10__ when appropriate.
2250 (ASM_SPEC): Pass -mg10 along to the assembler.
2251 * config/rl78/rl78.md (sel_rb): Disable for G10.
2252 * config/rl78/rl78.opt: Add -mg10 option.
2253 * config/rl78/t-rl78: Add -mg10 multilib.
2255 2013-09-16 Xinliang David Li <davidxl@google.com>
2257 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
2258 * omp-low.c (omp_max_vf): Ditto.
2259 (expand_omp_simd): Ditto.
2260 * tree-vectorizer.c (vectorize_loops): Ditto.
2261 (gate_vect_slp): Ditto.
2262 (gate_increase_alignment): Ditto.
2263 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
2264 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
2265 (gate_tree_vectorize): Name change.
2266 (tree_vectorize): Ditto.
2267 (pass_vectorize::gate): Call new function.
2268 (pass_vectorize::execute): Ditto.
2269 * opts.c: O3 default setting change.
2270 (finish_options): Check new flag.
2271 * doc/invoke.texi: Document new flags.
2272 * common.opt: New flags.
2274 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
2276 * doc/tm.texi.in (Cond Exec Macros): Remove node.
2277 (Condition Code): Don't reference it.
2278 * doc/tm.texi: Regenerate.
2280 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
2283 * lra-constraints.c (undo_optional_reloads): Consider all optional
2284 reload even if it did not get a hard reg.
2286 2013-09-16 Teresa Johnson <tejohnson@google.com>
2288 * dumpfile.c (dump_loc): Remove newline emission.
2289 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
2290 emission to dump_printf_loc calls where missing.
2291 (vect_mark_for_runtime_alias_test): Ditto.
2292 (vect_analyze_data_ref_dependence): Ditto.
2293 (vect_analyze_data_ref_dependences): Ditto.
2294 (vect_slp_analyze_data_ref_dependence): Ditto.
2295 (vect_slp_analyze_data_ref_dependences): Ditto.
2296 (vect_compute_data_ref_alignment): Ditto.
2297 (vect_update_misalignment_for_peel): Ditto.
2298 (vect_verify_datarefs_alignment): Ditto.
2299 (vector_alignment_reachable_p): Ditto.
2300 (vect_get_data_access_cost): Ditto.
2301 (vect_enhance_data_refs_alignment): Ditto.
2302 (vect_find_same_alignment_drs): Ditto.
2303 (vect_analyze_data_refs_alignment): Ditto.
2304 (vect_analyze_group_access): Ditto.
2305 (vect_analyze_data_ref_access): Ditto.
2306 (vect_analyze_data_ref_accesses): Ditto.
2307 (vect_prune_runtime_alias_test_list): Ditto.
2308 (vect_analyze_data_refs): Ditto.
2309 (vect_create_addr_base_for_vector_ref): Ditto.
2310 (vect_create_data_ref_ptr): Ditto.
2311 (vect_grouped_store_supported): Ditto.
2312 (vect_grouped_load_supported): Ditto.
2313 * value-prof.c (check_counter): Ditto.
2314 (check_ic_target): Ditto.
2315 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
2316 (vect_recog_widen_mult_pattern): Ditto.
2317 (vect_recog_widen_sum_pattern): Ditto.
2318 (vect_recog_over_widening_pattern): Ditto.
2319 (vect_recog_widen_shift_pattern): Ditto.
2320 (vect_recog_rotate_pattern): Ditto.
2321 (vect_recog_vector_vector_shift_pattern): Ditto.
2322 (vect_recog_divmod_pattern): Ditto.
2323 (vect_recog_mixed_size_cond_pattern): Ditto.
2324 (vect_recog_bool_pattern): Ditto.
2325 (vect_pattern_recog_1): Ditto.
2326 (vect_pattern_recog): Ditto.
2327 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
2328 (vect_is_simple_iv_evolution): Ditto.
2329 (vect_analyze_scalar_cycles_1): Ditto.
2330 (vect_get_loop_niters): Ditto.
2331 (vect_analyze_loop_1): Ditto.
2332 (vect_analyze_loop_form): Ditto.
2333 (vect_analyze_loop_operations): Ditto.
2334 (vect_analyze_loop_2): Ditto.
2335 (vect_analyze_loop): Ditto.
2336 (report_vect_op): Ditto.
2337 (vect_is_slp_reduction): Ditto.
2338 (vect_is_simple_reduction_1): Ditto.
2339 (vect_get_known_peeling_cost): Ditto.
2340 (vect_estimate_min_profitable_iters): Ditto.
2341 (vect_model_reduction_cost): Ditto.
2342 (vect_model_induction_cost): Ditto.
2343 (get_initial_def_for_induction): Ditto.
2344 (vect_create_epilog_for_reduction): Ditto.
2345 (vectorizable_reduction): Ditto.
2346 (vectorizable_induction): Ditto.
2347 (vectorizable_live_operation): Ditto.
2348 (vect_loop_kill_debug_uses): Ditto.
2349 (vect_transform_loop): Ditto.
2350 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
2351 (vect_stmt_relevant_p): Ditto.
2352 (process_use): Ditto.
2353 (vect_mark_stmts_to_be_vectorized): Ditto.
2354 (vect_model_simple_cost): Ditto.
2355 (vect_model_promotion_demotion_cost): Ditto.
2356 (vect_model_store_cost): Ditto.
2357 (vect_get_store_cost): Ditto.
2358 (vect_model_load_cost): Ditto.
2359 (vect_get_load_cost): Ditto.
2360 (vect_init_vector_1): Ditto.
2361 (vect_get_vec_def_for_operand): Ditto.
2362 (vect_finish_stmt_generation): Ditto.
2363 (vectorizable_call): Ditto.
2364 (vectorizable_conversion): Ditto.
2365 (vectorizable_assignment): Ditto.
2366 (vectorizable_shift): Ditto.
2367 (vectorizable_operation): Ditto.
2368 (vectorizable_store): Ditto.
2369 (vectorizable_load): Ditto.
2370 (vectorizable_condition): Ditto.
2371 (vect_analyze_stmt): Ditto.
2372 (vect_transform_stmt): Ditto.
2373 (vect_is_simple_use): Ditto.
2374 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
2375 (vect_can_advance_ivs_p): Ditto.
2376 (vect_update_ivs_after_vectorizer): Ditto.
2377 (vect_do_peeling_for_loop_bound): Ditto.
2378 (vect_gen_niters_for_prolog_loop): Ditto.
2379 (vect_update_inits_of_drs): Ditto.
2380 (vect_create_cond_for_alias_checks): Ditto.
2381 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
2382 (vect_build_slp_tree_1): Ditto.
2383 (vect_supported_load_permutation_p): Ditto.
2384 (vect_analyze_slp_instance): Ditto.
2385 (vect_analyze_slp): Ditto.
2386 (vect_make_slp_decision): Ditto.
2387 (vect_detect_hybrid_slp): Ditto.
2388 (vect_bb_vectorization_profitable_p): Ditto.
2389 (vect_slp_analyze_bb_1): Ditto.
2390 (vect_update_slp_costs_according_to_vf): Ditto.
2391 (vect_get_mask_element): Ditto.
2392 (vect_transform_slp_perm_load): Ditto.
2393 (vect_schedule_slp_instance): Ditto.
2394 (vect_schedule_slp): Ditto.
2395 (vect_slp_transform_bb): Ditto.
2396 * profile.c (read_profile_edge_counts): Ditto.
2397 (compute_branch_probabilities): Ditto.
2398 * coverage.c (get_coverage_counts): Ditto.
2400 2013-09-16 Diego Novillo <dnovillo@google.com>
2402 * tree-core.h: Add missing comment lines from refactoring of tree.h.
2404 2013-09-16 Jan Hubicka <jh@suse.cz>
2406 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
2407 abstract functions; for static functions check the presence of body.
2409 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2411 * config/aarch64/aarch64-simd-builtins.def (fma): New.
2412 * config/aarch64/aarch64-simd.md
2413 (aarch64_mla_elt<mode>): New.
2414 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2415 (aarch64_mls_elt<mode>): Likewise.
2416 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2417 (aarch64_fma4_elt<mode>): Likewise.
2418 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2419 (aarch64_fma4_elt_to_128v2df): Likewise.
2420 (aarch64_fma4_elt_to_64df): Likewise.
2421 (fnma<mode>4): Likewise.
2422 (aarch64_fnma4_elt<mode>): Likewise.
2423 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2424 (aarch64_fnma4_elt_to_128v2df): Likewise.
2425 (aarch64_fnma4_elt_to_64df): Likewise.
2426 * config/aarch64/iterators.md (VDQSF): New.
2427 * config/aarch64/arm_neon.h
2428 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
2429 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
2431 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2433 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
2434 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2435 (aarch64_mul3_elt_to_128df): Likewise.
2436 (aarch64_mul3_elt_to_64v2df): Likewise.
2437 * config/aarch64/iterators.md (VEL): Also handle DFmode.
2439 (VMUL_CHANGE_NLANES) Likewise.
2442 * config/aarch64/arm_neon.h
2443 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
2445 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2447 * config/aarch64/arm_neon.h
2448 (vcvtx_high_f32_f64): Fix parameters.
2450 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2451 Uros Bizjak <ubizjak@gmail.com>
2453 * config/alpha.c: Include tree-ssanames.h.
2455 2013-09-16 Richard Biener <rguenther@suse.de>
2457 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
2458 (dot_rdg_1): Handle control_dd.
2459 (create_edge_for_control_dependence): New function.
2460 (create_rdg_edges): Add control dependences if asked for.
2461 (build_rdg): Likewise.
2462 (generate_loops_for_partition): If there are not necessary
2463 control stmts remove all their dependencies.
2464 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
2465 (distribute_loop): Pass on control dependences.
2466 (tree_loop_distribution): Compute control dependences and remove
2467 restriction on number of loop nodes.
2469 2013-09-16 Jakub Jelinek <jakub@redhat.com>
2471 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
2474 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
2476 * cse.c (try_const_anchors): Punt on CC modes.
2478 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2480 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
2482 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
2485 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
2488 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2491 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
2493 (darwin_objc1_section): Likewise.
2494 (darwin_file_end): Emit Image Info section when required.
2496 2013-09-14 Jan Hubicka <jh@suse.cz>
2498 * tree-into-ssa.c (gate_into_ssa): New.
2499 (pass_data_build_ssa): Use it.
2500 * cgraph.h (expand_thunk): Update prototype.
2501 * cgraphunit.c (analyze_function): Expand thunks early.
2502 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
2503 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
2504 set lowered flag; do not add new function.
2505 (assemble_thunks_and_aliases): Update.
2506 * tree-ssa.c (gate_init_datastructures): New gate.
2507 (pass_data_init_datastructures): Use it.
2509 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2512 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
2513 xmm register set described in the psABI.
2515 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
2517 * dwarf2out.c (should_emit_struct_debug): Add check
2518 for type_decl variable is not NULL.
2520 2013-09-13 Jacek Caban <jacek@codeweavers.com>
2522 * config.gcc: Use new winnt-c.c target hooks
2523 * config/t-winnt: New file
2524 * config/winnt-c.c: New file
2525 * doc/tm.texi.in: Document new hook
2526 * doc/tm.texi: Regenerated
2528 2013-09-13 Jan Hubicka <jh@suse.cz>
2531 * ipa-inline.c (check_callers): New function.
2532 (check_caller_edge): Remove.
2533 (want_inline_function_to_all_callers_p): Also permit alises that are
2535 (inline_to_all_callers): Terminate the walk when devirtualization
2536 introduce new calls.
2538 2013-09-13 Jan Hubicka <jh@suse.cz>
2540 * ipa-inline-analysis.c (struct growth_data): Add node.
2541 (do_estimate_growth_1): Fix detection of recursion.
2543 2013-09-13 Jakub Jelinek <jakub@redhat.com>
2545 PR tree-optimization/58392
2546 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
2547 to avoid shadowing of outer loop variable. If
2548 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
2549 replace_by_duplicate_decl simduid of loops that have it set and
2550 set dest_cfun->has_simduid_loops and/or
2551 dest_cfun->has_force_vect_loops.
2552 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
2553 instead of maybe_lookup_decl.
2554 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
2555 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
2556 call. Copy over force_vect and copy and remap simduid. Set
2557 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
2558 (copy_cfg_body): Remove blocks_to_copy argument. Use
2559 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
2560 caller. Don't set cfun->has_simduid_loops and/or
2561 cfun->has_force_vect_loops here.
2562 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
2564 (expand_call_inline, tree_function_versioning): Adjust copy_body
2567 2013-09-13 Martin Jambor <mjambor@suse.cz>
2570 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
2571 the assert if the edge was a speculative one.
2573 2013-09-13 Richard Biener <rguenther@suse.de>
2575 * tree-data-ref.h (known_dependences_p): Move here ...
2576 * tree-loop-distribution.c (known_dependences_p): ... from here.
2577 (dump_rdg_component, debug_rdg_component): Remove.
2579 (generate_loops_for_partition): Use gimple_uid instead of
2580 relying on matching stmt visit order.
2581 (rdg_build_partitions): Take starting stmt vector.
2582 (ldist_gen): Merge into ...
2583 (distribute_loop): ... this function. Do not compute starting
2585 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
2587 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2589 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
2590 Set type attribute correctly. Set predicable_short_it attribute.
2591 (cmpsi_shiftsi): Remove %? from output template.
2593 2013-09-13 Richard Biener <rguenther@suse.de>
2595 * tree-loop-distribution.c (struct rdg_component,
2596 rdg_defs_used_in_other_loops_p, free_rdg_components,
2597 rdg_build_components): Remove.
2598 (stmts_from_loop): Do not record virtual PHIs.
2599 (generate_loops_for_partition): Skip virtual PHIs.
2600 (build_rdg_partition_for_component): Rename to ...
2601 (build_rdg_partition_for_vertex): ... this and adjust.
2602 (rdg_build_partitions): Take a vector of starting vertices
2603 instead of components. Remove unnecessary leftover handling.
2604 (ldist_gen): Do not build components or record other stores.
2605 (distribute_loop): Do not distribute loops containing stmts
2608 2013-09-13 Christian Bruel <christian.bruel@st.com>
2611 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
2613 2013-09-13 Kai Tietz <ktietz@redhat.com>
2615 * config.gcc: Separate cases for mingw and cygwin targets,
2616 and add 64-bit cygwin target case.
2618 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
2619 dll-export inline-functions.
2620 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
2623 2013-09-13 Jeff Law <law@redhat.com>
2627 2013-09-06 Jeff Law <law@redhat.com>
2629 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2630 edge implied equivalences into successor phis.
2632 2013-09-12 DJ Delorie <dj@redhat.com>
2634 * config/rl78/rl78-virt.md: Change from | to \; for asm line
2637 2013-09-12 Brooks Moses <bmoses@google.com>
2640 * Makefile.in: Do not install driver binaries in $(target)/bin.
2642 2013-09-12 DJ Delorie <dj@redhat.com>
2644 * config/rl78/rl78.opt (mrelax): New.
2645 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
2646 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
2648 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
2650 (rl78_expand_epilogue): Likewise.
2652 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
2655 * lra-eliminations.c (remove_reg_equal_offset_note): New.
2656 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
2657 pointer elimination with using remove_reg_equal_offset_note.
2659 2013-09-12 DJ Delorie <dj@redhat.com>
2661 * config/msp430/: New port.
2662 * config.gcc (msp430): Added.
2663 * doc/invoke.texi: Document MSP430 options.
2664 * doc/install.texi: Document msp430-elf
2665 * doc/md.texi: Document msp430-elf
2666 * doc/contrib.texi: Document msp430-elf
2668 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
2671 2013-09-12 Martin Jambor <mjambor@suse.cz>
2674 * ipa-prop.c (remove_described_reference): Give up if the edge in the
2675 reference descriptor is NULL.
2676 (ipa_edge_removal_hook): If owning a reference descriptor, set its
2679 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2681 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
2682 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
2683 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
2684 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
2685 make_temp_ssa_name): move to tree-ssanames.h
2686 * tree-ssa-alias.h: Move prototype.
2687 * tree-ssa.h: Include tree-ssanames.h.
2688 * tree-ssanames.c (FREE_SSANAMES): Move to here.
2689 * tree-ssanames.h: New. Move items from tree-flow*.h
2690 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
2692 2013-09-12 Richard Biener <rguenther@suse.de>
2694 PR tree-optimization/58404
2695 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
2696 propagate non-invariant addresses into dereferences wrapped
2697 in component references.
2699 2013-09-12 Richard Biener <rguenther@suse.de>
2701 PR tree-optimization/58402
2702 * passes.def: Move pass_late_warn_uninitialized later.
2704 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2706 * tree-ssa.h: New. Move content from tree-flow.h and
2708 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
2709 Move prototypes belonging to tree-ssa.c.
2710 * tree-flow-inline.h (redirect_edge_var_map_def,
2711 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
2713 * gimple.h: Adjust prototypes.
2714 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
2716 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
2717 * tree.h: Move prototype to tree-ssa.h.
2718 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
2719 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
2720 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
2721 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
2722 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
2723 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
2724 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
2725 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
2726 graphite-blocking.c, graphite-clast-to-gimple.c,
2727 graphite-dependences.c, graphite-interchange.c,
2728 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
2729 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
2730 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
2731 ipa-reference.c, ipa-split.c, ipa-utils.c,
2732 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
2733 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
2734 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
2735 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2736 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
2737 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
2738 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
2739 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
2740 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
2741 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
2742 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
2743 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2744 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
2745 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
2746 value-prof.c, var-tracking.c,
2747 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
2749 2013-09-12 Richard Biener <rguenther@suse.de>
2751 PR tree-optimization/58396
2752 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
2753 (build_rdg): Take a loop-nest parameter, fix memleaks.
2754 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
2756 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2758 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
2761 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
2763 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
2764 The FMA instruction names should have a 'v' prefix.
2766 2013-09-12 Richard Biener <rguenther@suse.de>
2768 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
2769 (dot_rdg): Use popen instead of system in optional code.
2770 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
2771 (already_processed_vertex_p): Adjust.
2772 (has_anti_or_output_dependence, predecessor_has_mem_write,
2773 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
2774 rdg_flag_uses): Remove.
2775 (rdg_flag_vertex): Simplify.
2776 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
2778 (build_rdg_partition_for_component): Process the first vertex
2779 of a component only.
2780 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
2782 2013-09-12 Alan Modra <amodra@gmail.com>
2784 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
2786 2013-09-11 DJ Delorie <dj@redhat.com>
2787 Nick Clifton <nickc@redhat.com>
2789 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
2790 (rl78_stack_based_mem): New.
2791 * config/rl78/constraints.md (Iv08): New.
2800 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
2802 (movsi): Change from expand to insn-and-split.
2803 (ashrsi3): Clobber AX.
2807 * config/rl78/rl78.md (CC_REG): Fix.
2808 (addsi3): Allow memory and immediate operands.
2809 (addsi3_internal): Split into...
2810 (addsi3_internal_virt): ...new, and ...
2811 (addsi3_internal_real): ...new.
2813 (subsi3_internal_virt): New.
2814 (subsi3_internal_real): New.
2815 (mulsi3): Add memory operand.
2816 (mulsi3_rl78): Likewise.
2817 (mulsi3_g13): Likewise.
2818 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
2819 (cbranchqi4_real): Add more constraint options.
2820 (cbranchhi4_real): Expand pattern.
2821 (cbranchhi4_real_signed): New.
2822 (cbranchhi4_real_inverted): New.
2823 (cbranchsi4_real_lt): New.
2824 (cbranchsi4_real_ge): New.
2825 (cbranchsi4_real_signed): New.
2826 (cbranchsi4_real): New.
2828 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
2830 (lshrsi3_virt): Likewise.
2831 (ashlsi3_virt): Likewise.
2832 (cbranchqi4_virt_signed): New.
2833 (cbranchhi4_virt_signed): New.
2834 (cbranchsi4_virt): New.
2835 * config/rl78/rl78.c: Whitespace fixes throughout.
2836 (move_elim_pass): New.
2837 (pass_data_rl78_move_elim): New.
2838 (pass_rl78_move_elim): New.
2839 (make_pass_rl78_move_elim): New.
2840 (rl78_devirt_info): Run devirt earlier.
2841 (rl78_move_elim_info): New.
2842 (rl78_asm_file_start): Register it.
2843 (rl78_split_movsi): New.
2844 (rl78_as_legitimate_address): Allow virtual base registers when
2846 (rl78_addr_space_convert): Remove spurious debug stuff.
2847 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
2848 (rl78_print_operand): More cases for not printing '#'.
2849 (rl78_expand_compare): Remove most of the logic.
2850 (content_memory): New.
2851 (clear_content_memory): New.
2852 (get_content_index): New.
2853 (get_content_name): New.
2854 (display_content_memory): New.
2855 (update_content): New.
2856 (record_content): New.
2857 (already_contains): New.
2858 (insn_ok_now): Re-recog insns with virtual registers.
2859 (add_postponed_content_update): New.
2860 (process_postponed_content_update): New.
2861 (gen_and_emit_move): New.
2862 (transcode_memory_rtx): Record new location content.
2863 Use gen_and_emit_move.
2864 (force_into_acc): New.
2865 (move_to_acc): Use gen_and_emit_move.
2866 (move_from_acc): Likewise.
2867 (move_acc_to_reg): Likewise.
2868 (move_to_x): Likewise.
2869 (move_to_hl): Likewise.
2870 (move_to_de): Likewise.
2871 (rl78_alloc_physical_registers_op1): Record location content.
2872 (has_constraint): New.
2873 (rl78_alloc_physical_registers_op2): Record location content.
2875 (rl78_alloc_physical_registers_ro1): Likewise.
2876 (rl78_alloc_physical_registers_cmp): Likewise.
2877 (rl78_alloc_physical_registers_umul): Likewise.
2878 (rl78_alloc_address_registers_macax): New.
2879 (rl78_alloc_physical_registers): Initialize and set location
2880 content memory as needed.
2881 (rl78_reorg): Make sure split2 is called.
2882 (rl78_rtx_costs): New.
2884 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
2886 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
2887 for (not (neg ...)) and (neg (not ...)) cases.
2889 2013-09-11 Richard Biener <rguenther@suse.de>
2892 * passes.def: Split critical edges before late uninit warning passes.
2893 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
2895 2013-09-11 Jakub Jelinek <jakub@redhat.com>
2897 PR tree-optimization/58385
2898 * fold-const.c (build_range_check): If both low and high are NULL,
2899 use omit_one_operand_loc to preserve exp side-effects.
2901 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2903 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
2905 2013-09-11 Richard Biener <rguenther@suse.de>
2907 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
2908 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
2909 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
2910 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
2911 stmts_from_loop, known_dependences_p, build_empty_rdg,
2912 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
2913 * tree-loop-distribution.c: ... here.
2914 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
2915 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
2916 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
2917 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
2918 * tree-loop-distribution.c: ... here.
2919 * tree-loop-distribution.c: Include gimple-pretty-print.h.
2920 (struct partition_s): Add loops member.
2921 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
2922 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
2923 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
2925 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2926 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2927 Sergey Lega <sergey.s.lega@intel.com>
2928 Anna Tikhonova <anna.tikhonova@intel.com>
2929 Ilya Tocar <ilya.tocar@intel.com>
2930 Andrey Turetskiy <andrey.turetskiy@intel.com>
2931 Ilya Verbin <ilya.verbin@intel.com>
2932 Kirill Yukhin <kirill.yukhin@intel.com>
2933 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2935 * config/i386/constraints.md (k): New.
2937 * config/i386/i386.c (const regclass_map): Add new mask registers.
2938 (dbx_register_map): Ditto.
2939 (dbx64_register_map): Ditto.
2940 (svr4_dbx_register_map): Ditto.
2941 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
2943 (ix86_preferred_reload_class): Disable constants for mask registers.
2944 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
2945 (ix86_hard_regno_mode_ok): Support new mask registers.
2946 (x86_order_regs_for_local_alloc): Ditto.
2947 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
2948 (FIXED_REGISTERS): Add new mask registers.
2949 (CALL_USED_REGISTERS): Ditto.
2950 (REG_ALLOC_ORDER): Ditto.
2951 (VALID_MASK_REG_MODE): New.
2952 (FIRST_MASK_REG): Ditto.
2953 (LAST_MASK_REG): Ditto.
2954 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
2955 (MAYBE_MASK_CLASS_P): New.
2956 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
2957 (REG_CLASS_CONTENTS): Ditto.
2958 (MASK_REGNO_P): New.
2959 (ANY_MASK_REG_P): Ditto.
2960 (HI_REGISTER_NAMES): Add new mask registers.
2961 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2962 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2964 (attribute "type"): Add mskmov, msklog.
2965 (attribute "length_immediate"): Support them.
2966 (attribute "memory"): Ditto.
2967 (attribute "prefix_0f"): Ditto.
2968 (*movhi_internal): Support new mask registers.
2969 (*movqi_internal): Ditto.
2970 (define_split): Split out clobber pattern is a logic
2971 insn on mask registers.
2972 (*k<logic><mode>): New.
2973 (*andhi_1): Extend to support mask regs.
2974 (*andqi_1): Extend to support mask regs.
2976 (define_split): Split and-not to and and not if operands
2978 (*<code><mode>_1): Separate HI mode to new pattern...
2979 (*<code>hi_1): This.
2980 (*<code>qi_1): Extend to support mask regs.
2982 (kortestzhi): Ditto.
2983 (kortestchi): Ditto.
2985 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2986 (*one_cmplhi2_1): ...Here, now with mask registers support.
2987 (*one_cmplqi2_1): Support new mask registers.
2988 (HI/QImode arithmetics splitter): Don't split if mask registers
2990 (HI/QImode not splitter): Ditto.
2991 * config/i386/predicated.md (mask_reg_operand): New.
2992 (general_reg_operand): Ditto.
2994 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2996 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2997 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2999 2013-09-10 Jeff Law <law@redhat.com>
3001 PR tree-optimization/58380
3002 * tree-ssa-threadupdate.c (thread_block): Recognize another case
3003 of threading through a buried loop header.
3005 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
3006 return value for single successor case.
3008 2013-09-10 Jan Hubicka <jh@suse.cz>
3010 * ipa-devirt.c (ipa_devirt): Enable with LTO.
3012 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
3015 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
3016 support conditional execution.
3017 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
3019 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
3021 * lra.c (lra): Clear lra_optional_reload_pseudos before every
3023 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
3024 Check destination too to check move insn.
3025 (undo_optional_reloads): Add check that the original peudo did not
3026 changed its allocation and the optional reload was inherited on last
3027 inheritance pass. Break loop after deciding to keep optional reload.
3028 (lra_undo_inheritance): Add check that inherited pseudo still in
3031 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
3033 * config/aarch64/aarch64.md (generic_sched): New.
3034 * config/aarch64/aarch64-generic.md (load): Make conditional
3035 on generic_sched attribute.
3036 (nonload): Likewise.
3038 2013-09-10 Jan Hubicka <jh@suse.cz>
3040 * lto-cgraph.c: Include ipa-utils.h.
3041 (compute_ltrans_boundary): Also add possible targets into the boundary.
3043 2013-09-10 Jan Hubicka <jh@suse.cz>
3045 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
3046 VAR_DECL of vtable rather than full expression.
3048 2013-09-10 Jan Hubicka <jh@suse.cz>
3049 Paolo Carlini <paolo.carlini@oracle.com>
3051 * cgraphunit.c (analyze_functions): Save input_location, set it
3052 to UNKNOWN_LOCATION and restore it at the end.
3054 2013-09-10 Martin Jambor <mjambor@suse.cz>
3056 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
3057 represented by a thunk.
3059 2013-09-10 Jeff Law <law@redhat.com>
3061 PR tree-optimization/58343
3062 * tree-ssa-threadupdate.c (thread_block): Identify and disable
3063 jump threading requests through loop headers buried in the middle
3064 of a jump threading path.
3066 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
3067 in return value/type.
3069 2013-09-10 Jakub Jelinek <jakub@redhat.com>
3071 PR rtl-optimization/58365
3072 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
3073 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
3076 2013-09-10 Richard Biener <rguenther@suse.de>
3078 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
3079 * tree-data-ref.c (create_rdg_vertices): Collect all data
3080 references, signal failure to the caller, use data-ref API.
3081 (build_rdg): Compute data references only once. Maintain lifetime
3082 of data references and data dependences from within RDG.
3083 (free_rdg): Free dependence relations.
3084 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
3085 inventing extra dependences.
3086 (distribute_loop): Update for RDG API changes.
3088 2013-09-10 Kai Tietz <ktietz@redhat.com>
3090 * doc/invoke.texi (fms-extensions): Document changed
3091 behavior for ms-abi targets.
3092 * config/i386/i386.c (ix86_option_override_internal):
3093 Set default value of option -fms-extension for ms-abi targets.
3095 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3097 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
3099 2013-09-10 Alan Modra <amodra@gmail.com>
3102 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
3104 2013-09-10 Alan Modra <amodra@gmail.com>
3106 * config/rs6000/predicates.md (add_cint_operand): New.
3107 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
3108 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
3109 using add_cint_operand.
3110 (largetoc_high_plus_aix): Likewise.
3112 2013-09-09 Jakub Jelinek <jakub@redhat.com>
3114 PR tree-optimization/58364
3115 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
3116 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
3117 the current range can't be an unconditional true or false.
3119 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
3121 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
3123 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
3125 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
3127 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
3130 * doc/invoke.texi (-Wdelete-incomplete): Document it.
3132 2013-09-09 Ian Bolton <ian.bolton@arm.com>
3134 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
3135 NO_REGS for immediate that can't be moved directly into FP_REGS.
3137 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3139 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
3140 comparison with negated operand.
3141 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
3144 2013-09-09 Richard Biener <rguenther@suse.de>
3147 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
3148 of a subloop record all its block as affecting loop-closed SSA form.
3150 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3152 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
3153 of an rtx/bitpos pair.
3154 (store_fixed_bit_field): Update accordingly.
3156 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3158 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
3160 * builtins.c (expand_errno_check): Likewise.
3161 * dwarf2cfi.c (init_return_column_size): Likewise.
3162 * except.c (sjlj_mark_call_sites): Likewise.
3163 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
3164 * lra-constraints.c (emit_inc): Likewise.
3165 * ree.c (combine_set_extension): Likewise.
3166 * regmove.c (fixup_match_2): Likewise.
3167 * reload1.c (inc_for_reload): Likewise.
3169 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3171 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
3172 (force_to_mode, simplify_shift_const_1, simplify_comparison):
3173 Use gen_int_mode with the mode of the associated simplify_* call.
3174 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
3175 * expmed.c (expand_shift_1): Likewise.
3176 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3177 * loop-iv.c (iv_number_of_iterations): Likewise.
3178 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3179 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
3181 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3183 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
3184 of the associated expand_* call.
3185 (asan_emit_stack_protection): Likewise.
3186 * builtins.c (round_trampoline_addr): Likewise.
3187 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
3188 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
3189 (emit_store_flag): Likewise.
3190 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
3191 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
3193 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3194 * ifcvt.c (noce_try_store_flag_constants): Likewise.
3195 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3196 * modulo-sched.c (generate_prolog_epilog): Likewise.
3197 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
3198 (expand_ctz, expand_ffs, expand_unop): Likewise.
3200 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3202 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
3203 of the associated gen_rtx_* call.
3204 * caller-save.c (init_caller_save): Likewise.
3205 * combine.c (find_split_point, make_extraction): Likewise.
3206 (make_compound_operation): Likewise.
3207 * dwarf2out.c (mem_loc_descriptor): Likewise.
3208 * explow.c (plus_constant, probe_stack_range): Likewise.
3209 * expmed.c (expand_mult_const): Likewise.
3210 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
3211 * reload1.c (init_reload): Likewise.
3212 * valtrack.c (cleanup_auto_inc_dec): Likewise.
3213 * var-tracking.c (adjust_mems): Likewise.
3214 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
3215 rather than gen_rtx_fmt_ee.
3217 2013-09-09 Jan Hubicka <jh@suse.cz>
3220 * value-prof.c (gimple_ic): Copy also abnormal edges.
3222 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3224 * asan.c (asan_shadow_cst): Use gen_int_mode.
3226 2013-09-08 Jan Hubicka <jh@suse.cz>
3228 * ipa-profile.c: Add toplevel comment.
3229 (ipa_propagate_frequency_1): Be more conservative when profile is read.
3230 (contains_hot_call_p): New function.
3231 (ipa_propagate_frequency): Set frequencies based on counts when
3233 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
3234 profile; do not tamper with profile after inlining if it is read.
3236 2013-09-08 Jan Hubicka <jh@suse.cz>
3238 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
3241 2013-09-08 Jan Hubicka <jh@suse.cz>
3243 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
3246 2013-09-08 Jeff Law <law@redhat.com>
3249 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
3252 2013-09-08 Andi Kleen <ak@linux.intel.com>
3254 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
3256 2013-09-08 Jan Hubicka <jh@suse.cz>
3258 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
3260 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
3262 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
3263 for non-debug insns.
3264 * lra.c (new_insn_reg): Take the containing insn as a parameter.
3265 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
3266 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
3269 2013-09-08 Jan Hubicka <jh@suse.cz>
3271 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
3272 targets and devirtualize to BUILT_IN_UNREACHABLE.
3273 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
3274 * ipa.c (walk_polymorphic_call_targets): New function.
3275 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
3276 functions; use the new timevar.
3277 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
3278 was removed from the program.
3279 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
3280 not consider it in the walk when its vtable is dead.
3281 (possible_polymorphic_call_targets_1): Pass anonymous flag to
3283 (devirt_variable_node_removal_hook): New function.
3284 (possible_polymorphic_call_targets): Also register
3285 devirt_variable_node_removal_hook.
3286 (ipa_devirt): Do not do non-speculative devirtualization.
3287 (gate_ipa_devirt): One execute if devirtualizing speculatively.
3289 2013-09-08 Jan Hubicka <jh@suse.cz>
3291 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
3292 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
3293 varpool_remove_variable_insertion_hook): Declare.
3294 * varpool.c (varpool_node_hook_list): New structure.
3295 (first_varpool_node_removal_hook,
3296 first_varpool_variable_insertion_hook): New variables.
3297 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
3298 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
3299 varpool_remove_variable_insertion_hook,
3300 varpool_call_variable_insertion_hooks): New functions.
3301 (varpool_remove_node): Use it.
3303 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
3306 * diagnostic.c (diagnostic_build_prefix): When s.file is
3307 "<built-in>" don't output line and column numbers.
3309 2013-09-06 Jan Hubicka <jh@suse.cz>
3311 * cgraphunit.c (expand_thunk): Get body before touching arguments.
3312 * lto-streamer-out.c: Stream thunks, too.
3313 * lto-streamer-in.c (input_function): Pop cfun here
3314 (lto_read_body): Instead of here.
3316 2013-09-06 Caroline Tice <cmtice@google.com>
3318 * doc/install.texi: Add documentation for the --enable-vtable-verify
3319 and the --disable-libvtv configure options.
3321 2013-09-06 Jeff Law <law@redhat.com>
3323 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3324 edge implied equivalences into successor phis.
3326 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
3328 * resource.c (mark_referenced_resources): Handle COND_EXEC.
3330 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
3332 * resource.c (mark_target_live_regs): Compute resources taking
3333 into account if a call is predicated or not.
3335 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3337 * toplev.c (output_stack_usage): Be prepared for suffixes created by
3338 the compiler in the function names.
3340 2013-09-06 Jan Hubicka <jh@suse.cz>
3343 * ipa-inline.c (has_caller_p): New function.
3344 (want_inline_function_to_all_callers_p): Use it.
3345 (sum_callers, inline_to_all_callers): Break out from ...
3346 (ipa_inline): ... here.
3348 2013-09-06 Jan Hubicka <jh@suse.cz>
3350 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
3351 only when AVX is enabled.
3353 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3355 * config/aarch64/aarch64.md
3356 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
3357 is fpsimd_<load/store>2.
3358 (load_pair<mode>): Likewise.
3359 (store_pair<mode>): Likewise.
3361 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3363 * config/arm/types.md (type): Add "mrs" type.
3364 * config/aarch64/aarch64.md
3365 (aarch64_load_tp_hard): Make type "mrs".
3367 (load_tp_hard): Make type "mrs".
3368 * config/arm/cortex-a15.md: Update with new attributes.
3369 * config/arm/cortex-a5.md: Update with new attributes.
3370 * config/arm/cortex-a53.md: Update with new attributes.
3371 * config/arm/cortex-a7.md: Update with new attributes.
3372 * config/arm/cortex-a8.md: Update with new attributes.
3373 * config/arm/cortex-a9.md: Update with new attributes.
3374 * config/arm/cortex-m4.md: Update with new attributes.
3375 * config/arm/cortex-r4.md: Update with new attributes.
3376 * config/arm/fa526.md: Update with new attributes.
3377 * config/arm/fa606te.md: Update with new attributes.
3378 * config/arm/fa626te.md: Update with new attributes.
3379 * config/arm/fa726te.md: Update with new attributes.
3381 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3383 * config/aarch64/aarch64.md
3384 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
3385 (*movtf_aarch64): Likewise.
3387 (thumb1_movdi_insn): Use "multiple" for type where more than one
3388 instruction is used for a move.
3389 (*arm32_movhf): Likewise.
3390 (*thumb_movdf_insn): Likewise.
3392 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3394 * config/arm/types.md (type): Rename fcpys to fmov.
3396 (*arm_movsi_vfp): Rename type fcpys as fmov.
3397 (*thumb2_movsi_vfp): Likewise
3398 (*movhf_vfp_neon): Likewise
3399 (*movhf_vfp): Likewise
3400 (*movsf_vfp): Likewise
3401 (*thumb2_movsf_vfp): Likewise
3402 (*movsfcc_vfp): Likewise
3403 (*thumb2_movsfcc_vfp): Likewise
3404 * config/aarch64/aarch64-simd.md
3405 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
3406 * config/aarch64/aarch64.md
3407 (*movsi_aarch64): Replace type mov_reg with fmovs.
3408 (*movdi_aarch64): Likewise
3409 (*movsf_aarch64): Likewise
3410 (*movdf_aarch64): Likewise
3412 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
3413 * config/arm/iwmmxt.md
3414 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
3415 * config/arm/arm1020e.md: Update with new attributes.
3416 * config/arm/cortex-a15-neon.md: Update with new attributes.
3417 * config/arm/cortex-a5.md: Update with new attributes.
3418 * config/arm/cortex-a53.md: Update with new attributes.
3419 * config/arm/cortex-a7.md: Update with new attributes.
3420 * config/arm/cortex-a8-neon.md: Update with new attributes.
3421 * config/arm/cortex-a9.md: Update with new attributes.
3422 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3423 * config/arm/cortex-r4f.md: Update with new attributes.
3424 * config/arm/marvell-pj4.md: Update with new attributes.
3425 * config/arm/vfp11.md: Update with new attributes.
3427 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3429 * config/aarch64/aarch64.md
3430 (*madd<mode>): Fix type attribute.
3431 (*maddsi_uxtw): Likewise.
3432 (*msub<mode>): Likewise.
3433 (*msubsi_uxtw): Likewise.
3434 (<su_optab>maddsidi4): Likewise.
3435 (<su_optab>msubsidi4): Likewise.
3437 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3439 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
3440 * config/arm/arm.md (core_cycles): Remove fdiv.
3441 * config/arm/vfp.md:
3442 (*sqrtsf2_vfp): Update for attribute changes.
3443 (*sqrtdf2_vfp): Likewise.
3444 * config/aarch64/aarch64.md:
3445 (sqrt<mode>2): Update for attribute changes.
3446 * config/arm/arm1020e.md: Update with new attributes.
3447 * config/arm/cortex-a15-neon.md: Update with new attributes.
3448 * config/arm/cortex-a5.md: Update with new attributes.
3449 * config/arm/cortex-a53.md: Update with new attributes.
3450 * config/arm/cortex-a7.md: Update with new attributes.
3451 * config/arm/cortex-a8-neon.md: Update with new attributes.
3452 * config/arm/cortex-a9.md: Update with new attributes.
3453 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3454 * config/arm/cortex-r4f.md: Update with new attributes.
3455 * config/arm/marvell-pj4.md: Update with new attributes.
3456 * config/arm/vfp11.md: Update with new attributes.
3458 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3460 * config/arm/types.md
3461 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
3462 * config/aarch64/aarch64.md
3463 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
3465 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
3466 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
3467 (float<GPI:mode><GPF:mode>2): Likewise.
3469 (*truncsisf2_vfp): Update with new attributes.
3470 (*truncsidf2_vfp): Likewise.
3471 (fixuns_truncsfsi2): Likewise.
3472 (fixuns_truncdfsi2): Likewise.
3473 (*floatsisf2_vfp): Likewise.
3474 (*floatsidf2_vfp): Likewise.
3475 (floatunssisf2): Likewise.
3476 (floatunssidf2): Likewise.
3477 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
3478 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
3479 * config/arm/arm1020e.md: Update with new attributes.
3480 * config/arm/cortex-a15-neon.md: Update with new attributes.
3481 * config/arm/cortex-a5.md: Update with new attributes.
3482 * config/arm/cortex-a53.md: Update with new attributes.
3483 * config/arm/cortex-a7.md: Update with new attributes.
3484 * config/arm/cortex-a8-neon.md: Update with new attributes.
3485 * config/arm/cortex-a9.md: Update with new attributes.
3486 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3487 * config/arm/cortex-r4f.md: Update with new attributes.
3488 * config/arm/marvell-pj4.md: Update with new attributes.
3489 * config/arm/vfp11.md: Update with new attributes.
3491 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3493 * config/aarch64/arm_neon.h
3494 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
3495 (vqtbx<1,2,3,4><q>_s8): Likewise.
3497 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3499 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
3500 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
3501 (add<mode>3): Add type attribute.
3502 (add<mode>3): Likewise.
3503 (usadd<mode>3): Likewise.
3504 (ssadd<mode>3): Likewise.
3505 (sub<mode>3): Likewise.
3506 (sub<mode>3): Likewise.
3507 (ussub<mode>3): Likewise.
3508 (sssub<mode>3): Likewise.
3509 (ssmulsa3): Likewise.
3510 (usmulusa3): Likewise.
3511 (arm_usatsihi): Likewise.
3513 (*movdi_vfp): Add types for all instructions.
3514 (*movdi_vfp_cortexa8): Likewise.
3515 (*movhf_vfp_neon): Likewise.
3516 (*movhf_vfp): Likewise.
3517 (*movdf_vfp): Likewise.
3518 (*thumb2_movdf_vfp): Likewise.
3519 (*thumb2_movdfcc_vfp): Likewise.
3520 * config/arm/arm.md: Add type attribute to all insn patterns.
3521 (*thumb1_adddi3): Add type attribute.
3522 (*arm_adddi3): Likewise.
3523 (*adddi_sesidi_di): Likewise.
3524 (*adddi_zesidi_di): Likewise.
3525 (*thumb1_addsi3): Likewise.
3526 (addsi3_compare0): Likewise.
3527 (*addsi3_compare0_scratch): Likewise.
3528 (*compare_negsi_si): Likewise.
3529 (cmpsi2_addneg): Likewise.
3530 (*addsi3_carryin_<optab>): Likewise.
3531 (*addsi3_carryin_alt2_<optab>): Likewise.
3532 (*addsi3_carryin_clobercc_<optab>): Likewise.
3533 (*subsi3_carryin): Likewise.
3534 (*subsi3_carryin_const): Likewise.
3535 (*subsi3_carryin_compare): Likewise.
3536 (*subsi3_carryin_compare_const): Likewise.
3537 (*arm_subdi3): Likewise.
3538 (*thumb_subdi3): Likewise.
3539 (*subdi_di_zesidi): Likewise.
3540 (*subdi_di_sesidi): Likewise.
3541 (*subdi_zesidi_di): Likewise.
3542 (*subdi_sesidi_di): Likewise.
3543 (*subdi_zesidi_ze): Likewise.
3544 (thumb1_subsi3_insn): Likewise.
3545 (*arm_subsi3_insn): Likewise.
3546 (*anddi3_insn): Likewise.
3547 (*anddi_zesidi_di): Likewise.
3548 (*anddi_sesdi_di): Likewise.
3549 (*ne_zeroextracts): Likewise.
3550 (*ne_zeroextracts): Likewise.
3551 (*ite_ne_zeroextr): Likewise.
3552 (*ite_ne_zeroextr): Likewise.
3553 (*anddi_notdi_di): Likewise.
3554 (*anddi_notzesidi): Likewise.
3555 (*anddi_notsesidi): Likewise.
3556 (andsi_notsi_si): Likewise.
3557 (thumb1_bicsi3): Likewise.
3558 (*iordi3_insn): Likewise.
3559 (*iordi_zesidi_di): Likewise.
3560 (*iordi_sesidi_di): Likewise.
3561 (*thumb1_iorsi3_insn): Likewise.
3562 (*xordi3_insn): Likewise.
3563 (*xordi_zesidi_di): Likewise.
3564 (*xordi_sesidi_di): Likewise.
3565 (*arm_xorsi3): Likewise.
3566 (*andsi_iorsi3_no): Likewise.
3567 (*smax_0): Likewise.
3568 (*smax_m1): Likewise.
3569 (*arm_smax_insn): Likewise.
3570 (*smin_0): Likewise.
3571 (*arm_smin_insn): Likewise.
3572 (*arm_umaxsi3): Likewise.
3573 (*arm_uminsi3): Likewise.
3574 (*minmax_arithsi): Likewise.
3575 (*minmax_arithsi_): Likewise.
3576 (*satsi_<SAT:code>): Likewise.
3577 (arm_ashldi3_1bit): Likewise.
3578 (arm_ashrdi3_1bit): Likewise.
3579 (arm_lshrdi3_1bit): Likewise.
3580 (*arm_negdi2): Likewise.
3581 (*thumb1_negdi2): Likewise.
3582 (*arm_negsi2): Likewise.
3583 (*thumb1_negsi2): Likewise.
3584 (*negdi_extendsid): Likewise.
3585 (*negdi_zero_extend): Likewise.
3586 (*arm_abssi2): Likewise.
3587 (*thumb1_abssi2): Likewise.
3588 (*arm_neg_abssi2): Likewise.
3589 (*thumb1_neg_abss): Likewise.
3590 (one_cmpldi2): Likewise.
3591 (extend<mode>di2): Likewise.
3592 (*compareqi_eq0): Likewise.
3593 (*arm_extendhisi2addsi): Likewise.
3594 (*arm_movdi): Likewise.
3595 (*thumb1_movdi_insn): Likewise.
3596 (*arm_movt): Likewise.
3597 (*thumb1_movsi_insn): Likewise.
3598 (pic_add_dot_plus_four): Likewise.
3599 (pic_add_dot_plus_eight): Likewise.
3600 (tls_load_dot_plus_eight): Likewise.
3601 (*thumb1_movhi_insn): Likewise.
3602 (*thumb1_movsf_insn): Likewise.
3603 (*movdf_soft_insn): Likewise.
3604 (*thumb_movdf_insn): Likewise.
3605 (cbranchsi4_insn): Likewise.
3606 (cbranchsi4_scratch): Likewise.
3607 (*negated_cbranchsi4): Likewise.
3608 (*tbit_cbranch): Likewise.
3609 (*tlobits_cbranch): Likewise.
3610 (*tstsi3_cbranch): Likewise.
3611 (*cbranchne_decr1): Likewise.
3612 (*addsi3_cbranch): Likewise.
3613 (*addsi3_cbranch_scratch): Likewise.
3614 (*arm_cmpdi_insn): Likewise.
3615 (*arm_cmpdi_unsig): Likewise.
3616 (*arm_cmpdi_zero): Likewise.
3617 (*thumb_cmpdi_zero): Likewise.
3618 (*deleted_compare): Likewise.
3619 (*mov_scc): Likewise.
3620 (*mov_negscc): Likewise.
3621 (*mov_notscc): Likewise.
3622 (*cstoresi_eq0_thumb1_insn): Likewise.
3623 (cstoresi_nltu_thumb1): Likewise.
3624 (cstoresi_ltu_thu): Likewise.
3625 (thumb1_addsi3_addgeu): Likewise.
3626 (*arm_jump): Likewise.
3627 (*thumb_jump): Likewise.
3628 (*check_arch2): Likewise.
3629 (arm_casesi_internal): Likewise.
3630 (thumb1_casesi_dispatch): Likewise.
3631 (*arm_indirect_jump): Likewise.
3632 (*thumb1_indirect_jump): Likewise.
3634 (*and_scc): Likewise.
3635 (*ior_scc): Likewise.
3636 (*compare_scc): Likewise.
3637 (*cond_move): Likewise.
3638 (*cond_arith): Likewise.
3639 (*cond_sub): Likewise.
3640 (*cmp_ite0): Likewise.
3641 (*cmp_ite1): Likewise.
3642 (*cmp_and): Likewise.
3643 (*cmp_ior): Likewise.
3644 (*ior_scc_scc): Likewise.
3645 (*ior_scc_scc_cmp): Likewise.
3646 (*and_scc_scc): Likewise.
3647 (*and_scc_scc_cmp): Likewise.
3648 (*and_scc_scc_nod): Likewise.
3649 (*negscc): Likewise.
3650 (movcond_addsi): Likewise.
3651 (movcond): Likewise.
3652 (*ifcompare_plus_move): Likewise.
3653 (*if_plus_move): Likewise.
3654 (*ifcompare_move_plus): Likewise.
3655 (*if_move_plus): Likewise.
3656 (*ifcompare_arith_arith): Likewise.
3657 (*if_arith_arith): Likewise.
3658 (*ifcompare_arith_move): Likewise.
3659 (*if_arith_move): Likewise.
3660 (*ifcompare_move_arith): Likewise.
3661 (*if_move_arith): Likewise.
3662 (*ifcompare_move_not): Likewise.
3663 (*if_move_not): Likewise.
3664 (*ifcompare_not_move): Likewise.
3665 (*if_not_move): Likewise.
3666 (*ifcompare_shift_move): Likewise.
3667 (*if_shift_move): Likewise.
3668 (*ifcompare_move_shift): Likewise.
3669 (*if_move_shift): Likewise.
3670 (*ifcompare_shift_shift): Likewise.
3671 (*ifcompare_not_arith): Likewise.
3672 (*ifcompare_arith_not): Likewise.
3673 (*if_arith_not): Likewise.
3674 (*ifcompare_neg_move): Likewise.
3675 (*if_neg_move): Likewise.
3676 (*ifcompare_move_neg): Likewise.
3677 (*if_move_neg): Likewise.
3678 (prologue_thumb1_interwork): Likewise.
3679 (*cond_move_not): Likewise.
3680 (*sign_extract_onebit): Likewise.
3681 (*not_signextract_onebit): Likewise.
3682 (stack_tie): Likewise.
3683 (align_4): Likewise.
3684 (align_8): Likewise.
3685 (consttable_end): Likewise.
3686 (consttable_1): Likewise.
3687 (consttable_2): Likewise.
3688 (consttable_4): Likewise.
3689 (consttable_8): Likewise.
3690 (consttable_16): Likewise.
3691 (*thumb1_tablejump): Likewise.
3692 (prefetch): Likewise.
3693 (force_register_use): Likewise.
3694 (thumb_eh_return): Likewise.
3695 (load_tp_hard): Likewise.
3696 (load_tp_soft): Likewise.
3697 (tlscall): Likewise.
3698 (*arm_movtas_ze): Likewise.
3699 (*arm_rev): Likewise.
3700 (*arm_revsh): Likewise.
3701 (*arm_rev16): Likewise.
3702 * config/arm/thumb2.md
3703 (*thumb2_smaxsi3): Likewise.
3704 (*thumb2_sminsi3): Likewise.
3705 (*thumb32_umaxsi3): Likewise.
3706 (*thumb2_uminsi3): Likewise.
3707 (*thumb2_negdi2): Likewise.
3708 (*thumb2_abssi2): Likewise.
3709 (*thumb2_neg_abss): Likewise.
3710 (*thumb2_movsi_insn): Likewise.
3711 (tls_load_dot_plus_four): Likewise.
3712 (*thumb2_movhi_insn): Likewise.
3713 (*thumb2_mov_scc): Likewise.
3714 (*thumb2_mov_negs): Likewise.
3715 (*thumb2_mov_negs): Likewise.
3716 (*thumb2_mov_nots): Likewise.
3717 (*thumb2_mov_nots): Likewise.
3718 (*thumb2_movsicc_): Likewise.
3719 (*thumb2_movsfcc_soft_insn): Likewise.
3720 (*thumb2_indirect_jump): Likewise.
3721 (*thumb2_and_scc): Likewise.
3722 (*thumb2_ior_scc): Likewise.
3723 (*thumb2_ior_scc_strict_it): Likewise.
3724 (*thumb2_cond_move): Likewise.
3725 (*thumb2_cond_arith): Likewise.
3726 (*thumb2_cond_ari): Likewise.
3727 (*thumb2_cond_sub): Likewise.
3728 (*thumb2_negscc): Likewise.
3729 (*thumb2_movcond): Likewise.
3730 (thumb2_casesi_internal): Likewise.
3731 (thumb2_casesi_internal_pic): Likewise.
3732 (*thumb2_alusi3_short): Likewise.
3733 (*thumb2_mov<mode>_shortim): Likewise.
3734 (*thumb2_addsi_short): Likewise.
3735 (*thumb2_subsi_short): Likewise.
3736 (thumb2_addsi3_compare0): Likewise.
3737 (*thumb2_cbz): Likewise.
3738 (*thumb2_cbnz): Likewise.
3739 (*thumb2_one_cmplsi2_short): Likewise.
3740 (*thumb2_negsi2_short): Likewise.
3741 (*orsi_notsi_si): Likewise.
3742 * config/arm/arm1020e.md: Update with new attributes.
3743 * config/arm/arm1026ejs.md: Update with new attributes.
3744 * config/arm/arm1136jfs.md: Update with new attributes.
3745 * config/arm/arm926ejs.md: Update with new attributes.
3746 * config/arm/cortex-a15.md: Update with new attributes.
3747 * config/arm/cortex-a5.md: Update with new attributes.
3748 * config/arm/cortex-a53.md: Update with new attributes.
3749 * config/arm/cortex-a7.md: Update with new attributes.
3750 * config/arm/cortex-a8.md: Update with new attributes.
3751 * config/arm/cortex-a9.md: Update with new attributes.
3752 * config/arm/cortex-m4.md: Update with new attributes.
3753 * config/arm/cortex-r4.md: Update with new attributes.
3754 * config/arm/fa526.md: Update with new attributes.
3755 * config/arm/fa606te.md: Update with new attributes.
3756 * config/arm/fa626te.md: Update with new attributes.
3757 * config/arm/fa726te.md: Update with new attributes.
3759 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3761 * config/aarch64/aarch64-simd.md
3762 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
3763 <vwx> iterator to ensure correct register choice.
3764 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3765 (aarch64_sqdmull_n<mode>): Likewise.
3766 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3767 * config/aarch64/arm_neon.h
3768 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
3769 (vml<as><q>_n_<su>16): Likewise.
3770 (vml<as>l_high_lane<q>_<su>16): Likewise.
3771 (vml<as>l_high_n_<su>16): Likewise.
3772 (vml<as>l_lane<q>_<su>16): Likewise.
3773 (vml<as>l_n_<su>16): Likewise.
3774 (vmul<q>_lane<q>_<su>16): Likewise.
3775 (vmul<q>_n_<su>16): Likewise.
3776 (vmull_lane<q>_<su>16): Likewise.
3777 (vmull_n_<su>16): Likewise.
3778 (vmull_high_lane<q>_<su>16): Likewise.
3779 (vmull_high_n_<su>16): Likewise.
3780 (vqrdmulh<q>_n_s16): Likewise.
3782 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
3784 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
3785 have the correct lane parameter.
3787 2013-09-06 Richard Biener <rguenther@suse.de>
3789 * cfganal.c (control_dependences::~control_dependences):
3790 Properly free all of the vector.
3792 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
3795 * config/i386/i386.c (ix86_conditional_register_usage):
3796 Proper initialize extended SSE registers.
3798 2013-09-06 Jan Hubicka <jh@suse.cz>
3800 PR tree-optimization/58311
3801 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
3803 2013-09-06 Jan Hubicka <jh@suse.cz>
3805 * Makefile.in (tree-sra.o): Update dependencies.
3806 * tree-sra.c: Include ipa-utils.h
3807 (scan_function): Use recursive_call_p.
3808 (has_caller_p): New function.
3809 (cgraph_for_node_and_aliases): Count also callers of aliases.
3811 2013-09-06 Jan Hubicka <jh@suse.cz>
3814 * cgraph.h (symtab_semantically_equivalent_p): Declare.
3815 * tree-tailcall.c: Include ipa-utils.h.
3816 (find_tail_calls): Use it.
3817 * ipa-pure-const.c (check_call): Likewise.
3818 * ipa-utils.c (recursive_call_p): New function.
3819 * ipa-utils.h (recursive_call_p): Dclare.
3820 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
3821 (symtab_semantically_equivalent_p): New function.
3822 * Makefile.in (tree-tailcall.o): Update dependencies.
3824 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3826 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
3829 2013-09-06 Richard Biener <rguenther@suse.de>
3831 * lto-streamer.h (lto_global_var_decls): Remove.
3832 * Makefile.in (OBJS): Remove lto-symtab.o.
3833 (lto-symtab.o): Remove.
3834 (GTFILES): Remove lto-symtab.c
3835 * lto-symtab.c: Move to lto/
3837 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3839 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
3840 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
3841 (UNSPEC_FPINT_RINT): New constant definitions.
3842 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
3843 definition with 2 attributes.
3844 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
3845 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
3848 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3850 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
3851 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
3852 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
3853 attribute to "z196_alone".
3854 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
3857 2013-09-06 Richard Biener <rguenther@suse.de>
3859 * basic-block.h (class control_dependences): New.
3860 * tree-ssa-dce.c (control_dependence_map): Remove.
3862 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
3863 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
3864 find_pdom, find_control_dependence, find_all_control_dependences):
3866 (mark_control_dependent_edges_necessary,
3867 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
3868 tree_dce_done, perform_tree_ssa_dce): Adjust.
3869 * cfganal.c (set_control_dependence_map_bit,
3870 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
3871 find_all_control_dependences): Move from tree-ssa-dce.c and
3872 implement as methods of control_dependences class.
3873 (control_dependences::control_dependences): New.
3874 (control_dependences::~control_dependences): Likewise.
3875 (control_dependences::get_edges_dependent_on): Likewise.
3876 (control_dependences::get_edge): Likewise.
3878 2013-09-04 Jan Hubicka <jh@suse.cz>
3880 * tree.c (types_same_for_odr): Drop overactive check.
3881 * ipa-devirt.c (hash_type_name): Likewise.
3883 2013-09-04 Jan Hubicka <jh@suse.cz>
3885 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
3886 (analyze_functions): ... here.
3888 2013-09-04 Jan Hubicka <jh@suse.cz>
3891 * cgraphunit.c (analyze_functions): Clear AUX fields
3892 after processing; initialize assembler name has.
3894 2013-09-05 Jeff Law <law@redhat.com>
3896 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
3897 from thread_around_empty_block. Record threading path into PATH.
3898 Recurse if threading through the initial block is successful.
3899 (thread_across_edge): Corresponding changes to slightly simplify.
3901 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3903 * config/aarch64/aarch64.md
3904 (type): Remove frecpe, frecps, frecpx.
3905 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
3906 fix to be a TARGET_SIMD instruction.
3907 (aarch64_frecps): Remove.
3908 * config/aarch64/aarch64-simd.md
3909 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
3910 (aarch64_frecps<mode>): Handle all float/vector of float modes.
3912 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3913 Sofiane Naci <sofiane.naci@arm.com>
3915 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
3916 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
3917 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
3918 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
3919 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
3920 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
3921 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
3922 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
3923 into "clz, "rbit". Rename "shift" to "shift_imm".
3924 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
3925 changes. Update for attribute changes all occurrences of arlo_* and
3927 * config/arm/arm-fixed.md: Update for attribute changes
3928 all occurrences of arlo_* types.
3929 * config/arm/thumb2.md: Update for attribute changes all occurrences
3931 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
3932 (cortexa7_older_only): Likewise.
3933 (cortexa7_younger): Likewise.
3934 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3935 (1020alu_shift_op): Likewise.
3936 (1020alu_shift_reg_op): Likewise.
3937 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3938 (alu_shift_op): Likewise.
3939 (alu_shift_reg_op): Likewise.
3940 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3941 (11_alu_shift_op): Likewise.
3942 (11_alu_shift_reg_op): Likewise.
3943 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3944 (9_alu_shift_reg_op): Likewise.
3945 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
3947 (cortex_a15_alu_shift): Likewise.
3948 (cortex_a15_alu_shift_reg): Likewise.
3949 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
3951 (cortex_a5_alu_shift): Likewise.
3952 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
3954 (cortex_a53_alu_shift): Likewise.
3955 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
3957 (cortex_a7_alu_reg): Likewise.
3958 (cortex_a7_alu_shift): Likewise.
3959 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3961 (cortex_a8_alu_shift): Likewise.
3962 (cortex_a8_alu_shift_reg): Likewise.
3963 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3964 (cortex_a9_dp_shift): Likewise.
3965 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3967 * config/arm/cortex-r4.md
3968 (cortex_r4_alu): Update for attribute changes.
3969 (cortex_r4_mov): Likewise.
3970 (cortex_r4_alu_shift_reg): Likewise.
3971 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3972 (526_alu_shift_op): Likewise.
3973 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3974 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3975 (626te_alu_shift_op): Likewise.
3976 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3977 (726te_alu_shift_op): Likewise.
3978 (726te_alu_shift_reg_op): Likewise.
3979 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3980 (mp626_alu_shift_op): Likewise.
3981 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3982 (pj4_alu_conds): Likewise.
3983 (pj4_shift): Likewise.
3984 (pj4_shift_conds): Likewise.
3985 (pj4_alu_shift): Likewise.
3986 (pj4_alu_shift_conds): Likewise.
3987 * config/aarch64/aarch64.md: Update for attribute change
3988 all occurrences of arlo_* and shift* types.
3990 2013-09-05 Mike Stump <mikestump@comcast.net>
3992 * tree.h: Move documentation for tree_function_decl to tree-core.h
3993 with the declaration.
3995 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3998 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3999 looking for widest mode.
4001 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
4003 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
4005 2013-09-05 Richard Biener <rguenther@suse.de>
4007 PR tree-optimization/58137
4008 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
4009 Do not create vectors of pointers.
4010 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
4011 types for the components of the vector initializer.
4012 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
4013 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
4015 2013-09-05 Martin Jambor <mjambor@suse.cz>
4017 * ipa-prop.c (remove_described_reference): Accept missing references,
4018 return false if that hppens, otherwise return true.
4019 (cgraph_node_for_jfunc): New function.
4020 (try_decrement_rdesc_refcount): Likewise.
4021 (try_make_edge_direct_simple_call): Use them.
4022 (ipa_edge_removal_hook): Remove references from rdescs.
4023 (ipa_edge_duplication_hook): Clone rdescs and their references
4024 when the new edge has the same caller as the old one.
4025 * cgraph.c (cgraph_resolve_speculation): Remove speculative
4026 reference before removing any edges.
4028 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
4030 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
4032 * thumb2.md (thumb2_storewb_parisi): New pattern.
4034 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
4036 * config/aarch64/aarch64-option-extensions.def: Add
4037 AARCH64_OPT_EXTENSION of 'crc'.
4038 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
4039 (AARCH64_ISA_CRC): Ditto.
4040 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
4041 description of the CRC extension.
4043 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
4045 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
4046 * config/rs6000/linux.h: Ditto.
4047 * alpha/linux.h: Ditto.
4048 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
4049 no_c99_libc_has_function.
4050 * config/c6x/uclinux-elf.h: Ditto.
4051 * config/lm32/uclinux-elf.h: Ditto.
4052 * config/m68k/uclinux.h: Ditto.
4053 * config/moxie/uclinux.h: Ditto.
4054 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
4055 (crisv32-*-linux*, cris-*-linux*): Ditto.
4056 * config/bfin/bfin.c: Include "tm_p.h".
4058 2013-09-05 Richard Biener <rguenther@suse.de>
4060 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
4061 check for a definition without a basic-block.
4063 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4064 Sofiane Naci <sofiane.naci@arm.com>
4066 * config/aarch64/aarch64.md
4067 (*movti_aarch64): Rename r_2_f and f_2_r.
4068 (*movsf_aarch64): Likewise.
4069 (*movdf_aarch64): Likewise.
4070 (*movtf_aarch64): Likewise.
4071 (aarch64_movdi_<mode>low): Likewise.
4072 (aarch64_movdi_<mode>high): Likewise.
4073 (aarch64_mov<mode>high_di): Likewise.
4074 (aarch64_mov<mode>low_di): Likewise.
4075 (aarch64_movtilow_tilow): Likewise.
4076 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
4077 values to config/arm/types.md
4078 (attribute "conds"): Update for attribute change.
4079 (anddi3_insn): Likewise.
4080 (iordi3_insn): Likewise.
4081 (xordi3_insn): Likewise.
4082 (one_cmpldi2): Likewise.
4083 * config/arm/types.md (type): Add Neon types.
4084 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
4085 use "type" attribute.
4086 (movmisalign<mode>_neon_store): Likewise.
4087 (movmisalign<mode>_neon_load): Likewise.
4088 (vec_set<mode>_internal): Likewise.
4089 (vec_setv2di_internal): Likewise.
4090 (vec_extract<mode>): Likewise.
4091 (vec_extractv2di): Likewise.
4092 (add<mode>3_neon): Likewise.
4093 (adddi3_neon): Likewise.
4094 (sub<mode>3_neon): Likewise.
4095 (subdi3_neon): Likewise.
4096 (mul<mode>3_neon): Likewise.
4097 (mul<mode>3add<mode>_neon): Likewise.
4098 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
4099 (fma<VCVTF:mode>4)): Likewise.
4100 (fma<VCVTF:mode>4_intrinsic): Likewise.
4101 (fmsub<VCVTF:mode>4)): Likewise.
4102 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4103 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4104 (ior<mode>3): Likewise.
4105 (and<mode>3): Likewise.
4106 (anddi3_neon): Likewise.
4107 (orn<mode>3_neon): Likewise.
4108 (orndi3_neon): Likewise.
4109 (bic<mode>3_neon): Likewise.
4110 (bicdi3_neon): Likewise.
4111 (xor<mode>3): Likewise.
4112 (one_cmpl<mode>2): Likewise.
4113 (abs<mode>2): Likewise.
4114 (neg<mode>2): Likewise.
4115 (umin<mode>3_neon): Likewise.
4116 (umax<mode>3_neon): Likewise.
4117 (smin<mode>3_neon): Likewise.
4118 (smax<mode>3_neon): Likewise.
4119 (vashl<mode>3): Likewise.
4120 (vashr<mode>3_imm): Likewise.
4121 (vlshr<mode>3_imm): Likewise.
4122 (ashl<mode>3_signed): Likewise.
4123 (ashl<mode>3_unsigned): Likewise.
4124 (neon_load_count): Likewise.
4125 (ashldi3_neon_noclobber): Likewise.
4126 (signed_shift_di3_neon): Likewise.
4127 (unsigned_shift_di3_neon): Likewise.
4128 (ashrdi3_neon_imm_noclobber): Likewise.
4129 (lshrdi3_neon_imm_noclobber): Likewise.
4130 (widen_ssum<mode>3): Likewise.
4131 (widen_usum<mode>3): Likewise.
4132 (quad_halves_<code>v4si): Likewise.
4133 (quad_halves_<code>v4sf): Likewise.
4134 (quad_halves_<code>v8hi): Likewise.
4135 (quad_halves_<code>v16qi): Likewise.
4136 (reduc_splus_v2di): Likewise.
4137 (neon_vpadd_internal<mode>): Likewise.
4138 (neon_vpsmin<mode>): Likewise.
4139 (neon_vpsmax<mode>): Likewise.
4140 (neon_vpumin<mode>): Likewise.
4141 (neon_vpumax<mode>): Likewise.
4142 (ss_add<mode>_neon): Likewise.
4143 (us_add<mode>_neon): Likewise.
4144 (ss_sub<mode>_neon): Likewise.
4145 (us_sub<mode>_neon): Likewise.
4146 (neon_vadd<mode>_unspec): Likewise.
4147 (neon_vaddl<mode>): Likewise.
4148 (neon_vaddw<mode>): Likewise.
4149 (neon_vhadd<mode>): Likewise.
4150 (neon_vqadd<mode>): Likewise.
4151 (neon_vaddhn<mode>): Likewise.
4152 (neon_vmul<mode>): Likewise.
4153 (neon_vmla<mode>): Likewise.
4154 (neon_vmlal<mode>): Likewise.
4155 (neon_vmls<mode>): Likewise.
4156 (neon_vmlsl<mode>): Likewise.
4157 (neon_vqdmulh<mode>): Likewise.
4158 (neon_vqdmlal<mode>): Likewise.
4159 (neon_vqdmlsl<mode>): Likewise.
4160 (neon_vmull<mode>): Likewise.
4161 (neon_vqdmull<mode>): Likewise.
4162 (neon_vsub<mode>_unspec): Likewise.
4163 (neon_vsubl<mode>): Likewise.
4164 (neon_vsubw<mode>): Likewise.
4165 (neon_vqsub<mode>): Likewise.
4166 (neon_vhsub<mode>): Likewise.
4167 (neon_vsubhn<mode>): Likewise.
4168 (neon_vceq<mode>): Likewise.
4169 (neon_vcge<mode>): Likewise.
4170 (neon_vcgeu<mode>): Likewise.
4171 (neon_vcgt<mode>): Likewise.
4172 (neon_vcgtu<mode>): Likewise.
4173 (neon_vcle<mode>): Likewise.
4174 (neon_vclt<mode>): Likewise.
4175 (neon_vcage<mode>): Likewise.
4176 (neon_vcagt<mode>): Likewise.
4177 (neon_vtst<mode>): Likewise.
4178 (neon_vabd<mode>): Likewise.
4179 (neon_vabdl<mode>): Likewise.
4180 (neon_vaba<mode>): Likewise.
4181 (neon_vabal<mode>): Likewise.
4182 (neon_vmax<mode>): Likewise.
4183 (neon_vmin<mode>): Likewise.
4184 (neon_vpaddl<mode>): Likewise.
4185 (neon_vpadal<mode>): Likewise.
4186 (neon_vpmax<mode>): Likewise.
4187 (neon_vpmin<mode>): Likewise.
4188 (neon_vrecps<mode>): Likewise.
4189 (neon_vrsqrts<mode>): Likewise.
4190 (neon_vqabs<mode>): Likewise.
4191 (neon_vqneg<mode>): Likewise.
4192 (neon_vcls<mode>): Likewise.
4193 (clz<mode>2): Likewise.
4194 (popcount<mode>2): Likewise.
4195 (neon_vrecpe): Likewise.
4196 (neon_vrsqrte): Likewise.
4197 (neon_vget_lane<mode>_sext_internal): Likewise.
4198 (neon_vget_lane<mode>_zext_internal): Likewise.
4199 (neon_vdup_n<mode>): Likewise.
4200 (neon_vdup_nv2di): Likewise.
4201 (neon_vdpu_lane<mode>_internal): Likewise.
4202 (neon_vswp<mode>): Likewise.
4203 (float<mode><V_cvtto>2): Likewise.
4204 (floatuns<mode><V_cvtto>2): Likewise.
4205 (fix_trunc<mode><V_cvtto>)2): Likewise
4206 (fixuns_trunc<mode><V_cvtto)2): Likewise.
4207 (neon_vcvt<mode>): Likewise.
4208 (neon_vcvtv4sfv4hf): Likewise.
4209 (neon_vcvtv4hfv4sf): Likewise.
4210 (neon_vcvt_n<mode>): Likewise.
4211 (neon_vmovn<mode>): Likewise.
4212 (neon_vqmovn<mode>): Likewise.
4213 (neon_vqmovun<mode>): Likewise.
4214 (neon_vmovl<mode>): Likewise.
4215 (neon_vmul_lane<mode>): Likewise.
4216 (neon_vmull_lane<mode>): Likewise.
4217 (neon_vqdmull_lane<mode>): Likewise.
4218 (neon_vqdmulh_lane<mode>): Likewise.
4219 (neon_vmla_lane<mode>): Likewise.
4220 (neon_vmlal_lane<mode>): Likewise.
4221 (neon_vqdmlal_lane<mode>): Likewise.
4222 (neon_vmls_lane<mode>): Likewise.
4223 (neon_vmlsl_lane<mode>): Likewise.
4224 (neon_vqdmlsl_lane<mode>): Likewise.
4225 (neon_vext<mode>): Likewise.
4226 (neon_vrev64<mode>): Likewise.
4227 (neon_vrev32<mode>): Likewise.
4228 (neon_vrev16<mode>): Likewise.
4229 (neon_vbsl<mode>_internal): Likewise.
4230 (neon_vshl<mode>): Likewise.
4231 (neon_vqshl<mode>): Likewise.
4232 (neon_vshr_n<mode>): Likewise.
4233 (neon_vshrn_n<mode>): Likewise.
4234 (neon_vqshrn_n<mode>): Likewise.
4235 (neon_vqshrun_n<mode>): Likewise.
4236 (neon_vshl_n<mode>): Likewise.
4237 (neon_vqshl_n<mode>): Likewise.
4238 (neon_vqshlu_n<mode>): Likewise.
4239 (neon_vshll_n<mode>): Likewise.
4240 (neon_vsra_n<mode>): Likewise.
4241 (neon_vsri_n<mode>): Likewise.
4242 (neon_vsli_n<mode>): Likewise.
4243 (neon_vtbl1v8qi): Likewise.
4244 (neon_vtbl2v8qi): Likewise.
4245 (neon_vtbl3v8qi): Likewise.
4246 (neon_vtbl4v8qi): Likewise.
4247 (neon_vtbx1v8qi): Likewise.
4248 (neon_vtbx2v8qi): Likewise.
4249 (neon_vtbx3v8qi): Likewise.
4250 (neon_vtbx4v8qi): Likewise.
4251 (neon_vtrn<mode>_internal): Likewise.
4252 (neon_vzip<mode>_internal): Likewise.
4253 (neon_vuzp<mode>_internal): Likewise.
4254 (neon_vld1<mode>): Likewise.
4255 (neon_vld1_lane<mode>): Likewise.
4256 (neon_vld1_dup<mode>): Likewise.
4257 (neon_vld1_dupv2di): Likewise.
4258 (neon_vst1<mode>): Likewise.
4259 (neon_vst1_lane<mode>): Likewise.
4260 (neon_vld2<mode>): Likewise.
4261 (neon_vld2_lane<mode>): Likewise.
4262 (neon_vld2_dup<mode>): Likewise.
4263 (neon_vst2<mode>): Likewise.
4264 (neon_vst2_lane<mode>): Likewise.
4265 (neon_vld3<mode>): Likewise.
4266 (neon_vld3qa<mode>): Likewise.
4267 (neon_vld3qb<mode>): Likewise.
4268 (neon_vld3_lane<mode>): Likewise.
4269 (neon_vld3_dup<mode>): Likewise.
4270 (neon_vst3<mode>): Likewise.
4271 (neon_vst3qa<mode>): Likewise.
4272 (neon_vst3qb<mode>): Likewise.
4273 (neon_vst3_lane<mode>): Likewise.
4274 (neon_vld4<mode>): Likewise.
4275 (neon_vld4qa<mode>): Likewise.
4276 (neon_vld4qb<mode>): Likewise.
4277 (neon_vld4_lane<mode>): Likewise.
4278 (neon_vld4_dup<mode>): Likewise.
4279 (neon_vst4<mode>): Likewise.
4280 (neon_vst4qa<mode>): Likewise.
4281 (neon_vst4qb<mode>): Likewise.
4282 (neon_vst4_lane<mode>): Likewise.
4283 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4284 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4285 (neon_vec_<US>mult_lo_<mode>): Likewise.
4286 (neon_vec_<US>mult_hi_<mode>): Likewise.
4287 (neon_vec_<US>shiftl_<mode>): Likewise.
4288 (neon_unpack<US>_<mode>): Likewise.
4289 (neon_vec_<US>mult_<mode>): Likewise.
4290 (vec_pack_trunc_<mode>): Likewise.
4291 (neon_vec_pack_trunk_<mode>): Likewise.
4292 (neon_vabd<mode>_2): Likewise.
4293 (neon_vabd<mode>_3): Likewise.
4294 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4295 (thumb2_movsi_vfp): Likewise.
4296 (movdi_vfp): Likewise.
4297 (movdi_vfp_cortexa8): Likewise.
4298 (movhf_vfp_neon): Likewise.
4299 (movhf_vfp): Likewiwse.
4300 (movsf_vfp): Likewiwse.
4301 (thumb2_movsf_vfp): Likewiwse.
4302 (movdf_vfp): Likewise.
4303 (thumb2_movdf_vfp): Likewise.
4304 (movsfcc_vfp): Likewise.
4305 (thumb2_movsfcc_vfp): Likewise.
4306 (movdfcc_vfp): Likewise.
4307 (thumb2_movdfcc_vfp): Likewise.
4308 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
4309 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
4310 (v10_v2c): Likewise.
4311 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
4313 (cortex_a15_neon_int_2): Likewise.
4314 (cortex_a15_neon_int_3): Likewise.
4315 (cortex_a15_neon_int_4): Likewise.
4316 (cortex_a15_neon_int_5): Likewise.
4317 (cortex_a15_neon_vqneg_vqabs): Likewise.
4318 (cortex_a15_neon_vmov): Likewise.
4319 (cortex_a15_neon_vaba): Likewise.
4320 (cortex_a15_neon_vaba_qqq): Likewise.
4321 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4322 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4323 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
4324 scalar_64_32_long_scalar): Likewise.
4325 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4326 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4327 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4328 lotype_qdd_64_32_long): Likewise.
4329 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4330 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4331 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4332 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4333 (cortex_a15_neon_shift_1): Likewise.
4334 (cortex_a15_neon_shift_2): Likewise.
4335 (cortex_a15_neon_shift_3): Likewise.
4336 (cortex_a15_neon_vshl_ddd): Likewise.
4337 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4338 (cortex_a15_neon_vsra_vrsra): Likewise.
4339 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4340 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4341 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4342 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4343 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4344 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4345 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4346 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4347 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4348 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4349 (cortex_a15_neon_bp_simple): Likewise.
4350 (cortex_a15_neon_bp_2cycle): Likewise.
4351 (cortex_a15_neon_bp_3cycle): Likewise.
4352 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4353 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4354 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4355 (cortex_a15_neon_vld2_4_regs): Likewise.
4356 (cortex_a15_neon_vld3_vld4): Likewise.
4357 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4358 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4359 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
4360 (cortex_a15_neon_vst3_vst4): Likewise.
4361 (cortex_a15_neon_vld1_vld2_lane): Likewise.
4362 (cortex_a15_neon_vld3_vld4_lane" 10
4363 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4364 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4365 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4366 (cortex_a15_neon_ldm_2): Likewise.0
4367 (cortex_a15_neon_stm_2): Likewise.
4368 (cortex_a15_neon_mcr): Likewise.
4369 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4370 (cortex_a15_neon_mrc): Likewise.
4371 (cortex_a15_neon_mrrc): Likewise.
4372 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4374 (cortex_a15_alu_shift): Likewise.
4375 (cortex_a15_alu_shift_reg): Likewise.
4376 (cortex_a15_mult32): Likewise.
4377 (cortex_a15_mult64): Likewise.
4378 (cortex_a15_block): Likewise.
4379 (cortex_a15_branch): Likewise.
4380 (cortex_a15_load1): Likewise.
4381 (cortex_a15_load3): Likewise.
4382 (cortex_a15_store1): Likewise.
4383 (cortex_a15_store3): Likewise.
4384 (cortex_a15_call): Likewise.
4385 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
4386 (cortex_a5_f2r): Likewise.
4387 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
4389 (cortex_a53_f2r): Likewise.
4390 * config/arm/cortex-a7.md
4391 (cortex_a7_branch): Update for attribute change.
4392 (cortex_a7_call): Likewise.
4393 (cortex_a7_alu_imm): Likewise.
4394 (cortex_a7_alu_reg): Likewise.
4395 (cortex_a7_alu_shift): Likewise.
4396 (cortex_a7_mul): Likewise.
4397 (cortex_a7_load1): Likewise.
4398 (cortex_a7_store1): Likewise.
4399 (cortex_a7_load2): Likewise.
4400 (cortex_a7_store2): Likewise.
4401 (cortex_a7_load3): Likewise.
4402 (cortex_a7_store3): Likewise.
4403 (cortex_a7_load4): Likewise.
4404 (cortex_a7_store4): Likewise.
4405 (cortex_a7_fpalu): Likewise.
4406 (cortex_a7_fconst): Likewise.
4407 (cortex_a7_fpmuls): Likewise.
4408 (cortex_a7_neon_mul): Likewise.
4409 (cortex_a7_fpmacs): Likewise.
4410 (cortex_a7_neon_mla: Likewise.
4411 (cortex_a7_fpmuld: Likewise.
4412 (cortex_a7_fpmacd: Likewise.
4413 (cortex_a7_fpfmad: Likewise.
4414 (cortex_a7_fdivs: Likewise.
4415 (cortex_a7_fdivd: Likewise.
4416 (cortex_a7_r2f: Likewise.
4417 (cortex_a7_f2r: Likewise.
4418 (cortex_a7_f_flags: Likewise.
4419 (cortex_a7_f_loads: Likewise.
4420 (cortex_a7_f_loadd: Likewise.
4421 (cortex_a7_f_stores: Likewise.
4422 (cortex_a7_f_stored: Likewise.
4423 (cortex_a7_neon): Likewise.
4424 * config/arm/cortex-a8-neon.md
4425 (cortex_a8_neon_mrc): Update for attribute change.
4426 (cortex_a8_neon_mrrc): Likewise.
4427 (cortex_a8_neon_int_1): Likewise.
4428 (cortex_a8_neon_int_2): Likewise.
4429 (cortex_a8_neon_int_3): Likewise.
4430 (cortex_a8_neon_int_4): Likewise.
4431 (cortex_a8_neon_int_5): Likewise.
4432 (cortex_a8_neon_vqneg_vqabs): Likewise.
4433 (cortex_a8_neon_vmov): Likewise.
4434 (cortex_a8_neon_vaba): Likewise.
4435 (cortex_a8_neon_vaba_qqq): Likewise.
4436 (cortex_a8_neon_vsma): Likewise.
4437 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4438 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4439 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4440 long_scalar): Likewise.
4441 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4442 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4443 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4444 long_scalar_qdd_64_32_long): Likewise.
4445 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4446 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4447 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4448 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4449 (cortex_a8_neon_shift_1): Likewise.
4450 (cortex_a8_neon_shift_2): Likewise.
4451 (cortex_a8_neon_shift_3): Likewise.
4452 (cortex_a8_neon_vshl_ddd): Likewise.
4453 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4454 (cortex_a8_neon_vsra_vrsra): Likewise.
4455 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4456 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4457 (cortex_a8_neon_fp_vsum): Likewise.
4458 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4459 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4460 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4461 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4462 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4463 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4464 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4465 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4466 (cortex_a8_neon_bp_simple): Likewise.
4467 (cortex_a8_neon_bp_2cycle): Likewise.
4468 (cortex_a8_neon_bp_3cycle): Likewise.
4469 (cortex_a8_neon_ldr): Likewise.
4470 (cortex_a8_neon_str): Likewise.
4471 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4472 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4473 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4474 (cortex_a8_neon_vld2_4_regs): Likewise.
4475 (cortex_a8_neon_vld3_vld4): Likewise.
4476 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4477 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4478 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4479 (cortex_a8_neon_vst3_vst4): Likewise.
4480 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4481 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4482 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4483 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4484 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4485 (cortex_a8_neon_mcr): Likewise.
4486 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4487 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4488 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
4490 (ca9_neon_mrrc): Likewise.
4491 (cortex_a9_neon_int_1): Likewise.
4492 (cortex_a9_neon_int_2): Likewise.
4493 (cortex_a9_neon_int_3): Likewise.
4494 (cortex_a9_neon_int_4): Likewise.
4495 (cortex_a9_neon_int_5): Likewise.
4496 (cortex_a9_neon_vqneg_vqabs): Likewise.
4497 (cortex_a9_neon_vmov): Likewise.
4498 (cortex_a9_neon_vaba): Likewise.
4499 (cortex_a9_neon_vaba_qqq): Likewise.
4500 (cortex_a9_neon_vsma): Likewise.
4501 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4502 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4503 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4504 long_scalar): Likewise.
4505 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4506 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4507 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4508 long_scalar_qdd_64_32_long): Likewise.
4509 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4510 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4511 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4512 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4513 (cortex_a9_neon_shift_1): Likewise.
4514 (cortex_a9_neon_shift_2): Likewise.
4515 (cortex_a9_neon_shift_3): Likewise.
4516 (cortex_a9_neon_vshl_ddd): Likewise.
4517 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4518 (cortex_a9_neon_vsra_vrsra): Likewise.
4519 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4520 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4521 (cortex_a9_neon_fp_vsum): Likewise.
4522 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4523 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4524 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4525 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4526 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4527 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4528 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4529 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4530 (cortex_a9_neon_bp_simple): Likewise.
4531 (cortex_a9_neon_bp_2cycle): Likewise.
4532 (cortex_a9_neon_bp_3cycle): Likewise.
4533 (cortex_a9_neon_ldr): Likewise.
4534 (cortex_a9_neon_str): Likewise.
4535 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4536 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4537 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4538 (cortex_a9_neon_vld2_4_regs): Likewise.
4539 (cortex_a9_neon_vld3_vld4): Likewise.
4540 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4541 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4542 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4543 (cortex_a9_neon_vst3_vst4): Likewise.
4544 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4545 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4546 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4547 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4548 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4549 (cortex_a9_neon_mcr): Likewise.
4550 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4551 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4552 (cortex_a9_fps): Likewise.
4553 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
4555 (cortex_m4_fmuls): Likewise.
4556 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
4558 (cortex_r4_mrc): Likewise.
4559 * config/arm/iterators.md: Update comment referring to neon_type.
4560 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
4561 (iwmmxt_movsi_insn): Likewise.
4562 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
4564 (pj4_core_to_vfp): Likewise.
4565 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
4567 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
4568 (vfp_fstore): Likewise.
4569 * doc/md.texi: Change references to neon_type to refer to type.
4571 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4573 * tree.h (DECL_BUILT_IN): Fix typo in comment.
4575 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
4577 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
4580 2013-09-04 Easwaran Raman <eraman@google.com>
4583 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
4584 (build_and_add_sum): Use it.
4585 (appears_later_in_bb): Simplify code.
4587 2013-09-04 Teresa Johnson <tejohnson@google.com>
4589 * dumpfile.c (dump_finish): Don't close stderr/stdout.
4591 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
4593 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
4595 2013-09-04 Jan Hubicka <jh@suse.cz>
4597 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
4598 * ipa-devirt.c: Include diganostic.h
4599 (odr_type_d): Add types and types_set.
4600 (hash_type_name): Work for types with vtables during LTO.
4601 (odr_hasher::remove): Fix comment; destroy types_set.
4602 (add_type_duplicate): New function,
4603 (get_odr_type): Use it.
4604 (dump_type_inheritance_graph): Dump type duplicates.
4605 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
4607 * tree.c (types_same_for_odr): Give exact answers on types with
4610 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4612 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
4613 explaining their differences.
4615 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
4617 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
4619 2013-09-03 Jeff Law <law@redhat.com>
4621 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
4622 when not threading through a joiner block. Pass joiner/no joiner
4623 state to register_jump_thread.
4624 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
4625 state from argument rather than implying on path length.
4626 Dump the entire jump thread path into debugging dump.
4627 * tree-flow.h (register_jump_thread): Update prototype.
4629 2013-08-29 Xinliang David Li <davidxl@google.com>
4631 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4632 Remove a trivial gcc_assert.
4634 2013-08-29 Xinliang David Li <davidxl@google.com>
4636 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
4637 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
4638 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4639 Delay base decl alignment adjustment.
4640 * tree-vectorizer.c (vect_destroy_datarefs): New function.
4641 * tree-vectorizer.h: New data structure.
4642 (set_dr_misalignment): New function.
4643 (dr_misalignment): Ditto.
4644 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
4645 (vectorizable_load): Ditto.
4646 (ensure_base_align): New function.
4647 (vectorize_loops): Add dbg_cnt support.
4648 (execute_vect_slp): Ditto.
4649 * dbgcnt.def: New debug counter.
4650 * Makefile: New dependency.
4652 2013-09-03 Meador Inge <meadori@codesourcery.com>
4656 2013-08-30 Meador Inge <meadori@codesourcery.com>
4658 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4660 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
4662 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
4663 function descriptor.
4665 2013-09-03 Richard Biener <rguenther@suse.de>
4667 * tree-affine.c (add_elt_to_tree): Fix association issue,
4668 avoid useless converts and make sure to always return a
4669 properly typed result.
4671 2013-09-03 Richard Biener <rguenther@suse.de>
4674 * fold-const.c (negate_expr_p): Fix division case.
4675 (negate_expr): Likewise.
4677 2013-09-03 Richard Biener <rguenther@suse.de>
4680 * tree-streamer-out.c: Include tm.h.
4681 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
4683 2013-09-03 Jan Hubicka <jh@suse.cz>
4685 * tree-profile.c (tree_profiling): Cleanup CFG when done.
4687 2013-09-03 Alan Modra <amodra@gmail.com>
4689 * config.gcc (powerpc*-*-linux*): Add support for little-endian
4690 multilibs to big-endian target and vice versa.
4691 * config/rs6000/t-linux64: Use := assignment on all vars.
4692 (MULTILIB_EXTRA_OPTS): Remove fPIC.
4693 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
4694 * config/rs6000/t-linux64le: New file.
4695 * config/rs6000/t-linux64bele: New file.
4696 * config/rs6000/t-linux64lebe: New file.
4698 2013-09-02 Jan Hubicka <jh@suse.cz>
4700 * ipa-inline-transform.c (inline_transform): Do not
4701 optimize_inline_calls when not optimizing.
4703 2013-09-02 Jan Hubicka <jh@suse.cz>
4705 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
4706 duplicated nodes for assembler names.
4707 * symtab.c (symtab_unregister_node): Do not attempt to unlink
4708 hard registers from assembler name hash.
4710 2013-09-02 Jan Hubicka <jh@suse.cz>
4712 * ipa-split.c (execute_split_functions): Split externally visible
4713 functions called once.
4715 2013-09-02 Martin Jambor <mjambor@suse.cz>
4718 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
4719 linked list. When finding the correct duplicate, also consider also
4720 the caller in additon to its inlined_to node.
4722 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
4724 * config/aarch64/aarch64-simd-builtins.def
4725 (dup_lane_scalar): Remove.
4726 * config/aarch64/aarch64-simd.md
4727 (aarch64_simd_dup): Add 'w->w' alternative.
4728 (aarch64_dup_lane<mode>): Allow for VALL.
4729 (aarch64_dup_lane_scalar<mode>): Remove.
4730 (aarch64_dup_lane_<vswap_width_name><mode>): New.
4731 (aarch64_get_lane_signed<mode>): Add w->w altenative.
4732 (aarch64_get_lane_unsigned<mode>): Likewise.
4733 (aarch64_get_lane<mode>): Likewise.
4734 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
4735 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
4736 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
4737 (VCON): Change container of V2SF.
4738 (vswap_width_name): Likewise.
4739 * config/aarch64/arm_neon.h
4740 (__aarch64_vdup_lane_any): New.
4741 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4742 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
4743 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4745 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
4748 * expr.c (emit_move_complex): Do not move complex FP values as parts if
4749 the source or the destination is a single hard register.
4751 2013-09-02 Richard Biener <rguenther@suse.de>
4754 * tree-scalar-evolution.c (instantiate_scev_name): Allow
4757 2013-09-02 Richard Biener <rguenther@suse.de>
4759 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
4760 arithmetic to sizetype.
4762 2013-09-02 Bin Cheng <bin.cheng@arm.com>
4764 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
4765 Find auto-increment use both before and after candidate.
4767 2013-09-02 Marek Polacek <polacek@redhat.com>
4769 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
4771 2013-09-01 Jan Hubicka <jh@suse.cz>
4773 * Makefile.in: Add ipa-profile.o
4774 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
4775 * cgraph.c (struct cgraph_propagate_frequency_data,
4776 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
4777 ipa-profile.c; replace cgraph_ by ipa_ prefix.
4778 * cgraph.h (cgraph_propagate_frequency): Remove.
4779 * ipa-inline-analysis.c: Include ipa-utils.h;
4780 drop duplicated cfgloop.h.
4781 (inline_update_callee_summaries): Update.
4782 * ipa-profile.c: New file.
4783 * ipa-utils.h (ipa_propagate_frequency): Declare.
4784 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
4785 data-streamer.h, value-prof.h.
4786 (symtab_remove_unreachable_nodes): Update profile.
4787 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
4788 account_time_size, cmp_counts, dump_histogram,
4789 ipa_profile_generate_summary, ipa_profile_write_summary,
4790 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
4791 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
4792 Move to ipa-profile.c.
4794 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
4796 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
4798 2013-09-01 Jan Hubicka <jh@suse.cz>
4800 * common.opt (fdevirtualize-speculatively): New function.
4801 * invoke.texi (fdevirtualize-speculatively): Document.
4802 * ipa-devirt.c: Include ipa-inline.h
4803 (likely_target_p): New function.
4804 (ipa_devirt): New function.
4805 (gate_ipa_devirt): New function.
4806 (pass_data_ipa_devirt): New static var.
4807 (pass_ipa_devirt): Likewise.
4808 (make_pass_ipa_devirt): New function.
4809 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
4810 (common_handle_option): Disable devirtualization when
4811 value range profiling is available.
4812 * passes.def (pass_ipa_devirt): Add.
4813 * timever.def (TV_IPA_DEVIRT): New timevar.
4814 * tree-pass.h (make_pass_ipa_devirt):
4816 2013-09-01 Iain Sandoe <iain@codesourcery.com>
4818 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
4819 include sanitize(undefined).
4821 2013-08-31 Diego Novillo <dnovillo@google.com>
4823 * Makefile.in (TREE_CORE_H): Define.
4825 (GTFILES): Add tree-core.h.
4826 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
4828 * tree-core.h: New file.
4829 Move all data structures, enum, typedefs, global
4830 declarations and constants from ...
4833 2013-08-31 Jan Hubicka <jh@suse.cz>
4835 * bulitins.c (expand_builtin): Do not early exit for gcov
4836 instrumented functions.
4838 2013-08-31 Marek Polacek <polacek@redhat.com>
4840 * ubsan.c: Include tm_p.h.
4842 2013-08-31 Jan Hubicka <jh@suse.cz>
4844 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
4847 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
4848 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
4849 * tree-cfg.c (verify_gimple_label): ... here.
4850 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
4851 (ipa_merge_profiles): New function.
4852 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
4853 (lto_input_function_body): Likewise.
4854 * ipa-utils.h (ipa_merge_profiles): Declare.
4855 * lto-streamer.h (lto_input_function_body): Update prototype.
4856 (emit_label_in_global_context_p): Remove.
4857 * lto-symtab.c: Include ipa-utils.h
4858 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4860 2013-08-31 Jan Hubicka <jh@suse.cz>
4862 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
4864 2013-08-31 Jan Hubicka <jh@suse.cz>
4866 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
4868 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
4870 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
4871 "cmp" RTX before signed_comparison_operator check to account
4874 2013-08-30 Jan Hubicka <jh@suse.cz>
4876 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
4877 (detect_type_change_1): Rename to ...
4878 (detect_type_change): ... this one; early return on non-polymorphic
4880 (detect_type_change_ssa): Add comp_type parameter; update
4881 use of detect_type_change.
4882 (compute_complex_assign_jump_func): Add param_type parameter;
4883 update use of detect_type_change_ssa.
4884 (compute_complex_ancestor_jump_func): Likewise.
4885 (ipa_get_callee_param_type): New function.
4886 (ipa_compute_jump_functions_for_edge): Compute parameter type;
4887 update calls to the jump function computation functions.
4889 2013-08-30 Teresa Johnson <tejohnson@google.com>
4890 Steven Bosscher <steven@gcc.gnu.org>
4892 * cfgrtl.c (fixup_new_cold_bb): New routine.
4893 (commit_edge_insertions): Invoke fixup_partitions.
4894 (find_partition_fixes): New routine.
4895 (fixup_partitions): Ditto.
4896 (verify_hot_cold_block_grouping): Update comments.
4897 (rtl_verify_edges): Invoke find_partition_fixes.
4898 (rtl_verify_bb_pointers): Update comments.
4899 (rtl_verify_bb_layout): Ditto.
4900 * basic-block.h (probably_never_executed_edge_p): Declare.
4901 (fixup_partitions): Ditto.
4902 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
4903 * bb-reorder.c (sanitize_hot_paths): New function.
4904 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
4906 * predict.c (probably_never_executed_edge_p): New routine.
4907 * cfg.c (check_bb_profile): Add partition insanity warnings.
4909 2013-08-30 Meador Inge <meadori@codesourcery.com>
4911 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4913 2013-08-30 Marek Polacek <polacek@redhat.com>
4915 * Makefile.in (ubsan.o): Add.
4916 (c-family/c-ubsan.o): Add.
4917 (builtins.o): Add ubsan.h dependency.
4918 * ubsan.h: New file.
4919 * ubsan.c: New file.
4920 * common.opt: Add -fsanitize=undefined option.
4921 (flag_sanitize): Add variable.
4922 (fsanitize=): Add option. Add Driver.
4923 (fsanitize=thread): Remove option.
4924 (fsanitize=address): Likewise.
4925 (static-libubsan): New option.
4926 * doc/invoke.texi: Document the new flag and -static-libubsan.
4927 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
4928 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
4929 * builtin-attrs.def (ATTR_COLD): Define.
4930 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4931 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
4932 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
4933 * flag-types.h (sanitize_code): New enum.
4934 * opts.c (common_handle_option): Parse command line arguments
4935 of -fsanitize=. Add -fsanitize=unreachable option.
4936 * varasm.c (get_variable_section): Adjust.
4937 (assemble_noswitch_variable): Likewise.
4938 (assemble_variable): Likewise.
4939 (output_constant_def_contents): Likewise.
4940 (categorize_decl_for_section): Likewise.
4941 (place_block_symbol): Likewise.
4942 (output_object_block): Likewise.
4943 * builtins.def: Likewise.
4944 * toplev.c (compile_file): Likewise.
4945 (process_options): Likewise.
4946 * cppbuiltin.c: Likewise.
4947 * tsan.c (tsan_pass): Likewise.
4948 (tsan_gate): Likewise.
4949 (tsan_gate_O0): Likewise.
4950 * cfgexpand.c (partition_stack_vars): Likewise.
4951 (expand_stack_vars): Likewise.
4952 (defer_stack_allocation): Likewise.
4953 (expand_used_vars): Likewise.
4954 * cfgcleanup.c (old_insns_match_p): Likewise.
4955 * asan.c (asan_finish_file): Likewise.
4956 (asan_instrument): Likewise.
4957 (gate_asan): Likewise.
4958 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4959 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4960 (asan_global_struct): Use pointer_sized_int_node instead
4961 calling build_nonstandard_integer_type.
4962 (initialize_sanitizer_builtins): Likewise.
4963 (asan_finish_file): Likewise.
4964 * gcc.c: Document %{%:function(args):X}.
4965 (static_spec_functions): Add sanitize.
4966 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4967 store funcval != NULL there.
4968 (do_spec_1): Adjust handle_spec_function caller.
4969 (handle_braces): Allow %:function(args) as condition.
4970 (sanitize_spec_function): New function.
4971 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4972 (LIBUBSAN_SPEC): Likewise.
4973 (LIBUBSAN_EARLY_SPEC): Likewise.
4974 (SANITIZER_SPEC): Handle libubsan.
4975 (SANITIZER_EARLY_SPEC): Likewise.
4976 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4977 instead of fsanitize=address.
4978 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4979 instead of fsanitize=address*.
4980 * builtins.c: Include ubsan.h.
4981 (fold_builtin_0): Instrument __builtin_unreachable.
4982 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4983 instead of flag_asan.
4984 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4985 (pointer_sized_int_node): Define.
4986 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4988 2013-08-30 Mike Stump <mikestump@comcast.net>
4990 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4993 2013-08-29 Jan Hubicka <jh@suse.cz>
4995 * cgraph.c (cgraph_function_body_availability): Handle weakref
4997 * passes.def: Remove pass_fixup_cfg.
4998 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4999 track when we need to remove functions.
5000 (gate_ipa_inline): Execute inlining always; add comment why.
5001 (pass_data_ipa_inline): Remove TODO_remove_functions.
5002 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
5003 do not produce summaries.
5004 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
5005 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
5006 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
5007 constant pool and vtable.
5009 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
5011 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
5012 New arm_neon.h's internal macros to specify 64-bit constants.
5013 Avoid using stdint.h's macros.
5015 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
5017 * recog.c (verify_changes): Verify that changes[i].old is non-zero
5018 before applying REG_P.
5020 2013-08-30 Jakub Jelinek <jakub@redhat.com>
5022 PR tree-optimization/58277
5023 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
5024 after seeing too many stmts with vdef in between dombb and current
5025 bb, invalidate everything.
5027 2013-08-30 Richard Biener <rguenther@suse.de>
5029 * fold-const.c (fold_single_bit_test): Fix overflow test.
5031 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
5033 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
5034 and which can live in a register, always retrieve the value on entry.
5035 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
5036 passed by invisible reference specially.
5037 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
5038 (vt_add_function_parameter): Correctly deal with a parameter passed by
5039 invisible reference.
5041 2013-08-30 Jan Hubicka <jh@suse.cz>
5043 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
5045 2013-08-30 Richard Biener <rguenther@suse.de>
5047 PR tree-optimization/58228
5048 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
5049 allow invariant loads in nested loop vectorization.
5051 2013-08-30 Richard Biener <rguenther@suse.de>
5053 PR tree-optimization/58223
5054 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
5055 (has_anti_or_output_dependence): ... this and adjust to also
5056 look for output dependences.
5057 (mark_nodes_having_upstream_mem_writes): Adjust.
5058 (rdg_flag_uses): Likewise.
5060 2013-08-30 Richard Biener <rguenther@suse.de>
5062 PR tree-optimization/58010
5063 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
5064 assert that we have a loop-closed PHI.
5066 2013-08-29 Jan Hubicka <jh@suse.cz>
5068 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
5069 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
5070 * lto-section-in.c (lto_free_function_in_decl_state): New function.
5071 (lto_free_function_in_decl_state_for_node): New function.
5073 2013-08-29 Xinliang David Li <davidxl@google.com>
5075 * loop-unroll.c (report_unroll_peel): Minor message change.
5076 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
5077 Emit alignment peeling message with default -fopt-info.
5078 (vect_loop_versioning): Emit loop version info message.
5079 * tree-vectorizer.c (vectorize_loops): Minor message change.
5080 (execute_vect_slp): Ditto.
5082 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
5084 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
5085 of the clone from the DECL_NAME of the original function.
5087 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
5089 * passes.c (register_pass): Add overload.
5090 * tree-pass.h (register_pass): Forward declare it. Add comment.
5092 2013-08-29 Jan Hubicka <jh@suse.cz>
5094 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
5095 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
5096 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
5097 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5098 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
5099 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
5100 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5101 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
5103 2013-08-29 Teresa Johnson <tejohnson@google.com>
5105 * dumpfile.c (dump_loc): Output column number.
5106 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
5107 * doc/invoke.texi: Document optall -fopt-info flag.
5108 * profile.c (read_profile_edge_counts): Use new dump framework.
5109 (compute_branch_probabilities): Ditto.
5110 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
5111 when pass not in any opt group.
5112 * pass_manager.h (pass_manager::get_pass_profile): New method.
5113 * value-prof.c (check_counter): Use new dump framework.
5114 (check_ic_target): Ditto.
5115 * coverage.c (get_coverage_counts): Ditto.
5116 (coverage_init): Setup new dump framework.
5118 2013-08-29 Richard Biener <rguenther@suse.de>
5120 PR tree-optimization/58246
5121 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
5122 handle the dominance check inside a basic-block.
5124 2013-08-29 Richard Biener <rguenther@suse.de>
5127 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
5128 of default defs that appear in abnormal PHI nodes.
5130 2013-08-29 Richard Biener <rguenther@suse.de>
5132 PR tree-optimization/57685
5133 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
5134 single-use operands to avoid exponential complexity.
5136 2013-08-28 Dehao Chen <dehao@google.com>
5138 * ipa-inline.c (edge_badness): Fix integer underflow.
5140 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
5142 * gtm-builtins.def (_ITM_free): Declare leaf.
5144 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5147 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
5148 (*tls_local_dynamic_base_64_largepic): Likewise.
5149 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
5150 Remove predicate from call operand.
5151 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
5152 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
5154 2013-08-28 Jeff Law <law@redhat.com>
5156 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
5157 checks for the number of predecessors and successors allowed.
5158 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
5159 which require copying a joiner block if there is a request which
5160 is a subpath that requires no joiner block copying.
5162 2013-08-28 Jan Hubicka <jh@suse.cz>
5164 * lto-streamer-out.c (DFS_write_tree_body): Drop
5165 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
5166 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
5167 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
5168 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
5169 Do not read DECL_ERROR_ISSUED.
5170 (unpack_ts_decl_with_vis_value_fields): Do not read
5172 (lto_input_ts_binfo_tree_pointers): Do not read
5173 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
5174 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
5175 write DECL_ERROR_ISSUED..
5176 (pack_ts_decl_with_vis_value_fields): Do not write
5178 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
5179 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
5180 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
5181 * tree.h (tree_decl_common): Update comment.
5182 (DECL_ERROR_ISSUED): Remove.
5184 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5187 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
5189 2013-08-28 Jan Hubicka <jh@suse.cz>
5191 * builtins.def (free): Declare leaf.
5193 2013-08-27 David Malcolm <dmalcolm@redhat.com>
5196 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
5197 * configure: Regenerate.
5199 2013-08-27 Martin Jambor <mjambor@suse.cz>
5201 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
5202 (ipa_ancestor_jf_data): Likewise.
5203 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
5204 (ipa_get_jf_pass_through_type_preserved): New function.
5205 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
5206 (ipa_get_jf_ancestor_type_preserved): New function.
5207 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
5208 (ipa_get_jf_ancestor_result): Likewise.
5209 (propagate_vals_accross_pass_through): Use
5210 ipa_get_jf_pass_through_result to do all the value mappings.
5211 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
5212 type_preserved flag.
5213 (ipa_set_jf_cst_copy): New function.
5214 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
5215 (ipa_set_jf_arith_pass_through): Likewise.
5216 (ipa_set_ancestor_jf): Likewise.
5217 (compute_complex_assign_jump_func): Set type_preserved instead of
5219 (ipa_compute_jump_functions_for_edge): Likewise.
5220 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
5221 (update_jump_functions_after_inlining): Update type_preserved.
5222 Explicitely create jump functions when combining one with pass_through.
5223 (ipa_write_jump_function): Stream the type_preserved flags.
5224 (ipa_read_jump_function): Likewise.
5226 2013-08-27 Jakub Jelinek <jakub@redhat.com>
5227 Aldy Hernandez <aldyh@redhat.com>
5229 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
5230 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
5231 * function.h (struct function): Add has_force_vect_loops and
5233 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
5234 * gimple.c (gimple_build_omp_critical): Add KIND argument and
5236 * gimple.def: Update CLAUSES comments.
5237 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
5238 (gimple_build_omp_for): Add argument to prototype.
5239 (gimple_omp_for_kind): New.
5240 (gimple_omp_for_set_kind): New.
5241 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
5242 GOVD_DATA_SHARE_CLASS.
5243 (enum omp_region_type): Add ORT_SIMD.
5244 (gimple_add_tmp_var): Handle ORT_SIMD.
5245 (gimplify_var_or_parm_decl): Same.
5246 (is_gimple_stmt): Same.
5247 (omp_firstprivatize_variable): Same.
5248 (omp_add_variable): Only use splay_tree_insert if lookup failed.
5249 (omp_notice_variable): Handle ORT_SIMD.
5250 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
5251 (omp_check_private): Handle ORT_SIMD.
5252 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
5254 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
5255 Handle OMP_CLAUSE_LASTPRIVATE.
5256 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
5258 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
5259 (gimplify_expr): Handle OMP_SIMD.
5260 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
5261 (expand_GOMP_SIMD_VF): New.
5262 (expand_GOMP_SIMD_LAST_LANE): New.
5263 * internal-fn.def (GOMP_SIMD_LANE): New.
5264 (GOMP_SIMD_VF): New.
5265 (GOMP_SIMD_LAST_LANE): New.
5266 * omp-low.c: Include target.h.
5267 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
5269 (check_omp_nesting_restrictions): Same.
5271 (lower_rec_simd_input_clauses): New.
5272 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
5274 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
5275 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
5276 (expand_omp_build_assign): New.
5277 (expand_omp_for_init_counts): New.
5278 (expand_omp_for_init_vars): New.
5279 (extract_omp_for_update_vars): New.
5280 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
5281 and rewrite accordingly.
5282 (expand_omp_simd): New.
5283 (expand_omp_for): Use expand_omp_simd.
5284 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
5285 (lower_omp_for): Do not lower the body.
5286 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
5288 * tree-flow.h (find_omp_clause): Remove prototype.
5289 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
5290 forcing vectorization of the loop, or if flag_tree_vectorize.
5291 (gate_tree_if_conversion): Similarly.
5292 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
5293 gimple_build_omp_for.
5294 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
5295 * tree-parloops (create_parallel_loop): Pass kind argument to
5296 gimple_build_omp_for.
5297 * tree-pretty-print.c (dump_omp_clause): Add cases for
5298 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
5299 OMP_CLAUSE__SIMDUID_.
5300 (dump_generic_node): Handle OMP_SIMD.
5301 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
5302 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
5303 unroll OMP_SIMD loops here.
5304 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
5305 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
5307 (vect_analyze_data_refs): Handle simd loops.
5308 * tree-vect-loop.c (vectorizable_live_operation): Handle
5310 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
5311 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
5312 (vectorizable_load): Same.
5313 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
5314 (struct simduid_to_vf): New.
5315 (simduid_to_vf::hash): New.
5316 (simduid_to-vf::equal): New.
5317 (struct simd_array_to_simduid): New.
5318 (simd_array_to_simduid::hash): New.
5319 (simd_array_to_simduid::equal): New.
5320 (adjust_simduid_builtins): New.
5321 (struct note_simd_array_uses_struct): New.
5322 (note_simd_array_uses_cb): New.
5323 (note_simd_array_uses): New.
5324 (vectorize_loops): Handle simd hints and adjust simd builtins
5326 * tree-vectorizer.h (struct _stmt_vec_info): Add
5327 simd_lane_access_p field.
5328 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
5329 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
5330 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
5331 (omp_clause_code_name): Same.
5332 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
5333 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
5334 * tree.def (OMP_SIMD): New entry.
5335 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
5336 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
5337 (OMP_CLAUSE_DECL): Adjust range for new clauses.
5338 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
5339 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
5340 (OMP_CLAUSE_LINEAR_STEP): New.
5341 (OMP_CLAUSE_SAFELEN_EXPR): New.
5342 (OMP_CLAUSE__SIMDUID__DECL): New.
5343 (find_omp_clause): New prototype.
5345 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
5347 * config/i386/driver-i386.c (host_detect_local_cpu): Update
5348 Haswell processor detection.
5350 2013-08-27 Christian Widmer <shadow@umbrox.de>
5353 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
5354 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
5355 AVX2 capable processors.
5357 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
5359 * config/aarch64/arm_neon.h: Replace all inline asm implementations
5360 of vget_low_* with implementations in terms of other intrinsics.
5362 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
5365 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
5366 values to -1, 0 and 1.
5368 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
5370 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
5371 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
5372 (<optab><mode>3_insn): Remove.
5373 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
5374 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
5375 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
5376 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
5377 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
5378 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
5379 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
5380 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
5381 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
5382 (ror<mode>3_insn): Likewise.
5383 * config/aarch64/predicates.md (aarch64_simd_register): New.
5385 2013-08-27 Richard Biener <rguenther@suse.de>
5387 PR tree-optimization/57521
5388 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
5389 one edge is non-critical.
5390 (find_phi_replacement_condition): Make sure to use a non-critical
5391 edge. Cleanup and remove old bug workarounds.
5392 (bb_postdominates_preds): Remove.
5393 (if_convertible_loop_p_1): Do not compute post-dominators.
5394 (combine_blocks): Do not free post-dominators.
5395 (main_tree_if_conversion): Likewise.
5396 (pass_data_if_conversion): Add TODO_verify_ssa.
5398 2013-08-27 DJ Delorie <dj@redhat.com>
5400 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
5402 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
5404 * function.c (assign_parm_find_data_types): Set passed_mode and
5405 nominal_mode to the TYPE_MODE of nominal_type for the built
5406 pointer type in case of the struct-pass-by-reference.
5408 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5410 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
5411 (UINT16_TYPE): Change default to "unsigned int".
5413 * config/avr/avr.opt (mfract-convert-truncate): New option.
5414 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
5415 is set, round negative fractional integers according to n1169
5416 when converting to integer types.
5418 2013-08-26 Jan Hubicka <jh@suse.cz>
5420 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
5421 methods can not be called indirectly when their address is not taken.
5423 2013-08-26 Jan Hubicka <jh@suse.cz>
5425 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
5428 2013-08-26 Jan Hubicka <jh@suse.cz>
5430 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
5433 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5435 * reload.c (find_valid_class): Allow classes that do not include
5436 FIRST_PSEUDO_REGISTER - 1.
5438 2013-08-26 Jan Hubicka <jh@suse.cz>
5440 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
5441 fix edge count/frequency when speculation failed; fix type check
5442 for the direct call.
5444 2013-08-26 Jan Hubicka <jh@suse.cz>
5446 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
5448 2013-08-26 Jan Hubicka <jh@suse.cz>
5450 * ipa-inline-transform.c (inline_transform): Be ready for basic block
5451 to be changed by edge redirection.
5453 2013-08-26 Jan Hubicka <jh@suse.cz>
5455 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
5456 formating; add sanity check.
5457 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
5458 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
5459 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
5460 (resolve_noninline_speculation): Update callee keys, too.
5462 2013-08-26 Jan Hubicka <jh@suse.cz>
5464 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
5465 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
5467 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5469 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
5472 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
5474 * config/i386/i386.c (ix86_debug_options): Remove prototype.
5475 (x86_64_elf_select_section): Ditto.
5476 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
5478 (ix86_pass_by_reference): Ditto.
5479 (output_set_got): Ditto.
5480 (ix86_unary_operator_ok): Ditto.
5481 (ix86_expand_builtin): Ditto.
5483 2013-08-23 Jan Hubicka <jh@suse.cz>
5485 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
5487 2013-08-23 Jan Hubicka <jh@suse.cz>
5489 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
5490 (tree_decl_with_vis): Add FINAL field.
5492 2013-08-23 Jeff Law <law@redhat.com>
5494 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
5495 the debugging dump when the expression is fully redundant.
5497 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5499 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
5500 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
5501 * pretty-print.c (pp_formatted_text_data): Likewise.
5502 (pp_write_text_to_stream): Likewise.
5503 (pp_write_text_as_dot_label_to_stream): Likewise.
5504 (pp_append_r): Likewise.
5505 (pp_format): Likewise.
5506 (pp_flush): Likewise.
5507 (pp_clear_output_area): Likewise.
5508 (pp_append_text): Likewise.
5509 (pp_formatted_text): Likewise.
5510 (pp_remaining_character_count_for_line): Likewise.
5511 (pp_newline): Likewise.
5512 (pp_character): Likewise.
5513 (output_buffer::~output_buffer): Define.
5514 (pretty_printer::~pretty_printer): Destruct output buffer.
5515 * pretty-print.h (output_buffer::~output_buffer): Declare.
5516 (pretty_printer::~pretty_printer): Declare virtual.
5518 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
5521 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
5522 HOST_WIDE_INT_M1U): New macros.
5523 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
5524 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
5525 unsigned -1 for lshift.
5526 * cse.c (cse_insn): Likewise.
5527 * double-int.c (rshift_double, lshift_double): Likewise.
5528 * builtins.c (fold_builtin_bitop): Likewise.
5529 * combine.c (force_to_mode): Likewise.
5530 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
5531 * simplify-rtx.c (simplify_const_unary_operation,
5532 simplify_const_binary_operation): Likewise.
5533 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
5534 check_va_list_escapes): Likewise.
5535 * rtlanal.c (nonzero_bits1): Likewise.
5536 * expmed.c (expand_smod_pow2): Likewise.
5537 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
5539 2013-08-23 Jan Hubicka <jh@suse.cz>
5541 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
5542 as having address taken.
5544 2013-08-23 Jan Hubicka <jh@suse.cz>
5546 * ipa-utils.h (method_class_type): Declare.
5547 * ipa-devirt.c (method_class_type): Export.
5549 * cgraphunit.c (analyze_functions): Do basic devirtualization;
5550 do not walk base classes of anonymous types.
5552 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
5554 PR rtl-optimization/58220
5556 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
5557 handle SEQUENCE insns properly.
5559 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5561 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
5563 (pp_newline_and_indent): Likewise.
5564 (pp_separate_with): Likewise.
5565 * pretty-print.c (pp_newline_and_flush): Define.
5566 (pp_newline_and_indent): Likewise.
5567 (pp_separate_with): Likewise.
5569 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5572 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
5573 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
5575 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
5577 * config/i386/predicates.md (ext_sse_reg_operand): New.
5578 * config/i386/i386.md (*movti_internal): Use
5579 predicate to determine if EVEX is needed.
5580 (*movsi_internal): Ditto.
5581 (*movdf_internal): Ditto.
5582 (*movsf_internal): Ditto.
5583 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
5585 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5587 PR tree-optimization/58209
5588 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
5589 (find_tail_calls): Give up for pointer result types if m is non-NULL.
5590 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
5591 emit POINTER_PLUS_EXPR.
5592 (create_tailcall_accumulator): For pointer result type accumulate in
5595 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
5597 * configure.ac: Add backslashes missing from the last change.
5598 * configure: Regenerate.
5600 2013-08-22 Jan Hubicka <jh@susue.cz>
5602 * ipa.c (function_and_variable_visibility): First remember function
5603 was global and then make it local.
5605 2013-08-22 Julian Brown <julian@codesourcery.com>
5607 * configure.ac: Add aarch64 to list of arches which use "nop" in
5609 * configure: Regenerate.
5611 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5613 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
5614 gnu_libc_has_function.
5615 * config/s390/tpf.h: Likewise.
5617 2013-08-22 Jan Hubicka <jh@susue.cz>
5619 * timevar.c (validate_phases): Add cast.
5621 2013-08-22 Jan Hubicka <jh@susue.cz>
5623 * timevar.c (validate_phases): Use size_t for memory.
5624 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
5626 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
5628 * pretty-print.h (output_buffer::output_buffer): Declare.
5629 (pretty_printer::pretty_printer): Likewise.
5630 (pp_construct): Remove.
5631 * pretty-print.c (output_buffer::output_buffer): Define.
5632 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
5633 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
5634 (print_gimple_expr): Likewise.
5635 (print_gimple_seq): Likewise.
5636 (gimple_dump_bb): Likewise.
5637 * sched-vis.c (dump_value_slim): Likewise.
5638 (dump_insn_slim): Likewise.
5639 (dump_rtl_slim): Likewise.
5640 (str_pattern_slim): Likewise.
5641 * tree-mudflap.c (mf_varname_tree): Likewise.
5642 * graph.c (print_graph_cfg): Likewise.
5643 (start_graph_dump): Likewise.
5644 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
5646 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
5647 pretty printer initialization.
5648 * coretypes.h (diagnostic_context): Remove superflous type alias
5650 (pretty_printer): Likewise. Declare directly as a class.
5651 (pretty_print_info): Remove declaration as class.
5652 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
5653 and pp_clear_output_area.
5654 (asan_add_global): Likewise.
5656 2013-08-22 Jan Hubicka <jh@suse.cz>
5658 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
5659 * ipa-utils.h (update_type_inheritance_graph): Declare.
5660 (possible_polymorphic_call_target_p): Declare.
5661 (possible_polymorphic_call_target_p): New.
5662 * ipa-devirt.c: Update toplevel comments.
5663 (cached_polymorphic_call_targets): Move up.
5664 (odr_type_d): Move ID down.
5665 (polymorphic_type_binfo_p): Update comment.
5666 (odr_hasher::remove): Likewise;
5667 (get_odr_type): Set anonymous_namespace.
5668 (dump_odr_type): Dump it.
5669 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
5670 (maybe_record_node): Record node in cached_polymorphic_call_targets.
5671 (record_binfo): Add comment.
5672 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
5674 (devirt_node_removal_hook): Do not iCE when cache is freed.
5675 (possible_polymorphic_call_target_p): New predicate.
5676 (update_type_inheritance_graph): New function.
5678 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
5679 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5680 Sergey Lega <sergey.s.lega@intel.com>
5681 Anna Tikhonova <anna.tikhonova@intel.com>
5682 Ilya Tocar <ilya.tocar@intel.com>
5683 Andrey Turetskiy <andrey.turetskiy@intel.com>
5684 Ilya Verbin <ilya.verbin@intel.com>
5685 Kirill Yukhin <kirill.yukhin@intel.com>
5686 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5688 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
5689 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
5690 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
5691 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
5692 (OPTION_MASK_ISA_AVX2_UNSET): Update.
5693 (OPTION_MASK_ISA_AVX512F_UNSET): New.
5694 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
5695 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
5696 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
5697 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
5698 OPT_mavx512pf, OPT_mavx512er cases.
5699 * config/i386/constraints.md (v): New constraint.
5700 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
5701 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
5702 (bit_AVX512CD): New.
5703 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5704 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
5705 * config/i386/i386-c.c (ix86_target_macros_internal):
5706 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
5708 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
5709 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
5710 * config/i386/i386.c (regclass_map, dbx_register_map)
5711 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
5712 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
5713 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
5715 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
5716 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
5717 -mavx512pf options. Fix formatting.
5718 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
5719 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
5720 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
5721 -mavx512cd, -mavx512pf options.
5722 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
5723 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
5724 (ix86_preferred_output_reload_class): Replace SSE_REGS with
5726 (ix86_hard_regno_mode_ok): Support 512-bit registers.
5727 (ix86_set_reg_reg_cost): Ditto.
5728 (x86_order_regs_for_local_alloc): Ditto.
5729 (MAX_VECT_LEN): Extend to 64-byte.
5730 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
5731 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
5732 (TARGET_AVX512ER, TARGET_AVX512CD): New.
5733 (BIGGEST_ALIGNMENT): Extend to 512-bits.
5734 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
5735 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
5736 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
5737 (SSE_REG_MODE_P): Support new modes.
5738 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
5739 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
5740 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
5741 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
5742 (REG_CLASS_CONTENTS): Add new registers.
5743 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
5744 (EXT_REX_SSE_REGNO_P): New.
5745 (HI_REGISTER_NAMES): Add new registers.
5746 * config/i386/i386.md: Define constants for new registers.
5747 (mode): Add new 512-bit modes.
5748 (prefix): Support evex prefix.
5749 (isa): Support avx512f, noavx512f, fma_avx512f.
5750 (ssemodesuffix): Add new 512-bit modes.
5752 (*movxi_internal_avx512f): Ditto.
5753 (*movdi_internal): Replace constraint "x" with the new constraint "v".
5755 (*movsi_internal): Likewise.
5756 (*movdf_internal): Likewise.
5757 (*movsf_internal): Likewise.
5758 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
5759 (<code><mode>3): Likewise.
5760 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
5762 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
5763 with the new constraint "v".
5764 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
5766 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
5767 with the new constraint "v".
5768 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
5769 (<sse2>_storedqu<avxsizesuffix>): Likewise.
5770 (*<plusminus_insn><mode>3): Likewise.
5771 (<sse>_vm<plusminus_insn><mode>3): Likewise.
5772 (*mul<mode>3): Likewise.
5773 (<sse>_vmmul<mode>3): Likewise.
5774 (<sse>_div<mode>3): Likewise.
5775 (<sse>_vmdiv<mode>3): Likewise.
5776 (<sse>_sqrt<mode>2): Likewise.
5777 (<sse>_vmsqrt<mode>2): Likewise.
5778 (*<code><mode>3_finite): Likewise.
5779 (*<code><mode>3) <smaxmin>: Likewise.
5780 (<sse>_vm<code><mode>3): Likewise.
5781 (*<code><mode>3) <any_logic>: Likewise.
5782 (*fma_fmadd_<mode>): Likewise.
5783 (*fma_fmsub_<mode>): Likewise.
5784 (*fma_fnmadd_<mode>): Likewise.
5785 (*fma_fnmsub_<mode>): Likewise.
5786 (*fma_fmaddsub_<mode>): Likewise.
5787 (*fma_fmsubadd_<mode>): Likewise.
5788 (*fmai_fmadd_<mode>): Likewise.
5789 (*fmai_fmsub_<mode>): Likewise.
5790 (*fmai_fnmadd_<mode>): Likewise.
5791 (*fmai_fnmsub_<mode>): Likewise.
5792 (sse_cvtsi2ss): Likewise.
5793 (sse_cvtsi2ssq): Likewise.
5794 (sse_cvtss2si): Likewise.
5795 (sse_cvtss2si_2): Likewise.
5796 (sse_cvtss2siq): Likewise.
5797 (sse_cvtss2siq_2): Likewise.
5798 (sse_cvttss2si): Likewise.
5799 (sse_cvtss2siq_2): Likewise.
5800 (float<sseintvecmodelower><mode>2): Likewise.
5801 (sse2_cvtsd2si_2): Likewise.
5802 (sse2_cvtsd2siq_2): Likewise.
5803 (*<plusminus_insn><mode>3): Likewise.
5804 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
5805 (*<sse4_1_avx2>_mul<mode>3): Likewise.
5806 (ashr<mode>3): Likewise.
5807 (<shift_insn><mode>3): Likewise.
5808 (avx2_<code><mode>3): Likewise.
5809 (*avx2_<code><mode>3): Likewise.
5810 (*andnot<mode>3): Likewise.
5811 (*<code><mode>3) <any_logic>: Likewise.
5812 (abs<mode>2): Likewise.
5813 (avx2_permvar<mode>): Likewise.
5814 (avx2_perm<mode>_1): Likewise.
5815 (*avx_vpermilp<mode>): Likewise.
5816 (avx_vpermilvar<mode>3): Likewise.
5817 (avx2_ashrv<mode>): Likewise.
5818 (avx2_<shift_insn>v<mode>): Likewise.
5819 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
5821 * doc/rtl.texi: Document XImode.
5823 2013-08-21 Jeff Law <law@redhat.com>
5825 * tree-flow.h (register_jump_thread): Pass vector of edges
5826 instead of each important edge.
5827 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
5828 thread path into a vector and pass that to register_jump_thread.
5829 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
5830 passed in edge vector to the current 3-edge form.
5833 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5835 * dce.c (fini_dce): Call df_analyze again just in case
5836 delete_unmarked_insns removed anything.
5838 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
5840 * reload.h (struct reg_equivs): Rename to ..
5841 (struct reg_equivs_s): .. this.
5843 2013-08-20 Martin Liska <marxin.liska@gmail.com>
5845 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
5847 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5849 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
5851 2013-08-21 Jeff Law <law@redhat.com>
5853 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
5854 simplify assignments too. If the RHS collapses to a singleton
5855 range, then return the value for the range.
5857 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
5859 * config/i386/sse.md (V16): Rename to...
5861 (mov<mode>): Update iterator name.
5862 (*mov<mode>_internal): Ditto.
5863 (push<mode>1): Ditto.
5864 (movmisalign<mode>): Ditto.
5866 2013-08-20 Jan Hubicka <jh@suse.cz>
5869 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
5870 entry for direct edges.
5871 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
5873 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5875 Revert my last two changes, r201865 and r201864:
5878 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5880 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5881 instances can own GC refs.
5883 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5884 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5885 (gcc::context::gt_pch_nx): Likewise.
5886 (gcc::context::gt_pch_nx): Likewise.
5887 * ggc.h (gt_ggc_mx <T>): New.
5888 (gt_pch_nx_with_op <T>): New.
5889 (gt_pch_nx <T>): New.
5890 * passes.c (opt_pass::gt_ggc_mx): New.
5891 (opt_pass::gt_pch_nx): New.
5892 (opt_pass::gt_pch_nx_with_op): New.
5893 (pass_manager::gt_ggc_mx): New.
5894 (pass_manager::gt_pch_nx): New.
5895 (pass_manager::gt_pch_nx_with_op): New.
5896 (pass_manager::operator new): Use
5897 ggc_internal_cleared_alloc_stat rather than xcalloc.
5898 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5899 (pass_manager::gt_ggc_mx): New.
5900 (pass_manager::gt_pch_nx): New.
5901 (pass_manager::gt_pch_nx_with_op): New.
5902 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5903 (opt_pass::operator new): New.
5904 (opt_pass::gt_ggc_mx): New.
5905 (opt_pass::gt_pch_nx): New.
5906 (opt_pass::gt_pch_nx_with_op): New.
5909 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5911 * Makefile.in (GTFILES): Add context.h.
5912 * context.c (gcc::context::operator new): New.
5913 (gcc::context::gt_ggc_mx): New.
5914 (gcc::context::gt_pch_nx): New.
5915 (gcc::context::gt_pch_nx): New.
5916 * context.h (gcc::context): Add GTY((user)) marking.
5917 (gcc::context::operator new): New.
5918 (gcc::context::gt_ggc_mx): New.
5919 (gcc::context::gt_pch_nx): New.
5920 (gcc::context::gt_pch_nx): New.
5921 (g): Add GTY marking.
5922 (gt_ggc_mx (gcc::context *)): New.
5923 (gt_pch_nx (gcc::context *)): New.
5924 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5925 void *cookie)): New.
5926 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5928 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5930 * dce.c (fini_dce): Call df_analyze again just in case
5931 delete_unmarked_insns removed anything.
5933 2013-08-20 Teresa Johnson <tejohnson@google.com>
5935 PR rtl-optimizations/57451
5936 * final.c (reemit_insn_block_notes): Prevent lexical blocks
5937 from crossing split section boundaries.
5939 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
5941 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
5942 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
5943 with MULTLIB_DEFAULTS.
5945 2013-08-20 Nick Clifton <nickc@redhat.com>
5947 * target.def (narrow_volatile_bitfield): Note that the default
5948 value is false, not !TARGET_STRICT_ALIGN.
5949 * doc/tm.texi: Regenerate.
5951 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
5953 Fix LIB_SPEC for systems without libpthread.
5955 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
5956 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5958 * config/i386/linux-common.h: Likewise.
5959 * config/mips/linux-common.h: Likewise.
5961 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5963 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5966 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5968 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5969 instances can own GC refs.
5971 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5972 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5973 (gcc::context::gt_pch_nx): Likewise.
5974 (gcc::context::gt_pch_nx): Likewise.
5975 * ggc.h (gt_ggc_mx <T>): New.
5976 (gt_pch_nx_with_op <T>): New.
5977 (gt_pch_nx <T>): New.
5978 * passes.c (opt_pass::gt_ggc_mx): New.
5979 (opt_pass::gt_pch_nx): New.
5980 (opt_pass::gt_pch_nx_with_op): New.
5981 (pass_manager::gt_ggc_mx): New.
5982 (pass_manager::gt_pch_nx): New.
5983 (pass_manager::gt_pch_nx_with_op): New.
5984 (pass_manager::operator new): Use
5985 ggc_internal_cleared_alloc_stat rather than xcalloc.
5986 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5987 (pass_manager::gt_ggc_mx): New.
5988 (pass_manager::gt_pch_nx): New.
5989 (pass_manager::gt_pch_nx_with_op): New.
5990 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5991 (opt_pass::operator new): New.
5992 (opt_pass::gt_ggc_mx): New.
5993 (opt_pass::gt_pch_nx): New.
5994 (opt_pass::gt_pch_nx_with_op): New.
5996 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5998 * Makefile.in (GTFILES): Add context.h.
5999 * context.c (gcc::context::operator new): New.
6000 (gcc::context::gt_ggc_mx): New.
6001 (gcc::context::gt_pch_nx): New.
6002 (gcc::context::gt_pch_nx): New.
6003 * context.h (gcc::context): Add GTY((user)) marking.
6004 (gcc::context::operator new): New.
6005 (gcc::context::gt_ggc_mx): New.
6006 (gcc::context::gt_pch_nx): New.
6007 (gcc::context::gt_pch_nx): New.
6008 (g): Add GTY marking.
6009 (gt_ggc_mx (gcc::context *)): New.
6010 (gt_pch_nx (gcc::context *)): New.
6011 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6012 void *cookie)): New.
6013 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6015 2013-08-20 Alan Modra <amodra@gmail.com>
6018 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
6019 (rs6000_emit_epilogue): Likewise.
6021 2013-08-19 Dehao Chen <dehao@google.com>
6023 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
6025 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
6026 Jakub Jelinek <jakub@redhat.com>
6028 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
6029 (BUILT_IN_FABSD64): Likewise.
6030 (BUILT_IN_FABSD128): Likewise.
6031 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
6032 (fold_builtin_1): Likewise.
6033 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
6034 destination and source operands.
6035 (*abstd2_fpr): Likewise.
6036 (*nabstd2_fpr): Likewise.
6038 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
6040 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
6043 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
6045 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
6048 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
6050 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
6051 * builtins.c (default_libc_has_function): New.
6052 (gnu_libc_has_function): Ditto.
6053 (no_c99_libc_has_function): Ditto.
6054 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
6055 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
6056 (fold_builtin_sincos): Likewise.
6057 (fold_builtin_cexp): Likewise.
6058 * builtins.def (DEF_C94_BUILTIN): Likewise.
6059 (DEF_C99_BUILTIN): Likewise.
6060 (DEF_C99_C90RES_BUILTIN): Likewise.
6061 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
6062 definitions to using this define.
6063 * config/darwin-protos.h (darwin_libc_has_function): New.
6064 * config/darwin.c (darwin_libc_has_function): Ditto.
6065 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
6066 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
6067 * config/darwin.h: Ditto.
6068 * config/elfos.h: Ditto.
6069 * config/freebsd.h: Ditto.
6070 * config/i386/cygming.h: Ditto.
6071 * config/i386/djgpp.h: Ditto.
6072 * config/i386/i386-interix.h: Ditto.
6073 * config/microblaze/microblaze.h: Ditto.
6074 * config/mmix/mmix.h: Ditto.
6075 * config/gnu-user.h: Ditto.
6076 * config/ia64/hpux.h: Ditto.
6077 * config/pa/pa-hpux.h: Ditto.
6078 * config/pdp11/pdp11.h: Ditto.
6079 * config/picochip/picochip.h: Ditto.
6080 * config/linux.h: Ditto.
6081 * config/netbsd.h: Ditto.
6082 * config/openbsd.h: Ditto.
6083 * config/rs6000/aix43.h: Ditto.
6084 * config/rs6000/aix51.h: Ditto.
6085 * config/rs6000/aix52.h: Ditto.
6086 * config/rs6000/aix53.h: Ditto.
6087 * config/rs6000/aix61.h: Ditto.
6088 * config/rs6000/darwin.h: Ditto.
6089 * config/rs6000/linux.h: Ditto.
6090 * config/rs6000/linux64.h: Ditto.
6091 * config/s390/tpf.h: Ditto.
6092 * config/sol2-10.h: Ditto.
6093 * config/sol2.h: Ditto.
6094 * config/vms/vms.h: Ditto.
6095 * config/vxworks.h: Ditto.
6096 * config/linux-android.c (linux_android_libc_has_function):
6097 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
6098 * config/linux-protos.h (linux_android_libc_has_function):
6100 * config/i386/i386.c (ix86_libc_has_function): New.
6101 * config/i386/i386-protos.h
6102 (ix86_libc_has_function): New declaration.
6103 * config/i386/i386.md
6104 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
6105 ("isinf<mode>2): Likewise.
6106 * convert.c (convert_to_integer): Using new target hook
6107 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
6108 TARGET_C99_FUNCTIONS.
6109 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
6110 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
6111 * coretypes.h (function_class): New enum for different
6112 classes of functions.
6113 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
6114 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
6115 (TARGET_HAS_SINCOS): Likewise.
6116 (TARGET_LIBC_HAS_FUNCTION): New.
6117 * doc/tm.texi: Regenerated.
6118 * targhooks.h (default_libc_has_function): New declaration.
6119 (no_c99_libc_has_function): Ditto.
6120 (gnu_libc_has_function): Ditto.
6121 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
6122 and TARGET_HAS_SINCOS.
6124 2013-08-18 Jan Hubicka <jh@suse.cz>
6126 * Makeifle-in (ipa-devirt.o): New.
6127 (GTFILES): Add ipa-utils.h and ipa-devirt.c
6128 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
6129 (analyze_functions): Look into possible targets of polymorphic call.
6130 * dumpfile.c (dump_files): Add type-inheritance dump.
6131 * dumpfile.h (TDI_inheritance): New.
6132 * ipa-devirt.c: New file.
6133 * ipa-utils.h (odr_type_d): Forward declare.
6134 (odr_type): New type.
6135 (build_type_inheritance_graph): Declare.
6136 (possible_polymorphic_call_targets): Declare and introduce inline
6137 variant when only edge is pased.
6138 (dump_possible_polymorphic_call_targets): Likewise.
6139 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
6140 * tree.c (type_in_anonymous_namespace_p): Break out from ...
6141 (types_same_for_odr): ... here.
6142 * tree.h (type_in_anonymous_namespace_p): Declare.
6144 2013-08-18 Jakub Jelinek <jakub@redhat.com>
6146 PR tree-optimization/58006
6147 * tree-parloops.c (take_address_of): Don't ICE if get_name
6149 (eliminate_local_variables_stmt): Remove clobber stmts.
6151 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
6153 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
6154 error message is issued for an alias to undefined symbol.
6156 2013-08-18 Jan Hubicka <jh@suse.cz>
6158 * cgraph.c (cgraph_create_indirect_edge): Discover
6159 polymorphic calls and record basic info into indirect_info.
6160 * gimple-fold.c (gimple_fold_call): When doing BINFO based
6161 devirtualization, ignore objc function calls.
6162 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
6163 call with no parm index info.
6164 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
6165 * tree.c (virtual_method_call_p): New function.
6166 * tree.h (virtual_method_call_p): Declare.
6168 2013-08-16 Jan Hubicka <jh@suse.cz>
6171 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
6173 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
6175 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
6178 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6180 * gengtype.c (type_for_name): Add special-case support for
6181 locating types within the "gcc::" namespace.
6182 (open_base_files): Emit a "using namespace gcc" directive.
6184 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
6187 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
6188 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
6190 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
6191 array instead of each individual operand as a separate argument.
6192 (emit_fusion_gpr_load): Likewise.
6193 (expand_fusion_gpr_load): Add new function declaration.
6195 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
6196 signature to have the operands passed as an array, instead of as
6197 separate arguments. Allow ZERO_EXTEND to be in the memory
6198 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
6199 depend on the register live/dead flags when peepholes are run.
6200 (expand_fusion_gpr_load): New function to be called from the
6201 peephole2 pass, to change the register that addis sets to be the
6203 (emit_fusion_gpr_load): Change the calling signature to have the
6204 operands passed as an array, instead of as separate arguments.
6205 Allow ZERO_EXTEND to be in the memory address, and also
6206 SIGN_EXTEND if -mpower8-fusion-sign.
6208 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
6210 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
6211 adjust the register addis loads up in the peephole2 pass. Do not
6212 depend on the register live/dead state when the peephole pass is done.
6214 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6216 * gengtype.c (create_user_defined_type): Ensure that the kind
6217 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
6218 declaration is seen before the GTY((user)) marking.
6220 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
6223 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
6225 2013-08-16 Jan Hubicka <jh@suse.cz>
6227 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
6228 arugment expected_type.
6229 (gimple_fold_call): Use it.
6230 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
6231 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
6232 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
6233 (try_make_edge_direct_virtual_call): Likewise.
6234 * tree.c (obj_type_ref_class): New.
6235 * tree.h (obj_type_ref_class): Use it.
6237 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
6239 * sched-vis.c (rtl_slim_pp_initialized): Remove.
6240 (rtl_slim_pp): Likewise.
6241 (init_rtl_slim_pretty_print): Likewise.
6242 (dump_value_slim): Don't call it. Use local pretty printer.
6243 (dump_insn_slim): Likewise.
6244 (dump_rtl_slim): Likewise.
6245 (str_pattern_slim): Likewise.
6246 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
6249 2013-08-16 Jakub Jelinek <jakub@redhat.com>
6251 PR tree-optimization/58164
6252 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
6253 walk gimple_goto_dest of GIMPLE_GOTO.
6255 PR tree-optimization/58165
6256 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
6257 bi_call must be the last stmt in a bb, don't split_block, instead
6258 use fallthru edge from it and give up if there is none.
6259 Release conds vector when returning early.
6261 2013-08-14 Xinliang David Li <davidxl@google.com>
6263 * config/i386/i386.c (ix86_option_override_internal):
6264 Remove unused variable and field.
6266 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6269 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6270 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
6271 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
6272 and Linux, correct BLKmode alignment when 128-bit alignment is
6273 required and compatibility flag is not set.
6274 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
6275 for zero-size arguments when compatibility flag is not set.
6277 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6279 PR tree-optimization/58145
6280 * tree-sra.c (build_ref_for_offset): If prev_base has
6281 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
6283 2013-08-14 Xinliang David Li <davidxl@google.com>
6285 * config/i386/i386.c (ix86_option_override_internal):
6286 Fix uninitialized variable error.
6288 2013-08-14 Xinliang David Li <davidxl@google.com>
6290 * config/i386/i386.opt: Define two new options.
6291 * config/i386/x86-tune.def: Add arch selector field in macros.
6292 * config/i386/i386.h: Adjust macro definition.
6293 * config/i386/i386.c (ix86_option_override_internal):
6295 (parse_mtune_ctrl_str): New function.
6296 (set_ix86_tune_features): New function.
6297 (ix86_function_specific_restore): Call the new helper function.
6299 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
6301 PR rtl-optimization/57662
6302 * sel-sched.c (code_motion_process_successors): When the current insn
6303 is removed after the recursive traversal, break from the loop.
6304 Add comments and debug printouts.
6306 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6307 Alexandre Oliva <aoliva@redhat.com>
6310 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
6311 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
6312 there also UNSPEC_PLTOFF.
6314 2013-08-14 Marek Polacek <polacek@redhat.com>
6316 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
6317 NUM_CONDITIONS bit positions.
6319 2013-08-13 Cary Coutant <ccoutant@google.com>
6321 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
6322 (attr_checksum): Hash vector contents instead of pointer.
6323 (attr_checksum_ordered): Likewise.
6325 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
6327 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
6328 when Pmode != word_mode. Add length_address attribute.
6329 (sse3_monitor_<mode>): Merge from sse3_monitor and
6330 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
6331 Pmode != word_mode. Update insn length attribute.
6332 * config/i386/i386.c (ix86_option_override_internal): Update
6333 ix86_gen_monitor selection for merged sse3_monitor insn.
6335 2013-08-13 Julian Brown <julian@codesourcery.com>
6337 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
6338 perform invalid legitimization on greater-than-word-size modes for
6341 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
6343 * ira.c (setup_class_translate_array): Use aclass instead of cl
6344 for classes not fully covered by allocno classes.
6346 2013-08-13 Jakub Jelinek <jakub@redhat.com>
6348 PR tree-optimization/57661
6349 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
6350 * tree-inline.c (tree_function_versioning): Initialize it.
6351 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
6352 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
6353 that is not being copied.
6356 * asan.c (instrument_strlen_call): Fix typo in comment.
6357 Use char * type even for the lhs of POINTER_PLUS_EXPR.
6359 2013-08-13 Steve Ellcey <sellcey@mips.com>
6361 * config/mips/mips.md (prefetch): Use lw instead of ld on
6362 loongson in 32bit mode.
6364 2013-08-13 Nick Clifton <nickc@redhat.com>
6366 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
6368 2013-08-13 Jan Hubicka <jh@suse.cz>
6370 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
6371 introduced edge; fix typo in sanity check.
6372 (cgraph_resolve_speculation): Export; improve diagnostic.
6373 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
6374 speculation at type mismatch.
6375 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
6376 (cgraph_resolve_speculation): Declare.
6377 (symtab_can_be_discarded): New function.
6378 * value-prof.c (gimple_ic_transform): Remove actual transform code.
6379 * ipa-inline-transform.c (speculation_removed): New global var.
6380 (clone_inlined_nodes): See if speculation can be removed.
6381 (inline_call): If speculations was removed, we growths may not match.
6382 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
6383 (speculation_useful_p): New function.
6384 (resolve_noninline_speculation): New function.
6385 (inline_small_functions): Resolve useless speculations.
6386 * ipa-inline.h (speculation_useful_p): Declare
6387 * ipa.c (can_replace_by_local_alias): Simplify.
6388 (ipa_profile): Produce speculative calls in non-lto, too;
6389 add simple cost model; produce local aliases.
6391 2013-08-13 David Malcolm <dmalcolm@redhat.com>
6393 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
6396 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
6398 * config/i386/i386.c (ix86_function_versions): Use error + inform.
6400 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
6402 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
6403 iterator instead of X87MODEF.
6405 2013-08-12 Perez Read <netfirewall@gmail.com>
6408 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
6409 operand 0 for intel asm alternative.
6410 (*movabs<mode>_2): Ditto for operand 1.
6412 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
6414 * config/aarch64/arm_none.h
6415 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6417 2013-08-12 Nick Clifton <nickc@redhat.com>
6419 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
6421 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6423 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
6424 expand for QI/HImode operand to produce more effictive code for
6425 unsigned char(short) --> float(double) conversion.
6427 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
6429 * doc/invoke.texi: Mention that -ftls-model does not force the final
6432 2013-08-12 Marek Polacek <polacek@redhat.com>
6433 Marc Glisse <marc.glisse@inria.fr>
6435 PR tree-optimization/57980
6436 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
6437 when creating -1 constant.
6439 2013-08-10 Jan Hubicka <jh@suse.cz>
6441 Workaround binutils PR14342.
6442 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
6443 (gimple_init_edge_profiler): Likewise.
6444 (gimple_gen_ic_func_profiler): Likewise.
6446 2013-08-09 Jan Hubicka <jh@suse.cz>
6448 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
6450 2013-08-09 Xinliang David Li <davidxl@google.com>
6452 * config/i386/stringop.def: New file.
6453 * config/i386/stringop.opt: New file.
6454 * config/i386/i386-opts.h: Include stringopt.def.
6455 * config/i386/i386.opt: Include stringopt.opt.
6456 * config/i386/i386.c (ix86_option_override_internal):
6457 Override default size based stringop inline strategies with options.
6458 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
6461 2013-08-09 Jan Hubicka <jh@suse.cz>
6463 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
6465 2013-08-09 Jan Hubicka <jh@suse.cz>
6467 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
6469 (dump_cgraph_node): Dump profile-id.
6470 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
6471 and common_target_probability.
6472 * lto-cgraph.c (lto_output_edge): Stream common targets.
6473 (lto_output_node): Stream profile ids.
6474 (input_node): Stream profile ids.
6475 (input_edge): Stream common targets.
6476 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
6477 * ipa.c: Include value-prof.h
6478 (ipa_profile_generate_summary): Turn indirect call statement histograms
6479 into common targets.
6480 (ipa_profile): Turn common targets into speculative edges.
6482 2013-08-09 Jan Hubicka <jh@suse.cz>
6484 * cgraph.h (cgraph_node): Add profile_id.
6485 * value-prof.c (cgraph_node_map): Turn into pointer_map.
6486 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
6487 (del_node_map): Update.
6488 (find_func_by_funcdef_no): Replace by ...
6489 (find_func_by_profile_id): ... this one.
6490 (gimple_ic_transform): Do not remove useful histograms when
6491 speculation is not done; dump info when indirect call removal
6493 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
6494 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
6495 (__gcov_indirect_call_profiler_v2): .. this one.
6496 * profile.h (init_node_map): Update.
6497 * coverage.c (coverage_compute_profile_id): New function.
6498 * coverage.h (coverage_compute_profile_id): Declare.
6499 * tree-profile.c (init_ic_make_global_vars): Make
6500 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
6501 (gimple_init_edge_profiler): Update prototype of
6502 __gcov_indirect_call_profiler.
6503 (gimple_gen_ic_func_profiler): Simplify.
6504 (tree_profiling): Use init_node_map
6506 2013-08-09 Jan Hubicka <jh@suse.cz>
6508 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
6509 non-speculative refs.
6510 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
6511 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
6512 (cgraph_set_call_stmt): Likewise.
6513 (cgraph_create_edge_1): Fix release checking compilatoin;
6515 (cgraph_free_edge): Free indirect info.
6516 (cgraph_turn_edge_to_speculative): New function.
6517 (cgraph_speculative_call_info): New function.
6518 (cgraph_make_edge_direct): Return direct edge; handle speculation.
6519 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
6520 (dump_cgraph_node): Dump speculation.
6521 (verify_edge_count_and_frequency): Accept speculative edges.
6522 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
6523 (verify_cgraph_node): Handle speculation.
6524 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
6525 (cgraph_set_call_stmt): Update prototype.
6526 (cgraph_make_edge_direct): Update prototype.
6527 (cgraph_speculative_call_info): Declare.
6528 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
6529 to change; update call of ipa_find_references.
6530 * ipa-ref.c (ipa_record_reference): Fix return value; clear
6531 lto_stmt_uid and speculative flags.
6532 (ipa_dump_references): Dump speculation.
6533 (ipa_clone_references): Clone speculative flag.
6534 (ipa_clone_referring): Likewise.
6535 (ipa_clone_ref): New function.
6536 (ipa_find_reference): Look into lto_stmt_uids
6537 (ipa_clear_stmts_in_references): Do not clear speculative calls.
6538 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
6539 (ipa_find_reference): Update declaration.
6540 (ipa_clone_ref): Declare.
6541 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
6542 stream speculative flag.
6543 (lto_output_ref): Stream statements uids and speculation.
6544 (input_ref): Likewise.
6545 (input_edge): Stream speuclation.
6546 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
6547 (cgraph_set_call_stmt_including_clones): Handle speculation.
6548 * ipa-inline.c (heap_edge_removal_hook): New function.
6549 (inline_small_functions): Register it.
6550 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
6551 also initialize refs.
6552 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
6554 (try_make_edge_direct_simple_call): Likewise.
6555 (try_make_edge_direct_simple_call): Likewise.
6556 (update_indirect_edges_after_inlining): Likewise.
6557 (remove_described_reference): Look proper lto_stmt_uid.
6558 (propagate_controlled_uses): Likewise.
6559 (propagate_controlled_uses): Liekwise.
6560 * tree-inline.c (copy_bb): Copy speculative edges.
6561 (redirect_all_calls): New function.
6562 (copy_cfg_body): Do redirection after loop info is updated.
6563 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
6565 2013-08-09 Jan Hubicka <jh@suse.cz>
6567 * lto-streamer-out.c (output_function): Renumber PHIs.
6568 * lto-streamer-in.c (input_function): Likewise.
6570 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
6572 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
6573 (get_lane_unsigned): Likewise.
6574 (dup_lane_scalar): Likewise.
6575 (get_lane): enable for VALL.
6576 * config/aarch64/aarch64-simd.md
6577 (aarch64_dup_lane_scalar<mode>): Remove.
6578 (aarch64_get_lane_signed<mode>): Likewise.
6579 (aarch64_get_lane_unsigned<mode>): Likewise.
6580 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
6581 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
6582 (aarch64_get_lane<mode>): Enable for all vector modes.
6583 (aarch64_get_lanedi): Remove misleading constraints.
6584 * config/aarch64/arm_neon.h
6585 (__aarch64_vget_lane_any): Define.
6586 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
6587 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
6588 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
6589 * config/aarch64/iterators.md (VDQQH): New.
6593 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
6595 * configure.ac: Add GAS check for LEON instructions on SPARC.
6596 * configure: Regenerate.
6597 * config.in: Likewise.
6598 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
6600 * config/sparc/sparc.opt (LEON, LEON3): New masks.
6601 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
6603 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
6604 (AS_LEON_FLAG): New macro.
6605 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
6606 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
6607 Deal with LEON and LEON3 for the memory model.
6608 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
6609 (atomic_compare_and_swap<mode>_1): Likewise.
6610 (*atomic_compare_and_swap<mode>_1): Likewise.
6612 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6614 * config/arm/neon.md (vcond): Fix floating-point vector
6615 comparisons against 0.
6617 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
6619 * lra-constraints.c (emit_spill_move): Remove assert.
6620 (process_alt_operands): Add more debugging
6621 output. Increase reject for spilling into memory. Decrease
6622 reject for reloading scratch.
6623 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
6625 2013-08-08 Steve Ellcey <sellcey@mips.com>
6627 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
6628 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
6629 micromips incompatible. Add nan2008.
6630 (MULTILIB_DIRNAMES): Add nan2008.
6631 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6632 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
6633 and micromips incompatible. Add nan2008.
6634 (MULTILIB_DIRNAMES): Add nan2008.
6635 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6637 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
6639 PR rtl-optimization/58079
6640 * combine.c (combine_simplify_rtx): Avoid using SUBST if
6641 simplify_comparison has widened a comparison with an integer.
6643 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6645 * config/arm/neon.md (movmisalign<mode>): Disable when we
6646 don't allow unaligned accesses.
6647 (*movmisalign<mode>_neon_store): Likewise.
6648 (*movmisalign<mode>_neon_load): Likewise.
6649 (*movmisalign<mode>_neon_store): Likewise.
6650 (*movmisalign<mode>_neon_load): Likewise.
6652 2013-08-08 Jan Hubicka <jh@suse.cz>
6654 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
6655 (make_pass_rebuild_cgraph_edges): Also clear references.
6656 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
6657 * ipa-inline-transform.c (inline_transform): Remove all references
6659 * cgraphunit.c (expand_function): Remove all references after
6661 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
6662 (ipa_find_reference): Rewrite to iterator.
6663 (remove_stmt_references): Likewise.
6664 (ipa_clear_stmts_in_references): New function.
6665 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
6666 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
6668 * ipa-split.c (split_function): Remove references in split function.
6670 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6673 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
6674 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
6676 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6679 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
6680 suggested mode for the assignment isn't compatible with the
6683 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
6686 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
6688 2013-08-07 Xinliang David Li <davidxl@google.com>
6690 * config/i386/i386.opt: New option -mtune-ctrl=.
6691 * config/i386/x86-tune.def: New file.
6692 * config/i386/i386.h: include x86-tune.def.
6693 * config/i386/i386.c (ix86_option_override_internal):
6694 Parsing -mtune-ctrl= option and set tune features.
6696 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
6699 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
6700 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
6703 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6705 * rtl.h (update_alignments): Declare.
6706 * final.c (grow_label_align): New function extracted from...
6707 (shorten_branches): ...here. Call it.
6708 (update_alignments): New function.
6709 * reorg.c (sibling_labels): New variable.
6710 (get_label_before): Add SIBLING parameter. If it is non-zero, push
6711 the new label along with it onto the sibling_labels vector.
6712 (fill_simple_delay_slots): Adjust call to get_label_before.
6713 (fill_slots_from_thread): Likewise.
6714 (relax_delay_slots): Likewise.
6715 (make_return_insns): Likewise.
6716 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
6718 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6720 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
6721 document its semantics.
6722 (diagnostic_report_diagnostic): Adjust accordingly.
6724 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6726 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
6727 (sparc_option_override): ...and port to new C++ pass API.
6728 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
6730 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
6732 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
6734 2013-08-06 Caroline Tice <cmtice@google.com>
6736 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
6737 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
6738 * tree-pass.h: Add pass_vtable_verify.
6739 * varasm.c (assemble_variable): Add code to properly set the comdat
6740 section and name for the .vtable_map_vars section.
6741 (assemble_vtyv_preinit_initializer): New function.
6742 (default_sectin_type_flags): Make sure .vtable_map_vars section has
6744 * output.h: Add function decl for assemble_vtv_preinit_initializer.
6745 * vtable-verify.c: New file.
6746 * vtable-verify.h: New file.
6747 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
6748 initialiation levels.
6749 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
6750 * passes.def: Insert pass_vtable_verify.
6751 * aclocal.m4: Reorder includes.
6752 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
6753 -fvtv-counts options.
6754 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
6755 as appropriate, if -fvtable-verify=... is used.
6756 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
6757 -fvtable-verify=... is used.
6758 * Makefile.in (OBJS): Add vtable-verify.o to list.
6759 (vtable-verify.o): Add new build rule.
6760 (GTFILES): Add vtable-verify.c to list.
6761 * common.opt (fvtable-verify=): New flag.
6762 (vtv_priority): Values for fvtable-verify= flag.
6763 (fvtv-counts): New flag.
6764 (fvtv-debug): New flag.
6765 * tree.h (save_vtable_map_decl): New extern function decl.
6767 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6769 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
6770 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
6771 (pass_data_rl78_devirt): ...new pass_data instance and...
6772 (make_pass_rl78_devirt): ...new function.
6773 (rl78_asm_file_start): Port pass registration to new C++ API.
6775 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6777 * coretypes.h (rtl_opt_pass): Add.
6778 (gcc::context): Add.
6779 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
6780 (epiphany_init): Port to new C++ pass API.
6781 (epiphany_optimize_mode_switching): Likewise.
6782 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
6783 (pass_manager::get_pass_mode_switching): New.
6784 (pass_manager::get_pass_peephole2): New.
6785 * mode-switching.c (pass_mode_switching): Add clone method.
6786 * recog.c (pass_peephole2): Add clone method.
6787 (pass_split_all_insns): Add clone method.
6789 2013-08-06 David Malcolm <dmalcolm@redhat.com>
6791 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
6792 (mips_option_override): ...here, porting to new C++ API for passes.
6794 2013-08-06 Jan Hubicka <jh@suse.cz>
6796 * cgraph.c (cgraph_get_body): New function based on lto.c
6798 * cgraph.h (cgraph_get_body): Declare.
6799 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
6801 * cgraphunit.c (expand_function): Get body prior expanding.
6802 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
6803 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
6805 * passes.c (do_per_function_toporder): Get body.
6806 * tree-inline.c (expand_call_inline): Get body prior inlining it.
6807 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
6809 2013-08-06 Martin Jambor <mjambor@suse.cz>
6812 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
6813 is not re-finalized. Rename second parameter to no_collect.
6815 2013-08-06 Martin Jambor <mjambor@suse.cz>
6818 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
6819 MEM_REF has proper alignment information.
6821 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
6824 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
6826 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
6827 rtx (*) (rtx, ...) with insn_gen_fn.
6828 * genoutput.c (output_insn_data): Cast gen_? function pointers to
6829 insn_gen_fn::stored_funcptr. Add initializer braces.
6831 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6833 Rewrite how instances of passes are cloned to remove assumptions
6834 about their sizes (thus allowing pass subclasses to have
6835 additional data fields, albeit non-GC-managed ones at this point).
6837 * passes.c (make_pass_instance): Now that passes have clone
6838 methods, rewrite this function to eliminate XNEW and memcpy
6839 calls that used hardcoded sizes. Since this function no longer
6840 creates pass instances, rename it to...
6841 (add_pass_instance): ...this. Document the old way that passes were
6842 numbered and flagged, and rework this function to continue using it.
6843 (next_pass_1): Add an initial_pass argument for use by
6845 (position_pass): When adding multiple instances of a pass, use
6846 the pass's clone method, rather than relying on the XNEW/memcpy
6847 within the former make_pass_instance (now add_pass_instance).
6848 (pass_manager::pass_manager): When invoking next_pass_1, also supply
6849 the initial instance of the current pass within the pass manager.
6851 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6853 This is the automated part of the conversion of passes from C
6854 structs to C++ classes.
6856 Patch autogenerated by refactor_passes.py from
6857 https://github.com/davidmalcolm/gcc-refactoring-scripts
6858 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
6860 * asan.c (pass_asan): Convert from a global struct to a subclass of
6861 gimple_opt_pass along with...
6862 (pass_data_asan): ...new pass_data instance and...
6863 (make_pass_asan): ...new function.
6864 (pass_asan_O0): Convert from a global struct to a subclass of
6865 gimple_opt_pass along with...
6866 (pass_data_asan_O0): ...new pass_data instance and...
6867 (make_pass_asan_O0): ...new function.
6868 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
6869 subclass of rtl_opt_pass along with...
6870 (pass_data_inc_dec): ...new pass_data instance and...
6871 (make_pass_inc_dec): ...new function.
6872 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
6873 a subclass of rtl_opt_pass along with...
6874 (pass_data_reorder_blocks): ...new pass_data instance and...
6875 (make_pass_reorder_blocks): ...new function.
6876 (pass_duplicate_computed_gotos): Convert from a global struct to a
6877 subclass of rtl_opt_pass along with...
6878 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
6879 (make_pass_duplicate_computed_gotos): ...new function.
6880 (pass_partition_blocks): Convert from a global struct to a subclass of
6881 rtl_opt_pass along with...
6882 (pass_data_partition_blocks): ...new pass_data instance and...
6883 (make_pass_partition_blocks): ...new function.
6884 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
6885 struct to a subclass of rtl_opt_pass along with...
6886 (pass_data_branch_target_load_optimize1): ...new pass_data instance
6888 (make_pass_branch_target_load_optimize1): ...new function.
6889 (pass_branch_target_load_optimize2): Convert from a global struct to a
6890 subclass of rtl_opt_pass along with...
6891 (pass_data_branch_target_load_optimize2): ...new pass_data instance
6893 (make_pass_branch_target_load_optimize2): ...new function.
6894 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
6895 of rtl_opt_pass along with...
6896 (pass_data_jump): ...new pass_data instance and...
6897 (make_pass_jump): ...new function.
6898 (pass_jump2): Convert from a global struct to a subclass of
6899 rtl_opt_pass along with...
6900 (pass_data_jump2): ...new pass_data instance and...
6901 (make_pass_jump2): ...new function.
6902 * cfgexpand.c (pass_expand): Convert from a global struct to a
6903 subclass of rtl_opt_pass along with...
6904 (pass_data_expand): ...new pass_data instance and...
6905 (make_pass_expand): ...new function.
6906 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
6907 of rtl_opt_pass along with...
6908 (pass_data_free_cfg): ...new pass_data instance and...
6909 (make_pass_free_cfg): ...new function.
6910 (pass_into_cfg_layout_mode): Convert from a global struct to a
6911 subclass of rtl_opt_pass along with...
6912 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
6913 (make_pass_into_cfg_layout_mode): ...new function.
6914 (pass_outof_cfg_layout_mode): Convert from a global struct to a
6915 subclass of rtl_opt_pass along with...
6916 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
6917 (make_pass_outof_cfg_layout_mode): ...new function.
6918 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
6919 struct to a subclass of gimple_opt_pass along with...
6920 (pass_data_build_cgraph_edges): ...new pass_data instance and...
6921 (make_pass_build_cgraph_edges): ...new function.
6922 (pass_rebuild_cgraph_edges): Convert from a global struct to a
6923 subclass of gimple_opt_pass along with...
6924 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
6925 (make_pass_rebuild_cgraph_edges): ...new function.
6926 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
6927 subclass of gimple_opt_pass along with...
6928 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
6930 (make_pass_remove_cgraph_callee_edges): ...new function.
6931 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
6932 struct to a subclass of rtl_opt_pass along with...
6933 (pass_data_stack_adjustments): ...new pass_data instance and...
6934 (make_pass_stack_adjustments): ...new function.
6935 * combine.c (pass_combine): Convert from a global struct to a subclass
6936 of rtl_opt_pass along with...
6937 (pass_data_combine): ...new pass_data instance and...
6938 (make_pass_combine): ...new function.
6939 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
6940 global struct to a subclass of rtl_opt_pass along with...
6941 (pass_data_compare_elim_after_reload): ...new pass_data instance
6943 (make_pass_compare_elim_after_reload): ...new function.
6944 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
6945 of rtl_opt_pass along with...
6946 (pass_data_rtl_cprop): ...new pass_data instance and...
6947 (make_pass_rtl_cprop): ...new function.
6948 * cse.c (pass_cse): Convert from a global struct to a subclass of
6949 rtl_opt_pass along with...
6950 (pass_data_cse): ...new pass_data instance and...
6951 (make_pass_cse): ...new function.
6952 (pass_cse2): Convert from a global struct to a subclass of
6953 rtl_opt_pass along with...
6954 (pass_data_cse2): ...new pass_data instance and...
6955 (make_pass_cse2): ...new function.
6956 (pass_cse_after_global_opts): Convert from a global struct to a
6957 subclass of rtl_opt_pass along with...
6958 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6959 (make_pass_cse_after_global_opts): ...new function.
6960 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6961 of rtl_opt_pass along with...
6962 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6963 (make_pass_ud_rtl_dce): ...new function.
6964 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6965 rtl_opt_pass along with...
6966 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6967 (make_pass_fast_rtl_dce): ...new function.
6968 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6969 a subclass of rtl_opt_pass along with...
6970 (pass_data_df_initialize_opt): ...new pass_data instance and...
6971 (make_pass_df_initialize_opt): ...new function.
6972 (pass_df_initialize_no_opt): Convert from a global struct to a
6973 subclass of rtl_opt_pass along with...
6974 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6975 (make_pass_df_initialize_no_opt): ...new function.
6976 (pass_df_finish): Convert from a global struct to a subclass of
6977 rtl_opt_pass along with...
6978 (pass_data_df_finish): ...new pass_data instance and...
6979 (make_pass_df_finish): ...new function.
6980 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6981 rtl_opt_pass along with...
6982 (pass_data_rtl_dse1): ...new pass_data instance and...
6983 (make_pass_rtl_dse1): ...new function.
6984 (pass_rtl_dse2): Convert from a global struct to a subclass of
6985 rtl_opt_pass along with...
6986 (pass_data_rtl_dse2): ...new pass_data instance and...
6987 (make_pass_rtl_dse2): ...new function.
6988 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6989 subclass of rtl_opt_pass along with...
6990 (pass_data_dwarf2_frame): ...new pass_data instance and...
6991 (make_pass_dwarf2_frame): ...new function.
6992 * except.c (pass_set_nothrow_function_flags): Convert from a global
6993 struct to a subclass of rtl_opt_pass along with...
6994 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6996 (make_pass_set_nothrow_function_flags): ...new function.
6997 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6998 subclass of rtl_opt_pass along with...
6999 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
7001 (make_pass_convert_to_eh_region_ranges): ...new function.
7002 * final.c (pass_compute_alignments): Convert from a global struct to a
7003 subclass of rtl_opt_pass along with...
7004 (pass_data_compute_alignments): ...new pass_data instance and...
7005 (make_pass_compute_alignments): ...new function.
7006 (pass_final): Convert from a global struct to a subclass of
7007 rtl_opt_pass along with...
7008 (pass_data_final): ...new pass_data instance and...
7009 (make_pass_final): ...new function.
7010 (pass_shorten_branches): Convert from a global struct to a subclass of
7011 rtl_opt_pass along with...
7012 (pass_data_shorten_branches): ...new pass_data instance and...
7013 (make_pass_shorten_branches): ...new function.
7014 (pass_clean_state): Convert from a global struct to a subclass of
7015 rtl_opt_pass along with...
7016 (pass_data_clean_state): ...new pass_data instance and...
7017 (make_pass_clean_state): ...new function.
7018 * function.c (pass_instantiate_virtual_regs): Convert from a global
7019 struct to a subclass of rtl_opt_pass along with...
7020 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
7021 (make_pass_instantiate_virtual_regs): ...new function.
7022 (pass_leaf_regs): Convert from a global struct to a subclass of
7023 rtl_opt_pass along with...
7024 (pass_data_leaf_regs): ...new pass_data instance and...
7025 (make_pass_leaf_regs): ...new function.
7026 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
7027 subclass of rtl_opt_pass along with...
7028 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
7030 (make_pass_thread_prologue_and_epilogue): ...new function.
7031 (pass_match_asm_constraints): Convert from a global struct to a
7032 subclass of rtl_opt_pass along with...
7033 (pass_data_match_asm_constraints): ...new pass_data instance and...
7034 (make_pass_match_asm_constraints): ...new function.
7035 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
7036 subclass of rtl_opt_pass along with...
7037 (pass_data_rtl_fwprop): ...new pass_data instance and...
7038 (make_pass_rtl_fwprop): ...new function.
7039 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
7040 rtl_opt_pass along with...
7041 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
7042 (make_pass_rtl_fwprop_addr): ...new function.
7043 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
7044 rtl_opt_pass along with...
7045 (pass_data_rtl_pre): ...new pass_data instance and...
7046 (make_pass_rtl_pre): ...new function.
7047 (pass_rtl_hoist): Convert from a global struct to a subclass of
7048 rtl_opt_pass along with...
7049 (pass_data_rtl_hoist): ...new pass_data instance and...
7050 (make_pass_rtl_hoist): ...new function.
7051 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
7052 subclass of gimple_opt_pass along with...
7053 (pass_data_lower_cf): ...new pass_data instance and...
7054 (make_pass_lower_cf): ...new function.
7055 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
7056 from a global struct to a subclass of gimple_opt_pass along with...
7057 (pass_data_strength_reduction): ...new pass_data instance and...
7058 (make_pass_strength_reduction): ...new function.
7059 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
7060 of rtl_opt_pass along with...
7061 (pass_data_rtl_ifcvt): ...new pass_data instance and...
7062 (make_pass_rtl_ifcvt): ...new function.
7063 (pass_if_after_combine): Convert from a global struct to a subclass of
7064 rtl_opt_pass along with...
7065 (pass_data_if_after_combine): ...new pass_data instance and...
7066 (make_pass_if_after_combine): ...new function.
7067 (pass_if_after_reload): Convert from a global struct to a subclass of
7068 rtl_opt_pass along with...
7069 (pass_data_if_after_reload): ...new pass_data instance and...
7070 (make_pass_if_after_reload): ...new function.
7071 * init-regs.c (pass_initialize_regs): Convert from a global struct to
7072 a subclass of rtl_opt_pass along with...
7073 (pass_data_initialize_regs): ...new pass_data instance and...
7074 (make_pass_initialize_regs): ...new function.
7075 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
7076 of ipa_opt_pass_d along with...
7077 (pass_data_ipa_cp): ...new pass_data instance and...
7078 (make_pass_ipa_cp): ...new function.
7079 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
7080 global struct to a subclass of gimple_opt_pass along with...
7081 (pass_data_inline_parameters): ...new pass_data instance and...
7082 (make_pass_inline_parameters): ...new function.
7083 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
7084 subclass of gimple_opt_pass along with...
7085 (pass_data_early_inline): ...new pass_data instance and...
7086 (make_pass_early_inline): ...new function.
7087 (pass_ipa_inline): Convert from a global struct to a subclass of
7088 ipa_opt_pass_d along with...
7089 (pass_data_ipa_inline): ...new pass_data instance and...
7090 (make_pass_ipa_inline): ...new function.
7091 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
7092 struct to a subclass of gimple_opt_pass along with...
7093 (pass_data_local_pure_const): ...new pass_data instance and...
7094 (make_pass_local_pure_const): ...new function.
7095 (pass_ipa_pure_const): Convert from a global struct to a subclass of
7096 ipa_opt_pass_d along with...
7097 (pass_data_ipa_pure_const): ...new pass_data instance and...
7098 (make_pass_ipa_pure_const): ...new function.
7099 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
7100 to a subclass of ipa_opt_pass_d along with...
7101 (pass_data_ipa_reference): ...new pass_data instance and...
7102 (make_pass_ipa_reference): ...new function.
7103 * ipa-split.c (pass_split_functions): Convert from a global struct to
7104 a subclass of gimple_opt_pass along with...
7105 (pass_data_split_functions): ...new pass_data instance and...
7106 (make_pass_split_functions): ...new function.
7107 (pass_feedback_split_functions): Convert from a global struct to a
7108 subclass of gimple_opt_pass along with...
7109 (pass_data_feedback_split_functions): ...new pass_data instance and...
7110 (make_pass_feedback_split_functions): ...new function.
7111 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
7112 global struct to a subclass of simple_ipa_opt_pass along with...
7113 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
7115 (make_pass_ipa_function_and_variable_visibility): ...new function.
7116 (pass_ipa_free_inline_summary): Convert from a global struct to a
7117 subclass of simple_ipa_opt_pass along with...
7118 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
7119 (make_pass_ipa_free_inline_summary): ...new function.
7120 (pass_ipa_whole_program_visibility): Convert from a global struct to a
7121 subclass of ipa_opt_pass_d along with...
7122 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
7124 (make_pass_ipa_whole_program_visibility): ...new function.
7125 (pass_ipa_profile): Convert from a global struct to a subclass of
7126 ipa_opt_pass_d along with...
7127 (pass_data_ipa_profile): ...new pass_data instance and...
7128 (make_pass_ipa_profile): ...new function.
7129 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
7130 ipa_opt_pass_d along with...
7131 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
7132 (make_pass_ipa_cdtor_merge): ...new function.
7133 * ira.c (pass_ira): Convert from a global struct to a subclass of
7134 rtl_opt_pass along with...
7135 (pass_data_ira): ...new pass_data instance and...
7136 (make_pass_ira): ...new function.
7137 (pass_reload): Convert from a global struct to a subclass of
7138 rtl_opt_pass along with...
7139 (pass_data_reload): ...new pass_data instance and...
7140 (make_pass_reload): ...new function.
7141 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
7142 subclass of rtl_opt_pass along with...
7143 (pass_data_cleanup_barriers): ...new pass_data instance and...
7144 (make_pass_cleanup_barriers): ...new function.
7145 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
7146 of rtl_opt_pass along with...
7147 (pass_data_loop2): ...new pass_data instance and...
7148 (make_pass_loop2): ...new function.
7149 (pass_rtl_loop_init): Convert from a global struct to a subclass of
7150 rtl_opt_pass along with...
7151 (pass_data_rtl_loop_init): ...new pass_data instance and...
7152 (make_pass_rtl_loop_init): ...new function.
7153 (pass_rtl_loop_done): Convert from a global struct to a subclass of
7154 rtl_opt_pass along with...
7155 (pass_data_rtl_loop_done): ...new pass_data instance and...
7156 (make_pass_rtl_loop_done): ...new function.
7157 (pass_rtl_move_loop_invariants): Convert from a global struct to a
7158 subclass of rtl_opt_pass along with...
7159 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
7160 (make_pass_rtl_move_loop_invariants): ...new function.
7161 (pass_rtl_unswitch): Convert from a global struct to a subclass of
7162 rtl_opt_pass along with...
7163 (pass_data_rtl_unswitch): ...new pass_data instance and...
7164 (make_pass_rtl_unswitch): ...new function.
7165 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
7166 subclass of rtl_opt_pass along with...
7167 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
7169 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7170 (pass_rtl_doloop): Convert from a global struct to a subclass of
7171 rtl_opt_pass along with...
7172 (pass_data_rtl_doloop): ...new pass_data instance and...
7173 (make_pass_rtl_doloop): ...new function.
7174 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
7175 a subclass of rtl_opt_pass along with...
7176 (pass_data_lower_subreg): ...new pass_data instance and...
7177 (make_pass_lower_subreg): ...new function.
7178 (pass_lower_subreg2): Convert from a global struct to a subclass of
7179 rtl_opt_pass along with...
7180 (pass_data_lower_subreg2): ...new pass_data instance and...
7181 (make_pass_lower_subreg2): ...new function.
7182 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
7183 struct to a subclass of ipa_opt_pass_d along with...
7184 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
7185 (make_pass_ipa_lto_gimple_out): ...new function.
7186 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
7187 of ipa_opt_pass_d along with...
7188 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
7189 (make_pass_ipa_lto_finish_out): ...new function.
7190 * mode-switching.c (pass_mode_switching): Convert from a global struct
7191 to a subclass of rtl_opt_pass along with...
7192 (pass_data_mode_switching): ...new pass_data instance and...
7193 (make_pass_mode_switching): ...new function.
7194 * modulo-sched.c (pass_sms): Convert from a global struct to a
7195 subclass of rtl_opt_pass along with...
7196 (pass_data_sms): ...new pass_data instance and...
7197 (make_pass_sms): ...new function.
7198 * omp-low.c (pass_expand_omp): Convert from a global struct to a
7199 subclass of gimple_opt_pass along with...
7200 (pass_data_expand_omp): ...new pass_data instance and...
7201 (make_pass_expand_omp): ...new function.
7202 (pass_lower_omp): Convert from a global struct to a subclass of
7203 gimple_opt_pass along with...
7204 (pass_data_lower_omp): ...new pass_data instance and...
7205 (make_pass_lower_omp): ...new function.
7206 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
7207 of gimple_opt_pass along with...
7208 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
7209 (make_pass_diagnose_omp_blocks): ...new function.
7210 * passes.c (pass_early_local_passes): Convert from a global struct to
7211 a subclass of simple_ipa_opt_pass along with...
7212 (pass_data_early_local_passes): ...new pass_data instance and...
7213 (make_pass_early_local_passes): ...new function.
7214 (pass_all_early_optimizations): Convert from a global struct to a
7215 subclass of gimple_opt_pass along with...
7216 (pass_data_all_early_optimizations): ...new pass_data instance and...
7217 (make_pass_all_early_optimizations): ...new function.
7218 (pass_all_optimizations): Convert from a global struct to a subclass
7219 of gimple_opt_pass along with...
7220 (pass_data_all_optimizations): ...new pass_data instance and...
7221 (make_pass_all_optimizations): ...new function.
7222 (pass_all_optimizations_g): Convert from a global struct to a subclass
7223 of gimple_opt_pass along with...
7224 (pass_data_all_optimizations_g): ...new pass_data instance and...
7225 (make_pass_all_optimizations_g): ...new function.
7226 (pass_rest_of_compilation): Convert from a global struct to a subclass
7227 of rtl_opt_pass along with...
7228 (pass_data_rest_of_compilation): ...new pass_data instance and...
7229 (make_pass_rest_of_compilation): ...new function.
7230 (pass_postreload): Convert from a global struct to a subclass of
7231 rtl_opt_pass along with...
7232 (pass_data_postreload): ...new pass_data instance and...
7233 (make_pass_postreload): ...new function.
7234 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
7235 subclass of rtl_opt_pass along with...
7236 (pass_data_gcse2): ...new pass_data instance and...
7237 (make_pass_gcse2): ...new function.
7238 * postreload.c (pass_postreload_cse): Convert from a global struct to
7239 a subclass of rtl_opt_pass along with...
7240 (pass_data_postreload_cse): ...new pass_data instance and...
7241 (make_pass_postreload_cse): ...new function.
7242 * predict.c (pass_profile): Convert from a global struct to a subclass
7243 of gimple_opt_pass along with...
7244 (pass_data_profile): ...new pass_data instance and...
7245 (make_pass_profile): ...new function.
7246 (pass_strip_predict_hints): Convert from a global struct to a subclass
7247 of gimple_opt_pass along with...
7248 (pass_data_strip_predict_hints): ...new pass_data instance and...
7249 (make_pass_strip_predict_hints): ...new function.
7250 * recog.c (pass_peephole2): Convert from a global struct to a subclass
7251 of rtl_opt_pass along with...
7252 (pass_data_peephole2): ...new pass_data instance and...
7253 (make_pass_peephole2): ...new function.
7254 (pass_split_all_insns): Convert from a global struct to a subclass of
7255 rtl_opt_pass along with...
7256 (pass_data_split_all_insns): ...new pass_data instance and...
7257 (make_pass_split_all_insns): ...new function.
7258 (pass_split_after_reload): Convert from a global struct to a subclass
7259 of rtl_opt_pass along with...
7260 (pass_data_split_after_reload): ...new pass_data instance and...
7261 (make_pass_split_after_reload): ...new function.
7262 (pass_split_before_regstack): Convert from a global struct to a
7263 subclass of rtl_opt_pass along with...
7264 (pass_data_split_before_regstack): ...new pass_data instance and...
7265 (make_pass_split_before_regstack): ...new function.
7266 (pass_split_before_sched2): Convert from a global struct to a subclass
7267 of rtl_opt_pass along with...
7268 (pass_data_split_before_sched2): ...new pass_data instance and...
7269 (make_pass_split_before_sched2): ...new function.
7270 (pass_split_for_shorten_branches): Convert from a global struct to a
7271 subclass of rtl_opt_pass along with...
7272 (pass_data_split_for_shorten_branches): ...new pass_data instance
7274 (make_pass_split_for_shorten_branches): ...new function.
7275 * ree.c (pass_ree): Convert from a global struct to a subclass of
7276 rtl_opt_pass along with...
7277 (pass_data_ree): ...new pass_data instance and...
7278 (make_pass_ree): ...new function.
7279 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
7280 subclass of rtl_opt_pass along with...
7281 (pass_data_stack_regs): ...new pass_data instance and...
7282 (make_pass_stack_regs): ...new function.
7283 (pass_stack_regs_run): Convert from a global struct to a subclass of
7284 rtl_opt_pass along with...
7285 (pass_data_stack_regs_run): ...new pass_data instance and...
7286 (make_pass_stack_regs_run): ...new function.
7287 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
7288 subclass of rtl_opt_pass along with...
7289 (pass_data_cprop_hardreg): ...new pass_data instance and...
7290 (make_pass_cprop_hardreg): ...new function.
7291 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
7292 subclass of rtl_opt_pass along with...
7293 (pass_data_reginfo_init): ...new pass_data instance and...
7294 (make_pass_reginfo_init): ...new function.
7295 * regmove.c (pass_regmove): Convert from a global struct to a subclass
7296 of rtl_opt_pass along with...
7297 (pass_data_regmove): ...new pass_data instance and...
7298 (make_pass_regmove): ...new function.
7299 * regrename.c (pass_regrename): Convert from a global struct to a
7300 subclass of rtl_opt_pass along with...
7301 (pass_data_regrename): ...new pass_data instance and...
7302 (make_pass_regrename): ...new function.
7303 * reorg.c (pass_delay_slots): Convert from a global struct to a
7304 subclass of rtl_opt_pass along with...
7305 (pass_data_delay_slots): ...new pass_data instance and...
7306 (make_pass_delay_slots): ...new function.
7307 (pass_machine_reorg): Convert from a global struct to a subclass of
7308 rtl_opt_pass along with...
7309 (pass_data_machine_reorg): ...new pass_data instance and...
7310 (make_pass_machine_reorg): ...new function.
7311 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
7312 of rtl_opt_pass along with...
7313 (pass_data_sched): ...new pass_data instance and...
7314 (make_pass_sched): ...new function.
7315 (pass_sched2): Convert from a global struct to a subclass of
7316 rtl_opt_pass along with...
7317 (pass_data_sched2): ...new pass_data instance and...
7318 (make_pass_sched2): ...new function.
7319 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
7320 to a subclass of rtl_opt_pass along with...
7321 (pass_data_stack_ptr_mod): ...new pass_data instance and...
7322 (make_pass_stack_ptr_mod): ...new function.
7323 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
7324 to a subclass of rtl_opt_pass along with...
7325 (pass_data_rtl_store_motion): ...new pass_data instance and...
7326 (make_pass_rtl_store_motion): ...new function.
7327 * tracer.c (pass_tracer): Convert from a global struct to a subclass
7328 of gimple_opt_pass along with...
7329 (pass_data_tracer): ...new pass_data instance and...
7330 (make_pass_tracer): ...new function.
7331 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
7332 to a subclass of gimple_opt_pass along with...
7333 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
7334 (make_pass_diagnose_tm_blocks): ...new function.
7335 (pass_lower_tm): Convert from a global struct to a subclass of
7336 gimple_opt_pass along with...
7337 (pass_data_lower_tm): ...new pass_data instance and...
7338 (make_pass_lower_tm): ...new function.
7339 (pass_tm_init): Convert from a global struct to a subclass of
7340 gimple_opt_pass along with...
7341 (pass_data_tm_init): ...new pass_data instance and...
7342 (make_pass_tm_init): ...new function.
7343 (pass_tm_mark): Convert from a global struct to a subclass of
7344 gimple_opt_pass along with...
7345 (pass_data_tm_mark): ...new pass_data instance and...
7346 (make_pass_tm_mark): ...new function.
7347 (pass_tm_edges): Convert from a global struct to a subclass of
7348 gimple_opt_pass along with...
7349 (pass_data_tm_edges): ...new pass_data instance and...
7350 (make_pass_tm_edges): ...new function.
7351 (pass_tm_memopt): Convert from a global struct to a subclass of
7352 gimple_opt_pass along with...
7353 (pass_data_tm_memopt): ...new pass_data instance and...
7354 (make_pass_tm_memopt): ...new function.
7355 (pass_ipa_tm): Convert from a global struct to a subclass of
7356 simple_ipa_opt_pass along with...
7357 (pass_data_ipa_tm): ...new pass_data instance and...
7358 (make_pass_ipa_tm): ...new function.
7359 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
7360 subclass of gimple_opt_pass along with...
7361 (pass_data_call_cdce): ...new pass_data instance and...
7362 (make_pass_call_cdce): ...new function.
7363 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
7364 subclass of gimple_opt_pass along with...
7365 (pass_data_build_cfg): ...new pass_data instance and...
7366 (make_pass_build_cfg): ...new function.
7367 (pass_split_crit_edges): Convert from a global struct to a subclass of
7368 gimple_opt_pass along with...
7369 (pass_data_split_crit_edges): ...new pass_data instance and...
7370 (make_pass_split_crit_edges): ...new function.
7371 (pass_warn_function_return): Convert from a global struct to a
7372 subclass of gimple_opt_pass along with...
7373 (pass_data_warn_function_return): ...new pass_data instance and...
7374 (make_pass_warn_function_return): ...new function.
7375 (pass_warn_function_noreturn): Convert from a global struct to a
7376 subclass of gimple_opt_pass along with...
7377 (pass_data_warn_function_noreturn): ...new pass_data instance and...
7378 (make_pass_warn_function_noreturn): ...new function.
7379 (pass_warn_unused_result): Convert from a global struct to a subclass
7380 of gimple_opt_pass along with...
7381 (pass_data_warn_unused_result): ...new pass_data instance and...
7382 (make_pass_warn_unused_result): ...new function.
7383 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
7384 a subclass of gimple_opt_pass along with...
7385 (pass_data_merge_phi): ...new pass_data instance and...
7386 (make_pass_merge_phi): ...new function.
7387 * tree-complex.c (pass_lower_complex): Convert from a global struct to
7388 a subclass of gimple_opt_pass along with...
7389 (pass_data_lower_complex): ...new pass_data instance and...
7390 (make_pass_lower_complex): ...new function.
7391 (pass_lower_complex_O0): Convert from a global struct to a subclass of
7392 gimple_opt_pass along with...
7393 (pass_data_lower_complex_O0): ...new pass_data instance and...
7394 (make_pass_lower_complex_O0): ...new function.
7395 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
7396 subclass of gimple_opt_pass along with...
7397 (pass_data_lower_eh): ...new pass_data instance and...
7398 (make_pass_lower_eh): ...new function.
7399 (pass_refactor_eh): Convert from a global struct to a subclass of
7400 gimple_opt_pass along with...
7401 (pass_data_refactor_eh): ...new pass_data instance and...
7402 (make_pass_refactor_eh): ...new function.
7403 (pass_lower_resx): Convert from a global struct to a subclass of
7404 gimple_opt_pass along with...
7405 (pass_data_lower_resx): ...new pass_data instance and...
7406 (make_pass_lower_resx): ...new function.
7407 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
7408 of gimple_opt_pass along with...
7409 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
7410 (make_pass_lower_eh_dispatch): ...new function.
7411 (pass_cleanup_eh): Convert from a global struct to a subclass of
7412 gimple_opt_pass along with...
7413 (pass_data_cleanup_eh): ...new pass_data instance and...
7414 (make_pass_cleanup_eh): ...new function.
7415 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
7416 to a subclass of simple_ipa_opt_pass along with...
7417 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
7418 (make_pass_ipa_lower_emutls): ...new function.
7419 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
7420 a subclass of gimple_opt_pass along with...
7421 (pass_data_if_conversion): ...new pass_data instance and...
7422 (make_pass_if_conversion): ...new function.
7423 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
7424 subclass of gimple_opt_pass along with...
7425 (pass_data_build_ssa): ...new pass_data instance and...
7426 (make_pass_build_ssa): ...new function.
7427 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
7428 global struct to a subclass of gimple_opt_pass along with...
7429 (pass_data_loop_distribution): ...new pass_data instance and...
7430 (make_pass_loop_distribution): ...new function.
7431 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
7432 subclass of gimple_opt_pass along with...
7433 (pass_data_mudflap_1): ...new pass_data instance and...
7434 (make_pass_mudflap_1): ...new function.
7435 (pass_mudflap_2): Convert from a global struct to a subclass of
7436 gimple_opt_pass along with...
7437 (pass_data_mudflap_2): ...new pass_data instance and...
7438 (make_pass_mudflap_2): ...new function.
7439 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
7440 subclass of gimple_opt_pass along with...
7441 (pass_data_mudflap_1): ...new pass_data instance and...
7442 (make_pass_mudflap_1): ...new function.
7443 (pass_mudflap_2): Convert from a global struct to a subclass of
7444 gimple_opt_pass along with...
7445 (pass_data_mudflap_2): ...new pass_data instance and...
7446 (make_pass_mudflap_2): ...new function.
7447 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
7448 gimple_opt_pass along with...
7449 (pass_data_nrv): ...new pass_data instance and...
7450 (make_pass_nrv): ...new function.
7451 (pass_return_slot): Convert from a global struct to a subclass of
7452 gimple_opt_pass along with...
7453 (pass_data_return_slot): ...new pass_data instance and...
7454 (make_pass_return_slot): ...new function.
7455 * tree-object-size.c (pass_object_sizes): Convert from a global struct
7456 to a subclass of gimple_opt_pass along with...
7457 (pass_data_object_sizes): ...new pass_data instance and...
7458 (make_pass_object_sizes): ...new function.
7459 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
7460 global struct to a subclass of gimple_opt_pass along with...
7461 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
7463 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7464 (pass_fixup_cfg): Convert from a global struct to a subclass of
7465 gimple_opt_pass along with...
7466 (pass_data_fixup_cfg): ...new pass_data instance and...
7467 (make_pass_fixup_cfg): ...new function.
7468 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
7469 (make_pass_mudflap_1): ...new function.
7470 (pass_mudflap_2): Replace declaration with that of...
7471 (make_pass_mudflap_2): ...new function.
7472 (pass_asan): Replace declaration with that of...
7473 (make_pass_asan): ...new function.
7474 (pass_asan_O0): Replace declaration with that of...
7475 (make_pass_asan_O0): ...new function.
7476 (pass_tsan): Replace declaration with that of...
7477 (make_pass_tsan): ...new function.
7478 (pass_tsan_O0): Replace declaration with that of...
7479 (make_pass_tsan_O0): ...new function.
7480 (pass_lower_cf): Replace declaration with that of...
7481 (make_pass_lower_cf): ...new function.
7482 (pass_refactor_eh): Replace declaration with that of...
7483 (make_pass_refactor_eh): ...new function.
7484 (pass_lower_eh): Replace declaration with that of...
7485 (make_pass_lower_eh): ...new function.
7486 (pass_lower_eh_dispatch): Replace declaration with that of...
7487 (make_pass_lower_eh_dispatch): ...new function.
7488 (pass_lower_resx): Replace declaration with that of...
7489 (make_pass_lower_resx): ...new function.
7490 (pass_build_cfg): Replace declaration with that of...
7491 (make_pass_build_cfg): ...new function.
7492 (pass_early_tree_profile): Replace declaration with that of...
7493 (make_pass_early_tree_profile): ...new function.
7494 (pass_cleanup_eh): Replace declaration with that of...
7495 (make_pass_cleanup_eh): ...new function.
7496 (pass_sra): Replace declaration with that of...
7497 (make_pass_sra): ...new function.
7498 (pass_sra_early): Replace declaration with that of...
7499 (make_pass_sra_early): ...new function.
7500 (pass_early_ipa_sra): Replace declaration with that of...
7501 (make_pass_early_ipa_sra): ...new function.
7502 (pass_tail_recursion): Replace declaration with that of...
7503 (make_pass_tail_recursion): ...new function.
7504 (pass_tail_calls): Replace declaration with that of...
7505 (make_pass_tail_calls): ...new function.
7506 (pass_tree_loop): Replace declaration with that of...
7507 (make_pass_tree_loop): ...new function.
7508 (pass_tree_loop_init): Replace declaration with that of...
7509 (make_pass_tree_loop_init): ...new function.
7510 (pass_lim): Replace declaration with that of...
7511 (make_pass_lim): ...new function.
7512 (pass_tree_unswitch): Replace declaration with that of...
7513 (make_pass_tree_unswitch): ...new function.
7514 (pass_predcom): Replace declaration with that of...
7515 (make_pass_predcom): ...new function.
7516 (pass_iv_canon): Replace declaration with that of...
7517 (make_pass_iv_canon): ...new function.
7518 (pass_scev_cprop): Replace declaration with that of...
7519 (make_pass_scev_cprop): ...new function.
7520 (pass_empty_loop): Replace declaration with that of...
7521 (make_pass_empty_loop): ...new function.
7522 (pass_record_bounds): Replace declaration with that of...
7523 (make_pass_record_bounds): ...new function.
7524 (pass_graphite): Replace declaration with that of...
7525 (make_pass_graphite): ...new function.
7526 (pass_graphite_transforms): Replace declaration with that of...
7527 (make_pass_graphite_transforms): ...new function.
7528 (pass_if_conversion): Replace declaration with that of...
7529 (make_pass_if_conversion): ...new function.
7530 (pass_loop_distribution): Replace declaration with that of...
7531 (make_pass_loop_distribution): ...new function.
7532 (pass_vectorize): Replace declaration with that of...
7533 (make_pass_vectorize): ...new function.
7534 (pass_slp_vectorize): Replace declaration with that of...
7535 (make_pass_slp_vectorize): ...new function.
7536 (pass_complete_unroll): Replace declaration with that of...
7537 (make_pass_complete_unroll): ...new function.
7538 (pass_complete_unrolli): Replace declaration with that of...
7539 (make_pass_complete_unrolli): ...new function.
7540 (pass_parallelize_loops): Replace declaration with that of...
7541 (make_pass_parallelize_loops): ...new function.
7542 (pass_loop_prefetch): Replace declaration with that of...
7543 (make_pass_loop_prefetch): ...new function.
7544 (pass_iv_optimize): Replace declaration with that of...
7545 (make_pass_iv_optimize): ...new function.
7546 (pass_tree_loop_done): Replace declaration with that of...
7547 (make_pass_tree_loop_done): ...new function.
7548 (pass_ch): Replace declaration with that of...
7549 (make_pass_ch): ...new function.
7550 (pass_ccp): Replace declaration with that of...
7551 (make_pass_ccp): ...new function.
7552 (pass_phi_only_cprop): Replace declaration with that of...
7553 (make_pass_phi_only_cprop): ...new function.
7554 (pass_build_ssa): Replace declaration with that of...
7555 (make_pass_build_ssa): ...new function.
7556 (pass_build_alias): Replace declaration with that of...
7557 (make_pass_build_alias): ...new function.
7558 (pass_build_ealias): Replace declaration with that of...
7559 (make_pass_build_ealias): ...new function.
7560 (pass_dominator): Replace declaration with that of...
7561 (make_pass_dominator): ...new function.
7562 (pass_dce): Replace declaration with that of...
7563 (make_pass_dce): ...new function.
7564 (pass_dce_loop): Replace declaration with that of...
7565 (make_pass_dce_loop): ...new function.
7566 (pass_cd_dce): Replace declaration with that of...
7567 (make_pass_cd_dce): ...new function.
7568 (pass_call_cdce): Replace declaration with that of...
7569 (make_pass_call_cdce): ...new function.
7570 (pass_merge_phi): Replace declaration with that of...
7571 (make_pass_merge_phi): ...new function.
7572 (pass_split_crit_edges): Replace declaration with that of...
7573 (make_pass_split_crit_edges): ...new function.
7574 (pass_pre): Replace declaration with that of...
7575 (make_pass_pre): ...new function.
7576 (pass_profile): Replace declaration with that of...
7577 (make_pass_profile): ...new function.
7578 (pass_strip_predict_hints): Replace declaration with that of...
7579 (make_pass_strip_predict_hints): ...new function.
7580 (pass_lower_complex_O0): Replace declaration with that of...
7581 (make_pass_lower_complex_O0): ...new function.
7582 (pass_lower_complex): Replace declaration with that of...
7583 (make_pass_lower_complex): ...new function.
7584 (pass_lower_vector): Replace declaration with that of...
7585 (make_pass_lower_vector): ...new function.
7586 (pass_lower_vector_ssa): Replace declaration with that of...
7587 (make_pass_lower_vector_ssa): ...new function.
7588 (pass_lower_omp): Replace declaration with that of...
7589 (make_pass_lower_omp): ...new function.
7590 (pass_diagnose_omp_blocks): Replace declaration with that of...
7591 (make_pass_diagnose_omp_blocks): ...new function.
7592 (pass_expand_omp): Replace declaration with that of...
7593 (make_pass_expand_omp): ...new function.
7594 (pass_expand_omp_ssa): Replace declaration with that of...
7595 (make_pass_expand_omp_ssa): ...new function.
7596 (pass_object_sizes): Replace declaration with that of...
7597 (make_pass_object_sizes): ...new function.
7598 (pass_strlen): Replace declaration with that of...
7599 (make_pass_strlen): ...new function.
7600 (pass_fold_builtins): Replace declaration with that of...
7601 (make_pass_fold_builtins): ...new function.
7602 (pass_stdarg): Replace declaration with that of...
7603 (make_pass_stdarg): ...new function.
7604 (pass_early_warn_uninitialized): Replace declaration with that of...
7605 (make_pass_early_warn_uninitialized): ...new function.
7606 (pass_late_warn_uninitialized): Replace declaration with that of...
7607 (make_pass_late_warn_uninitialized): ...new function.
7608 (pass_cse_reciprocals): Replace declaration with that of...
7609 (make_pass_cse_reciprocals): ...new function.
7610 (pass_cse_sincos): Replace declaration with that of...
7611 (make_pass_cse_sincos): ...new function.
7612 (pass_optimize_bswap): Replace declaration with that of...
7613 (make_pass_optimize_bswap): ...new function.
7614 (pass_optimize_widening_mul): Replace declaration with that of...
7615 (make_pass_optimize_widening_mul): ...new function.
7616 (pass_warn_function_return): Replace declaration with that of...
7617 (make_pass_warn_function_return): ...new function.
7618 (pass_warn_function_noreturn): Replace declaration with that of...
7619 (make_pass_warn_function_noreturn): ...new function.
7620 (pass_cselim): Replace declaration with that of...
7621 (make_pass_cselim): ...new function.
7622 (pass_phiopt): Replace declaration with that of...
7623 (make_pass_phiopt): ...new function.
7624 (pass_forwprop): Replace declaration with that of...
7625 (make_pass_forwprop): ...new function.
7626 (pass_phiprop): Replace declaration with that of...
7627 (make_pass_phiprop): ...new function.
7628 (pass_tree_ifcombine): Replace declaration with that of...
7629 (make_pass_tree_ifcombine): ...new function.
7630 (pass_dse): Replace declaration with that of...
7631 (make_pass_dse): ...new function.
7632 (pass_nrv): Replace declaration with that of...
7633 (make_pass_nrv): ...new function.
7634 (pass_rename_ssa_copies): Replace declaration with that of...
7635 (make_pass_rename_ssa_copies): ...new function.
7636 (pass_sink_code): Replace declaration with that of...
7637 (make_pass_sink_code): ...new function.
7638 (pass_fre): Replace declaration with that of...
7639 (make_pass_fre): ...new function.
7640 (pass_check_data_deps): Replace declaration with that of...
7641 (make_pass_check_data_deps): ...new function.
7642 (pass_copy_prop): Replace declaration with that of...
7643 (make_pass_copy_prop): ...new function.
7644 (pass_vrp): Replace declaration with that of...
7645 (make_pass_vrp): ...new function.
7646 (pass_uncprop): Replace declaration with that of...
7647 (make_pass_uncprop): ...new function.
7648 (pass_return_slot): Replace declaration with that of...
7649 (make_pass_return_slot): ...new function.
7650 (pass_reassoc): Replace declaration with that of...
7651 (make_pass_reassoc): ...new function.
7652 (pass_rebuild_cgraph_edges): Replace declaration with that of...
7653 (make_pass_rebuild_cgraph_edges): ...new function.
7654 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
7655 (make_pass_remove_cgraph_callee_edges): ...new function.
7656 (pass_build_cgraph_edges): Replace declaration with that of...
7657 (make_pass_build_cgraph_edges): ...new function.
7658 (pass_local_pure_const): Replace declaration with that of...
7659 (make_pass_local_pure_const): ...new function.
7660 (pass_tracer): Replace declaration with that of...
7661 (make_pass_tracer): ...new function.
7662 (pass_warn_unused_result): Replace declaration with that of...
7663 (make_pass_warn_unused_result): ...new function.
7664 (pass_diagnose_tm_blocks): Replace declaration with that of...
7665 (make_pass_diagnose_tm_blocks): ...new function.
7666 (pass_lower_tm): Replace declaration with that of...
7667 (make_pass_lower_tm): ...new function.
7668 (pass_tm_init): Replace declaration with that of...
7669 (make_pass_tm_init): ...new function.
7670 (pass_tm_mark): Replace declaration with that of...
7671 (make_pass_tm_mark): ...new function.
7672 (pass_tm_memopt): Replace declaration with that of...
7673 (make_pass_tm_memopt): ...new function.
7674 (pass_tm_edges): Replace declaration with that of...
7675 (make_pass_tm_edges): ...new function.
7676 (pass_split_functions): Replace declaration with that of...
7677 (make_pass_split_functions): ...new function.
7678 (pass_feedback_split_functions): Replace declaration with that of...
7679 (make_pass_feedback_split_functions): ...new function.
7680 (pass_strength_reduction): Replace declaration with that of...
7681 (make_pass_strength_reduction): ...new function.
7682 (pass_ipa_lower_emutls): Replace declaration with that of...
7683 (make_pass_ipa_lower_emutls): ...new function.
7684 (pass_ipa_function_and_variable_visibility): Replace declaration with
7686 (make_pass_ipa_function_and_variable_visibility): ...new function.
7687 (pass_ipa_tree_profile): Replace declaration with that of...
7688 (make_pass_ipa_tree_profile): ...new function.
7689 (pass_early_local_passes): Replace declaration with that of...
7690 (make_pass_early_local_passes): ...new function.
7691 (pass_ipa_whole_program_visibility): Replace declaration with that
7693 (make_pass_ipa_whole_program_visibility): ...new function.
7694 (pass_ipa_lto_gimple_out): Replace declaration with that of...
7695 (make_pass_ipa_lto_gimple_out): ...new function.
7696 (pass_ipa_increase_alignment): Replace declaration with that of...
7697 (make_pass_ipa_increase_alignment): ...new function.
7698 (pass_ipa_inline): Replace declaration with that of...
7699 (make_pass_ipa_inline): ...new function.
7700 (pass_ipa_free_lang_data): Replace declaration with that of...
7701 (make_pass_ipa_free_lang_data): ...new function.
7702 (pass_ipa_free_inline_summary): Replace declaration with that of...
7703 (make_pass_ipa_free_inline_summary): ...new function.
7704 (pass_ipa_cp): Replace declaration with that of...
7705 (make_pass_ipa_cp): ...new function.
7706 (pass_ipa_reference): Replace declaration with that of...
7707 (make_pass_ipa_reference): ...new function.
7708 (pass_ipa_pure_const): Replace declaration with that of...
7709 (make_pass_ipa_pure_const): ...new function.
7710 (pass_ipa_pta): Replace declaration with that of...
7711 (make_pass_ipa_pta): ...new function.
7712 (pass_ipa_lto_finish_out): Replace declaration with that of...
7713 (make_pass_ipa_lto_finish_out): ...new function.
7714 (pass_ipa_tm): Replace declaration with that of...
7715 (make_pass_ipa_tm): ...new function.
7716 (pass_ipa_profile): Replace declaration with that of...
7717 (make_pass_ipa_profile): ...new function.
7718 (pass_ipa_cdtor_merge): Replace declaration with that of...
7719 (make_pass_ipa_cdtor_merge): ...new function.
7720 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
7722 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7723 (pass_init_datastructures): Replace declaration with that of...
7724 (make_pass_init_datastructures): ...new function.
7725 (pass_fixup_cfg): Replace declaration with that of...
7726 (make_pass_fixup_cfg): ...new function.
7727 (pass_expand): Replace declaration with that of...
7728 (make_pass_expand): ...new function.
7729 (pass_instantiate_virtual_regs): Replace declaration with that of...
7730 (make_pass_instantiate_virtual_regs): ...new function.
7731 (pass_rtl_fwprop): Replace declaration with that of...
7732 (make_pass_rtl_fwprop): ...new function.
7733 (pass_rtl_fwprop_addr): Replace declaration with that of...
7734 (make_pass_rtl_fwprop_addr): ...new function.
7735 (pass_jump): Replace declaration with that of...
7736 (make_pass_jump): ...new function.
7737 (pass_jump2): Replace declaration with that of...
7738 (make_pass_jump2): ...new function.
7739 (pass_lower_subreg): Replace declaration with that of...
7740 (make_pass_lower_subreg): ...new function.
7741 (pass_cse): Replace declaration with that of...
7742 (make_pass_cse): ...new function.
7743 (pass_fast_rtl_dce): Replace declaration with that of...
7744 (make_pass_fast_rtl_dce): ...new function.
7745 (pass_ud_rtl_dce): Replace declaration with that of...
7746 (make_pass_ud_rtl_dce): ...new function.
7747 (pass_rtl_dce): Replace declaration with that of...
7748 (make_pass_rtl_dce): ...new function.
7749 (pass_rtl_dse1): Replace declaration with that of...
7750 (make_pass_rtl_dse1): ...new function.
7751 (pass_rtl_dse2): Replace declaration with that of...
7752 (make_pass_rtl_dse2): ...new function.
7753 (pass_rtl_dse3): Replace declaration with that of...
7754 (make_pass_rtl_dse3): ...new function.
7755 (pass_rtl_cprop): Replace declaration with that of...
7756 (make_pass_rtl_cprop): ...new function.
7757 (pass_rtl_pre): Replace declaration with that of...
7758 (make_pass_rtl_pre): ...new function.
7759 (pass_rtl_hoist): Replace declaration with that of...
7760 (make_pass_rtl_hoist): ...new function.
7761 (pass_rtl_store_motion): Replace declaration with that of...
7762 (make_pass_rtl_store_motion): ...new function.
7763 (pass_cse_after_global_opts): Replace declaration with that of...
7764 (make_pass_cse_after_global_opts): ...new function.
7765 (pass_rtl_ifcvt): Replace declaration with that of...
7766 (make_pass_rtl_ifcvt): ...new function.
7767 (pass_into_cfg_layout_mode): Replace declaration with that of...
7768 (make_pass_into_cfg_layout_mode): ...new function.
7769 (pass_outof_cfg_layout_mode): Replace declaration with that of...
7770 (make_pass_outof_cfg_layout_mode): ...new function.
7771 (pass_loop2): Replace declaration with that of...
7772 (make_pass_loop2): ...new function.
7773 (pass_rtl_loop_init): Replace declaration with that of...
7774 (make_pass_rtl_loop_init): ...new function.
7775 (pass_rtl_move_loop_invariants): Replace declaration with that of...
7776 (make_pass_rtl_move_loop_invariants): ...new function.
7777 (pass_rtl_unswitch): Replace declaration with that of...
7778 (make_pass_rtl_unswitch): ...new function.
7779 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
7780 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7781 (pass_rtl_doloop): Replace declaration with that of...
7782 (make_pass_rtl_doloop): ...new function.
7783 (pass_rtl_loop_done): Replace declaration with that of...
7784 (make_pass_rtl_loop_done): ...new function.
7785 (pass_web): Replace declaration with that of...
7786 (make_pass_web): ...new function.
7787 (pass_cse2): Replace declaration with that of...
7788 (make_pass_cse2): ...new function.
7789 (pass_df_initialize_opt): Replace declaration with that of...
7790 (make_pass_df_initialize_opt): ...new function.
7791 (pass_df_initialize_no_opt): Replace declaration with that of...
7792 (make_pass_df_initialize_no_opt): ...new function.
7793 (pass_reginfo_init): Replace declaration with that of...
7794 (make_pass_reginfo_init): ...new function.
7795 (pass_inc_dec): Replace declaration with that of...
7796 (make_pass_inc_dec): ...new function.
7797 (pass_stack_ptr_mod): Replace declaration with that of...
7798 (make_pass_stack_ptr_mod): ...new function.
7799 (pass_initialize_regs): Replace declaration with that of...
7800 (make_pass_initialize_regs): ...new function.
7801 (pass_combine): Replace declaration with that of...
7802 (make_pass_combine): ...new function.
7803 (pass_if_after_combine): Replace declaration with that of...
7804 (make_pass_if_after_combine): ...new function.
7805 (pass_ree): Replace declaration with that of...
7806 (make_pass_ree): ...new function.
7807 (pass_partition_blocks): Replace declaration with that of...
7808 (make_pass_partition_blocks): ...new function.
7809 (pass_match_asm_constraints): Replace declaration with that of...
7810 (make_pass_match_asm_constraints): ...new function.
7811 (pass_regmove): Replace declaration with that of...
7812 (make_pass_regmove): ...new function.
7813 (pass_split_all_insns): Replace declaration with that of...
7814 (make_pass_split_all_insns): ...new function.
7815 (pass_fast_rtl_byte_dce): Replace declaration with that of...
7816 (make_pass_fast_rtl_byte_dce): ...new function.
7817 (pass_lower_subreg2): Replace declaration with that of...
7818 (make_pass_lower_subreg2): ...new function.
7819 (pass_mode_switching): Replace declaration with that of...
7820 (make_pass_mode_switching): ...new function.
7821 (pass_sms): Replace declaration with that of...
7822 (make_pass_sms): ...new function.
7823 (pass_sched): Replace declaration with that of...
7824 (make_pass_sched): ...new function.
7825 (pass_ira): Replace declaration with that of...
7826 (make_pass_ira): ...new function.
7827 (pass_reload): Replace declaration with that of...
7828 (make_pass_reload): ...new function.
7829 (pass_clean_state): Replace declaration with that of...
7830 (make_pass_clean_state): ...new function.
7831 (pass_branch_prob): Replace declaration with that of...
7832 (make_pass_branch_prob): ...new function.
7833 (pass_value_profile_transformations): Replace declaration with that
7835 (make_pass_value_profile_transformations): ...new function.
7836 (pass_postreload_cse): Replace declaration with that of...
7837 (make_pass_postreload_cse): ...new function.
7838 (pass_gcse2): Replace declaration with that of...
7839 (make_pass_gcse2): ...new function.
7840 (pass_split_after_reload): Replace declaration with that of...
7841 (make_pass_split_after_reload): ...new function.
7842 (pass_branch_target_load_optimize1): Replace declaration with that
7844 (make_pass_branch_target_load_optimize1): ...new function.
7845 (pass_thread_prologue_and_epilogue): Replace declaration with that
7847 (make_pass_thread_prologue_and_epilogue): ...new function.
7848 (pass_stack_adjustments): Replace declaration with that of...
7849 (make_pass_stack_adjustments): ...new function.
7850 (pass_peephole2): Replace declaration with that of...
7851 (make_pass_peephole2): ...new function.
7852 (pass_if_after_reload): Replace declaration with that of...
7853 (make_pass_if_after_reload): ...new function.
7854 (pass_regrename): Replace declaration with that of...
7855 (make_pass_regrename): ...new function.
7856 (pass_cprop_hardreg): Replace declaration with that of...
7857 (make_pass_cprop_hardreg): ...new function.
7858 (pass_reorder_blocks): Replace declaration with that of...
7859 (make_pass_reorder_blocks): ...new function.
7860 (pass_branch_target_load_optimize2): Replace declaration with that
7862 (make_pass_branch_target_load_optimize2): ...new function.
7863 (pass_leaf_regs): Replace declaration with that of...
7864 (make_pass_leaf_regs): ...new function.
7865 (pass_split_before_sched2): Replace declaration with that of...
7866 (make_pass_split_before_sched2): ...new function.
7867 (pass_compare_elim_after_reload): Replace declaration with that of...
7868 (make_pass_compare_elim_after_reload): ...new function.
7869 (pass_sched2): Replace declaration with that of...
7870 (make_pass_sched2): ...new function.
7871 (pass_stack_regs): Replace declaration with that of...
7872 (make_pass_stack_regs): ...new function.
7873 (pass_stack_regs_run): Replace declaration with that of...
7874 (make_pass_stack_regs_run): ...new function.
7875 (pass_df_finish): Replace declaration with that of...
7876 (make_pass_df_finish): ...new function.
7877 (pass_compute_alignments): Replace declaration with that of...
7878 (make_pass_compute_alignments): ...new function.
7879 (pass_duplicate_computed_gotos): Replace declaration with that of...
7880 (make_pass_duplicate_computed_gotos): ...new function.
7881 (pass_variable_tracking): Replace declaration with that of...
7882 (make_pass_variable_tracking): ...new function.
7883 (pass_free_cfg): Replace declaration with that of...
7884 (make_pass_free_cfg): ...new function.
7885 (pass_machine_reorg): Replace declaration with that of...
7886 (make_pass_machine_reorg): ...new function.
7887 (pass_cleanup_barriers): Replace declaration with that of...
7888 (make_pass_cleanup_barriers): ...new function.
7889 (pass_delay_slots): Replace declaration with that of...
7890 (make_pass_delay_slots): ...new function.
7891 (pass_split_for_shorten_branches): Replace declaration with that of...
7892 (make_pass_split_for_shorten_branches): ...new function.
7893 (pass_split_before_regstack): Replace declaration with that of...
7894 (make_pass_split_before_regstack): ...new function.
7895 (pass_convert_to_eh_region_ranges): Replace declaration with that
7897 (make_pass_convert_to_eh_region_ranges): ...new function.
7898 (pass_shorten_branches): Replace declaration with that of...
7899 (make_pass_shorten_branches): ...new function.
7900 (pass_set_nothrow_function_flags): Replace declaration with that of...
7901 (make_pass_set_nothrow_function_flags): ...new function.
7902 (pass_dwarf2_frame): Replace declaration with that of...
7903 (make_pass_dwarf2_frame): ...new function.
7904 (pass_final): Replace declaration with that of...
7905 (make_pass_final): ...new function.
7906 (pass_rtl_seqabstr): Replace declaration with that of...
7907 (make_pass_rtl_seqabstr): ...new function.
7908 (pass_release_ssa_names): Replace declaration with that of...
7909 (make_pass_release_ssa_names): ...new function.
7910 (pass_early_inline): Replace declaration with that of...
7911 (make_pass_early_inline): ...new function.
7912 (pass_inline_parameters): Replace declaration with that of...
7913 (make_pass_inline_parameters): ...new function.
7914 (pass_update_address_taken): Replace declaration with that of...
7915 (make_pass_update_address_taken): ...new function.
7916 (pass_convert_switch): Replace declaration with that of...
7917 (make_pass_convert_switch): ...new function.
7918 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
7919 to a subclass of simple_ipa_opt_pass along with...
7920 (pass_data_ipa_tree_profile): ...new pass_data instance and...
7921 (make_pass_ipa_tree_profile): ...new function.
7922 * tree-sra.c (pass_sra_early): Convert from a global struct to a
7923 subclass of gimple_opt_pass along with...
7924 (pass_data_sra_early): ...new pass_data instance and...
7925 (make_pass_sra_early): ...new function.
7926 (pass_sra): Convert from a global struct to a subclass of
7927 gimple_opt_pass along with...
7928 (pass_data_sra): ...new pass_data instance and...
7929 (make_pass_sra): ...new function.
7930 (pass_early_ipa_sra): Convert from a global struct to a subclass of
7931 gimple_opt_pass along with...
7932 (pass_data_early_ipa_sra): ...new pass_data instance and...
7933 (make_pass_early_ipa_sra): ...new function.
7934 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
7935 subclass of gimple_opt_pass along with...
7936 (pass_data_ccp): ...new pass_data instance and...
7937 (make_pass_ccp): ...new function.
7938 (pass_fold_builtins): Convert from a global struct to a subclass of
7939 gimple_opt_pass along with...
7940 (pass_data_fold_builtins): ...new pass_data instance and...
7941 (make_pass_fold_builtins): ...new function.
7942 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
7943 subclass of gimple_opt_pass along with...
7944 (pass_data_copy_prop): ...new pass_data instance and...
7945 (make_pass_copy_prop): ...new function.
7946 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
7947 global struct to a subclass of gimple_opt_pass along with...
7948 (pass_data_rename_ssa_copies): ...new pass_data instance and...
7949 (make_pass_rename_ssa_copies): ...new function.
7950 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
7951 subclass of gimple_opt_pass along with...
7952 (pass_data_dce): ...new pass_data instance and...
7953 (make_pass_dce): ...new function.
7954 (pass_dce_loop): Convert from a global struct to a subclass of
7955 gimple_opt_pass along with...
7956 (pass_data_dce_loop): ...new pass_data instance and...
7957 (make_pass_dce_loop): ...new function.
7958 (pass_cd_dce): Convert from a global struct to a subclass of
7959 gimple_opt_pass along with...
7960 (pass_data_cd_dce): ...new pass_data instance and...
7961 (make_pass_cd_dce): ...new function.
7962 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7963 subclass of gimple_opt_pass along with...
7964 (pass_data_dominator): ...new pass_data instance and...
7965 (make_pass_dominator): ...new function.
7966 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7967 gimple_opt_pass along with...
7968 (pass_data_phi_only_cprop): ...new pass_data instance and...
7969 (make_pass_phi_only_cprop): ...new function.
7970 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7971 subclass of gimple_opt_pass along with...
7972 (pass_data_dse): ...new pass_data instance and...
7973 (make_pass_dse): ...new function.
7974 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7975 a subclass of gimple_opt_pass along with...
7976 (pass_data_forwprop): ...new pass_data instance and...
7977 (make_pass_forwprop): ...new function.
7978 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7979 struct to a subclass of gimple_opt_pass along with...
7980 (pass_data_tree_ifcombine): ...new pass_data instance and...
7981 (make_pass_tree_ifcombine): ...new function.
7982 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7983 subclass of gimple_opt_pass along with...
7984 (pass_data_ch): ...new pass_data instance and...
7985 (make_pass_ch): ...new function.
7986 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7987 subclass of gimple_opt_pass along with...
7988 (pass_data_tree_loop): ...new pass_data instance and...
7989 (make_pass_tree_loop): ...new function.
7990 (pass_tree_loop_init): Convert from a global struct to a subclass of
7991 gimple_opt_pass along with...
7992 (pass_data_tree_loop_init): ...new pass_data instance and...
7993 (make_pass_tree_loop_init): ...new function.
7994 (pass_lim): Convert from a global struct to a subclass of
7995 gimple_opt_pass along with...
7996 (pass_data_lim): ...new pass_data instance and...
7997 (make_pass_lim): ...new function.
7998 (pass_tree_unswitch): Convert from a global struct to a subclass of
7999 gimple_opt_pass along with...
8000 (pass_data_tree_unswitch): ...new pass_data instance and...
8001 (make_pass_tree_unswitch): ...new function.
8002 (pass_predcom): Convert from a global struct to a subclass of
8003 gimple_opt_pass along with...
8004 (pass_data_predcom): ...new pass_data instance and...
8005 (make_pass_predcom): ...new function.
8006 (pass_vectorize): Convert from a global struct to a subclass of
8007 gimple_opt_pass along with...
8008 (pass_data_vectorize): ...new pass_data instance and...
8009 (make_pass_vectorize): ...new function.
8010 (pass_graphite): Convert from a global struct to a subclass of
8011 gimple_opt_pass along with...
8012 (pass_data_graphite): ...new pass_data instance and...
8013 (make_pass_graphite): ...new function.
8014 (pass_graphite_transforms): Convert from a global struct to a subclass
8015 of gimple_opt_pass along with...
8016 (pass_data_graphite_transforms): ...new pass_data instance and...
8017 (make_pass_graphite_transforms): ...new function.
8018 (pass_check_data_deps): Convert from a global struct to a subclass of
8019 gimple_opt_pass along with...
8020 (pass_data_check_data_deps): ...new pass_data instance and...
8021 (make_pass_check_data_deps): ...new function.
8022 (pass_iv_canon): Convert from a global struct to a subclass of
8023 gimple_opt_pass along with...
8024 (pass_data_iv_canon): ...new pass_data instance and...
8025 (make_pass_iv_canon): ...new function.
8026 (pass_scev_cprop): Convert from a global struct to a subclass of
8027 gimple_opt_pass along with...
8028 (pass_data_scev_cprop): ...new pass_data instance and...
8029 (make_pass_scev_cprop): ...new function.
8030 (pass_record_bounds): Convert from a global struct to a subclass of
8031 gimple_opt_pass along with...
8032 (pass_data_record_bounds): ...new pass_data instance and...
8033 (make_pass_record_bounds): ...new function.
8034 (pass_complete_unroll): Convert from a global struct to a subclass of
8035 gimple_opt_pass along with...
8036 (pass_data_complete_unroll): ...new pass_data instance and...
8037 (make_pass_complete_unroll): ...new function.
8038 (pass_complete_unrolli): Convert from a global struct to a subclass of
8039 gimple_opt_pass along with...
8040 (pass_data_complete_unrolli): ...new pass_data instance and...
8041 (make_pass_complete_unrolli): ...new function.
8042 (pass_parallelize_loops): Convert from a global struct to a subclass
8043 of gimple_opt_pass along with...
8044 (pass_data_parallelize_loops): ...new pass_data instance and...
8045 (make_pass_parallelize_loops): ...new function.
8046 (pass_loop_prefetch): Convert from a global struct to a subclass of
8047 gimple_opt_pass along with...
8048 (pass_data_loop_prefetch): ...new pass_data instance and...
8049 (make_pass_loop_prefetch): ...new function.
8050 (pass_iv_optimize): Convert from a global struct to a subclass of
8051 gimple_opt_pass along with...
8052 (pass_data_iv_optimize): ...new pass_data instance and...
8053 (make_pass_iv_optimize): ...new function.
8054 (pass_tree_loop_done): Convert from a global struct to a subclass of
8055 gimple_opt_pass along with...
8056 (pass_data_tree_loop_done): ...new pass_data instance and...
8057 (make_pass_tree_loop_done): ...new function.
8058 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
8059 struct to a subclass of gimple_opt_pass along with...
8060 (pass_data_cse_reciprocals): ...new pass_data instance and...
8061 (make_pass_cse_reciprocals): ...new function.
8062 (pass_cse_sincos): Convert from a global struct to a subclass of
8063 gimple_opt_pass along with...
8064 (pass_data_cse_sincos): ...new pass_data instance and...
8065 (make_pass_cse_sincos): ...new function.
8066 (pass_optimize_bswap): Convert from a global struct to a subclass of
8067 gimple_opt_pass along with...
8068 (pass_data_optimize_bswap): ...new pass_data instance and...
8069 (make_pass_optimize_bswap): ...new function.
8070 (pass_optimize_widening_mul): Convert from a global struct to a
8071 subclass of gimple_opt_pass along with...
8072 (pass_data_optimize_widening_mul): ...new pass_data instance and...
8073 (make_pass_optimize_widening_mul): ...new function.
8074 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
8075 subclass of gimple_opt_pass along with...
8076 (pass_data_phiopt): ...new pass_data instance and...
8077 (make_pass_phiopt): ...new function.
8078 (pass_cselim): Convert from a global struct to a subclass of
8079 gimple_opt_pass along with...
8080 (pass_data_cselim): ...new pass_data instance and...
8081 (make_pass_cselim): ...new function.
8082 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
8083 subclass of gimple_opt_pass along with...
8084 (pass_data_phiprop): ...new pass_data instance and...
8085 (make_pass_phiprop): ...new function.
8086 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
8087 subclass of gimple_opt_pass along with...
8088 (pass_data_pre): ...new pass_data instance and...
8089 (make_pass_pre): ...new function.
8090 (pass_fre): Convert from a global struct to a subclass of
8091 gimple_opt_pass along with...
8092 (pass_data_fre): ...new pass_data instance and...
8093 (make_pass_fre): ...new function.
8094 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
8095 subclass of gimple_opt_pass along with...
8096 (pass_data_reassoc): ...new pass_data instance and...
8097 (make_pass_reassoc): ...new function.
8098 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
8099 subclass of gimple_opt_pass along with...
8100 (pass_data_sink_code): ...new pass_data instance and...
8101 (make_pass_sink_code): ...new function.
8102 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
8103 subclass of gimple_opt_pass along with...
8104 (pass_data_strlen): ...new pass_data instance and...
8105 (make_pass_strlen): ...new function.
8106 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
8107 struct to a subclass of gimple_opt_pass along with...
8108 (pass_data_build_alias): ...new pass_data instance and...
8109 (make_pass_build_alias): ...new function.
8110 (pass_build_ealias): Convert from a global struct to a subclass of
8111 gimple_opt_pass along with...
8112 (pass_data_build_ealias): ...new pass_data instance and...
8113 (make_pass_build_ealias): ...new function.
8114 (pass_ipa_pta): Convert from a global struct to a subclass of
8115 simple_ipa_opt_pass along with...
8116 (pass_data_ipa_pta): ...new pass_data instance and...
8117 (make_pass_ipa_pta): ...new function.
8118 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
8119 subclass of gimple_opt_pass along with...
8120 (pass_data_uncprop): ...new pass_data instance and...
8121 (make_pass_uncprop): ...new function.
8122 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
8123 global struct to a subclass of gimple_opt_pass along with...
8124 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
8125 (make_pass_late_warn_uninitialized): ...new function.
8126 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
8127 to a subclass of gimple_opt_pass along with...
8128 (pass_data_init_datastructures): ...new pass_data instance and...
8129 (make_pass_init_datastructures): ...new function.
8130 (pass_early_warn_uninitialized): Convert from a global struct to a
8131 subclass of gimple_opt_pass along with...
8132 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
8133 (make_pass_early_warn_uninitialized): ...new function.
8134 (pass_update_address_taken): Convert from a global struct to a
8135 subclass of gimple_opt_pass along with...
8136 (pass_data_update_address_taken): ...new pass_data instance and...
8137 (make_pass_update_address_taken): ...new function.
8138 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
8139 struct to a subclass of gimple_opt_pass along with...
8140 (pass_data_release_ssa_names): ...new pass_data instance and...
8141 (make_pass_release_ssa_names): ...new function.
8142 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
8143 subclass of gimple_opt_pass along with...
8144 (pass_data_stdarg): ...new pass_data instance and...
8145 (make_pass_stdarg): ...new function.
8146 * tree-switch-conversion.c (pass_convert_switch): Convert from a
8147 global struct to a subclass of gimple_opt_pass along with...
8148 (pass_data_convert_switch): ...new pass_data instance and...
8149 (make_pass_convert_switch): ...new function.
8150 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
8151 to a subclass of gimple_opt_pass along with...
8152 (pass_data_tail_recursion): ...new pass_data instance and...
8153 (make_pass_tail_recursion): ...new function.
8154 (pass_tail_calls): Convert from a global struct to a subclass of
8155 gimple_opt_pass along with...
8156 (pass_data_tail_calls): ...new pass_data instance and...
8157 (make_pass_tail_calls): ...new function.
8158 * tree-vect-generic.c (pass_lower_vector): Convert from a global
8159 struct to a subclass of gimple_opt_pass along with...
8160 (pass_data_lower_vector): ...new pass_data instance and...
8161 (make_pass_lower_vector): ...new function.
8162 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
8163 gimple_opt_pass along with...
8164 (pass_data_lower_vector_ssa): ...new pass_data instance and...
8165 (make_pass_lower_vector_ssa): ...new function.
8166 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
8167 to a subclass of gimple_opt_pass along with...
8168 (pass_data_slp_vectorize): ...new pass_data instance and...
8169 (make_pass_slp_vectorize): ...new function.
8170 (pass_ipa_increase_alignment): Convert from a global struct to a
8171 subclass of simple_ipa_opt_pass along with...
8172 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
8173 (make_pass_ipa_increase_alignment): ...new function.
8174 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
8175 gimple_opt_pass along with...
8176 (pass_data_vrp): ...new pass_data instance and...
8177 (make_pass_vrp): ...new function.
8178 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
8179 subclass of simple_ipa_opt_pass along with...
8180 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
8181 (make_pass_ipa_free_lang_data): ...new function.
8182 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
8183 gimple_opt_pass along with...
8184 (pass_data_tsan): ...new pass_data instance and...
8185 (make_pass_tsan): ...new function.
8186 (pass_tsan_O0): Convert from a global struct to a subclass of
8187 gimple_opt_pass along with...
8188 (pass_data_tsan_O0): ...new pass_data instance and...
8189 (make_pass_tsan_O0): ...new function.
8190 * var-tracking.c (pass_variable_tracking): Convert from a global
8191 struct to a subclass of rtl_opt_pass along with...
8192 (pass_data_variable_tracking): ...new pass_data instance and...
8193 (make_pass_variable_tracking): ...new function.
8194 * web.c (pass_web): Convert from a global struct to a subclass of
8195 rtl_opt_pass along with...
8196 (pass_data_web): ...new pass_data instance and...
8197 (make_pass_web): ...new function.
8198 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
8199 declaration with that of...
8200 (make_pass_mode_switch_use): ...new function.
8201 (pass_resolve_sw_modes): Replace declaration with that of...
8202 (make_pass_resolve_sw_modes): ...new function.
8203 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
8204 from a global struct to a subclass of rtl_opt_pass along with...
8205 (pass_data_mode_switch_use): ...new pass_data instance and...
8206 (make_pass_mode_switch_use): ...new function.
8207 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
8208 from a global struct to a subclass of rtl_opt_pass along with...
8209 (pass_data_resolve_sw_modes): ...new pass_data instance and...
8210 (make_pass_resolve_sw_modes): ...new function.
8211 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
8212 struct to a subclass of rtl_opt_pass along with...
8213 (pass_data_insert_vzeroupper): ...new pass_data instance and...
8214 (make_pass_insert_vzeroupper): ...new function.
8215 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
8216 global struct to a subclass of rtl_opt_pass along with...
8217 (pass_data_work_around_errata): ...new pass_data instance and...
8218 (make_pass_work_around_errata): ...new function.
8219 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
8220 struct to a subclass of rtl_opt_pass along with...
8221 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
8222 (make_pass_mips_machine_reorg2): ...new function.
8224 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8226 * passes.c (pass_manager::operator new): New.
8228 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8230 Handwritten part of conversion of passes to C++ classes.
8232 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
8233 (toplev.o): Add dep on PASS_MANAGER_H.
8234 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
8235 of early local pases to reflect this moving from a global to a
8236 member of gcc::pass_manager.
8237 (cgraph_add_new_function): Likewise.
8238 * lto-cgraph.c (lto_output_node): Update for conversion of
8239 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
8240 * passes.c (opt_pass::clone): New.
8241 (opt_pass::gate): New.
8242 (opt_pass::execute): New.
8243 (opt_pass::opt_pass): New.
8244 (pass_manager::execute_early_local_passes): New.
8245 (pass_manager::execute_pass_mode_switching): new.
8246 (finish_optimization_passes): Convert to...
8247 (pass_manager::finish_optimization_passes): ...this.
8248 (finish_optimization_passes): Update for conversion of passes to
8250 (register_dump_files_1): Use has_gate since we cannot portably
8251 check a vtable entry against NULL.
8252 (dump_one_pass): Likewise.
8253 (ipa_write_summaries_2): Likewise.
8254 (ipa_write_optimization_summaries_1): Likewise.
8255 (ipa_read_summaries_1): Likewise.
8256 (ipa_read_optimization_summaries_1): Likewise.
8257 (execute_ipa_stmt_fixups): Likewise.
8258 (pass_manager::pass_manager): Rewrite pass-creation, invoking
8259 pass-creation functions rather than wiring up globals, and
8260 storing the results in fields of pass_manager generated using
8262 (pass_manager::dump_profile_report): Update for conversion of
8263 passes to C++ classes.
8264 (pass_manager::execute_ipa_summary_passes): Likewise.
8265 (execute_one_ipa_transform_pass): Likewise.
8266 (execute_one_pass): Use has_gate and has_execute since we cannot
8267 portably check a vtable entry against NULL.
8268 * pass_manager.h (pass_manager::finish_optimization_passes): New.
8269 (pass_manager): Use pass-instances.def to add fields for the
8270 various pass instances.
8271 * toplev.c (finalize): Update for move of
8272 finish_optimization_passes to a method of gcc::pass_manager.
8273 * toplev.h (finish_optimization_passes): Move to method of class
8275 * tree-pass.h (struct pass_data): New.
8276 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
8277 (opt_pass::gate): Convert to virtual function.
8278 (opt_pass::~opt_pass): New.
8279 (opt_pass::clone): New.
8280 (opt_pass::execute): Convert to virtual function.
8281 (opt_pass::opt_pass): New.
8282 (opt_pass::ctxt_): new.
8283 (gimple_opt_pass): Convert to subclass of opt_pass.
8284 (gimple_opt_pass::gimple_opt_pass): New.
8285 (rtl_opt_pass): Convert to subclass of opt_pass.
8286 (rtl_opt_pass::rtl_opt_pass): New.
8287 (ipa_opt_pass_d): Convert to subclass of opt_pass.
8288 (ipa_opt_pass_d::ipa_opt_pass_d): New.
8289 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
8290 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
8291 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
8292 invocation of pass_mode_switching to reflect this moving from a
8293 global to a member of gcc::pass_manager.
8294 (ix86_option_override): Rework how pass_insert_vzeroupper is
8295 added to the pass_manager to reflect autogenerated changes.
8296 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
8298 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
8300 PR rtl-optimization/57708
8301 * recog.c (peep2_find_free_register): Validate all regs in a
8304 2013-08-05 Jan Hubicka <jh@suse.cz>
8307 * cgraph.c (verify_cgraph_node): Accept local flags from other
8309 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
8310 (function_and_variable_visibility): Likewise.
8311 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
8313 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8315 * graph.c (init_graph_slim_pretty_print): Remove.
8316 (print_graph_cfg): Do not call it. Use local pretty printer.
8317 (start_graph_dump): Likewise.
8319 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8321 * gimple-pretty-print.c (buffer): Remove.
8322 (initialized): Likewise.
8323 (maybe_init_pretty_print): Likewise.
8324 (print_gimple_stmt): Do not call it. Use non-static local
8325 pretty_printer variable.
8326 (print_gimple_expr): Likewise.
8327 (print_gimple_seq): Likewise.
8328 (gimple_dump_bb): Likewise.
8330 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8332 * asan.c (asan_pp): Remove.
8333 (asan_pp_initialized): Likewise.
8334 (asan_pp_initialize): Likewise.
8335 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
8336 (asan_emit_stack_protection): Tidy. Use local pretty printer.
8337 (asan_add_global): Likewise.
8339 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
8341 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
8342 * diagnostic.h (diagnostic_flush_buffer): Adjust.
8343 * pretty-print.c (pp_formatted_text_data): Likewise.
8344 (pp_indent): Rename from pp_base_indent.
8345 (pp_format): Rename from pp_base_format.
8346 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
8347 (pp_format_verbatim): Rename from pp_base_format_verbatim.
8348 (pp_flush): Rename from pp_base_flush.
8349 (pp_set_line_maximum_length): Rename from
8350 pp_base_set_line_maximum_length.
8351 (pp_clear_output_area): Rename from pp_base_clear_output_area.
8352 (pp_set_prefix): Rename from pp_base_set_prefix.
8353 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
8354 (pp_emit_prefix): Rename from pp_base_emit_prefix.
8355 (pp_append_text): Rename from pp_base_append_text.
8356 (pp_formatted_text): Rename from pp_base_formatted_text.
8357 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
8358 (pp_remaining_character_count_for_line): Rename from
8359 pp_base_remaining_character_count_for_line.
8360 (pp_newline): Rename from pp_base_newline.
8361 (pp_character): Rename from pp_base_character.
8362 (pp_string): Rename from pp_base_string.
8363 (pp_maybe_space): Rename from pp_base_maybe_space.
8364 * asan.c (asan_pp_string): Adjust.
8365 (asan_emit_stack_protection): Likewise.
8366 (asan_add_global): Likewise.
8367 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
8368 * tree-mudflap.c (mf_varname_tree): Likewise.
8369 * tree-pretty-print.c (pp_tree_identifier): Rename from
8370 pp_base_tree_identifier.
8371 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
8372 Declare as function.
8374 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8376 * pretty-print.h (pp_bar_bar): New.
8377 (pp_ampersand_ampersand): Likewise.
8378 (pp_less_equal): Likewise.
8379 (pp_greater_equal): Likewise.
8380 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
8381 printer functions instead of pp_string or operators and punctuators.
8382 (dump_gimple_call): Likewise.
8383 (dump_gimple_omp_for): Likewise.
8384 (dump_gimple_transaction): Likewise.
8385 (dump_gimple_phi): Likewise.
8386 (pp_gimple_stmt_1): Likewise.
8387 * sched-vis.c (print_insn): Likewise.
8388 * tree-mudflap.c (mf_varname_tree): Likewise.
8389 * tree-pretty-print.c (dump_block_node): Likewise.
8390 (dump_generic_node): Likewise.
8392 2013-08-02 Jan Hubicka <jh@suse.cz>
8394 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
8396 * lto-streamer-out.c (tree_is_indexable): Results decls and
8397 parm decls are not indexable.
8398 (DFS_write_tree_body): Do not follow args and results.
8399 (hash_tree): Likewise.
8400 (output_functions): Rearrange so struct function is needed
8401 only when real body is output; be able to also ouptut abstract
8402 functions; output DECL_ARGUMENTS and DECL_RESULT.
8403 (lto_output): When not in WPA, ale store abstract functions.
8404 (write_symbol): Do not care about RESULT_DECL.
8405 (output_symbol_p): Handle correctly sbtract decls.
8406 * lto-streamer-in.c (input_function): Rearrange so struct
8407 function can be NULL at entry; allow streaming of
8408 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
8409 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
8410 sanity check during LTO.
8411 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
8412 RESULT_DECl and DECL_ARGUMENTS.
8413 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
8416 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8418 * pretty-print.h (pp_underscore): New.
8420 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
8421 printer functions instead of pp_character.
8422 (dump_binary_rhs): Likewise.
8423 (dump_ternary_rhs): Likewise.
8424 (dump_gimple_call_args): Likewise.
8425 (pp_points_to_solution): Likewise.
8426 (dump_gimple_call): Likewise.
8427 (dump_gimple_switch): Likewise.
8428 (dump_gimple_cond): Likewise.
8429 (dump_gimple_bind): Likewise.
8430 (dump_gimple_try): Likewise.
8431 (dump_gimple_omp_for): Likewise.
8432 (dump_gimple_omp_continue): Likewise.
8433 (dump_gimple_omp_single): Likewise.
8434 (dump_gimple_omp_sections): Likewise.
8435 (dump_gimple_omp_block): Likewise.
8436 (dump_gimple_omp_critical): Likewise.
8437 (dump_gimple_transaction): Likewise.
8438 (dump_gimple_asm): Likewise.
8439 (dump_gimple_phi): Likewise.
8440 (dump_gimple_omp_parallel): Likewise.
8441 (dump_gimple_omp_task): Likewise.
8442 (dump_gimple_omp_atomic_load): Likewise.
8443 (dump_gimple_omp_atomic_store): Likewise.
8444 (dump_gimple_mem_ops): Likewise.
8445 (pp_gimple_stmt_1): Likewise.
8446 (pp_cfg_jump): Likewise.
8447 (dump_implicit_edges): Likewise.
8448 (gimple_dump_bb_for_graph): Likewise.
8449 * graph.c (draw_cfg_node): Likewise.
8450 * langhooks.c (lhd_print_error_function): Likewise.
8451 * sched-vis.c (print_exp): Likewise.
8452 (print_value): Likewise.
8453 (print_pattern): Likewise.
8454 (print_insn): Likewise.
8455 (rtl_dump_bb_for_graph): Likewise.
8456 * tree-pretty-print.c (dump_function_declaration): Likewise.
8457 (dump_array_domain): Likewise.
8458 (dump_omp_clause): Likewise.
8459 (dump_location): Likewise.
8460 (dump_generic_node): Likewise.
8461 (print_struct_decl): Likewise.
8462 * diagnostic.c (diagnostic_show_locus): Use pp_space.
8464 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8466 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
8467 candidate table when replacing a candidate statement.
8468 (replace_rhs_if_not_dup): Likewise.
8469 (replace_one_candidate): Likewise.
8471 2013-08-02 Jan Hubicka <jh@suse.cz>
8472 Martin Liska <marxin.liska@gmail.com>
8474 * cgraphunit.c (add_new_function): Fix logic when adding from
8476 (assemble_thunk): Rename to ...
8477 (expand_thunk); .. this one; export; get it working with
8478 general functions; make produced gimple valid.
8479 * cgraph.h (expand_thunk): Declare.
8481 2013-08-02 Jan Hubicka <jh@suse.cz>
8483 * ipa-cp.c (gather_context_independent_values): Use
8484 ipa_get_param_move_cost.
8485 (get_replacement_map): Remove PARAM; move parameter folding
8487 (create_specialized_node): Update.
8488 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
8489 assert that we have gimple body; update move_cost.
8490 (count_formal_params): Assert that we have gimple body.
8491 (ipa_dump_param): New function.
8492 (ipa_alloc_node_params): Break out from ...
8493 (ipa_initialize_node_params): ... here.
8494 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
8495 (ipa_write_node_info): Stream move costs.
8496 (ipa_read_node_info): Read move costs.
8497 (ipa_update_after_lto_read): Do not recompute node params.
8498 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
8499 (ipa_get_param): Check we are not in WPA.
8500 (ipa_get_param_move_cost): New.
8501 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
8502 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
8503 parm numbers to be present.
8505 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8507 PR rtl-optimization/58048
8508 * lra-constraints.c (process_alt_operands): Don't check asm
8509 operand on register.
8511 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
8513 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
8514 the implied StoreLoad barrier for atomic operations if before.
8516 2013-08-02 Jan Hubicka <jh@suse.cz>
8517 Martin Liska <marxin.liska@gmail.com>
8519 * cgraph.c (cgraph_function_body_availability): Do not check
8521 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
8522 symtab_node_availability): Declare.
8523 * ipa.c (can_replace_by_local_alias): New.
8524 (function_and_variable_visibility): Use it.
8525 * symtab.c (symtab_for_node_and_aliases,
8526 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
8528 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8530 PR rtl-optimization/57963
8531 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
8532 (lra_constraints): Use them.
8534 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
8536 * config/arm/types.md (define_attr "type"): Add "load_acq"
8538 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
8540 (cortex_a53_store1): Likewise.
8542 2013-08-01 Jan Hubicka <jh@suse.cz>
8544 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
8545 partitions are not needed.
8547 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
8549 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
8550 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
8551 MAYBE_NON_Q_CLASS_P where appropriate.
8553 2013-08-01 Jan Hubicka <jh@suse.cz>
8555 * cgraph.h (release_function_body): Declare.
8556 * tree.c (free_lang_data_in_decl): Free, parameters and return values
8557 of unused delcarations.
8559 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8561 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
8562 RTL form when subtracting a constant.
8564 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8566 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
8567 Generate canonical plus rtx with negated immediate instead of minus
8569 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
8571 2013-08-01 Jan Hubicka <jh@suse.cz>
8573 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
8574 (cgraph_release_function_body): Likewise.
8575 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
8576 * cgraph.h (cgrpah_node): Rename abstract_and_needed
8577 to used_as_abstract_origin.
8578 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
8579 symbols used as abstract origins.
8580 * cgraphunit.c (analyze_functions): Update.
8581 * ipa.c (symtab_remove_unreachable_nodes): Recompute
8582 used_as_abstract_origin.
8583 * tree-inline.c (tree_function_versioning): Update
8584 used_as_abstract_origin; be ready for DECL_RESULT and
8585 DECL_ARGUMENTS to be NULL.
8587 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
8588 for abstract functions.
8589 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
8592 2013-08-01 Jan Hubicka <jh@suse.cz>
8594 * profile.c (compute_value_histograms): Fix thinko.
8596 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8598 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
8599 aarch-common-protos.h to extra_headers.
8600 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8601 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
8602 * config/aarch64/t-aarch64 (aarch-common.o): Define.
8604 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8606 * config/aarch64/aarch64.md (define_attr "type"): Delete.
8607 Include "../arm/types.md". Define "type" attribute for all patterns.
8608 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
8611 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
8613 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
8614 to support power8 load fusion.
8615 (fusion_gpr_mem_load): Likewise.
8617 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
8619 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
8620 declarations for power8 load fusion.
8621 (emit_fusion_gpr_load): Likewise.
8623 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
8624 tuning for power8, turn on fusion mode by default. Turn on sign
8625 extending fusion mode if normal fusion mode is on, and we are at
8627 (fusion_gpr_load_p): New function, return true if we can fuse an
8628 addis instruction with a dependent load to a GPR.
8629 (emit_fusion_gpr_load): Emit the instructions for power8 load
8632 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
8633 (VSX load fusion peepholes): New peepholes to fuse together an
8634 addi instruction with a VSX load instruction.
8636 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
8637 peepholes to fuse an addis instruction with a load to a GPR base
8638 register. If we are supporting sign extending fusions, convert
8639 sign extending loads to zero extending loads and add an explicit
8642 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8644 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
8645 aarch-common-protos.h to extra_headers.
8646 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8647 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
8648 (arm_early_store_addr_dep): Likewise.
8649 (arm_no_early_alu_shift_dep): Likewise.
8650 (arm_no_early_alu_shift_value_dep): Likewise.
8651 (arm_no_early_mul_dep): Likewise.
8652 (arm_no_early_store_addr_dep): Likewise.
8653 (arm_mac_accumulator_is_mul_result): Likewise.
8654 (arm_mac_accumulator_is_result): Likewise.
8655 * config/arm/aarch-common.c: ... here. New file.
8656 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
8658 (arm_early_store_addr_dep): Likewise.
8659 (arm_no_early_alu_shift_dep): Likewise.
8660 (arm_no_early_alu_shift_value_dep): Likewise.
8661 (arm_no_early_mul_dep): Likewise.
8662 (arm_no_early_store_addr_dep): Likewise.
8663 (arm_mac_accumulator_is_mul_result): Likewise.
8664 (arm_mac_accumulator_is_result): Likewise.
8665 * config/arm/aarch-common-protos.h: ... here. New file.
8666 * config/arm/t-arm (aarch-common.o): Define.
8668 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8670 * config/arm/arm.md: Include new file "types.md".
8671 (define_attr "type"): Move from here to ...
8672 (define_attr "mul32"): Likewise.
8673 (define_attr "mul64"): Likewise.
8674 * config/arm/types.md: ... here. New file.
8676 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
8678 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
8679 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
8681 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8683 * gen-pass-instances.awk: Fix offset of substr().
8685 2013-07-31 David Malcolm <dmalcolm@redhat.com>
8687 * Makefile.in (pass-instances.def): New.
8688 (passes.o): Replace dependency on passes.def with one on
8691 * gen-pass-instances.awk: New.
8693 * passes.c (pass_manager::pass_manager): Use pass-instances.def
8694 rather than passes.def, updating local definition of NEXT_PASS
8695 macro to add an extra NUM parameter (currently unused).
8697 2013-07-30 David Malcolm <dmalcolm@redhat.com>
8699 * Makefile.in (PASS_MANAGER_H): New.
8700 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
8701 (passes.o): Likewise.
8702 (statistics.o): Likewise.
8703 (cgraphunit.o): Likewise.
8704 (context.o): Depend on PASS_MANAGER_H.
8706 * pass_manager.h: New.
8708 * cgraphunit.c (cgraph_add_new_function): Update for moves
8709 of globals to fields of pass_manager.
8710 (analyze_function): Likewise.
8711 (expand_function): Likewise.
8712 (ipa_passes): Likewise.
8713 (compile): Likewise.
8715 * context.c (context::context): New.
8716 * context.h (context::context): New.
8717 (context::get_passes): New.
8718 (context::passes_): New.
8720 * lto-cgraph.c (input_node): Update for moves of globals to
8721 fields of pass_manager.
8723 * passes.c (all_passes): Remove, in favor of a field of the
8724 same name within the new class pass_manager.
8725 (all_small_ipa_passes): Likewise.
8726 (all_lowering_passes): Likewise.
8727 (all_regular_ipa_passes): Likewise.
8728 (all_late_ipa_passes): Likewise.
8729 (all_lto_gen_passes): Likewise.
8730 (passes_by_id): Likewise.
8731 (passes_by_id_size): Likewise.
8732 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8733 the new class pass_manager.
8734 (set_pass_for_id): Convert to...
8735 (pass_manager::set_pass_for_id): ...method.
8736 (get_pass_for_id): Convert to...
8737 (pass_manager::get_pass_for_id): ...method.
8738 (register_one_dump_file): Move body of implementation into...
8739 (pass_manager::register_one_dump_file): ...here.
8740 (register_dump_files_1): Convert to...
8741 (pass_manager::register_dump_files_1): ...method.
8742 (register_dump_files): Convert to...
8743 (pass_manager::register_dump_files): ...method.
8744 (create_pass_tab): Update for moves of globals to fields of
8746 (dump_passes): Move body of implementation into...
8747 (pass_manager::dump_passes): ...here.
8748 (register_pass): Move body of implementation into...
8749 (pass_manager::register_pass): ...here.
8750 (init_optimization_passes): Convert into...
8751 (pass_manager::pass_manager): ...constructor for new
8752 pass_manager class, and initialize the pass_lists array.
8753 (check_profile_consistency): Update for moves of globals to
8754 fields of pass_manager.
8755 (dump_profile_report): Move body of implementation into...
8756 (pass_manager::dump_profile_report): ...here.
8757 (ipa_write_summaries_1): Update for moves of pass lists from
8758 being globals to fields of pass_manager.
8759 (ipa_write_optimization_summaries): Likewise.
8760 (ipa_read_summaries): Likewise.
8761 (ipa_read_optimization_summaries): Likewise.
8762 (execute_all_ipa_stmt_fixups): Likewise.
8764 * statistics.c (statistics_fini): Update for moves of globals to
8765 fields of pass_manager.
8767 * toplev.c (general_init): Replace call to
8768 init_optimization_passes with construction of the pass_manager
8771 * tree-pass.h (all_passes): Remove, in favor of a field of the
8772 same name within the new class pass_manager.
8773 (all_small_ipa_passes): Likewise.
8774 (all_lowering_passes): Likewise.
8775 (all_regular_ipa_passes): Likewise.
8776 (all_lto_gen_passes): Likewise.
8777 (all_late_ipa_passes): Likewise.
8778 (passes_by_id): Likewise.
8779 (passes_by_id_size): Likewise.
8780 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8781 the new class pass_manager.
8782 (get_pass_for_id): Remove.
8784 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8786 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
8789 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8791 * arm.md (mulhi3): New expand pattern.
8793 2013-07-30 Jan Hubicka <jh@suse.cz>
8794 Martin Liska <marxin.liska@gmail.com>
8796 * profile.c (compute_value_histograms): Do not ICE when
8797 there is mismatch only on some counters.
8799 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8801 PR rtl-optimization/57637
8802 * function.c (move_insn_for_shrink_wrap): Also check the
8803 GEN set of the LIVE problem for the liveness analysis
8804 if it exists, otherwise give up.
8806 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8808 PR tree-optimization/57993
8809 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
8810 replaced statement in the candidate table.
8811 (phi_add_costs): Return infinite cost when the hidden basis does
8812 not dominate all phis on which the candidate is dependent.
8813 (replace_one_candidate): Record replaced statement in the
8816 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
8818 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
8819 (ashlv2si3): New expander.
8820 (*ashlv2si3_i): New define_insn_and_split.
8821 * predicates.md (float_operation): Allow patterns with three
8824 PR rtl-optimization/58021
8825 * mode-switching.c (create_pre_exit): Always split off preceding
8826 insns if we are not at the basic block head.
8828 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
8830 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
8831 (UCLIBC_DYNAMIC_LINKER): New macro.
8832 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
8834 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
8835 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
8837 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
8838 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
8839 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
8840 for SF and DF modes. Use ieee_quad_format for TF mode.
8841 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
8842 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
8843 (mips_option_override): Handle `-mnan=legacy'.
8844 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
8845 `-mabs=2008' and `-mnan=2008'.
8846 (OPTION_DEFAULT_SPECS): Add "nan" default.
8847 (ASM_SPEC): Handle `-mnan='.
8848 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
8849 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
8850 comment accordingly.
8851 (neg<mode>2): Likewise.
8852 * config/mips/mips.opt (mabs, mnan): New options.
8853 * doc/install.texi (Configuration): Document `--with-nan=' option.
8854 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
8856 (MIPS Options): Document them.
8857 * config.gcc <mips*-*-*>: Handle `--with-nan='.
8858 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
8859 * configure: Regenerate.
8860 * config.in: Regenerate.
8862 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8864 * config/i386/i386.md (float post-reload splitters): Do not check
8865 for subregs of SSE registers.
8867 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8868 H.J. Lu <hongjiu.lu@intel.com>
8872 * config/i386/i386.md (post-reload splitter
8873 to avoid partial SSE reg dependency stalls): New pattern.
8875 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
8877 * config/s390/s390.md ("movcc"): Swap load and store instructions.
8879 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
8881 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
8882 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
8884 2013-07-26 Cary Coutant <ccoutant@google.com>
8886 * dwarf2out.c (die_checksum_ordered): Don't include template
8887 instantiations in signature.
8888 (is_template_parameter): New function.
8889 (is_template_instantiation): New function.
8890 (generate_skeleton_bottom_up): Don't include template instantiations
8892 (generate_skeleton): Likewise.
8893 (break_out_comdat_types): Move recursive call to break out nested
8895 (prune_unused_types_mark_generic_parms_dies): Call
8896 is_template_parameter.
8898 2013-07-26 Ian Bolton <ian.bolton@arm.com>
8900 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
8901 uses vector registers.
8902 * config/aarch64/iterators.md: Add attributes rtn and vas.
8904 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8905 Richard Earnshaw <richard.earnshaw@arm.com>
8907 * combine.c (simplify_comparison): Re-canonicalize operands
8909 * config/arm/arm.md (movcond_addsi): New splitter.
8911 2013-07-25 Sterling Augustine <saugustine@google.com>
8913 * dwarf2out.c (size_of_pubnames): Move code to...
8914 (include_pubname_in_output): ...here. New.
8915 (want_pubnames): Rearrange.
8916 (output_pubnames): Call include_pubname_in_output. Move assertion.
8918 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8920 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
8922 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8925 * doc/extend.texi: Remove obsolete builtins. Fix
8926 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
8928 2013-07-25 Jan Hubicka <jh@suse.cz>
8930 * cgraph.c (release_function_body): Break out from ...
8931 (cgraph_release_function_body): ... this one; also release DECL_RESULT
8933 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
8934 old_tree in the map.
8935 (create_specialized_node): Update.
8936 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
8938 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
8939 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
8940 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
8941 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
8944 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8946 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
8947 addsi3_carryin_alt2_<optab>): Correct output template.
8949 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8951 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
8952 Adjust for arm_restrict_it.
8953 Remove trailing whitespace.
8955 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
8957 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8958 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8960 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8962 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8964 PR rtl-optimization/57960
8965 * lra-constraints.c (process_alt_operands): Use the right mode
8966 when checking strict_low.
8968 2013-07-25 Jan Hubicka <jh@suse.cz>
8970 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8971 * cgraph.c (cgraph_remove_node): Do not release function body
8972 when in cgraph streaming.
8973 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8974 in other partitions are not considered reachable; fix handling of
8977 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8979 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8981 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8986 * config/arm/arm.md ("*sibcall_insn): Replace use of
8987 Ss with US. Adjust output for v5 and v4t.
8988 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8990 * config/arm/constraints.md ("Ss"): Rename to US.
8992 2013-07-25 Terry Guo <terry.guo@arm.com>
8994 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8995 shift_add/shift_sub0/shift_sub1 RTXs.
8997 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8998 Anton Blanchard <anton@au1.ibm.com>
9000 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
9001 (altivec_vpks<VI_char>ss): Likewise.
9002 (altivec_vpks<VI_char>us): Likewise.
9003 (altivec_vpku<VI_char>us): Likewise.
9004 (altivec_vpku<VI_char>um): Likewise.
9006 2013-07-24 David Malcolm <dmalcolm@redhat.com>
9008 Introduce context class.
9010 * Makefile.in (CONTEXT_H): New.
9011 (OBJS): Add context.o.
9012 (toplev.o): Add CONTEXT_H to dependencies.
9015 * toplev.c (general_init): Create the singleton gcc::context instance.
9021 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
9023 PR rtl-optimization/57968
9024 * mode-switching.c (create_pre_exit): Allow instructions that
9025 don't set a return register to need a non-exit mode.
9027 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9028 Anton Blanchard <anton@au1.ibm.com>
9030 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
9031 operands to vperm for little endian.
9032 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
9033 of lvsl to create the control mask for a vperm for little endian.
9035 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9036 Anton Blanchard <anton@au1.ibm.com>
9038 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
9039 two operands for little-endian.
9041 2013-07-23 Steve Ellcey <sellcey@mips.com>
9043 * config/mips/mips.c (mips_case_values_threshold): New.
9044 (TARGET_CASE_VALUES_THRESHOLD): Define.
9046 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9047 Anton Blanchard <anton@au1.ibm.com>
9049 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
9050 selection of field for vector splat in little endian mode.
9052 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9054 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
9055 expanders to rs6000.md.
9056 (ior<mode>3): Likewise.
9057 (and<mode>3): Likewise.
9058 (one_cmpl<mode>2): Likewise.
9059 (nor<mode>3): Likewise.
9060 (andc<mode>3): Likewise.
9061 (eqv<mode>3): Likewise.
9062 (nand<mode>3): Likewise.
9063 (orc<mode>3): Likewise.
9065 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
9068 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
9069 to split multi-word logical operations.
9070 (rs6000_split_logical_di): Likewise.
9071 (rs6000_split_logical): Likewise.
9073 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
9074 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
9075 and allow TImode operations in 32-bit.
9076 (vsx_and<mode>3_64bit): Likewise.
9077 (vsx_ior<mode>3_32bit): Likewise.
9078 (vsx_ior<mode>3_64bit): Likewise.
9079 (vsx_xor<mode>3_32bit): Likewise.
9080 (vsx_xor<mode>3_64bit): Likewise.
9081 (vsx_one_cmpl<mode>2_32bit): Likewise.
9082 (vsx_one_cmpl<mode>2_64bit): Likewise.
9083 (vsx_nor<mode>3_32bit): Likewise.
9084 (vsx_nor<mode>3_64bit): Likewise.
9085 (vsx_andc<mode>3_32bit): Likewise.
9086 (vsx_andc<mode>3_64bit): Likewise.
9087 (vsx_eqv<mode>3_32bit): Likewise.
9088 (vsx_eqv<mode>3_64bit): Likewise.
9089 (vsx_nand<mode>3_32bit): Likewise.
9090 (vsx_nand<mode>3_64bit): Likewise.
9091 (vsx_orc<mode>3_32bit): Likewise.
9092 (vsx_orc<mode>3_64bit): Likewise.
9094 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
9095 logical types in GPRs.
9097 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
9098 logical insns to rs6000.md, and allow TImode operations in
9100 (altivec_ior<mode>3): Likewise.
9101 (altivec_xor<mode>3): Likewise.
9102 (altivec_one_cmpl<mode>2): Likewise.
9103 (altivec_nor<mode>3): Likewise.
9104 (altivec_andc<mode>3): Likewise.
9106 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
9107 attributes for moving the 128-bit logical operations into
9109 (BOOL_REGS_OUTPUT): Likewise.
9110 (BOOL_REGS_OP1): Likewise.
9111 (BOOL_REGS_OP2): Likewise.
9112 (BOOL_REGS_UNARY): Likewise.
9113 (BOOL_REGS_AND_CR0): Likewise.
9114 (one_cmpl<mode>2): Add support for DI logical operations on
9115 32-bit, splitting the operations to 32-bit.
9119 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
9120 changes to combine the 32/64-bit code, allow logical operations on
9121 TI mode in 32-bit, and to use similar match_operator patterns like
9122 scalar mode uses. Combine the Altivec and VSX code for logical
9123 operations, and move it here.
9124 (ior<mode>3, 128-bit types): Likewise.
9125 (xor<mode>3, 128-bit types): Likewise.
9126 (one_cmpl<mode>3, 128-bit types): Likewise.
9127 (nor<mode>3, 128-bit types): Likewise.
9128 (andc<mode>3, 128-bit types): Likewise.
9129 (eqv<mode>3, 128-bit types): Likewise.
9130 (nand<mode>3, 128-bit types): Likewise.
9131 (orc<mode>3, 128-bit types): Likewise.
9132 (and<mode>3_internal): Likewise.
9133 (bool<mode>3_internal): Likewise.
9134 (boolc<mode>3_internal1): Likewise.
9135 (boolc<mode>3_internal2): Likewise.
9136 (boolcc<mode>3_internal1): Likewise.
9137 (boolcc<mode>3_internal2): Likewise.
9138 (eqv<mode>3_internal1): Likewise.
9139 (eqv<mode>3_internal2): Likewise.
9140 (one_cmpl1<mode>3_internal): Likewise.
9142 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9144 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9145 Rename flag_stack_usage to flag_stack_usage_info.
9147 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9149 * config/microblaze/sync.md: New file.
9150 * config/microblaze/microblaze.md: Include sync.md
9151 * config/microblaze/microblaze.c: Add print_operand 'y'.
9152 * config/microblaze/constraints.md: Add memory_contraint
9153 'Q' which is a single register.
9155 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
9157 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
9159 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9161 * reload.c (find_reloads): Exit loop once we find this operand
9162 cannot be reloaded somehow for this alternative.
9164 * reload.c (find_reloads): Exit loop once we find a hard register.
9166 * rtlanal.c (computed_jump_p): Exit loop once we find label
9169 * i386.c (ix86_pad_returns): Exit loop after setting replace.
9171 * cfgloopmanip.c (remove_path): Exit loop after setting
9174 * gensupport.c (subst_dup): Avoid loop if code is not
9175 MATCH_DUP nor MATCH_OP_DUP.
9177 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
9179 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
9181 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9183 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
9184 true for SP_REGNUM if mode == ptr_mode.
9185 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
9186 with value R0_REGNUM + 31.
9188 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9190 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
9191 pad pointer-typed argument downward.
9193 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9195 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
9196 and __ILP32__ when the ILP32 model is in use.
9198 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9200 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
9201 (aarch64_load_symref_appropriately): In the case of
9202 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
9203 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
9204 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
9205 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
9206 if 'mode' doesn't equal to 'ptr_mode'.
9207 (aarch64_output_mi_thunk): Add an assertion on the alignment of
9208 'vcall_offset'; change to call aarch64_emit_move differently depending
9209 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
9210 to calculate the upper bound of 'vcall_offset'.
9211 (aarch64_cannot_force_const_mem): Change to also return true if
9213 (aarch64_legitimize_reload_address): In the case of large
9214 displacements, add new local variable 'xmode' and an assertion
9215 based on it; change to use 'xmode' to generate the new rtx and
9217 (aarch64_asm_trampoline_template): Change to generate the template
9218 differently depending on TARGET_ILP32 or not; change to use
9219 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
9220 (aarch64_trampoline_size): Removed.
9221 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
9222 and replace immediate literals with it. Change to use 'ptr_mode'
9223 instead of 'DImode' and call convert_memory_address if the mode
9224 of 'fnaddr' doesn't equal to 'ptr_mode'.
9225 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
9227 (aarch64_elf_asm_destructor): Likewise.
9228 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
9229 on TARGET_ILP32 instead of aarch64_trampoline_size.
9230 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
9231 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
9232 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
9233 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
9234 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
9235 (storewb_pair<GPI:mode>_<P:mode>): ... this.
9236 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
9237 depending on the value of 'mode'.
9238 (add_losym_<mode>): New.
9239 (ldr_got_small_<mode>): New, based on ldr_got_small.
9240 (ldr_got_small): Remove.
9241 (ldr_got_small_sidi): New.
9242 * config/aarch64/iterators.md (P): New.
9243 (PTR): Change to 'ptr_mode' in the condition.
9245 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9247 * config.gcc (aarch64*-*-*): Support --with-abi.
9248 (aarch64*-*-elf): Support --with-multilib-list.
9249 (aarch64*-*-linux*): Likewise.
9250 (supported_defaults): Add abi to aarch64*-*-*.
9251 * configure.ac: Mention AArch64 for --with-multilib-list.
9252 * configure: Re-generated.
9253 * config/aarch64/biarchilp32.h: New file.
9254 * config/aarch64/biarchlp64.h: New file.
9255 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
9257 (MULTILIB_DEFAULTS): Ditto.
9258 (DRIVER_SELF_SPECS): Ditto.
9259 (ASM_SPEC): Update to also substitute -mabi.
9260 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
9261 file whose name depends on -mabi= and -mbig-endian.
9262 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
9264 (POINTER_SIZE): New define.
9265 (POINTERS_EXTEND_UNSIGNED): Ditto.
9266 (enum aarch64_abi_type): New enumeration tag.
9267 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
9268 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
9269 (TARGET_ILP32): New define.
9270 * config/aarch64/aarch64.opt (mabi): New.
9272 (ilp32, lp64): New values for -mabi.
9273 * config/aarch64/t-aarch64 (comma): New define.
9274 (MULTILIB_OPTIONS): Ditto.
9275 (MULTILIB_DIRNAMES): Ditto.
9276 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
9277 * doc/invoke.texi: Document -mabi for AArch64.
9279 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
9281 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
9283 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9284 Anton Blanchard <anton@au1.ibm.com>
9286 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
9287 endianness when selecting field to splat.
9289 2013-07-22 Eric Christopher <echristo@gmail.com>
9291 * dwarf2out.c (die_odr_checksum): New function to use
9292 CHECKSUM_ macros and ULEB128 for DIE tag.
9293 (generate_type_signature): Use.
9295 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
9297 * config.gcc (sparc*-*-*): Accept leon3 processor.
9298 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
9299 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
9300 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
9301 * config/sparc/sparc.opt (enum processor_type): Add leon3.
9302 (mfix-ut699): Adjust comment.
9303 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
9304 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
9305 (CPP_CPU_SPEC): Likewise.
9306 (ASM_CPU_SPEC): Likewise.
9307 * config/sparc/sparc.c (leon3_cost): New constant.
9308 (sparc_option_override): Add leon3 support.
9309 (mem_ref): New function.
9310 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
9311 (sparc_do_work_around_errata): Look into the instruction in the delay
9312 slot and adjust accordingly. Add fix for the data cache nullify issues
9313 of the UT699. Change insertion position for the NOP.
9314 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
9315 (leon3_load): New reservation.
9316 (leon_store): Bump latency to 2.
9317 (grfpu): New automaton.
9318 (grfpu_alu): New unit.
9319 (grfpu_ds): Likewise.
9320 (leon_fp_alu): Adjust.
9321 (leon_fp_mult): Delete.
9322 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
9323 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
9324 * config/sparc/sparc.md (cpu): Add leon3.
9325 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
9327 (atomic_test_and_set): Likewise.
9330 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
9332 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
9333 default for R5900 targets.
9334 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
9335 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
9336 * config/mips/mips.c (mips_option_override): Report an error for
9337 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
9338 for -march=r5900 -mhard-float.
9340 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9342 * df-problems.c (can_move_insns_across): Exit loop once we
9343 find a non-fixed, non-global register.
9345 * ipa-pure-const.c (propagate_nothrow): Exit loop after
9348 * omega.c (omega_eliminate_red): Break after setting red_found.
9349 (omega_problem_has_red_equations): Similarly after setting found.
9350 (omega_query_variable): Similarly after setting coupled.
9352 2013-07-22 Marek Polacek <polacek@redhat.com>
9354 * gimplify.c: Don't include gimple.h twice.
9356 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9358 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
9359 instead of TARGET_THUMB1.
9360 (Pz): New constraint.
9361 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
9363 (compare_negsi_si): Likewise.
9364 (compare_addsi2_op0): Likewise.
9365 (compare_addsi2_op1): Likewise.
9366 (addsi3_carryin_<optab>): Likewise.
9367 (addsi3_carryin_alt2_<optab>): Likewise.
9368 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
9369 for arm_restrict_it.
9370 (subsi3_carryin): Likewise.
9371 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
9372 (minmax_arithsi): Disable for arm_restrict_it.
9373 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
9374 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
9375 (satsi_<SAT:code>_shift): Likewise.
9376 (arm_shiftsi3): Add alternative for 16-bit encoding.
9377 (arm32_movhf): Disable for arm_restrict_it.
9378 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
9379 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
9381 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
9383 * config/arm/arm.md (attribute "insn"): Delete.
9384 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
9385 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
9386 (not_shiftsi): Update for attribute change.
9387 (not_shiftsi_compare0): Likewise.
9388 (not_shiftsi_compare0_scratch): Likewise.
9389 (arm_one_cmplsi2): Likewise.
9390 (thumb1_one_cmplsi2): Likewise.
9391 (notsi_compare0): Likewise.
9392 (notsi_compare0_scratch): Likewise.
9393 (thumb1_movdi_insn): Likewise.
9394 (arm_movsi_insn): Likewise.
9395 (movhi_insn_arch4): Likewise.
9396 (movhi_bytes): Likewise.
9397 (arm_movqi_insn): Likewise.
9398 (thumb1_movqi_insn): Likewise.
9399 (arm32_movhf): Likewise.
9400 (thumb1_movhf): Likewise.
9401 (arm_movsf_soft_insn): Likewise.
9402 (thumb1_movsf_insn): Likewise.
9403 (thumb_movdf_insn): Likewise.
9404 (movsicc_insn): Likewise.
9405 (movsfcc_soft_insn): Likewise.
9406 (and_scc): Likewise.
9407 (cond_move): Likewise.
9408 (if_move_not): Likewise.
9409 (if_not_move): Likewise.
9410 (if_shift_move): Likewise.
9411 (if_move_shift): Likewise.
9412 (if_shift_shift): Likewise.
9413 (if_not_arith): Likewise.
9414 (if_arith_not): Likewise.
9415 (cond_move_not): Likewise.
9416 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
9417 (neon_mov<mode>): Likewise.
9418 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
9419 (thumb2_movsi_vfp): Likewise.
9420 (movsf_vfp): Likewise.
9421 (thumb2_movsf_vfp): Likewise.
9422 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9424 (cortexa7_older_only): Likewise.
9425 (cortexa7_younger): Likewise.
9426 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
9427 (1020alu_shift_op): Likewise.
9428 (1020alu_shift_reg_op): Likewise.
9429 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
9430 (alu_shift_op): Likewise.
9431 (alu_shift_reg_op): Likewise.
9432 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
9433 (11_alu_shift_op): Likewise.
9434 (11_alu_shift_reg_op): Likewise.
9435 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
9436 (9_alu_shift_reg_op): Likewise.
9437 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9439 (cortex_a15_alu_shift): Likewise.
9440 (cortex_a15_alu_shift_reg): Likewise.
9441 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
9442 (cortex_a5_alu_shift): Likewise.
9443 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
9445 (cortex_a53_alu_shift): Likewise.
9446 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9448 (cortex_a7_alu_reg): Likewise.
9449 (cortex_a7_alu_shift): Likewise.
9450 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9451 (cortex_a8_alu_shift): Likewise.
9452 (cortex_a8_alu_shift_reg): Likewise.
9453 (cortex_a8_mov): Likewise.
9454 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9455 (cortex_a9_dp_shift): Likewise.
9456 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
9457 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
9458 (cortex_r4_mov): Likewise.
9459 (cortex_r4_alu_shift): Likewise.
9460 (cortex_r4_alu_shift_reg): Likewise.
9461 * config/arm/fa526.md (526_alu_op): Update for attribute change.
9462 (526_alu_shift_op): Likewise.
9463 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
9464 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
9465 (626te_alu_shift_op): Likewise.
9466 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
9467 (726te_alu_op): Likewise.
9468 (726te_alu_shift_op): Likewise.
9469 (726te_alu_shift_reg_op): Likewise.
9470 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
9471 (mp626_alu_shift_op): Likewise.
9472 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
9473 (pj4_alu_e1_conds): Likewise.
9474 (pj4_alu): Likewise.
9475 (pj4_alu_conds): Likewise.
9476 (pj4_shift): Likewise.
9477 (pj4_shift_conds): Likewise.
9478 (pj4_alu_shift): Likewise.
9479 (pj4_alu_shift_conds): Likewise.
9481 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9483 * config/arm/predicates.md (shiftable_operator_strict_it):
9485 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
9486 Disable cond_exec version for arm_restrict_it.
9487 (thumb2_smaxsi3): Convert to generate cond_exec.
9488 (thumb2_sminsi3): Likewise.
9489 (thumb32_umaxsi3): Likewise.
9490 (thumb2_uminsi3): Likewise.
9491 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
9492 (thumb2_neg_abssi2): Likewise.
9493 (thumb2_mov_scc): Add alternative for 16-bit encoding.
9494 (thumb2_movsicc_insn): Adjust alternatives.
9495 (thumb2_mov_negscc): Disable for arm_restrict_it.
9496 (thumb2_mov_negscc_strict_it): New pattern.
9497 (thumb2_mov_notscc_strict_it): New pattern.
9498 (thumb2_mov_notscc): Disable for arm_restrict_it.
9499 (thumb2_ior_scc): Likewise.
9500 (thumb2_ior_scc_strict_it): New pattern.
9501 (thumb2_cond_move): Adjust for arm_restrict_it.
9502 (thumb2_cond_arith): Disable for arm_restrict_it.
9503 (thumb2_cond_arith_strict_it): New pattern.
9504 (thumb2_cond_sub): Adjust for arm_restrict_it.
9505 (thumb2_movcond): Likewise.
9506 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
9507 (thumb2_zero_extendhisi2_v6): Likewise.
9508 (thumb2_zero_extendqisi2_v6): Likewise.
9509 (orsi_notsi_si): Likewise.
9510 (orsi_not_shiftsi_si): Likewise.
9512 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
9514 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
9515 instruction sequence is 1 byte shorter.
9517 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
9519 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
9520 it is not needed after split.
9522 2013-07-20 Iain Sandoe <iain@codesourcery.com>
9525 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
9526 second label for nonlocal goto receivers. Don't output pic base labels
9527 unless we're producing PIC; mark that action unreachable().
9528 (ix86_save_reg): If the function contains a nonlocal label, save the
9530 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
9531 * config/darwin.c (emitted_pic_label_num): New GTY.
9532 (update_pic_label_number_if_needed): New.
9533 (machopic_output_function_base_name): Adjust for nonlocal receiver
9535 (machopic_should_output_picbase_label): New.
9536 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
9537 (nonlocal_goto_receiver): New insn and split.
9539 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
9541 * config/aarch64/aarch64-builtins.c
9542 (aarch64_fold_builtin): Fold abs in all modes.
9543 * config/aarch64/aarch64-simd-builtins.def
9544 (abs): Enable for all modes.
9545 * config/aarch64/arm_neon.h
9546 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
9547 (vabs_f64): Add missing intrinsic.
9549 2013-07-19 Ian Bolton <ian.bolton@arm.com>
9551 * config/aarch64/arm_neon.h (vabs_s64): New function
9553 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
9556 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
9557 * config/avr/avr.md (adjust_len): Add `round'.
9558 * config/avr/avr-protos.h (avr_out_round): New prototype.
9559 (avr_out_plus): Add `out_label' argument.
9560 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
9561 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
9562 Handle the case where `insn' is just a pattern.
9563 (avr_out_bitop): Handle the case where `insn' is just a pattern.
9564 (avr_out_round): New function.
9565 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
9567 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
9569 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9570 Add check for flag_stack_usage to handle -fstack-usage support
9572 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
9574 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
9575 interaction for new Power8 flags and VSX.
9577 2013-07-18 Sriraman Tallam <tmsriram@google.com>
9580 * tree-inline.c (expand_call_inline): Emit errors during
9581 early_inlining only if optimization is not turned on.
9583 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9587 * passes.c (init_optimization_passes): Move the construction of
9588 the pass hierarchy into a new passes.def file.
9590 * Makefile.in (passes.o): Add dependency on passes.def.
9592 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9594 * passes.c (init_optimization_passes): Introduce macros for
9595 constructing the tree of passes (INSERT_PASSES_AFTER,
9596 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
9597 TERMINATE_PASS_LIST).
9599 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
9600 Wei Mi <wmi@google.com>
9602 PR rtl-optimization/57878
9603 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
9605 (reload_pseudo_compare_func): Check nregs first for reload
9608 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9610 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
9612 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
9614 * read-rtl.c (validate_const_int): Once an invalid character is
9615 seen, quit the loop.
9617 * gengtype.c (write_roots): Similarly once we find the "deletable"
9618 or "if_marked" option.
9620 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9622 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
9623 "xtab" and "sat". Move value "clz" from here to ...
9624 (attriubte "type"): ... here.
9625 (satsi_<SAT:code>): Delete "insn" attribute.
9626 (satsi_<SAT:code>_shift): Likewise.
9627 (arm_zero_extendqisi2addsi): Likewise.
9628 (arm_extendqisi2addsi): Likewise.
9629 (clzsi2): Update for attribute changes.
9630 (rbitsi2): Likewise.
9631 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
9633 (arm_usatsihi): Likewise.
9634 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9636 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9638 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
9639 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
9640 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
9641 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
9642 in alphabetical order.
9643 (attribute "core_cycles"): Update for attribute changes.
9644 (arm_addsi3): Likewise.
9645 (addsi3_compare0): Likewise.
9646 (addsi3_compare0_scratch): Likewise.
9647 (addsi3_compare_op1): Likewise.
9648 (addsi3_compare_op2): Likewise.
9649 (compare_addsi2_op0): Likewise.
9650 (compare_addsi2_op1): Likewise.
9651 (addsi3_carryin_shift_<optab>): Likewise.
9652 (subsi3_carryin_shift): Likewise.
9653 (rsbsi3_carryin_shift): Likewise.
9654 (arm_subsi3_insn): Likewise.
9655 (subsi3_compare0): Likewise.
9656 (subsi3_compare): Likewise.
9657 (arm_andsi3_insn): Likewise.
9658 (thumb1_andsi3_insn): Likewise.
9659 (andsi3_compare0): Likewise.
9660 (andsi3_compare0_scratch): Likewise.
9661 (zeroextractsi_compare0_scratch
9662 (andsi_not_shiftsi_si): Likewise.
9663 (iorsi3_insn): Likewise.
9664 (iorsi3_compare0): Likewise.
9665 (iorsi3_compare0_scratch): Likewise.
9666 (arm_xorsi3): Likewise.
9667 (thumb1_xorsi3_insn): Likewise.
9668 (xorsi3_compare0): Likewise.
9669 (xorsi3_compare0_scratch): Likewise.
9670 (satsi_<SAT:code>_shift): Likewise.
9672 (arm_shiftsi3): Likewise.
9673 (shiftsi3_compare0): Likewise.
9674 (not_shiftsi): Likewise.
9675 (not_shiftsi_compare0): Likewise.
9676 (not_shiftsi_compare0_scratch): Likewise.
9677 (arm_one_cmplsi2): Likewise.
9678 (thumb_one_complsi2): Likewise.
9679 (notsi_compare0): Likewise.
9680 (notsi_compare0_scratch): Likewise.
9681 (thumb1_zero_extendhisi2): Likewise.
9682 (arm_zero_extendhisi2): Likewise.
9683 (arm_zero_extendhisi2_v6): Likewise.
9684 (arm_zero_extendhisi2addsi): Likewise.
9685 (thumb1_zero_extendqisi2): Likewise.
9686 (thumb1_zero_extendqisi2_v6): Likewise.
9687 (arm_zero_extendqisi2): Likewise.
9688 (arm_zero_extendqisi2_v6): Likewise.
9689 (arm_zero_extendqisi2addsi): Likewise.
9690 (thumb1_extendhisi2): Likewise.
9691 (arm_extendhisi2): Likewise.
9692 (arm_extendhisi2_v6): Likewise.
9693 (arm_extendqisi): Likewise.
9694 (arm_extendqisi_v6): Likewise.
9695 (arm_extendqisi2addsi): Likewise.
9696 (thumb1_extendqisi2): Likewise.
9697 (thumb1_movdi_insn): Likewise.
9698 (arm_movsi_insn): Likewise.
9699 (movsi_compare0): Likewise.
9700 (movhi_insn_arch4): Likewise.
9701 (movhi_bytes): Likewise.
9702 (arm_movqi_insn): Likewise.
9703 (thumb1_movqi_insn): Likewise.
9704 (arm32_movhf): Likewise.
9705 (thumb1_movhf): Likewise.
9706 (arm_movsf_soft_insn): Likewise.
9707 (thumb1_movsf_insn): Likewise.
9708 (movdf_soft_insn): Likewise.
9709 (thumb_movdf_insn): Likewise.
9710 (arm_cmpsi_insn): Likewise.
9711 (cmpsi_shiftsi): Likewise.
9712 (cmpsi_shiftsi_swp): Likewise.
9713 (arm_cmpsi_negshiftsi_si): Likewise.
9714 (movsicc_insn): Likewise.
9715 (movsfcc_soft_insn): Likewise.
9716 (arith_shiftsi): Likewise.
9717 (arith_shiftsi_compare0
9718 (arith_shiftsi_compare0_scratch
9719 (sub_shiftsi): Likewise.
9720 (sub_shiftsi_compare0
9721 (sub_shiftsi_compare0_scratch
9722 (and_scc): Likewise.
9723 (cond_move): Likewise.
9724 (if_plus_move): Likewise.
9725 (if_move_plus): Likewise.
9726 (if_move_not): Likewise.
9727 (if_not_move): Likewise.
9728 (if_shift_move): Likewise.
9729 (if_move_shift): Likewise.
9730 (if_shift_shift): Likewise.
9731 (if_not_arith): Likewise.
9732 (if_arith_not): Likewise.
9733 (cond_move_not): Likewise.
9734 (thumb1_ashlsi3): Set type attribute.
9735 (thumb1_ashrsi3): Likewise.
9736 (thumb1_lshrsi3): Likewise.
9737 (thumb1_rotrsi3): Likewise.
9738 (shiftsi3_compare0_scratch): Likewise.
9739 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
9740 (neon_mov<mode>): Likewise.
9741 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
9743 (thumb2_movsi_insn): Likewise.
9744 (thumb2_cmpsi_neg_shiftsi): Likewise.
9745 (thumb2_extendqisi_v6): Likewise.
9746 (thumb2_zero_extendhisi2_v6): Likewise.
9747 (thumb2_zero_extendqisi2_v6): Likewise.
9748 (thumb2_shiftsi3_short): Likewise.
9749 (thumb2_addsi3_compare0_scratch): Likewise.
9750 (orsi_not_shiftsi_si): Likewise.
9751 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9752 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
9754 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9755 (1020alu_shift_op): Likewise.
9756 (1020alu_shift_reg_op): Likewise.
9757 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9758 (alu_shift_op): Likewise.
9759 (alu_shift_reg_op): Likewise.
9760 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9761 (11_alu_shift_op): Likewise.
9762 (11_alu_shift_reg_op): Likewise.
9763 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9764 (9_alu_shift_reg_op): Likewise.
9765 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9767 (cortex_a15_alu_shift): Likewise.
9768 (cortex_a15_alu_shift_reg): Likewise.
9769 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
9771 (cortex_a5_alu_shift): Likewise.
9772 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
9774 (cortex_a53_alu_shift): Likewise.
9775 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9777 (cortex_a7_alu_reg): Likewise.
9778 (cortex_a7_alu_shift): Likewise.
9779 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
9781 (cortex_a8_alu_shift): Likewise.
9782 (cortex_a8_alu_shift_reg): Likewise.
9783 (cortex_a8_mov): Likewise.
9784 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9785 (cortex_a9_dp_shift): Likewise.
9786 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
9788 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
9790 (cortex_r4_mov): Likewise.
9791 (cortex_r4_alu_shift): Likewise.
9792 (cortex_r4_alu_shift_reg): Likewise.
9793 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9794 (526_alu_shift_op): Likewise.
9795 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9796 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9797 (626te_alu_shift_op): Likewise.
9798 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
9799 (726te_alu_op): Likewise.
9800 (726te_alu_shift_op): Likewise.
9801 (726te_alu_shift_reg_op): Likewise.
9802 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9803 (mp626_alu_shift_op): Likewise.
9804 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
9805 (pj4_alu_e1_conds): Likewise.
9806 (pj4_alu): Likewise.
9807 (pj4_alu_conds): Likewise.
9808 (pj4_shift): Likewise.
9809 (pj4_shift_conds): Likewise.
9810 (pj4_alu_shift): Likewise.
9811 (pj4_alu_shift_conds): Likewise.
9812 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9814 (cortexa7_older_only): Likewise.
9815 (cortexa7_younger): Likewise.
9817 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9819 * ipa-pure-const.c (generate_summary): Rename to...
9820 (pure_const_generate_summary): ... this.
9822 2013-07-17 Iain Sandoe <iain@codesourcery.com>
9824 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
9826 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
9829 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
9832 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9834 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
9835 enabled without -march=zEC12.
9836 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
9839 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9841 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
9842 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
9843 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
9844 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
9845 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
9846 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
9847 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
9849 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
9850 and ISA_HAS_NMADD3_NMSUB3.
9851 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
9852 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
9853 (nmsub4<mode>, nmsub3<mode>): Likewise.
9854 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
9856 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9858 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
9859 TARGET_MIPS5400 checking.
9861 2013-07-16 Jakub Jelinek <jakub@redhat.com>
9862 Peter Bergner <bergner@vnet.ibm.com>
9864 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
9865 registers in the comment.
9866 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
9867 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
9868 rather than FIRST_PSEUDO_REGISTERS.
9870 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
9872 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
9873 enable extra ISA flags with TARGET_HTM.
9875 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9877 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
9880 2013-07-15 Cong Hou <congh@google.com>
9882 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
9883 in compare function for sorting.
9885 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
9887 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
9888 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
9889 * config/rs6000/rs6000.opt: Add -mhtm option.
9890 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
9891 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
9892 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9893 __HTM__ if the HTM instructions are available.
9894 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
9895 htm_spr_reg_operand): New define_predicates.
9896 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
9897 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
9899 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
9900 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
9901 HTM builtin functions.
9902 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
9903 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
9904 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
9905 (rs6000_builtin_mask_calculate): Likewise.
9906 (rs6000_option_override_internal): Likewise.
9907 (bdesc_htm): Add new HTM builtin support.
9908 (htm_spr_num): New function.
9909 (htm_spr_regno): Likewise.
9910 (rs6000_htm_spr_icode): Likewise.
9911 (htm_expand_builtin): Likewise.
9912 (htm_init_builtins): Likewise.
9913 (rs6000_expand_builtin): Add support for HTM builtin functions.
9914 (rs6000_init_builtins): Likewise.
9915 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
9917 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
9918 (TARGET_HTM, MASK_HTM): Define macros.
9919 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
9920 (FIXED_REGISTERS): Likewise.
9921 (CALL_USED_REGISTERS): Likewise.
9922 (CALL_REALLY_USED_REGISTERS): Likewise.
9923 (REG_ALLOC_ORDER): Likewise.
9924 (enum reg_class): Likewise.
9925 (REG_CLASS_NAMES): Likewise.
9926 (REG_CLASS_CONTENTS): Likewise.
9927 (REGISTER_NAMES): Likewise.
9928 (ADDITIONAL_REGISTER_NAMES): Likewise.
9929 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
9930 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
9931 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
9932 * config/rs6000/htm.md: New file.
9933 * config/rs6000/htmintrin.h: New file.
9934 * config/rs6000/htmxlintrin.h: New file.
9936 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
9938 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9939 Define SYMBOL_TINY_GOT, update comment.
9940 * config/aarch64/aarch64.c
9941 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
9942 (aarch64_expand_mov_immediate): Likewise.
9943 (aarch64_print_operand): Likewise.
9944 (aarch64_classify_symbol): Likewise.
9945 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
9946 (ldr_got_tiny): Define.
9948 2013-07-13 Tobias Grosser <tobias@grosser.es>
9950 PR tree-optimization/54094
9951 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
9952 scheduling dimension for the parallelism check from the polyhedral
9953 information in the AST.
9954 * graphite-dependences.c (carries_deps): Do not assume the schedule is
9957 2013-07-13 Jason Merrill <jason@redhat.com>
9959 * print-tree.c (debug_vec_tree): Use debug_raw.
9960 (debug_raw (vec<tree, va_gc> &)): New.
9961 (debug_raw (vec<tree, va_gc> *)): New.
9962 * tree.h: Declare them.
9964 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9966 * ifcvt.c (ifcvt_after_combine): New static variable.
9967 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9969 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9970 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9971 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9973 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9975 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9977 2013-07-12 Michael Matz <matz@suse.de>
9980 * convert.c (convert_to_real): Reject non-float inner types.
9982 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9984 * config/aarch64/aarch64-protos.h
9985 (aarch64_simd_immediate_valid_for_move): Remove.
9986 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9987 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9989 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9991 2013-07-11 Steve Ellcey <sellcey@mips.com>
9993 * config/mips/mips.c (mips_conditional_register_usage): Do not
9994 use t[0-7] registers in MIPS16 mode when optimizing for size.
9996 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9998 * config/i386/i386.c (dispatch_function_versions): Fix array
9999 indexing of function_version_info to match actual_versions.
10001 2013-07-11 Teresa Johnson <tejohnson@google.com>
10003 * vec.h (struct va_gc): Move release out-of-line.
10004 (va_gc::release): Call ggc_free on released vec.
10006 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10008 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
10009 Require GOT register as additional operand in UNSPEC.
10010 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
10011 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10012 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10013 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
10014 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
10015 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
10016 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
10018 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
10021 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
10022 name seen by assembler/linker rather if available.
10024 2013-07-11 Andreas Schwab <schwab@suse.de>
10026 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
10028 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
10030 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
10032 2013-07-10 Joseph Myers <joseph@codesourcery.com>
10034 * doc/tm.texi.in: Move hook documentation to ....
10035 * target.def: ... here.
10037 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
10038 text on @hook line.
10039 * doc/tm.texi: Regenerate.
10041 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
10044 * doc/invoke.texi: Document -Wconditionally-supported.
10046 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10049 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
10052 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10055 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
10056 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
10057 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
10058 Remove duplicate devices.
10059 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
10060 * config/avr/t-multilib: Regenerate.
10061 * config/avr/avr-tables.opt: Regenerate.
10062 * doc/avr-mmcu.texi: Regenerate.
10064 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10067 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
10069 2013-07-10 Graham Stott <graham.stott@btinternet.com>
10071 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
10072 the cost of MULT when optimizing for size.
10074 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10076 * config/cr16/cr16-protos.h: Don't include target.h.
10078 2013-07-09 Joseph Myers <joseph@codesourcery.com>
10080 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
10081 adjust register size for TDmode and TFmode for VSX registers.
10083 2013-07-08 Kai Tietz <ktietz@redhat.com>
10086 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
10087 hook_bool_const_tree_true.
10089 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10091 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
10092 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
10093 * config/s390/s390.md: Define FPR*_REGNUM constants.
10094 Fix FPR2_REGNUM constant (18 -> 17).
10095 ("*trunc<BFP:mode><DFP_ALL:mode>2")
10096 ("*trunc<DFP_ALL:mode><BFP:mode>2")
10097 ("trunc<BFP:mode><DFP_ALL:mode>2")
10098 ("trunc<DFP_ALL:mode><BFP:mode>2")
10099 ("*extend<BFP:mode><DFP_ALL:mode>2")
10100 ("*extend<DFP_ALL:mode><BFP:mode>2")
10101 ("extend<BFP:mode><DFP_ALL:mode>2")
10102 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
10105 2013-07-08 Graham Stott <graham.stott@btinternet.com>
10107 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
10109 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10111 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
10112 and cfun_fpr_bit_p to cfun_fpr_save_p.
10113 (s390_frame_area, s390_register_info, s390_frame_info)
10114 (s390_emit_prologue, s390_emit_epilogue)
10115 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
10117 * config/s390/s390.h: Define *_REGNUM macros for floating point
10120 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
10122 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
10124 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
10126 PR rtl-optimization/57786
10127 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
10128 and break out of the loop when it is set to false.
10130 2013-07-08 Jakub Jelinek <jakub@redhat.com>
10133 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
10134 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
10135 (const_int 63)) 0)).
10136 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
10137 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
10138 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
10140 PR rtl-optimization/57829
10141 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
10142 mask bits outside of mode are just sign-extension from mode to HWI.
10144 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10146 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
10147 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
10148 adjust_address instead of change_address to keep info about alignment.
10149 (emit_strmov): Remove.
10150 (emit_memmov): New function.
10151 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
10152 (expand_movmem_epilogue): Likewise and return updated rtx for
10154 (expand_constant_movmem_prologue): Likewise and return updated rtx for
10155 destination and source.
10156 (decide_alignment): Refactor, handle vector_loop.
10157 (ix86_expand_movmem): Likewise.
10158 (ix86_expand_setmem): Likewise.
10159 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
10161 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
10163 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
10164 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
10166 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10168 * config/i386/sse.md (sse_movlhps): Change alternative 3
10169 of operand 2 to "m".
10171 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10174 * config/i386/sse.md (iptr): New mode attribute.
10175 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
10176 (<sse>_vm<plusminus_insn><mode>3): Ditto.
10177 (<sse>_vmmul<mode>3): Ditto.
10178 (<sse>_vmdiv<mode>3): Ditto.
10179 (sse_vmrcpv4sf2): Ditto.
10180 (<sse>_vmsqrt<mode>2): Ditto.
10181 (sse_vmrsqrtv4sf2): Ditto.
10182 (<sse>_vm<code><mode>3): Ditto.
10183 (avx_vmcmp<mode>3): Ditto.
10184 (<sse>_vmmaskcmp<mode>3): Ditto.
10185 (<sse>_comi): Ditto.
10186 (<sse>_ucomi): Ditto.
10187 (*xop_vmfrcz_<mode>): Ditto.
10188 (*fmai_fmadd_<mode>): Ditto.
10189 (*fmai_fmsub_<mode>): Ditto.
10190 (*fmai_fnmadd_<mode>): Ditto.
10191 (*fmai_fnmsub_<mode>): Ditto.
10192 (*fma4i_vmfmadd_<mode>): Ditto.
10193 (*fma4i_vmfmsub_<mode>): Ditto.
10194 (*fma4i_vmfnmadd_<mode>): Ditto.
10195 (*fma4i_vmfnmsub_<mode>): Ditto.
10196 (*xop_vmfrcz_<mode>): Ditto.
10197 (sse_cvtps2pi): Ditto.
10198 (sse_cvttps2pi): Ditto.
10199 (sse_cvtss2si): Ditto.
10200 (sse_cvtss2si_2): Ditto.
10201 (sse_cvtss2siq_2): Ditto.
10202 (sse_cvttss2si): Ditto.
10203 (sse_cvttss2siq): Ditto.
10204 (sse_cvtsd2si): Ditto.
10205 (sse_cvtsd2si_2): Ditto.
10206 (sse_cvtsd2siq_2): Ditto.
10207 (sse_cvttsd2si): Ditto.
10208 (sse_cvttsd2siq): Ditto.
10209 (sse_cvtsd2ss): Ditto.
10210 (sse_cvtss2sd): Ditto.
10211 (avx2_pbroadcast<mode>): Ditto.
10212 (avx2_pbroadcast<mode>_1): Ditto.
10213 (*avx_vperm_broadcast_v4sf): Ditto.
10215 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
10216 (sse_movlhps): Ditto.
10217 (sse_storehps): Ditto.
10218 (sse_loadhps): Ditto.
10219 (sse_storelps): Ditto.
10220 (sse_loadlps): Ditto.
10221 (*vec_concatv4sf): Ditto.
10222 (*vec_interleave_highv2df): Ditto.
10223 (*vec_interleave_lowv2df): Ditto.
10224 (*vec_extractv2df_1_sse): Ditto.
10225 (*vec_extractv2df_0_sse): Ditto.
10226 (sse2_storelpd): Ditto.
10227 (sse2_loadlpd): Ditto.
10228 (sse2_movsd): Ditto.
10229 (*vec_concatv4si): Ditto.
10230 (vec_concatv2di): Ditto.
10232 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
10233 for Intel asm dialect.
10234 (mmx_punpcklwd): Ditto.
10235 (mmx_punpckldq): Ditto.
10237 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
10238 for intel assembler dialect.
10240 2013-07-06 Jakub Jelinek <jakub@redhat.com>
10243 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
10244 for BUILT_IN_C{LZ,LRSB}*.
10245 * tree.h (CASE_INT_FN): Add FN##IMAX case.
10246 * tree-vrp.c (extract_range_basic): Handle
10247 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
10248 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
10249 fall thru to code calling set_value*.
10250 * builtins.c (expand_builtin): Remove *IMAX cases.
10251 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
10252 if width is bigger than 2*HWI.
10254 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
10256 PR rtl-optimization/55342
10257 * lra-int.h (lra_subreg_reload_pseudos): New.
10258 * lra.c: Add undoing optional reloads to the block diagram.
10259 (lra_subreg_reload_pseudos): New.
10260 (lra_optional_reload_pseudos): Change comments.
10261 (lra): Init and clear lra_subreg_reload_pseudos. Clear
10262 lra_optional_reload_pseudos after undo transformations.
10263 * lra-assigns.c (pseudo_prefix_title): New.
10264 (lra_setup_reg_renumber): Use it.
10265 (spill_for): Ditto. Check subreg reload pseudos too.
10266 (assign_by_spills): Consider subreg reload pseudos too.
10267 * lra-constraints.c (simplify_operand_subreg): Use
10268 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
10269 (curr_insn_transform): Recognize and do optional reloads.
10270 (undo_optional_reloads): New.
10271 (lra_undo_inheritance): Call undo_optional_reloads.
10273 2013-07-05 Thomas Quinot <quinot@adacore.com>
10275 * tree-complex.c (expand_complex_operations_1): Fix typo.
10277 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
10279 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
10280 (tune_params): New member 'const vec_costs'.
10281 * config/aarch64/aarch64.c (generic_vector_cost): New.
10282 (generic_tunings): New member 'generic_vector_cost'.
10283 (aarch64_builtin_vectorization_cost): New.
10284 (aarch64_add_stmt_cost): New.
10285 (TARGET_VECTORIZE_ADD_STMT_COST): New.
10286 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
10288 2013-07-03 Jakub Jelinek <jakub@redhat.com>
10291 * config/i386/predicates.md (vsib_address_operand): Disallow
10292 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
10294 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
10296 PR middle-end/55030
10297 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
10298 expand_builtin_setjmp_receiver.
10299 (expand_label): Adjust, call expand_builtin_setjmp_receiver
10300 with NULL for the label parameter.
10301 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
10302 the frame-pointer. Adjust comments.
10303 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
10304 only if LABEL is non-NULL.
10306 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
10308 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
10309 (ARM_ABI_AAPCS64): Ditto.
10311 (ARM_DEFAULT_ABI): Ditto.
10313 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
10315 * config/aarch64/aarch64-builtins.c
10316 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
10317 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
10319 * config/aarch64/aarch64-simd.md
10320 (aarch64_ld1<VALL:mode>): New.
10321 (aarch64_st1<VALL:mode>): Likewise.
10322 * config/aarch64/arm_neon.h
10323 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
10325 2013-07-02 Sriraman Tallam <tmsriram@google.com>
10327 * config/i386/i386.c (gate_insert_vzeroupper): Check if
10329 (ix86_option_override_internal):Turn on all -mavx target flags by
10330 default as they are dependent on AVX anyway.
10332 2013-07-02 Cary Coutant <ccoutant@google.com>
10334 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
10335 deterministic hash.
10336 (loc_checksum_ordered): Likewise.
10337 (hash_loc_operands): Remove inline keyword.
10339 2013-07-02 Jakub Jelinek <jakub@redhat.com>
10341 PR tree-optimization/57741
10342 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
10343 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
10344 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
10345 Allow REAL_CST step_exprs if flag_associative_math.
10346 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
10348 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10350 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
10352 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10354 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
10356 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10358 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
10360 (iorsi3_insn): Likewise.
10361 (arm_xorsi3): Likewise.
10363 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
10365 * arm.md (attribute "wtype"): Delete. Move attribute values from here
10367 (attribute "type"): ... here, and prefix with "wmmx_".
10368 (attribute "core_cycles"): Update for attribute changes.
10369 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
10370 (tbcstv4hi): Likewise.
10371 (tbcstv2si): Likewise.
10372 (iwmmxt_iordi3): Likewise.
10373 (iwmmxt_xordi3): Likewise.
10374 (iwmmxt_anddi3): Likewise.
10375 (iwmmxt_nanddi3): Likewise.
10376 (iwmmxt_arm_movdi): Likewise.
10377 (iwmmxt_movsi_insn): Likewise.
10378 (mov<mode>_internal): Likewise.
10379 (and<mode>3_iwmmxt): Likewise.
10380 (ior<mode>3_iwmmxt): Likewise.
10381 (xor<mode>3_iwmmxt): Likewise.
10382 (add<mode>3_iwmmxt): Likewise.
10383 (ssaddv8qi3): Likewise.
10384 (ssaddv4hi3): Likewise.
10385 (ssaddv2si3): Likewise.
10386 (usaddv8qi3): Likewise.
10387 (usaddv4hi3): Likewise.
10388 (usaddv2si3): Likewise.
10389 (sub<mode>3_iwmmxt): Likewise.
10390 (sssubv8qi3): Likewise.
10391 (sssubv4hi3): Likewise.
10392 (sssubv2si3): Likewise.
10393 (ussubv8qi3): Likewise.
10394 (ussubv4hi3): Likewise.
10395 (ussubv2si3): Likewise.
10396 (mulv4hi3_iwmmxt): Likewise.
10397 (smulv4hi3_highpart): Likewise.
10398 (umulv4hi3_highpart): Likewise.
10399 (iwmmxt_wmacs): Likewise.
10400 (iwmmxt_wmacsz): Likewise.
10401 (iwmmxt_wmacu): Likewise.
10402 (iwmmxt_wmacuz): Likewise.
10403 (iwmmxt_clrdi): Likewise.
10404 (iwmmxt_clrv8qi): Likewise.
10405 (iwmmxt_clr4hi): Likewise.
10406 (iwmmxt_clr2si): Likewise.
10407 (iwmmxt_uavgrndv8qi3): Likewise.
10408 (iwmmxt_uavgrndv4hi3): Likewise.
10409 (iwmmxt_uavgv8qi3): Likewise.
10410 (iwmmxt_uavgv4hi3): Likewise.
10411 (iwmmxt_tinsrb): Likewise.
10412 (iwmmxt_tinsrh): Likewise.
10413 (iwmmxt_tinsrw): Likewise.
10414 (iwmmxt_textrmub): Likewise.
10415 (iwmmxt_textrmsb): Likewise.
10416 (iwmmxt_textrmuh): Likewise.
10417 (iwmmxt_textrmsh): Likewise.
10418 (iwmmxt_textrmw): Likewise.
10419 (iwmxxt_wshufh): Likewise.
10420 (eqv8qi3): Likewise.
10421 (eqv4hi3): Likewise.
10422 (eqv2si3): Likewise.
10423 (gtuv8qi3): Likewise.
10424 (gtuv4hi3): Likewise.
10425 (gtuv2si3): Likewise.
10426 (gtv8qi3): Likewise.
10427 (gtv4hi3): Likewise.
10428 (gtv2si3): Likewise.
10429 (smax<mode>3_iwmmxt): Likewise.
10430 (umax<mode>3_iwmmxt): Likewise.
10431 (smin<mode>3_iwmmxt): Likewise.
10432 (umin<mode>3_iwmmxt): Likewise.
10433 (iwmmxt_wpackhss): Likewise.
10434 (iwmmxt_wpackwss): Likewise.
10435 (iwmmxt_wpackdss): Likewise.
10436 (iwmmxt_wpackhus): Likewise.
10437 (iwmmxt_wpackwus): Likewise.
10438 (iwmmxt_wpackdus): Likewise.
10439 (iwmmxt_wunpckihb): Likewise.
10440 (iwmmxt_wunpckihh): Likewise.
10441 (iwmmxt_wunpckihw): Likewise.
10442 (iwmmxt_wunpckilb): Likewise.
10443 (iwmmxt_wunpckilh): Likewise.
10444 (iwmmxt_wunpckilw): Likewise.
10445 (iwmmxt_wunpckehub): Likewise.
10446 (iwmmxt_wunpckehuh): Likewise.
10447 (iwmmxt_wunpckehuw): Likewise.
10448 (iwmmxt_wunpckehsb): Likewise.
10449 (iwmmxt_wunpckehsh): Likewise.
10450 (iwmmxt_wunpckehsw): Likewise.
10451 (iwmmxt_wunpckelub): Likewise.
10452 (iwmmxt_wunpckeluh): Likewise.
10453 (iwmmxt_wunpckeluw): Likewise.
10454 (iwmmxt_wunpckelsb): Likewise.
10455 (iwmmxt_wunpckelsh): Likewise.
10456 (iwmmxt_wunpckelsw): Likewise.
10457 (ror<mode>3): Likewise.
10458 (ashr<mode>3_iwmmxt): Likewise.
10459 (lshr<mode>3_iwmmxt): Likewise.
10460 (ashl<mode>3_iwmmxt): Likewise.
10461 (ror<mode>3_di): Likewise.
10462 (ashr<mode>3_di): Likewise.
10463 (lshr<mode>3_di): Likewise.
10464 (ashl<mode>3_di): Likewise.
10465 (iwmmxt_wmadds): Likewise.
10466 (iwmmxt_wmaddu): Likewise.
10467 (iwmmxt_tmia): Likewise.
10468 (iwmmxt_tmiaph): Likewise.
10469 (iwmmxt_tmiabb): Likewise.
10470 (iwmmxt_tmiatb): Likewise.
10471 (iwmmxt_tmiabt): Likewise.
10472 (iwmmxt_tmiatt): Likewise.
10473 (iwmmxt_tmovmskb): Likewise.
10474 (iwmmxt_tmovmskh): Likewise.
10475 (iwmmxt_tmovmskw): Likewise.
10476 (iwmmxt_waccb): Likewise.
10477 (iwmmxt_wacch): Likewise.
10478 (iwmmxt_waccw): Likewise.
10479 (iwmmxt_waligni): Likewise.
10480 (iwmmxt_walignr): Likewise.
10481 (iwmmxt_walignr0): Likewise.
10482 (iwmmxt_walignr1): Likewise.
10483 (iwmmxt_walignr2): Likewise.
10484 (iwmmxt_walignr3): Likewise.
10485 (iwmmxt_wsadb): Likewise.
10486 (iwmmxt_wsadh): Likewise.
10487 (iwmmxt_wsadbz): Likewise.
10488 (iwmmxt_wsadhz): Likewise.
10489 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
10490 (iwmmxt_wabsdiffb): Likewise.
10491 (iwmmxt_wabsdiffh): Likewise.
10492 (iwmmxt_wabsdiffw): Likewise.
10493 (iwmmxt_waddsubhx): Likewise
10494 (iwmmxt_wsubaddhx): Likewise.
10495 (addc<mode>3): Likewise.
10496 (iwmmxt_avg4): Likewise.
10497 (iwmmxt_avg4r): Likewise.
10498 (iwmmxt_wmaddsx): Likewise.
10499 (iwmmxt_wmaddux): Likewise.
10500 (iwmmxt_wmaddsn): Likewise.
10501 (iwmmxt_wmaddun): Likewise.
10502 (iwmmxt_wmulwsm): Likewise.
10503 (iwmmxt_wmulwum): Likewise.
10504 (iwmmxt_wmulsmr): Likewise.
10505 (iwmmxt_wmulumr): Likewise.
10506 (iwmmxt_wmulwsmr): Likewise.
10507 (iwmmxt_wmulwumr): Likewise.
10508 (iwmmxt_wmulwl): Likewise.
10509 (iwmmxt_wqmulm): Likewise.
10510 (iwmmxt_wqmulwm): Likewise.
10511 (iwmmxt_wqmulmr): Likewise.
10512 (iwmmxt_wqmulwmr): Likewise.
10513 (iwmmxt_waddbhusm): Likewise.
10514 (iwmmxt_waddbhusl): Likewise.
10515 (iwmmxt_wqmiabb): Likewise.
10516 (iwmmxt_wqmiabt): Likewise.
10517 (iwmmxt_wqmiatb): Likewise.
10518 (iwmmxt_wqmiatt): Likewise.
10519 (iwmmxt_wqmiabbn): Likewise.
10520 (iwmmxt_wqmiabtn): Likewise.
10521 (iwmmxt_wqmiatbn): Likewise.
10522 (iwmmxt_wqmiattn): Likewise.
10523 (iwmmxt_wmiabb): Likewise.
10524 (iwmmxt_wmiabt): Likewise.
10525 (iwmmxt_wmiatb): Likewise.
10526 (iwmmxt_wmiatt): Likewise.
10527 (iwmmxt_wmiabbn): Likewise.
10528 (iwmmxt_wmiabtn): Likewise.
10529 (iwmmxt_wmiatbn): Likewise.
10530 (iwmmxt_wmiattn): Likewise.
10531 (iwmmxt_wmiawbb): Likewise.
10532 (iwmmxt_wmiawbt): Likewise.
10533 (iwmmxt_wmiawtb): Likewise.
10534 (iwmmxt_wmiawtt): Likewise.
10535 (iwmmxt_wmiawbbn): Likewise.
10536 (iwmmxt_wmiawbtn): Likewise.
10537 (iwmmxt_wmiawtbn): Likewise.
10538 (iwmmxt_wmiawttn): Likewise.
10539 (iwmmxt_wmerge): Likewise.
10540 (iwmmxt_tandc<mode>3): Likewise.
10541 (iwmmxt_torc<mode>3): Likewise.
10542 (iwmmxt_torvsc<mode>3): Likewise.
10543 (iwmmxt_textrc<mode>3): Likewise.
10544 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
10545 (wmmxt_pack): Likewise.
10546 (wmmxt_mult_c1): Likewise.
10547 (wmmxt_mult_c2): Likewise.
10548 (wmmxt_alu_c1): Likewise.
10549 (wmmxt_alu_c2): Likewise.
10550 (wmmxt_alu_c3): Likewise.
10551 (wmmxt_transfer_c1): Likewise.
10552 (wmmxt_transfer_c2): Likewise.
10553 (wmmxt_transfer_c3): Likewise.
10554 (marvell_f_iwmmxt_wstr): Likewise.
10555 (marvell_f_iwmmxt_wldr): Likewise.
10557 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
10559 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
10561 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10564 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10565 * lra-constraints.c (need_for_split_p): Check call used hard regs
10566 living through calls.
10568 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
10569 call used regs for call insn.
10571 2013-06-28 Jakub Jelinek <jakub@redhat.com>
10574 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
10575 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
10577 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10579 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
10581 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10583 * lra-constraints.c (need_for_split_p): Check call used hard regs
10584 living through calls.
10586 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
10589 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
10590 to tie with any other modes. Eliminate Altivec vector mode tests,
10591 since these are a subset of ALTIVEC or VSX vector modes. Simplify
10592 code, to return 0 if testing MODE2 for a condition, if we've
10593 already tested MODE1 for the same condition.
10595 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10597 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
10600 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10602 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10603 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
10605 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10607 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
10609 (aarch64_symbolic_constant_p): Remove.
10610 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
10611 static. Fix line length and white space.
10612 (aarch64_symbolic_constant_p): Remove.
10613 * config/aarch64/predicates.md (aarch64_valid_symref):
10614 Use aarch64_classify_symbol_expression.
10616 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10618 * config/arm/constraints.md (Ts): New constraint.
10619 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
10621 (compare_scc): Use "Ts" constraint for operand 0.
10622 (ior_scc_scc): Likewise.
10623 (and_scc_scc): Likewise.
10624 (and_scc_scc_nodom): Likewise.
10625 (ior_scc_scc_cmp): Likewise for operand 7.
10626 (and_scc_scc_cmp): Likewise.
10627 * config/arm/thumb2.md (thumb2_movsi_insn):
10628 Add alternatives for 16-bit encodings.
10629 (thumb2_movhi_insn): Likewise.
10630 (thumb2_movsicc_insn): Likewise.
10631 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
10632 (thumb2_negscc): Use "Ts" constraint.
10633 Move mvn instruction outside cond_exec block.
10634 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
10635 for 16-bit encodings.
10637 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10639 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
10641 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
10642 (mulsi3subsi): Likewise.
10643 (mulsidi3adddi): Likewise.
10644 (mulsidi3_v6): Likewise.
10645 (umulsidi3_v6): Likewise.
10646 (umulsidi3adddi_v6): Likewise.
10647 (smulsi3_highpart_v6): Likewise.
10648 (umulsi3_highpart_v6): Likewise.
10649 (mulhisi3tb): Likewise.
10650 (mulhisi3bt): Likewise.
10651 (mulhisi3tt): Likewise.
10652 (maddhisi4): Likewise.
10653 (maddhisi4tb): Likewise.
10654 (maddhisi4tt): Likewise.
10655 (maddhidi4): Likewise.
10656 (maddhidi4tb): Likewise.
10657 (maddhidi4tt): Likewise.
10658 (zeroextractsi_compare0_scratch): Likewise.
10659 (insv_zero): Likewise.
10660 (insv_t2): Likewise.
10661 (anddi_notzesidi_di): Likewise.
10662 (anddi_notsesidi_di): Likewise.
10663 (andsi_notsi_si): Likewise.
10664 (iordi_zesidi_di): Likewise.
10665 (xordi_zesidi_di): Likewise.
10666 (andsi_iorsi3_notsi): Likewise.
10667 (smax_0): Likewise.
10668 (smax_m1): Likewise.
10669 (smin_0): Likewise.
10670 (not_shiftsi): Likewise.
10671 (unaligned_loadsi): Likewise.
10672 (unaligned_loadhis): Likewise.
10673 (unaligned_loadhiu): Likewise.
10674 (unaligned_storesi): Likewise.
10675 (unaligned_storehi): Likewise.
10676 (extv_reg): Likewise.
10677 (extzv_t2): Likewise.
10678 (divsi3): Likewise.
10679 (udivsi3): Likewise.
10680 (arm_zero_extendhisi2addsi): Likewise.
10681 (arm_zero_extendqisi2addsi): Likewise.
10682 (compareqi_eq0): Likewise.
10683 (arm_extendhisi2_v6): Likewise.
10684 (arm_extendqisi2addsi): Likewise.
10685 (arm_movt): Likewise.
10686 (thumb2_ldrd): Likewise.
10687 (thumb2_ldrd_base): Likewise.
10688 (thumb2_ldrd_base_neg): Likewise.
10689 (thumb2_strd): Likewise.
10690 (thumb2_strd_base): Likewise.
10691 (thumb2_strd_base_neg): Likewise.
10692 (arm_negsi2): Add alternative for 16-bit encoding.
10693 (arm_one_cmplsi2): Likewise.
10695 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10697 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
10698 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
10699 (movdfcc): Likewise.
10700 * config/arm/vfp.md (*thumb2_movsf_vfp):
10701 Disable predication for arm_restrict_it.
10702 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
10703 (*thumb2_movdfcc_vfp): Likewise.
10704 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
10705 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
10706 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
10707 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
10708 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
10709 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
10710 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
10711 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
10712 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
10713 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
10714 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
10715 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
10716 Disable predication for arm_restrict_it.
10718 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
10720 * config/i386/bmiintrin.h (_bextr_u32): New.
10721 (_bextr_u64): Ditto.
10723 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
10725 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
10726 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
10727 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
10728 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
10729 * config/mips/n32-elf.h: ...this new file.
10731 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
10734 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
10735 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
10737 2013-06-27 Catherine Moore <clm@codesourcery.com>
10739 * config/mips/mips-tables.opt: Regenerate.
10740 * config/mips/mips-cpus.def: Add m14ke and m14kec.
10741 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
10742 * doc/invoke.texi: Add -m14kc.
10744 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10747 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
10748 constraints of operand 1 and 2.
10751 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
10752 to match RTL canonicalization. Swap predicates and
10753 constraints of operand 1 and 2.
10755 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
10757 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
10758 Process OP_INOUT regs for splitting too.
10760 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10762 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
10763 decl before the loop, initialize to NULL.
10764 (vectorizable_load): Initialize ptr_incr to NULL.
10766 2013-06-27 Martin Jambor <mjambor@suse.cz>
10769 * ipa-ref.h (ipa_maybe_record_reference): Declare.
10770 * ipa-ref.c (ipa_maybe_record_reference): New function.
10771 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
10772 * ipa-cp.c (create_specialized_node): Record potential references from
10774 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
10776 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10778 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
10779 parameter 'mode' of type 'enum machine_mode mode'; change to pass
10780 'mode' to force_reg.
10781 (aarch64_add_offset): Update calls to aarch64_force_temporary.
10782 (aarch64_expand_mov_immediate): Likewise.
10784 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10786 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
10787 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
10789 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10791 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
10792 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10793 (struct machine_function): Add tbegin_p.
10794 (s390_canonicalize_comparison): Fold CC mode compares to
10795 conditional jump if possible.
10796 (s390_emit_jump): Return the emitted jump.
10797 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
10798 Handle CCRAWmode compares.
10799 (s390_option_override): Default to -mhtm if available.
10800 (s390_reg_clobbered_rtx): Handle floating point regs as well.
10801 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
10802 FPRs instead of df_regs_ever_live_p.
10803 (s390_optimize_nonescaping_tx): New function.
10804 (s390_init_frame_layout): Extend clobbered_regs array to cover
10806 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
10807 (s390_expand_tbegin): New function.
10808 (enum s390_builtin): New enum definition.
10809 (code_for_builtin): New array definition.
10810 (s390_init_builtins): New function.
10811 (s390_expand_builtin): New function.
10812 (TARGET_INIT_BUILTINS): Define.
10813 (TARGET_EXPAND_BUILTIN): Define.
10814 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
10815 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
10816 (s390_alc_comparison): Likewise.
10817 * config/s390/s390-modes.def: Add CCRAWmode.
10818 * config/s390/s390.h (processor_flags): Add PF_TX.
10819 (TARGET_CPU_HTM): Define macro.
10820 (TARGET_HTM): Define macro.
10821 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
10822 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
10823 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10824 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
10825 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
10826 (TBEGIN_MASK, TBEGINC_MASK): New constants.
10827 ("*cc_to_int"): Move up.
10828 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
10829 constants other than 0.
10830 ("*ccraw_to_int"): New insn and splitter definition.
10831 ("tbegin", "tbegin_nofloat", "tbegin_retry")
10832 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
10833 ("tx_assist"): New expander.
10834 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
10835 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
10836 * config/s390/s390.opt: Add -mhtm option.
10837 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
10838 * config/s390/htmxlintrin.h: New file.
10839 * config/s390/htmintrin.h: New file.
10840 * config/s390/s390intrin.h: New file.
10841 * doc/extend.texi: Document htm builtins.
10842 * config.gcc: Add the new header files to extra_headers.
10844 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
10846 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
10847 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
10849 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
10850 Pat Haugen <pthaugen@us.ibm.com>
10851 Peter Bergner <bergner@vnet.ibm.com>
10853 * config/rs6000/power8.md: New.
10854 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
10855 setting for power8 entry.
10856 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
10857 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
10858 test for Power4/Power5 only.
10859 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
10861 (force_new_group): Adjust comment.
10862 * config/rs6000/rs6000.md: Include power8.md.
10864 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
10866 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
10867 * config/arm/arm-protos.h (arm_max_conditional_execute): New
10869 (tune_params): Update comment.
10870 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
10871 (arm_max_conditional_execute): New function.
10872 (thumb2_final_prescan_insn): Use max_insn_skipped and
10873 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
10875 2013-06-25 Jakub Jelinek <jakub@redhat.com>
10877 PR tree-optimization/57705
10878 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
10879 SSA_NAME step, provided that it is not defined inside the loop.
10880 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
10881 (get_initial_def_for_induction): Handle SSA_NAME IV step.
10883 2013-06-25 Martin Jambor <mjambor@suse.cz>
10885 PR middle-end/57670
10886 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
10887 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
10889 (ipa_note_param_call): Initialize member_ptr flag.
10890 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
10891 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
10892 (ipa_write_indirect_edge_info): Stream member_ptr flag.
10893 (ipa_read_indirect_edge_info): Likewise.
10895 2013-06-25 Richard Biener <rguenther@suse.de>
10897 PR middle-end/56977
10898 * passes.c (init_optimization_passes): Move pass_fold_builtins
10899 and pass_dce earlier with -Og.
10901 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
10903 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
10904 <BIT_FIELD_REF>: Remove trailing TAB.
10905 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
10908 2013-06-24 Martin Jambor <mjambor@suse.cz>
10910 PR tree-optimization/57358
10911 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
10912 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
10913 (ipa_analyze_params_uses): Generate pessimistic info when true.
10915 2013-06-24 Martin Jambor <mjambor@suse.cz>
10917 PR tree-optimization/57539
10918 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
10919 global.inlined_to of the new node to it. All callers changed.
10920 * ipa-inline-transform.c (clone_inlined_nodes): New variable
10921 inlining_into, pass it to cgraph_clone_node.
10922 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
10923 ipa_free_edge_args_substructures.
10924 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
10925 rdesc linked list. Do not assert rdesc edges have inlined caller.
10926 Assert we have found an rdesc in the rdesc list.
10928 2013-06-24 Richard Biener <rguenther@suse.de>
10930 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
10931 (pointer_set_lookup): Declare.
10932 (class pointer_map): New template class implementing a
10933 generic pointer to T map.
10934 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
10935 pointer_map<T>::contains, pointer_map<T>::insert,
10936 pointer_map<T>::traverse): New functions.
10937 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
10938 (pointer_set_lookup): New function.
10939 (pointer_set_contains): Use pointer_set_lookup.
10940 (pointer_set_insert): Likewise.
10941 (insert_aux): Remove.
10942 (struct pointer_map_t): Embed a pointer_set_t.
10943 (pointer_map_create): Adjust.
10944 (pointer_map_destroy): Likewise.
10945 (pointer_map_contains): Likewise.
10946 (pointer_map_insert): Likewise.
10947 (pointer_map_traverse): Likewise.
10948 * tree-streamer.h (struct streamer_tree_cache_d): Use a
10949 pointer_map<unsigned> instead of a pointer_map_t.
10950 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
10951 (streamer_tree_cache_lookup): Likewise.
10952 (streamer_tree_cache_create): Likewise.
10953 (streamer_tree_cache_delete): Likewise.
10954 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
10955 pointer_map<unsigned> instead of a pointer_map_t.
10956 (lto_init_tree_ref_encoder): Adjust.
10957 (lto_destroy_tree_ref_encoder): Likewise.
10958 * lto-section-out.c (lto_output_decl_index): Likewise.
10959 (lto_record_function_out_decl_state): Likewise.
10960 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10962 2013-06-24 Richard Biener <rguenther@suse.de>
10964 PR tree-optimization/57488
10965 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10967 2013-06-24 Alan Modra <amodra@gmail.com>
10969 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10970 (gen_easy_altivec_constant): Likewise.
10971 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10972 easy_vector_constant_msb): Likewise.
10974 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10977 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10978 add missing return true.
10980 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10983 * config/sh/predicates.md (general_extend_operand): Invoke
10984 general_movsrc_operand for memory operands.
10985 (general_movsrc_operand): Allow reg+reg addressing, do not use
10986 general_operand for memory operands.
10988 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10990 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10991 when current target options does not apply.
10992 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10993 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10994 * config/i386/bmiintrin.h: Pass appropriate target
10995 attributes to header.
10996 * config/i386/mmintrin.h: Ditto.
10997 * config/i386/nmmintrin.h: Ditto.
10998 * config/i386/avx2intrin.h: Ditto.
10999 * config/i386/fxsrintrin.h: Ditto.
11000 * config/i386/tbmintrin.h: Ditto.
11001 * config/i386/xsaveintrin.h: Ditto.
11002 * config/i386/f16cintrin.h: Ditto.
11003 * config/i386/xtestintrin.h: Ditto.
11004 * config/i386/xsaveoptintrin.h: Ditto.
11005 * config/i386/bmi2intrin.h: Ditto.
11006 * config/i386/lzcntintrin.h: Ditto.
11007 * config/i386/smmintrin.h: Ditto.
11008 * config/i386/wmmintrin.h: Ditto.
11009 * config/i386/x86intrin.h: Remove all header include guards.
11010 * config/i386/prfchwintrin.h: Ditto.
11011 * config/i386/pmmintrin.h: Ditto.
11012 * config/i386/tmmintrin.h: Ditto.
11013 * config/i386/xmmintrin.h: Ditto.
11014 * config/i386/popcntintrin.h: Ditto.
11015 * config/i386/rdseedintrin.h: Ditto.
11016 * config/i386/ammintrin.h: Ditto.
11017 * config/i386/emmintrin.h: Ditto.
11018 * config/i386/immintrin.h: Remove all header include guards.
11019 * config/i386/fma4intrin.h: Ditto.
11020 * config/i386/lwpintrin.h: Ditto.
11021 * config/i386/xopintrin.h: Ditto.
11022 * config/i386/ia32intrin.h: Ditto.
11023 * config/i386/avxintrin.h: Ditto.
11024 * config/i386/rtmintrin.h: Ditto.
11025 * config/i386/fmaintrin.h: Ditto.
11026 * config/i386/mm3dnow.h: Ditto.
11028 2013-06-22 Sriraman Tallam <tmsriram@google.com>
11030 * common/config/i386/i386-common.c: Handle LZCNT.
11032 2013-06-22 Andi Kleen <ak@linux.intel.com>
11034 * doc/extend.texi: Use __atomic_store_n instead of
11035 __atomic_store in HLE example.
11037 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
11039 * config/sh/sh.c: Remove <cstdlib> workaround.
11041 2013-06-21 Andi Kleen <ak@linux.intel.com>
11043 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
11045 2013-06-21 Andi Kleen <ak@linux.intel.com>
11047 * doc/extend.texi: Document that __atomic_clear and
11048 __atomic_test_and_set should only be used with bool.
11050 2013-06-20 Jan Hubicka <jh@suse.cz>
11052 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
11053 types_same_for_odr.
11054 * tree.c (decls_same_for_odr): New function.
11055 (same_for_edr): New function.
11056 (types_same_for_odr): New function.
11057 (get_binfo_at_offset): Use it.
11058 * tree.h (types_same_for_odr): Declare.
11060 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
11061 Jason Merrill <jason@redhat.com>
11063 * system.h: Include <cstdlib> as well as <stdlib.h>.
11065 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
11068 * config/i386/i386.c (construct_container): Report error if
11069 long double is used with disabled x87 float returns.
11071 2013-06-20 Jan Hubicka <jh@suse.cz>
11073 * lto-cgraph.c (input_symtab): Do not set cgraph state.
11075 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
11077 PR rtl-optimization/57425
11078 PR rtl-optimization/57569
11079 * alias.c (write_dependence_p): Remove parameters mem_mode and
11080 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
11081 Changed all callers.
11082 (canon_anti_dependence): Get comments and semantics in sync.
11083 Add parameter mem_canonicalized. Changed all callers.
11084 * rtl.h (canon_anti_dependence): Update prototype.
11086 2013-06-20 Richard Biener <rguenther@suse.de>
11088 * data-streamer-in.c (streamer_read_uhwi): Optimize single
11089 byte case, inline streamer_read_uchar and defer section
11092 2013-06-20 Richard Biener <rguenther@suse.de>
11094 PR tree-optimization/57584
11095 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
11096 SSA names into the expanded expression that take part in
11097 abnormal coalescing.
11099 2013-06-19 Sharad Singhai <singhai@google.com>
11101 * gcov.c (print_usage): Handle new option.
11102 (process_args): Ditto.
11103 (get_gcov_intermediate_filename): New function.
11104 (output_intermediate_file): New function.
11105 (output_gcov_file): New function
11106 (generate_results): Handle new option.
11107 (release_function): Relase demangled name.
11108 (read_graph_file): Handle demangled name.
11109 (output_lines): Ditto.
11110 * doc/gcov.texi: Document gcov intermediate format.
11112 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
11115 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
11116 (lra_emit_add): Use the functions. Add comment about Y as an
11119 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
11122 * collect2.c (collect_atexit): New.
11123 (collect_exit): Delete.
11124 (main): Register collect_atexit with atexit.
11125 (collect_wait): Change collect_exit to exit.
11127 * collect2.h (collect_exit): Delete.
11128 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
11130 2013-06-19 Wei Mi <wmi@google.com>
11132 PR rtl-optimization/57518
11133 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
11134 if regno is used in paradoxical subreg.
11135 (update_equiv_regs): Check pdx_subregs[regno] before
11136 set a reg to be equivalent with a mem.
11138 2013-06-19 Matthias Klose <doko@ubuntu.com>
11141 * file-find.h (find_a_file): Add a mode parameter.
11142 * file-find.c (find_a_file): Likewise.
11143 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
11144 with X_OK for the executables.
11145 * collect2.c (main): Call find_a_file with X_OK.
11147 2013-06-19 Steve Ellcey <sellcey@mips.com>
11150 * config/mips/mips.md (casesi_internal_mips16_<mode>):
11151 Use NEXT_INSN instead of next_real_insn.
11153 2013-06-19 Jan Hubicka <jh@suse.cz>
11155 * cgraph.h (const_value_known_p): Replace by ...
11156 (ctor_for_folding): .. this one.
11157 * cgraphunit.c (process_function_and_variable_attributes): Use it.
11158 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
11159 * expr.c (expand_expr_real_1): Likewise.
11160 (string_constant): Likewise.
11161 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
11162 * ipa.c (process_references): Likewise.
11163 (symtab_remove_unreachable_nodes): Likewise.
11164 * ipa-inline-analysis.c (param_change_prob): Likewise.
11165 * gimple-fold.c (canonicalize_constructor_val): Likewise.
11166 (get_base_constructor): Likwise.
11167 * varpool.c (varpool_remove_node): Likewise.
11168 (varpool_remove_initializer): LIkewise.
11169 (dump_varpool_node): LIkwise.
11170 (const_value_known_p): Rewrite to ...
11171 (ctor_for_folding): ... this one.
11173 2013-06-19 Jakub Jelinek <jakub@redhat.com>
11176 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
11177 PERSONALITY in $PATH derived prefixes.
11179 2013-06-19 Jeff Law <law@redhat.com>
11181 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
11184 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
11185 (simplify_bitwise_binary): Use it to simpify certain binary ops on
11188 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
11190 * config/arm/vfp.md: Move VFP instruction classification documentation
11192 * config/arm/arm.md: ... here. Update instruction classification
11195 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
11197 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
11198 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
11199 pattern. Use more efficient sequences on ARMv5 and Thumb2.
11201 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
11204 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11205 with NEXT_INSN. Use tablejump_p to check for jump table data
11208 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
11211 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
11212 that now in C++ the value is correct per the C++ standards.
11214 2013-06-19 Richard Biener <rguenther@suse.de>
11216 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
11217 for global context.
11219 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11222 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11225 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11226 with next_active_insn.
11228 2013-06-18 Sriraman Tallam <tmsriram@google.com>
11230 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
11231 functions are inlined during failures to flag an error.
11232 * tree-inline.c (expand_call_inline): Allow the error to be flagged
11233 in early inline pass.
11235 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
11237 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
11240 2013-06-18 Julian Brown <julian@codesourcery.com>
11242 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
11243 Permit virtual register pre-reload if !strict.
11244 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
11246 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
11248 * config/arm/neon.md (movmisalign<mode>): Use
11249 neon_perm_struct_or_reg_operand instead of
11250 neon_struct_or_register_operand.
11251 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
11252 neon_permissive_struct_operand instead of neon_struct_operand.
11253 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
11254 neon_vector_mem_operand.
11255 * config/arm/predicates.md (neon_struct_operand): Adjust call to
11256 neon_vector_mem_operand.
11257 (neon_permissive_struct_operand): New.
11258 (neon_struct_or_register_operand): Rename to...
11259 (neon_perm_struct_or_reg_operand): This. Adjust call to
11260 neon_vector_mem_operand.
11262 2013-06-18 Richard Biener <rguenther@suse.de>
11264 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
11265 * lto-streamer.h: Include pointer-set.h.
11266 (struct lto_decl_slot): Remove.
11267 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
11268 Remove next_index entry.
11269 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
11270 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
11271 (lto_init_tree_ref_encoder): Adjust.
11272 (lto_destroy_tree_ref_encoder): Likewise.
11273 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
11274 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
11275 (lto_output_decl_index): Adjust.
11276 (lto_new_out_decl_state): Likewise.
11277 (lto_record_function_out_decl_state): Likewise.
11278 * lto-streamer-out.c (copy_function): Likewise.
11280 2013-06-18 Richard Biener <rguenther@suse.de>
11282 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
11283 * cgraphunit.c: Include cfgloop.h.
11284 (init_lowered_empty_function): Initialize the loop tree.
11285 (assemble_thunk): Insert new BBs into loops.
11287 2013-06-18 Richard Biener <rguenther@suse.de>
11289 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
11290 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
11291 the map from cache entry to cache index optional.
11292 (streamer_tree_cache_replace_tree): Adjust accordingly.
11293 (streamer_tree_cache_append): Likewise.
11294 (streamer_tree_cache_delete): Likewise.
11295 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
11296 streamer cache map from cache entry to cache index.
11297 * lto-streamer-out.c (create_output_block): Adjust.
11299 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
11301 * config/arm/arm.md (attribute "insn"): Move multiplication and
11302 division attributes to...
11303 (attribute "type"): ... here. Remove mult.
11304 (attribute "mul32"): New attribute.
11305 (attribute "mul64"): Add umaal.
11306 (*arm_mulsi3): Update attributes.
11307 (*arm_mulsi3_v6): Likewise.
11308 (*thumb_mulsi3): Likewise.
11309 (*thumb_mulsi3_v6): Likewise.
11310 (*mulsi3_compare0): Likewise.
11311 (*mulsi3_compare0_v6): Likewise.
11312 (*mulsi_compare0_scratch): Likewise.
11313 (*mulsi_compare0_scratch_v6): Likewise.
11314 (*mulsi3addsi): Likewise.
11315 (*mulsi3addsi_v6): Likewise.
11316 (*mulsi3addsi_compare0): Likewise.
11317 (*mulsi3addsi_compare0_v6): Likewise.
11318 (*mulsi3addsi_compare0_scratch): Likewise.
11319 (*mulsi3addsi_compare0_scratch_v6): Likewise.
11320 (*mulsi3subsi): Likewise.
11321 (*mulsidi3adddi): Likewise.
11322 (*mulsi3addsi_v6): Likewise.
11323 (*mulsidi3adddi_v6): Likewise.
11324 (*mulsidi3_nov6): Likewise.
11325 (*mulsidi3_v6): Likewise.
11326 (*umulsidi3_nov6): Likewise.
11327 (*umulsidi3_v6): Likewise.
11328 (*umulsidi3adddi): Likewise.
11329 (*umulsidi3adddi_v6): Likewise.
11330 (*smulsi3_highpart_nov6): Likewise.
11331 (*smulsi3_highpart_v6): Likewise.
11332 (*umulsi3_highpart_nov6): Likewise.
11333 (*umulsi3_highpart_v6): Likewise.
11334 (mulhisi3): Likewise.
11335 (*mulhisi3tb): Likewise.
11336 (*mulhisi3bt): Likewise.
11337 (*mulhisi3tt): Likewise.
11338 (maddhisi4): Likewise.
11339 (*maddhisi4tb): Likewise.
11340 (*maddhisi4tt): Likewise.
11341 (maddhidi4): Likewise.
11342 (*maddhidi4tb): Likewise.
11343 (*maddhidi4tt): Likewise.
11344 (divsi3): Likewise.
11345 (udivsi3): Likewise.
11346 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
11347 (thumb2_mulsi_short_compare0): Likewise.
11348 (thumb2_mulsi_short_compare0_scratch): Likewise.
11349 * config/arm/arm1020e.md (1020mult1): Update attribute change.
11350 (1020mult2): Likewise.
11351 (1020mult3): Likewise.
11352 (1020mult4): Likewise.
11353 (1020mult5): Likewise.
11354 (1020mult6): Likewise.
11355 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
11357 (cortex_a15_mult64): Likewise.
11358 (cortex_a15_sdiv): Likewise.
11359 (cortex_a15_udiv): Likewise.
11360 * config/arm/arm1026ejs.md (mult1): Update attribute change.
11366 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
11367 (pj4_ir_div): Likewise.
11368 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
11369 (11_mult2): Likewise.
11370 (11_mult3): Likewise.
11371 (11_mult4): Likewise.
11372 (11_mult5): Likewise.
11373 (11_mult6): Likewise.
11374 (11_mult7): Likewise.
11375 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
11376 (cortex_a8_mla): Likewise.
11377 (cortex_a8_mull): Likewise.
11378 (cortex_a8_smulwy): Likewise.
11379 (cortex_a8_smlald): Likewise.
11380 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
11381 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
11382 (cortex_r4_mul_3): Likewise.
11383 (cortex_r4_mla_4): Likewise.
11384 (cortex_r4_mla_3): Likewise.
11385 (cortex_r4_smlald): Likewise.
11386 (cortex_r4_mull): Likewise.
11387 (cortex_r4_sdiv): Likewise.
11388 (cortex_r4_udiv): Likewise.
11389 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
11390 (cortex_a7_idiv): Likewise.
11391 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
11392 (9_mult2): Likewise.
11393 (9_mult3): Likewise.
11394 (9_mult4): Likewise.
11395 (9_mult5): Likewise.
11396 (9_mult6): Likewise.
11397 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
11398 (cortex_a53_sdiv): Likewise.
11399 (cortex_a53_udiv): Likewise.
11400 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
11401 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
11402 (mp626_mult2): Likewise.
11403 (mp626_mult3): Likewise.
11404 (mp626_mult4): Likewise.
11405 * config/arm/fa526.md (526_mult1): Update attribute change.
11406 (526_mult2): Likewise.
11407 * config/arm/arm-generic.md (mult): Update attribute change.
11408 (mult_ldsched_strongarm): Likewise.
11409 (mult_ldsched): Likewise.
11410 (multi_cycle): Likewise.
11411 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
11412 * config/arm/fa606te.md (606te_mult1): Update attribute change.
11413 (606te_mult2): Likewise.
11414 (606te_mult3): Likewise.
11415 (606te_mult4): Likewise.
11416 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
11417 (cortex_a9_mac16): Likewise.
11418 (cortex_a9_multiply): Likewise.
11419 (cortex_a9_mac): Likewise.
11420 (cortex_a9_multiply_long): Likewise.
11421 * config/arm/fa626te.md (626te_mult1): Update attribute change.
11422 (626te_mult2): Likewise.
11423 (626te_mult3): Likewise.
11424 (626te_mult4): Likewise.
11426 2013-06-18 Richard Biener <rguenther@suse.de>
11429 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
11431 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11434 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11435 with next_active_insn.
11437 2013-06-18 Alan Modra <amodra@gmail.com>
11439 * config/rs6000/rs6000.h (enum data_align): New.
11440 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
11441 (DATA_ABI_ALIGNMENT): Define.
11442 (CONSTANT_ALIGNMENT): Correct comment.
11443 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
11444 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
11446 2013-06-17 David Malcolm <dmalcolm@redhat.com>
11448 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
11449 ATTRIBUTE_UNUSED marking.
11451 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
11453 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
11454 alternative and update.
11455 (aarch64_dup_lanedi): Delete.
11456 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
11457 * config/aarch64/aarch64-simd-builtins.def: Update.
11459 2013-06-17 Richard Biener <rguenther@suse.de>
11461 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
11462 (lto_input_scc): Declare.
11463 (lto_input_tree_1): Likewise.
11464 (struct lto_stats_d): Add num_tree_bodies_output and
11465 num_pickle_refs_output.
11466 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
11467 (lto_read_tree_1): Split out from ...
11468 (lto_read_tree): ... this.
11469 (lto_input_scc): New function.
11470 (lto_input_tree_1): Split out from ...
11471 (lto_input_tree): ... this. Handle LTO_tree_scc.
11472 (lto_data_in_create): Create the streamer cache without hashes.
11473 * lto-streamer-out.c (create_output_block): Create the streamer
11474 cache with hashes when not doing WPA.
11475 (lto_write_tree_1): Split out from ...
11476 (lto_write_tree): ... this.
11477 (get_symbol_initial_value): New function.
11478 (lto_output_tree_1): Split out from ...
11479 (lto_output_tree): ... this. Write trees as series of SCCs
11480 using a DFS walk via DFS_write_tree.
11481 (struct sccs, struct scc_entry): New types.
11482 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
11483 (DFS_write_tree_body): New function.
11484 (DFS_write_tree): Likewise.
11485 (hash_tree): Likewise.
11486 (scc_entry_compare): Likewise.
11487 (hash_scc): Likewise.
11488 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
11489 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
11490 TREE_CHAIN as regular reference.
11491 (streamer_read_integer_cst): Remove.
11492 (streamer_get_pickled_tree): Adjust.
11493 * tree-streamer-out.c (streamer_write_chain): Disable streaming
11494 of DECL_EXTERNALs in BLOCK_VARS for now.
11495 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
11497 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
11498 Add hash value argument and record that if hashes are recorded
11500 (streamer_tree_cache_insert_1): Adjust.
11501 (streamer_tree_cache_insert): Likewise.
11502 (streamer_tree_cache_insert_at): Rename to ...
11503 (streamer_tree_cache_replace_tree): ... this and adjust.
11504 (streamer_tree_cache_append): Adjust.
11505 (record_common_node): Likewise.
11506 (streamer_tree_cache_create): Add argument whether to
11507 record hash values together with trees.
11508 (streamer_tree_cache_delete): Adjust.
11509 * tree-streamer.h (struct streamer_tree_cache_d): Add
11511 (streamer_read_integer_cst): Remove.
11512 (streamer_tree_cache_insert): Adjust.
11513 (streamer_tree_cache_append): Likewise.
11514 (streamer_tree_cache_insert_at): Rename to ...
11515 (streamer_tree_cache_replace_tree): ... this and adjust.
11516 (streamer_tree_cache_create): Add argument whether to record hashes.
11517 (streamer_tree_cache_get): Rename to ...
11518 (streamer_tree_cache_get_tree): ... this.
11519 (streamer_tree_cache_get_hash): New function.
11520 * tree.c (cache_integer_cst): New function.
11521 * tree.h (cache_integer_cst): Declare.
11522 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
11523 * lto-symtab.c (lto_varpool_replace_node): Only release
11524 DECL_INITIAL of non-prevailing decls.
11525 * varpool.c (varpool_remove_initializer): Do not release
11526 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
11528 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11530 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
11531 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
11532 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
11533 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
11534 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
11535 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
11536 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
11537 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
11538 instead of TARGET_64BIT.
11539 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
11540 Require ISA_HAS_<D>DIV.
11542 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
11544 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
11545 (mips*-*-linux*): Move default with_llsc setting to where other
11547 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
11549 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
11550 Likewise. Remove default with_tune setting. Move default float
11551 setting to its own block. Handle with_llsc in the same block as above.
11553 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
11555 PR rtl-optimization/57425
11556 PR rtl-optimization/57569
11557 * alias.c (write_dependence_p): Add new parameters mem_mode,
11558 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
11559 Changed all callers.
11560 (canon_anti_dependence): New function.
11561 * cse.c (check_dependence): Use canon_anti_dependence.
11562 * cselib.c (cselib_invalidate_mem): Likewise.
11563 * rtl.h (canon_anti_dependence): Declare.
11565 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11567 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
11568 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
11569 ".set mips3" for 64-bit targets.
11571 2013-06-15 Dehao Chen <dehao@google.com>
11573 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
11574 * gimple-low.c (gimple_check_call_matching_types): Likewise.
11575 (gimple_check_call_args): Likewise.
11576 * value-prof.c (check_ic_target): Likewise.
11577 * ipa-inline.c (early_inliner): Likewise.
11578 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
11579 * cgraph.c (cgraph_create_edge_1): Likewise.
11580 (cgraph_make_edge_direct): Likewise.
11582 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
11585 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
11586 rs6000_output_move_128bit to handle emitting quad memory
11587 operations. Set attribute length to 8 bytes.
11589 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
11591 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
11593 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
11594 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
11595 (aarch64_<su>mlsl<mode>): Likewise.
11597 2013-06-14 Mike Stump <mikestump@comcast.net>
11599 * Makefile.in (TARGET_H): Add insn-codes.h.
11601 2013-06-14 Alan Modra <amodra@gmail.com>
11603 PR middle-end/57134
11604 PR middle-end/57586
11605 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
11606 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
11607 bitfield expansion when EXPAND_MEMORY.
11608 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
11610 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
11612 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
11613 test for clearing quad memory on 32-bit later.
11615 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11617 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
11618 (fold_negate_expr): Likewise.
11619 (fold_real_zero_addition_p): Handle vectors.
11620 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
11622 2013-06-14 Alan Modra <amodra@gmail.com>
11624 * varasm.c (force_const_mem): Revert 2013-06-07 change.
11626 2013-06-13 Jan Hubicka <jh@suse.cz>
11628 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11629 Local comdats are not externally visible.
11630 * symtab.c (dump_symtab_base): Dump externally visible.
11631 (verify_symtab_base): Verify back links in the symtab hash.
11633 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11635 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
11636 CONVERT_EXPR as equal nodes.
11638 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11640 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
11642 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11644 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
11645 Generalize to complex and vector.
11646 * tree.c (build_all_ones_cst): New function.
11647 * tree.h (build_all_ones_cst): Declare it.
11649 2013-06-13 Alan Modra <amodra@gmail.com>
11651 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
11652 * config/rs6000/rs6000.md (signbittf2): New insn.
11653 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
11654 (abstf2_internal, cmptf_internal2): Likewise.
11655 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
11657 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
11658 Pat Haugen <pthaugen@us.ibm.com>
11659 Peter Bergner <bergner@vnet.ibm.com>
11661 * config/rs6000/rs6000.c (emit_load_locked): Add support for
11662 power8 byte, half-word, and quad-word atomic instructions.
11663 (emit_store_conditional): Likewise.
11664 (rs6000_expand_atomic_compare_and_swap): Likewise.
11665 (rs6000_expand_atomic_op): Likewise.
11667 * config/rs6000/sync.md (larx): Add new modes for power8.
11669 (AINT): New mode iterator to include TImode as well as normal
11670 integer modes on power8.
11671 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
11672 that VSX registers are not considered. Use AINT mode iterator
11673 instead of INT1 to allow inclusion of quad word atomic operations
11675 (load_locked<mode>): Likewise.
11676 (store_conditional<mode>): Likewise.
11677 (atomic_compare_and_swap<mode>): Likewise.
11678 (atomic_exchange<mode>): Likewise.
11679 (atomic_nand<mode>): Likewise.
11680 (atomic_fetch_<fetchop_name><mode>): Likewise.
11681 (atomic_nand_fetch<mode>): Likewise.
11682 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
11684 (ATOMIC): On power8, add QImode, HImode modes.
11685 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
11686 modes that promote to SImode.
11687 (load_lockedti): Convert TImode arguments to PTImode, so that we
11688 get a guaranteed even/odd register pair.
11689 (load_lockedpti): Likewise.
11690 (store_conditionalti): Likewise.
11691 (store_conditionalpti): Likewise.
11693 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
11694 atomic load/store instructions.
11697 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
11699 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
11701 (insn_count): New attribute, with most cases extracted from...
11702 (length): ...here. Redefine most cases in terms of insn_count.
11703 (single_insn): Delete.
11704 (can_delay): Use insn_count to check for single instructions.
11705 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
11706 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
11707 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
11708 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
11709 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
11710 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
11711 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
11712 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
11713 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
11714 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
11715 rather than "length".
11716 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
11717 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
11718 Use "insn_count" rather than "length".
11719 * config/mips/mips-dsp.md
11720 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
11721 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
11724 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
11726 PR tree-optimization/57361
11727 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
11729 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
11731 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
11733 (aarch64_simd_combine<mode>): New instruction expansion.
11734 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
11735 function prototype.
11736 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
11737 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
11739 2013-06-12 Jan Hubicka <jh@suse.cz>
11741 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
11742 decl has when in streaming stage.
11743 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
11744 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
11746 2013-06-12 Roland Stigge <stigge@antcom.de>
11749 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
11751 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11753 PR tree-optimization/57537
11754 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
11755 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
11757 2013-06-12 Richard Biener <rguenther@suse.de>
11759 * data-streamer.h (streamer_write_char_stream): CSE
11760 obs->current_pointer.
11761 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
11762 streamer_write_char_stream manually and optimize the resulting loop.
11763 (streamer_write_hwi_stream): Likewise.
11765 2013-06-12 Jan Hubicka <jh@suse.cz>
11767 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
11768 * cgraph.h (varpool_create_empty_node): Declare.
11769 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
11771 * symtab.c (symtab_unregister_node): Be lax about missin entries
11773 (symtab_get_node): Update comment.
11774 * varpool.c (varpool_create_empty_node): Break out from ...
11775 (varpool_node_for_decl): ... here.
11776 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
11778 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
11780 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
11781 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
11782 part. Use straight-line flow at the end.
11783 <COMPONENT_REF>: Remove superfluous else.
11784 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
11786 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11789 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
11790 target hook even for !TREE_PUBLIC decls. If no resolution info
11791 is available, return false for common and external decls.
11793 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11795 * config/rl78/constraints.md (U): New constraint.
11796 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
11799 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
11802 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
11803 to allow returning address to AT_PLATFORM name.
11805 2013-06-11 Jan Hubicka <jh@suse.cz>
11807 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
11808 * cgraph.h (symtab_node_base): Add weakref flag.
11809 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
11810 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
11811 (output_weakrefs): Use weakref flag.
11812 * fold-const.c (simple_operand_p): Handle WEAK.
11813 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
11814 * ipa.c (varpool_externally_visible_p): Drop weakref.
11815 (function_and_variable_visibility): Update comment; fix weakref
11816 sanity checks; do not clear DECL_WEAK on them.
11817 * lto-cgraph.c (lto_output_node): update.
11818 (lto_output_varpool_node): Update.
11819 (input_overwrite_node): Update.
11820 (input_node): Update.
11821 (input_varpool_node): Update.
11822 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
11823 (lto_symtab_merge_symbols): Add sanity check.
11824 (lto_symtab_prevailing_decl): Do not special case weakrefs.
11825 * passes.c (rest_of_decl_compilation): Set static flag, too.
11826 * symtab.c (dump_symtab_base): Dump weakref.
11827 (verify_symtab_base): Sanity check weakrefs.
11828 (symtab_make_decl_local): Remove duplicated code.
11829 (symtab_alias_ultimate_target): Simplify.
11830 * varpool.c (varpool_create_variable_alias): Set weakref flag.
11832 2013-06-11 Tom de Vries <tom@codesourcery.com>
11834 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
11835 sequence_vect == NULL.
11837 2013-06-11 DJ Delorie <dj@redhat.com>
11839 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
11840 (rl78_unwind_word_mode): New.
11842 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11844 * final.c (debug_prefix_maps): Make static.
11846 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11848 * function.c (initial_trampoline): Remove stray copy.
11850 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
11852 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
11854 2013-06-11 Martin Jambor <mjambor@suse.cz>
11856 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
11857 within bounds at the beginning of the function.
11859 2013-06-11 Alan Modra <amodra@gmail.com>
11861 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
11863 (get_named_section): Don't NULL !DECL_P decl.
11865 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
11867 * doc/invoke.texi (core-avx2): Document.
11869 (atom): Updated with MOVBE.
11871 2013-06-11 Richard Biener <rguenther@suse.de>
11873 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
11875 2013-06-11 Anton Blanchard <anton@samba.org>
11877 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
11878 correct shift value in little-endian mode.
11880 2013-06-11 Jakub Jelinek <jakub@redhat.com>
11883 * varasm.c (get_variable_align): Move #endif to the right place.
11885 2013-06-10 Cary Coutant <ccoutant@google.com>
11887 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
11888 for hash so that hash table traversal order is deterministic.
11890 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
11891 Pat Haugen <pthaugen@us.ibm.com>
11892 Peter Bergner <bergner@vnet.ibm.com>
11894 * config/rs6000/vector.md (GPR move splitter): Do not split moves
11895 of vectors in GPRS if they are direct moves or quad word load or
11898 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
11900 (direct_move_p): Likewise.
11901 (quad_load_store_p): Likewise.
11903 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
11904 classes into bins based on the physical register type.
11905 (reg_class_to_reg_type): Likewise.
11906 (IS_STD_REG_TYPE): Likewise.
11907 (IS_FP_VECT_REG_TYPE): Likewise.
11908 (reload_fpr_gpr): Arrays to determine what insn to use if we can
11909 use direct move instructions.
11910 (reload_gpr_vsx): Likewise.
11911 (reload_vsx_gpr): Likewise.
11912 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
11913 information that is a simplification of register classes. Also
11914 precalculate direct move reload helpers.
11915 (direct_move_p): New function to return true if the operation can
11916 be done as a direct move instruciton.
11917 (quad_load_store_p): New function to return true if the operation
11918 is a quad memory operation.
11919 (rs6000_legitimize_address): If quad memory, only allow register
11920 indirect for TImode addresses.
11921 (rs6000_legitimate_address_p): Likewise.
11922 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
11923 (rs6000_reload_register_type): Likewise.
11924 (register_to_reg_type): Return register type.
11925 (rs6000_secondary_reload_simple_move): New helper function for
11926 secondary reload and secondary memory needed to identify anything
11927 that is a simple move, and does not need reloading.
11928 (rs6000_secondary_reload_direct_move): New helper function for
11929 secondary reload to identify cases that can be done with several
11930 instructions via the direct move instructions.
11931 (rs6000_secondary_reload_move): New helper function for secondary
11932 reload to identify moves between register types that can be done.
11933 (rs6000_secondary_reload): Add support for quad memory operations
11934 and for direct move.
11935 (rs6000_secondary_memory_needed): Likewise.
11936 (rs6000_debug_secondary_memory_needed): Change argument names.
11937 (rs6000_output_move_128bit): New function to return the move to
11938 use for 128-bit moves, including knowing about the various
11939 limitations of quad memory operations.
11941 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
11942 memory operations. call rs6000_output_move_128bit for the actual
11943 instruciton(s) to generate.
11944 (vsx_movti_64bit): Likewise.
11946 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
11947 (UNSPEC_P8V_MTVSRWZ): Likewise.
11948 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
11949 (UNSPEC_P8V_MTVSRD): Likewise.
11950 (UNSPEC_P8V_XXPERMDI): Likewise.
11951 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
11952 (UNSPEC_FUSION_GPR): Likewise.
11953 (FMOVE128_GPR): New iterator for direct move.
11954 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
11955 (f32_sv): Likewise.
11956 (f32_dm): Likewise.
11957 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11958 loads and direct move instructions.
11959 (zero_extendsidi2_lfiwzx): Likewise.
11960 (extendsidi2_lfiwax): Likewise.
11961 (extendsidi2_nocell): Likewise.
11962 (floatsi<mode>2_lfiwax): Likewise.
11963 (lfiwax): Likewise.
11964 (floatunssi<mode>2_lfiwzx): Likewise.
11965 (lfiwzx): Likewise.
11966 (fix_trunc<mode>_stfiwx): Likewise.
11967 (fixuns_trunc<mode>_stfiwx): Likewise.
11968 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11969 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11970 (parity<mode>2_cmpb): Set length/type attr.
11971 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11972 for 'mr.' to fast_compare.
11973 (bpermd_<mode>): Change type attr to popcnt.
11974 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11975 (p8_mtvsrwz_1): Likewise.
11976 (p8_mtvsrwz_2): Likewise.
11977 (reload_fpr_from_gpr<mode>): Likewise.
11978 (p8_mtvsrd_1): Likewise.
11979 (p8_mtvsrd_2): Likewise.
11980 (p8_xxpermdi_<mode>): Likewise.
11981 (reload_vsx_from_gpr<mode>): Likewise.
11982 (reload_vsx_from_gprsf): Likewise.
11983 (p8_mfvsrd_3_<mode>): LIkewise.
11984 (reload_gpr_from_vsx<mode>): Likewise.
11985 (reload_gpr_from_vsxsf): Likewise.
11986 (p8_mfvsrd_4_disf): Likewise.
11987 (multi-word GPR splits): Do not split direct moves or quad memory
11990 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11992 * tree-into-ssa.c (interesting_blocks): Make static.
11994 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11997 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11998 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11999 Use DATA_ABI_ALIGNMENT for that case instead if defined.
12000 (get_variable_align): New function.
12001 (get_variable_section, emit_bss, emit_common,
12002 assemble_variable_contents, place_block_symbol): Use
12003 get_variable_align instead of DECL_ALIGN.
12004 (assemble_noswitch_variable): Add align argument, use it
12005 instead of DECL_ALIGN.
12006 (assemble_variable): Adjust caller. Use get_variable_align
12007 instead of DECL_ALIGN.
12008 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
12010 (DATA_ABI_ALIGNMENT): Define.
12011 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
12012 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
12013 opt is false, only return the psABI mandated alignment increase.
12014 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
12015 (DATA_ABI_ALIGNMENT): ... this.
12016 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
12017 (DATA_ABI_ALIGNMENT): ... this.
12018 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
12019 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
12020 (DATA_ABI_ALIGNMENT): ... this.
12021 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
12022 * doc/tm.texi: Regenerated.
12024 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
12026 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
12027 cmp_code to construct REG_EQUAL note.
12029 2013-06-09 Jakub Jelinek <jakub@redhat.com>
12032 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
12033 that operands[2] doesn't overlap with operands[0].
12035 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
12036 Jan Hubicka <jh@suse.cz>
12038 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
12039 hack to mark symbols as used.
12041 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
12043 PR rtl-optimization/57559
12044 * lra-constraints.c (process_alt_operands): Don't discourage
12045 memory with known offset for offsetable memory constraint.
12046 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
12048 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
12050 * varasm.c (struct oc_local_state): Reorder fields.
12051 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
12052 and adjust accordingly.
12053 (output_constructor): Reorder initialization code and adjust call to
12054 output_constructor_bitfield.
12056 2013-06-07 Jan Hubicka <jh@suse.cz>
12058 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
12060 2013-06-07 David Malcolm <dmalcolm@redhat.com>
12062 * tree-object-size.c (unknown): Make const.
12064 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12066 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
12067 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
12068 for last alternative in the cpu_facility attribute.
12070 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12073 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
12074 (xordi3): Change operand 2 constraint to arm_xordi_operand.
12075 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
12076 * config/arm/constraints.md (Dg): New constraint.
12077 * config/arm/neon.md (xordi3_neon): Remove.
12078 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
12079 * config/arm/predicates.md (arm_xordi_operand): New predicate.
12081 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12083 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
12084 Clean up alternatives.
12086 2013-06-07 Alan Modra <amodra@gmail.com>
12088 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
12091 2013-06-07 Alan Modra <amodra@gmail.com>
12093 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
12095 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12097 * config/arm/constraints.md (Df): New constraint.
12098 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
12099 Correct length attribute for last two alternatives.
12101 2013-06-07 Alan Modra <amodra@gmail.com>
12103 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12104 override user -mfp-in-toc.
12105 (offsettable_ok_by_alignment): Consider just the current access
12106 rather than the whole object, unless BLKmode. Handle
12107 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
12108 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
12109 for -mcmodel=medium.
12110 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
12111 override user -mfp-in-toc or -msum-in-toc. Default to
12112 -mno-fp-in-toc for -mcmodel=medium.
12114 2013-06-06 DJ Delorie <dj@redhat.com>
12116 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
12117 TARGET_VALID_POINTER_MODE.
12119 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12120 Pat Haugen <pthaugen@us.ibm.com>
12121 Peter Bergner <bergner@vnet.ibm.com>
12123 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
12124 Document new power8 builtins.
12126 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
12127 condition code register, to allow 128-bit logical operations to be
12128 done in the VSX or GPR registers.
12129 (nor<mode>3): Use the canonical form for nor.
12130 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
12131 vclz*, and vpopcnt* vector instructions.
12132 (nand<mode>3): Likewise.
12133 (orc<mode>3): Likewise.
12134 (clz<mode>2): LIkewise.
12135 (popcount<mode>2): Likewise.
12137 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
12138 that only the GPRs are recognized.
12140 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12141 support for new power8 builtins.
12143 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
12145 (xscvdpspn): Likewise.
12151 (vpopcnt): Likewise.
12152 (vpopcntb): Likewise.
12153 (vpopcnth): Likewise.
12154 (vpopcntw): Likewise.
12155 (vpopcntd): Likewise.
12157 (vmrgew): Likewise.
12158 (vmrgow): Likewise.
12160 (eqv_v16qi3): Likewise.
12161 (eqv_v8hi3): Likewise.
12162 (eqv_v4si3): Likewise.
12163 (eqv_v2di3): Likewise.
12164 (eqv_v4sf3): Likewise.
12165 (eqv_v2df3): Likewise.
12167 (nand_v16qi3): Likewise.
12168 (nand_v8hi3): Likewise.
12169 (nand_v4si3): Likewise.
12170 (nand_v2di3): Likewise.
12171 (nand_v4sf3): Likewise.
12172 (nand_v2df3): Likewise.
12174 (orc_v16qi3): Likewise.
12175 (orc_v8hi3): Likewise.
12176 (orc_v4si3): Likewise.
12177 (orc_v2di3): Likewise.
12178 (orc_v4sf3): Likewise.
12179 (orc_v2df3): Likewise.
12181 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
12182 allow power8 quad mode in 64-bit.
12183 (rs6000_builtin_vectorized_function): Add support to vectorize
12184 ISA 2.07 count leading zeros, population count builtins.
12185 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
12186 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
12187 (builtin_function_type): Add vgbbd builtin function which takes an
12189 (altivec_expand_vec_perm_const): Add support for new power8 merge
12192 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
12193 that does not include TImdoe for use with 32-bit.
12194 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
12196 (UNSPEC_VSX_CVDPSPN): Likewise.
12197 (vsx_xscvdpspn): Likewise.
12198 (vsx_xscvspdpn): Likewise.
12199 (vsx_xscvdpspn_scalar): Likewise.
12200 (vsx_xscvspdpn_directmove): Likewise.
12201 (vsx_and<mode>3): Split logical operations into 32-bit and
12202 64-bit. Add support to do logical operations on TImode as well as
12203 VSX vector types. Allow logical operations to be done in either
12204 VSX registers or in general purpose registers in 64-bit mode. Add
12205 splitters if GPRs were used. For AND, add clobber of CCmode to
12206 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
12208 (vsx_and<mode>3_32bit): Likewise.
12209 (vsx_and<mode>3_64bit): Likewise.
12210 (vsx_ior<mode>3): Likewise.
12211 (vsx_ior<mode>3_32bit): Likewise.
12212 (vsx_ior<mode>3_64bit): Likewise.
12213 (vsx_xor<mode>3): Likewise.
12214 (vsx_xor<mode>3_32bit): Likewise.
12215 (vsx_xor<mode>3_64bit): Likewise.
12216 (vsx_one_cmpl<mode>2): Likewise.
12217 (vsx_one_cmpl<mode>2_32bit): Likewise.
12218 (vsx_one_cmpl<mode>2_64bit): Likewise.
12219 (vsx_nor<mode>3): Likewise.
12220 (vsx_nor<mode>3_32bit): Likewise.
12221 (vsx_nor<mode>3_64bit): Likewise.
12222 (vsx_andc<mode>3): Likewise.
12223 (vsx_andc<mode>3_32bit): Likewise.
12224 (vsx_andc<mode>3_64bit): Likewise.
12225 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
12226 and xxlorc instructions.
12227 (vsx_eqv<mode>3_64bit): Likewise.
12228 (vsx_nand<mode>3_32bit): Likewise.
12229 (vsx_nand<mode>3_64bit): Likewise.
12230 (vsx_orc<mode>3_32bit): Likewise.
12231 (vsx_orc<mode>3_64bit): Likewise.
12233 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
12235 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
12237 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
12238 (p8_vmrgow): Likewise.
12239 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
12240 GPRs to be split under VSX.
12241 (p8v_clz<mode>2): Add power8 count leading zero support.
12242 (p8v_popcount<mode>2): Add power8 population count support.
12243 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
12246 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
12249 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
12251 (vec_nand): Likewise.
12252 (vec_vclz): Likewise.
12253 (vec_vclzb): Likewise.
12254 (vec_vclzd): Likewise.
12255 (vec_vclzh): Likewise.
12256 (vec_vclzw): Likewise.
12257 (vec_vgbbd): Likewise.
12258 (vec_vmrgew): Likewise.
12259 (vec_vmrgow): Likewise.
12260 (vec_vpopcnt): Likewise.
12261 (vec_vpopcntb): Likewise.
12262 (vec_vpopcntd): Likewise.
12263 (vec_vpopcnth): Likewise.
12264 (vec_vpopcntw): Likewise.
12266 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12268 PR rtl-optimization/57468
12269 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
12272 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12274 PR rtl-optimization/57459
12275 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
12276 type when setting live regs.
12278 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12280 * config/s390/s390.opt (mlra): New option.
12281 * config/s390/s390.c (s390_decompose_address): Check displacement
12282 for all registers for LRA.
12283 (s390_secondary_reload): Don't used secondary reloads for LRA.
12284 (s390_lra_p): New function.
12285 (TARGET_LRA_P): Define.
12286 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
12287 of attribute cpu_facility to zarch for the last alternative.
12288 (*cmpmem_short): Ditto.
12290 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
12292 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
12293 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
12294 (arm_expand_prologue): Likewise.
12296 2013-06-06 Teresa Johnson <tejohnson@google.com>
12299 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
12300 as this is now done by redirect_edge_and_branch_force.
12301 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
12302 barriers, and fix interaction with splitting.
12303 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
12304 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
12305 reflect changes made in the routine.
12306 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
12307 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
12308 since this is called in cfglayout mode, and replace partition fixup
12309 with assert as that is now done by force_nonfallthru_and_redirect.
12310 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
12311 already be marked with region crossing note.
12312 (insert_section_boundary_note): Make non-static, gate on flag
12313 has_bb_partition, rewrite to also check for multiple partitions.
12314 (rest_of_handle_reorder_blocks): Remove call to
12315 insert_section_boundary_note, now done later during free_cfg.
12316 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
12317 * bb-reorder.h (insert_section_boundary_note): Declare.
12318 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
12319 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
12320 invoke insert_section_boundary_note.
12321 (try_redirect_by_replacing_jump): Remove unnecessary
12322 check for region crossing note.
12323 (fixup_partition_crossing): New function.
12324 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
12325 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
12326 in non-cfglayout mode.
12327 (force_nonfallthru_and_redirect): Fixup partition boundaries,
12328 remove old code that tried to do this. Emit barrier correctly
12329 when we are in cfglayout mode.
12330 (last_bb_in_partition): New function.
12331 (rtl_split_edge): Correctly fixup partition boundaries.
12332 (commit_one_edge_insertion): Remove old code that tried to
12333 fixup region crossing edge since this is now handled in
12334 split_block, and set up insertion point correctly since
12335 block may now end in a jump.
12336 (verify_hot_cold_block_grouping): Guard against checking when not in
12337 linearized RTL mode.
12338 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
12340 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
12341 rtl_verify_flow_info, so not called in cfglayout mode.
12342 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
12343 (fixup_reorder_chain): Remove old code that attempted to fixup region
12344 crossing note as this is now handled in force_nonfallthru_and_redirect.
12345 (duplicate_insn_chain): Don't duplicate switch section notes.
12346 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
12348 * basic-block.h (emit_barrier_after_bb): Declare.
12350 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12352 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
12353 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
12354 arm_usatsihi): Adjust alternatives for arm_restrict_it.
12356 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12358 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
12360 * config/arm/ldmstm.md: Regenerate.
12362 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12364 * config/arm/sync.md (atomic_loaddi_1):
12365 Disable predication for arm_restrict_it.
12366 (arm_load_exclusive<mode>): Likewise.
12367 (arm_load_exclusivesi): Likewise.
12368 (arm_load_exclusivedi): Likewise.
12369 (arm_load_acquire_exclusive<mode>): Likewise.
12370 (arm_load_acquire_exclusivesi): Likewise.
12371 (arm_load_acquire_exclusivedi): Likewise.
12372 (arm_store_exclusive<mode>): Likewise.
12373 (arm_store_exclusive<mode>): Likewise.
12374 (arm_store_release_exclusivedi): Likewise.
12375 (arm_store_release_exclusive<mode>): Likewise.
12377 2013-06-06 Richard Biener <rguenther@suse.de>
12379 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
12381 (lto_tag_is_tree_code_p): Adjust.
12382 (lto_tag_is_gimple_code_p): Likewise.
12383 (lto_gimple_code_to_tag): Likewise.
12384 (lto_tag_to_gimple_code): Likewise.
12385 (lto_tree_code_to_tag): Likewise.
12386 (lto_tag_to_tree_code): Likewise.
12387 * data-streamer.h (streamer_write_hwi_in_range): Use
12388 uhwi streaming to stream the normalized range.
12389 (streamer_read_hwi_in_range): Likewise.
12391 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12393 * config/arm/arm.md (enabled_for_depr_it): New attribute.
12394 (predicable_short_it): Likewise.
12395 (predicated): Likewise.
12396 (enabled): Handle above.
12397 (define_cond_exec): Set predicated attribute to yes.
12399 2013-06-05 Mike Stump <mikestump@comcast.net>
12401 * gdbinit.in (__FUNCTION__): Add.
12403 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
12405 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
12406 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
12408 2013-06-05 Jan Hubicka <jh@suse.cz>
12410 * varasm.c (mark_decl_referenced): Revert the removal until targets
12413 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
12415 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
12416 instead of mark_decl_referenced.
12418 2013-06-05 Jan Hubicka <jh@suse.cz>
12420 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
12421 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
12422 and symtab_used_from_object_file_p.
12423 (cgraph_make_node_local_1): Clear forced_by_abi.
12424 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
12425 * cgraph.h (symtab_node_base): Add forced_by_abi.
12426 (decide_is_variable_needed): Remove.
12427 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
12428 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
12429 (decide_is_symbol_needed): ... this one; handle symbols in general;
12430 always analyze virtuals; honnor forced_by_abi.
12431 (cgraph_finalize_function): Update.
12432 (varpool_finalize_decl): Update.
12433 (symbol_defined_and_needed): Remove.
12434 (analyze_functions): Update.
12435 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12436 output_refs, input_overwrite_node): Handle forced_by_abi.
12437 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
12438 (address_taken_from_non_vtable_p): ... this one.
12439 (comdat_can_be_unshared_p_1): New function.
12440 (cgraph_comdat_can_be_unshared_p): Rename to ...
12441 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
12442 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
12443 (function_and_variable_visibility): Clear forced_by_abi as needed.
12444 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
12445 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
12446 * symtab.c (dump_symtab_base): Dump forced_by_abi.
12447 * varpool.c (decide_is_variable_needed): Remove.
12449 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12451 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
12452 (arm_option_override): Override arm_restrict_it where appropriate.
12453 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
12454 * config/arm/arm.opt (mrestrict-it): New command-line option.
12455 * doc/invoke.texi: Document -mrestrict-it.
12457 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12459 * tsan.c (tsan_atomic_table): Make const.
12461 2013-06-05 Richard Biener <rguenther@suse.de>
12463 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
12464 index associated with the tree we are supposed to replace.
12465 * tree-streamer-out.c (pack_ts_base_value_fields): Output
12466 TREE_ASM_WRITTEN as zero for everything but SSA names.
12468 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12470 * tree-ssa-structalias.c (call_stmt_vars): Make static.
12472 2013-06-04 Jan Hubicka <jh@suse.cz>
12474 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
12475 (input_node, input_varpool_node): Handle correctly external same
12477 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
12478 nodes at ltrans stage.
12480 2013-06-04 Jan Hubicka <jh@suse.cz>
12482 * ipa-inline.c (update_caller_keys): Fix availability test.
12483 (update_callee_keys): Likewise.
12484 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
12485 to follow ELF standard.
12487 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
12489 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
12490 (mips64r5900el-*-elf*): New configurations.
12491 * config/mips/mips-cpus.def (r5900): New processor.
12492 * config/mips/mips-tables.opt: Regenerate.
12493 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
12494 (mips_issue_rate): Handle PROCESSOR_R5900.
12495 (mips_reorg_process_insns): Force reorder mode for the R5900.
12496 * config/mips/mips.h (TARGET_MIPS5900): Define.
12497 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
12499 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
12501 * config/mips/mips.md (processor): Add r5900.
12502 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
12504 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12506 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
12507 into function to generate MOVI instruction.
12508 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
12509 (aarch64_preferred_simd_mode): Turn into wrapper.
12510 (aarch64_output_scalar_simd_mov_immediate): New function.
12511 * config/aarch64/aarch64-protos.h: Add prototype for above.
12513 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12515 * config/aarch64/aarch64.c (simd_immediate_info): Remove
12516 element_char member.
12517 (sizetochar): Return signed char.
12518 (aarch64_simd_valid_immediate): Remove elchar and other
12519 unnecessary variables.
12520 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
12521 Calculate element_char as required.
12522 * config/aarch64/aarch64-protos.h: Update and move prototype
12523 for aarch64_output_simd_mov_immediate.
12524 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
12527 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12529 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
12530 information completed by aarch64_simd_valid_immediate.
12531 (aarch64_legitimate_constant_p): Update arguments.
12532 (aarch64_simd_valid_immediate): Work with struct rather than many
12534 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
12535 (aarch64_simd_make_constant): Update arguments.
12536 (aarch64_output_simd_mov_immediate): Work with struct rather than
12537 many pointers. Output immediate directly rather than as operand.
12538 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12540 * config/aarch64/constraints.md (Dn): Update arguments.
12542 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12544 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
12546 (aarch64_simd_immediate_valid_for_move): Remove.
12547 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
12548 (aarch64_simd_make_constant): Update call.
12549 (aarch64_output_simd_mov_immediate): Update call.
12550 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12552 * config/aarch64/constraints.md (Dn): Update call.
12554 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12556 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
12557 return type to bool for prototype.
12558 (aarch64_legitimate_constant_p): Check for true instead of not -1.
12559 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
12560 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
12562 2013-06-04 Catherine Moore <clm@codesourcery.com>
12564 * config/mips/mips.opt (meva): New.
12565 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
12566 (ASM_SPEC): Handle -meva.
12567 * doc/invoke.texi (meva): Document.
12569 2013-06-04 Alan Modra <amodra@gmail.com>
12571 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
12574 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12576 * rtl.def: Add extra fourth optional field to define_cond_exec.
12577 * gensupport.c (process_one_cond_exec): Process attributes from
12579 * doc/md.texi: Document fourth field in define_cond_exec.
12581 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
12583 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
12584 out the processing order as in store_bit_field_1.
12586 2013-06-04 Jan Hubicka <jh@suse.cz>
12588 PR middle-end/57500
12589 * cgraphunit.c (cgraph_process_same_body_aliases): Create
12590 non-VAR_DECL node if it does not exist yet.
12592 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
12594 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
12595 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
12596 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
12597 target_cpu_default setting.
12599 2013-06-03 Teresa Johnson <tejohnson@google.com>
12601 * dumpfile.c (opt_info_switch_p): Change -fopt-info
12602 default to -fopt-info=optimized instead of all.
12603 * doc/invoke.texi: Ditto.
12604 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
12605 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
12606 (execute_vect_slp): Emit BB vectorization success under
12607 MSG_OPTIMIZED_LOCATIONS.
12608 * tree-vect-slp.c (vect_slp_transform_bb): Change
12609 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
12610 * tree-vect-loop.c (vect_transform_loop): Ditto.
12612 2013-06-03 Jason Merrill <jason@redhat.com>
12615 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12616 Use TARGET_EXPR for C++.
12618 2013-06-03 Jakub Jelinek <jakub@redhat.com>
12620 PR rtl-optimization/57268
12621 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
12622 if DEBUG_INSN_P (insn).
12625 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12627 PR rtl-optimization/57268
12628 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12629 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12631 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12633 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
12634 (ix86_avoid_lea_for_addr): Likewise.
12635 (exact_dependency_1): Likewise.
12636 (ix86_adjust_cost): Likewise.
12637 (swap_top_of_ready_list): Fix formatting and !reload_completed check
12639 (do_reorder_for_imul): Fix typo, formatting and
12640 !reload_completed check removed.
12641 (ix86_sched_reorder): Fix typo and formatting.
12642 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
12645 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
12647 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
12649 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
12651 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
12652 <STRING_CST>: Likewise.
12653 <VECTOR_CST>: Likewise.
12655 2013-06-01 Janus Weil <janus@gcc.gnu.org>
12656 Mikael Morin <mikael@gcc.gnu.org>
12658 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
12659 * config.in: Regenerated.
12660 * configure: Regenerated.
12662 2013-06-01 Jan Hubicka <jh@suse.cz>
12664 PR middle-end/57366
12665 * cgraphunit.c (compile): When weakref is not supported,
12666 set up transparent aliases before final output pass.
12667 * varasm.c (assemble_alias): Do not try to do it here.
12669 2013-06-01 Jan Hubicka <jh@suse.cz>
12671 PR middle-end/57467
12672 * passes.c (for_per_function): Skip unanalyzed functions.
12674 2013-06-01 Jan Hubicka <jh@suse.cz>
12676 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
12677 (lto_symtab_merge_symbols_1): ... this one.
12678 (lto_symtab_merge_cgraph_nodes): Rename to ...
12679 (lto_symtab_merge_symbols): ... this one; simplify.
12680 * cgraph.c (same_body_aliases_done): Rename to ...
12681 (cpp_implicit_aliases_done): ... this one.
12682 (cgraph_create_function_alias): Update.
12683 (cgraph_same_body_alias): Update.
12684 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
12685 (verify_edge_corresponds_to_fndecl): Simplify.
12686 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
12687 (cgraph_node): Remove same_body_alias.
12688 (varpool_node): Remove alias_of and extra_name_alias.
12689 (same_body_aliases_done): Rename to ..
12690 (cpp_implicit_aliases_done): ... this one.
12691 (symtab_alias_ultimate_target): Add default parameter.
12692 (symtab_resolve_alias): New function.
12693 (fixup_same_cpp_alias_visibility): Declare.
12694 (cgraph_function_node): Add default parameter.
12695 (cgraph_node_asm_name): Likewise.
12696 (cgraph_function_or_thunk_node): Add default parameter; do
12697 not ICE when it is NULL.
12698 (varpool_variable_node): Likewise.
12699 * tree-emutls.c (create_emultls_var): Update.
12700 (ipa_lower_emutls): Update.
12701 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
12702 (cgraph_reset_node): Reset alias info.
12703 (cgraph_finalize_function): Update.
12704 (fixup_same_cpp_alias_visibility): Move to symtab.c.
12705 (analyze_function): Simplify.
12706 (cgraph_process_same_body_aliases): Simplify.
12707 (analyze_functions): Fixup same body aliases.
12708 (handle_alias_pairs): Simplify.
12709 (assemble_thunk): Update.
12710 (assemble_thunks_and_aliases): Update.
12711 (output_weakrefs): Rewrite.
12712 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
12713 (lto_output_varpool_node): Likewise.
12714 (compute_ltrans_boundary): Remve assert.
12715 (get_alias_symbol): New functoin.
12716 (input_node): Rewrite alias handling.
12717 (input_varpool_node): Likewise.
12718 * ipa-pure-const.c (propagate_pure_const): Fix formating.
12719 * ipa.c (process_references): Handle weakrefs correctly.
12720 (symtab_remove_unreachable_nodes): Likewise.
12721 * trans-mem.c (get_cg_data): Update.
12722 (ipa_tm_create_version_alias): Update.
12723 (ipa_tm_execute): Update.
12724 * symtab.c (dump_symtab_base): Dump aliases.
12725 (verify_symtab_base): Verify aliases.
12726 (symtab_node_availability): New function.
12727 (symtab_alias_ultimate_target): Simplify.
12728 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
12729 handle all the fixup cases.
12730 (symtab_resolve_alias): New function.
12731 * passes.c (ipa_write_summaries): Handle weakrefs.
12732 * varpool.c (varpool_analyze_node): Simplify.
12733 (assemble_aliases): Update.
12734 (varpool_create_variable_alias): Simplify.
12735 (varpool_extra_name_alias): Simplify.
12736 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
12737 (lto_symtab_merge_symbols): ... this one.
12739 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
12742 PR rtl-optimization/57268
12743 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12744 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12746 2013-06-01 Tobias Burnus <burnus@net-b.de>
12748 Partially reverted:
12749 2013-05-31 Tobias Burnus <burnus@net-b.de>
12751 PR middle-end/57073
12752 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
12755 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12757 PR rtl-optimization/57268
12758 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12759 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12761 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12763 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
12764 unordered comparison operators when -fno-trapping-math is in effect
12766 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
12767 and implement unordered comparison operators properly on the e500.
12769 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12771 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
12772 for constant scalar integers.
12773 (simplify_relational_operation_1): Likewise.
12775 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
12777 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
12778 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
12781 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12782 Igor Zamyatin <igor.zamyatin@intel.com>
12784 Silvermont (SLM) architecture performance tuning.
12785 * config/i386/i386.h (enum ix86_tune_indices): Add
12786 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
12787 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
12789 * config/i386/i386.c (initial_ix86_tune_features)
12790 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
12791 (ix86_lea_outperforms): Handle Silvermont tuning.
12792 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
12794 (ix86_use_lea_for_mov): Likewise.
12795 (ix86_avoid_lea_for_addr): Likewise.
12796 (ix86_lea_for_add_ok): Likewise.
12797 (exact_dependency_1): New function.
12798 (exact_store_load_dependency): Likewise.
12799 (ix86_adjust_cost): Handle Silvermont tuning.
12800 (do_reoder_for_imul): Likewise.
12801 (swap_top_of_ready_list): New function.
12802 (ix86_sched_reorder): Changed to handle Silvermont tuning.
12804 * config/i386/i386.md (peepholes that split memory operand in fp
12807 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12809 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
12810 Remove un-necessary braces.
12812 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12814 * config/aarch64/aarch64.c (aarch64_classify_symbol):
12815 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
12817 2013-05-31 Tobias Burnus <burnus@net-b.de>
12819 PR middle-end/57073
12820 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
12822 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12825 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
12826 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
12827 * config/arm/neon.md (iordi3_neon): Remove.
12828 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
12829 * config/arm/predicates.md (imm_for_neon_logic_operand):
12830 Move to earlier in the file.
12831 (neon_logic_op2): Likewise.
12832 (arm_iordi_operand_neon): New predicate.
12834 2013-05-31 Richard Biener <rguenther@suse.de>
12836 PR tree-optimization/57478
12837 PR tree-optimization/57453
12838 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
12841 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12843 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
12844 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
12846 2013-05-30 Tobias Burnus <burnus@net-b.de>
12847 Thomas Koenig <tkoenig@gcc.gnu.org>
12849 PR middle-end/57073
12850 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
12851 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
12853 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
12855 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
12857 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
12859 * target.def (register_usage_leveling_p): New hook.
12860 * targhooks.c (default_register_usage_leveling_p): New.
12861 * targhooks.h (default_register_usage_leveling_p): New prototype.
12862 * lra-assigns.c (register_usage_leveling_p): Use the hook.
12863 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
12864 * doc/tm.texi: Update.
12865 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
12867 2013-05-30 Ian Bolton <ian.bolton@arm.com>
12869 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
12870 (*insv_reg<mode>): New define_insn.
12872 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
12874 PR rtl-optimization/57439
12875 * postreload.c (move2add_valid_value_p): Check that we have
12876 a zero subreg_regno_offset when accessing the register in
12877 the requested mode.
12879 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12880 Igor Zamyatin <igor.zamyatin@intel.com>
12882 Silvermont (SLM) architecture pipeline model, tuning and
12884 * config.gcc: Add slm config options and target.
12886 * config/i386/slm.md: New.
12888 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
12890 * config/i386/i386-c.c (ix86_target_macros_internal): New case
12892 (ix86_target_macros_internal): Likewise.
12894 * config/i386/i386.c (slm_cost): New cost.
12895 (m_SLM): New macro flag.
12896 (initial_ix86_tune_features): Set m_SLM.
12897 (x86_accumulate_outgoing_args): Likewise.
12898 (x86_arch_always_fancy_math_387): Likewise.
12899 (processor_target_table): Add slm cost.
12900 (cpu_names): Add slm cpu name.
12901 (x86_option_override_internal): Set SLM ISA.
12902 (ix86_issue_rate): New case PROCESSOR_SLM.
12903 (ia32_multipass_dfa_lookahead): Likewise.
12904 (fold_builtin_cpu): Add slm.
12906 * config/i386/i386.h (TARGET_SLM): New target macro.
12907 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
12908 (processor_type): Add PROCESSOR_SLM.
12910 * config/i386/i386.md (cpu): Add new value "slm".
12911 (slm.md): Include slm.md.
12913 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
12914 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12916 * config/arm/arm-protos.h: Add and update function protos.
12917 * config/arm/arm.c (use_simple_return_p): New added.
12918 (thumb2_expand_return): Check simple_return flag.
12919 * config/arm/arm.md: Add simple_return and conditional simple_return.
12920 * config/arm/iterators.md: Add iterator for return and simple_return.
12922 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12924 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
12925 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
12926 (arm_emit_vfp_multi_reg_pop): Likewise.
12927 (thumb2_emit_ldrd_pop): Likewise.
12928 (arm_expand_epilogue): Add misc REG_CFA notes.
12929 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
12931 2013-05-29 Lawrence Crowl <crowl@google.com>
12933 * config/arm/t-arm: Update for below.
12935 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
12936 Change type to hash_table. Update dependent calls and types.
12938 * config/i386/t-cygming: Update for below.
12940 * config/i386/t-interix: Update for below.
12942 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
12943 Change type to hash_table. Update dependent calls and types.
12944 (i386_find_on_wrapper_list::wrappers): Likewise.
12946 * config/ia64/t-ia64: Update for below.
12948 * config/ia64/ia64.c (bundle_state_table):
12949 Change type to hash_table. Update dependent calls and types.
12951 * config/mips/mips.c (mips_reorg_process_insns::htab):
12952 Change type to hash_table. Update dependent calls and types.
12954 * config/sol2.c (solaris_comdat_htab):
12955 Change type to hash_table. Update dependent calls and types.
12957 * config/t-sol2: Update for above.
12959 2013-05-29 Teresa Johnson <tejohnson@google.com>
12961 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12962 functions are not yet marked as defined.
12964 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12965 Pat Haugen <pthaugen@us.ibm.com>
12966 Peter Bergner <bergner@vnet.ibm.com>
12968 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12972 (vrotl<mode>3): Likewise.
12973 (vashl<mode>3): Likewise.
12974 (vlshr<mode>3): Likewise.
12975 (vashr<mode>3): Likewise.
12977 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12978 support for power8 V2DI builtins.
12980 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12981 power8 V2DI builtins.
12982 (vupkhsw): Likewise.
12983 (vupklsw): Likewise.
12984 (vaddudm): Likewise.
12985 (vminsd): Likewise.
12986 (vmaxsd): Likewise.
12987 (vminud): Likewise.
12988 (vmaxud): Likewise.
12989 (vpkudum): Likewise.
12990 (vpksdss): Likewise.
12991 (vpkudus): Likewise.
12992 (vpksdus): Likewise.
12997 (vsubudm): Likewise.
12998 (vcmpequd): Likewise.
12999 (vcmpgtsd): Likewise.
13000 (vcmpgtud): Likewise.
13001 (vcmpequd_p): Likewise.
13002 (vcmpgtsd_p): Likewise.
13003 (vcmpgtud_p): Likewise.
13004 (vupkhsw): Likewise.
13005 (vupklsw): Likewise.
13006 (vaddudm): Likewise.
13007 (vmaxsd): Likewise.
13008 (vmaxud): Likewise.
13009 (vminsd): Likewise.
13010 (vminud): Likewise.
13011 (vpksdss): Likewise.
13012 (vpksdus): Likewise.
13013 (vpkudum): Likewise.
13014 (vpkudus): Likewise.
13019 (vsubudm): Likewise.
13021 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
13022 support for power8 V2DI instructions.
13024 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
13025 power8 V2DI instructions. Combine pack and unpack insns to use an
13026 iterator for each mode. Check whether a particular mode supports
13027 Altivec instructions instead of just checking TARGET_ALTIVEC.
13028 (UNSPEC_VPKUWUM): Likewise.
13029 (UNSPEC_VPKSHSS): Likewise.
13030 (UNSPEC_VPKSWSS): Likewise.
13031 (UNSPEC_VPKUHUS): Likewise.
13032 (UNSPEC_VPKSHUS): Likewise.
13033 (UNSPEC_VPKUWUS): Likewise.
13034 (UNSPEC_VPKSWUS): Likewise.
13035 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
13036 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
13037 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
13038 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
13039 (UNSPEC_VUPKHSB): Likewise.
13040 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
13041 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
13042 (UNSPEC_VUPKHSH): Likewise.
13043 (UNSPEC_VUPKLSB): Likewise.
13044 (UNSPEC_VUPKLSH): Likewise.
13046 (VI_char): Likewise.
13047 (VI_scalar): Likewise.
13048 (VI_unit): Likewise.
13050 (VP_small): Likewise.
13051 (VP_small_lc): Likewise.
13052 (VU_char): Likewise.
13053 (add<mode>3): Likewise.
13054 (altivec_vaddcuw): Likewise.
13055 (altivec_vaddu<VI_char>s): Likewise.
13056 (altivec_vadds<VI_char>s): Likewise.
13057 (sub<mode>3): Likewise.
13058 (altivec_vsubcuw): Likewise.
13059 (altivec_vsubu<VI_char>s): Likewise.
13060 (altivec_vsubs<VI_char>s): Likewise.
13061 (altivec_vavgs<VI_char>): Likewise.
13062 (altivec_vcmpbfp): Likewise.
13063 (altivec_eq<mode>): Likewise.
13064 (altivec_gt<mode>): Likewise.
13065 (altivec_gtu<mode>): Likewise.
13066 (umax<mode>3): Likewise.
13067 (smax<mode>3): Likewise.
13068 (umin<mode>3): Likewise.
13069 (smin<mode>3): Likewise.
13070 (altivec_vpkuhum): Likewise.
13071 (altivec_vpkuwum): Likewise.
13072 (altivec_vpkshss): Likewise.
13073 (altivec_vpkswss): Likewise.
13074 (altivec_vpkuhus): Likewise.
13075 (altivec_vpkshus): Likewise.
13076 (altivec_vpkuwus): Likewise.
13077 (altivec_vpkswus): Likewise.
13078 (altivec_vpks<VI_char>ss): Likewise.
13079 (altivec_vpks<VI_char>us): Likewise.
13080 (altivec_vpku<VI_char>us): Likewise.
13081 (altivec_vpku<VI_char>um): Likewise.
13082 (altivec_vrl<VI_char>): Likewise.
13083 (altivec_vsl<VI_char>): Likewise.
13084 (altivec_vsr<VI_char>): Likewise.
13085 (altivec_vsra<VI_char>): Likewise.
13086 (altivec_vsldoi_<mode>): Likewise.
13087 (altivec_vupkhsb): Likewise.
13088 (altivec_vupkhs<VU_char>): Likewise.
13089 (altivec_vupkls<VU_char>): Likewise.
13090 (altivec_vupkhsh): Likewise.
13091 (altivec_vupklsb): Likewise.
13092 (altivec_vupklsh): Likewise.
13093 (altivec_vcmpequ<VI_char>_p): Likewise.
13094 (altivec_vcmpgts<VI_char>_p): Likewise.
13095 (altivec_vcmpgtu<VI_char>_p): Likewise.
13096 (abs<mode>2): Likewise.
13097 (vec_unpacks_hi_v16qi): Likewise.
13098 (vec_unpacks_hi_v8hi): Likewise.
13099 (vec_unpacks_lo_v16qi): Likewise.
13100 (vec_unpacks_hi_<VP_small_lc>): Likewise.
13101 (vec_unpacks_lo_v8hi): Likewise.
13102 (vec_unpacks_lo_<VP_small_lc>): Likewise.
13103 (vec_pack_trunc_v8h): Likewise.
13104 (vec_pack_trunc_v4si): Likewise.
13105 (vec_pack_trunc_<mode>): Likewise.
13107 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
13109 (vec_vmaxsd): Likewise.
13110 (vec_vmaxud): Likewise.
13111 (vec_vminsd): Likewise.
13112 (vec_vminud): Likewise.
13113 (vec_vpksdss): Likewise.
13114 (vec_vpksdus): Likewise.
13115 (vec_vpkudum): Likewise.
13116 (vec_vpkudus): Likewise.
13117 (vec_vrld): Likewise.
13118 (vec_vsld): Likewise.
13119 (vec_vsrad): Likewise.
13120 (vec_vsrd): Likewise.
13121 (vec_vsubudm): Likewise.
13122 (vec_vupkhsw): Likewise.
13123 (vec_vupklsw): Likewise.
13125 2013-05-29 Jan Hubicka <jh@suse.cz>
13127 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
13128 flags; reorder rest of fields in more consistent way.
13129 (varpool_node): Remove analyzed, finalized and alias.
13130 (cgraph_ndoe): Likewise.
13131 (symtab_alias_ultimate_target): New function.
13132 (cgraph_function_node): Move offline.
13133 (cgraph_reset_node): Declare.
13134 (cgraph_comdat_can_be_unshared_p): Remove.
13135 (varpool_remove_initializer): Declare.
13136 (varpool_first_defined_variable, varpool_next_defined_variable
13137 cgraph_first_defined_function, cgraph_next_defined_function): Update.
13138 (cgraph_function_with_gimple_body_p): Update.
13139 (varpool_all_refs_explicit_p): Update.
13140 (symtab_alias_target): New function.
13141 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
13142 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
13143 (cgraph_function_or_thunk_node): Simplify using
13144 symtab_alias_ultimate_target.
13145 (varpool_variable_node): Likewise.
13146 * cgraph.c (cgraph_create_function_alias): Update.
13147 (cgraph_add_thunk): Update.
13148 (cgraph_remove_node): Update.
13149 (dump_cgraph_node): Do not dump removed flags.
13150 (cgraph_function_body_availability): Update.
13151 (cgraph_propagate_frequency): Update.
13152 (verify_cgraph_node): Check sanity of local flag.
13153 (cgraph_function_node): Move here from cgraph.h; revamp for
13154 cgraph_function_or_thunk_node.
13155 * lto-symtab.c (lto_varpool_replace_node): Update.
13156 (lto_symtab_resolve_can_prevail_p): Update.
13157 (lto_symtab_merge_cgraph_nodes): Update.
13158 * ipa-cp.c (determine_versionability, initialize_node_lattices,
13159 propagate_constants_accross_call, devirtualization_time_bonus,
13160 ipcp_propagate_stage): Update.
13161 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
13162 * ipa-inline-transform.c (clone_inlined_nodes,
13163 preserve_function_body_p): Update.
13164 * ipa-reference.c (propagate): Update.
13165 (write_node_summary_p): Update.
13166 * toplev.c (wrapup_global_declaration_2): Update.
13167 * cgraphunit.c (cgraph_analyze_function): Rename to ...
13168 (analyze_function) ... this one.
13169 (cgraph_process_new_functions): Update.
13170 (cgraph_reset_node): Export.
13171 (cgraph_finalize_function): Update.
13172 (cgraph_add_new_function): Update.
13173 (process_function_and_variable_attributes): Update.
13174 (varpool_finalize_decl): Update.
13175 (symbol_finalized): Remove.
13176 (symbol_finalized_and_needed): Rename to ...
13177 (symbol_defined_and_needed): ... update.
13178 (cgraph_analyze_functions): Update.
13179 (handle_alias_pairs): Update.
13180 (mark_functions_to_output): Update.
13181 (assemble_thunk): Update.
13182 (output_in_order): Update.
13183 (output_weakrefs): Update.
13184 (finalize_compilation_unit): Update.
13185 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
13186 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
13187 input_node, input_varpool_node): Update.
13188 * dbxout.c (dbxout_expand_expr): Update.
13189 * cgraphclones.c (cgraph_clone_node): Update.
13190 (cgraph_copy_node_for_versioning): Update.
13191 (cgraph_materialize_clone): Update.
13192 (cgraph_materialize_all_clones): Update.
13193 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
13194 propagate_pure_const, propagate_nothrow): Update.
13195 * lto-streamer-out.c (lto_output, write_symbol): Update.
13196 * ipa-utils.c (ipa_reverse_postorder): Update.
13197 * ipa-inline.c (can_inline_edge_p): Update.
13198 (update_caller_keys, ipa_inline): Update.
13199 * dwarf2out.c (reference_to_unused,
13200 premark_types_used_by_global_vars_helper): Update.
13201 * tree-eh.c (tree_could_trap_p): Update.
13202 * ipa-split.c (consider_split, execute_split_functions): Update.
13203 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
13204 has_addr_references_p): Update; move ahead in file for better
13206 (process_references): Simplify.
13207 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
13208 bodies are removed.
13209 (cgraph_comdat_can_be_unshared_p): Make static.
13210 (cgraph_externally_visible_p): Update.
13211 (varpool_externally_visible_p): Update.
13212 (function_and_variable_visibility): Update.
13213 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
13214 ipa_tm_mark_force_output_node): Update.
13215 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
13216 estimate_edge_devirt_benefit, inline_generate_summary,
13217 inline_write_summary): Update.
13218 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
13219 * ipa-prop.c (ipa_compute_jump_functions): Update.
13220 (ipa_print_node_params, ipa_prop_read_section,
13221 ipa_update_after_lto_read, read_replacements_section): Update.
13222 * varasm.c (mark_decl_referenced): Update.
13223 (assemble_alias, dump_tm_clone_pairs): Update.
13224 * tree-inline.c (copy_bb): Update.
13225 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
13227 * symtab.c (dump_symtab_base): Print new flags.
13228 (verify_symtab_base): Verify new flags.
13229 (symtab_alias_ultimate_target): New function.
13230 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
13231 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
13233 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
13235 * i386.c (ix86_get_function_versions_dispatcher,
13236 ix86_generate_version_dispatcher_body): Update.
13237 (fold_builtin_cpu): Use varpool_add_new_variable.
13238 * varpool.c (varpool_remove_initializer): Break out from ...
13239 (varpool_remove_node): ... this one.
13240 (dump_varpool_node, varpool_node_for_asm,
13241 cgraph_variable_initializer_availability, varpool_analyze_node,
13242 varpool_assemble_decl, varpool_remove_unreferenced_decls,
13243 varpool_finalize_named_section_flags, varpool_create_variable_alias):
13246 2013-05-29 Jan Hubicka <jh@suse.cz>
13248 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
13250 2013-05-29 Easwaran Raman <eraman@google.com>
13252 PR tree-optimization/57442
13253 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
13254 when control exits the main loop.
13256 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
13258 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
13260 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
13261 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
13262 * rx/t-rx: Add rx100 under multi library matches option for nofpu
13265 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13267 PR tree-optimization/57441
13268 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
13269 Don't limit size of incr_vec to number of candidates.
13271 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
13273 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
13274 and mips16 directories.
13275 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
13276 (MULTILIB_DIRNAMES): Ditto.
13277 (MULTILIB_EXCEPTIONS): Add new exceptions.
13278 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
13279 (MULTILIB_DIRNAMES): Ditto.
13280 (MULTILIB_EXCEPTIONS): Add new exceptions.
13282 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13283 Marcus Shawcroft <marcus.shawcroft@arm.com>
13285 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
13286 SYMBOL_TINY_ABSOLUTE.
13287 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
13288 SYMBOL_TINY_ABSOLUTE.
13289 (aarch64_expand_mov_immediate): Likewise.
13290 (aarch64_classify_symbol): Likewise.
13291 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
13292 Permit SYMBOL_TINY_ABSOLUTE.
13293 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
13295 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13296 Marcus Shawcroft <marcus.shawcroft@arm.com>
13298 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
13299 Refactor if/switch. Replace gcc_assert with if.
13301 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13303 * config/i386/i386.c (initial_ix86_tune_features): Enable
13304 FP Reassociation for AMD bdver1 and bdver2.
13306 2013-05-29 Martin Jambor <mjambor@suse.cz>
13308 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
13309 and IMAGPART_EXPR do not occur within other handled_components.
13311 2013-05-29 Richard Biener <rguenther@suse.de>
13313 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
13314 access on whether the use is in the BB we currently try to
13316 (vect_bb_vectorization_profitable_p): Pass the BB we currently
13317 vectorize to vect_bb_slp_scalar_cost.
13319 2013-05-29 Richard Biener <rguenther@suse.de>
13321 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
13322 computing scalar cost offsetted by stmts that are kept live
13324 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
13325 for computation of scalar cost.
13327 2013-05-28 Steve Ellcey <sellcey@mips.com>
13329 * config/mips/mips-cpus.def (mips32r2): Change processor type.
13331 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13333 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
13334 array notation built-in reduction functions.
13335 * doc/passes.texi (Passes): Added documentation about changes done
13337 * doc/invoke.texi (C Dialect Options): Added documentation about
13338 the -fcilkplus flag.
13339 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
13340 (BUILTINS_DEF): Depend on cilkplus.def.
13341 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
13342 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
13343 * cilkplus.def: New file.
13345 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
13347 PR rtl-optimization/57439
13348 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
13350 2013-05-28 Easwaran Raman <eraman@google.com>
13352 PR tree-optimization/57337
13353 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
13354 (find_insert_point): Correctly identify the insertion point
13355 when two statements with the same UID is compared.
13357 2013-05-28 Richard Biener <rguenther@suse.de>
13359 PR tree-optimization/56787
13360 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
13361 from the list of data references.
13362 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
13364 (vect_analyze_loop_operations): Likewise.
13365 (vect_transform_loop): Remove clobbers.
13367 2013-05-28 Martin Jambor <mjambor@suse.cz>
13369 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
13370 and REALPART_EXPRs have scalar type.
13372 2013-05-28 Richard Biener <rguenther@suse.de>
13374 PR tree-optimization/57411
13375 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
13377 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
13378 virtual operand propagation.
13380 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13382 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
13383 destination register for bmasksi_vis.
13384 (vector_init_bshuffle): Likewise.
13385 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
13387 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13389 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
13390 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
13391 mode if the instruction isn't available in the original mode.
13392 * config/sparc/sparc.opt (mfix-ut699): New option.
13393 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
13394 (divdf3): Turn into expander.
13395 (divdf3_nofix): New insn.
13396 (divdf3_fix): Likewise.
13397 (divsf3): Disable if -mfix-ut699.
13398 (sqrtdf2): Turn into expander.
13399 (sqrtdf2_nofix): New insn.
13400 (sqrtdf2_fix): Likewise.
13401 (sqrtsf2): Disable if -mfix-ut699.
13403 2013-05-27 Richard Biener <rguenther@suse.de>
13405 PR middle-end/57412
13406 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
13407 block for the new loop.
13409 2013-05-27 Richard Biener <rguenther@suse.de>
13411 PR tree-optimization/57343
13412 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
13413 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
13414 (number_of_iterations_cond): Do not build the folded tree.
13416 2013-05-27 Richard Biener <rguenther@suse.de>
13419 PR middle-end/57381
13420 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13421 OEP_CONSTANT_ADDRESS_OF retained.
13423 PR tree-optimization/57417
13424 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
13425 for unchanged base.
13426 (set_ssa_val_to): Compare addresses using
13427 get_addr_base_and_unit_offset.
13429 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
13431 PR rtl-optimization/56833
13432 * postreload.c (move2add_record_mode): New function.
13433 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
13434 (move2add_use_add2_insn): Use move2add_record_sym_value.
13435 (move2add_use_add3_insn): Likewise.
13436 (reload_cse_move2add): Use move2add_valid_value_p and
13437 move2add_record_mode. Invalidate call-clobbered and REG_INC
13438 affected regs by setting reg_mode to VOIDmode.
13439 (move2add_note_store): Don't pretend the inside of a SUBREG is
13440 the actual destination. Invalidate single/leading registers by
13441 setting reg_mode to VOIDmode.
13442 Use move2add_record_sym_value, move2add_valid_value_p and
13443 move2add_record_mode.
13445 2013-05-27 Richard Biener <rguenther@suse.de>
13447 PR tree-optimization/57396
13448 * tree-affine.c (double_int_constant_multiple_p): Properly
13449 return false for val == 0 and div != 0.
13451 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13453 * config/mips/mips.h: Use #elif in preprocessor conditions.
13455 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13458 * config/mips/constraints.md (kl): New constraint.
13459 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
13460 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
13461 constraint for operand 0. Split after CSE for MIPS16. Emit a move
13462 from LO for MIPS16.
13463 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
13465 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13468 * config/mips/mips.c (mips_can_inline_p): New function.
13469 (TARGET_CAN_INLINE_P): Define.
13471 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
13473 * sched-int.h (ds_t, dw_t): Make unsigned int.
13474 Fix documentation that describes how all the ds_t bits are used.
13475 Reserve the last bit for delayed-branch scheduling.
13476 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
13477 (BITS_PER_DEP_WEAK): Fix definition and documentation.
13478 (gen_dep_weak_1): Remove prototype.
13479 * sched-deps.c (get_dep_weak_1): Make static.
13480 * target.def (speculate_insn, needs_block_p, gen_spec_check,
13481 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
13482 * doc/tm.texi: Regenerate.
13483 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
13485 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
13488 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
13490 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
13491 Sandra Loosemore <sandra@codesourcery.com>
13493 * config.gcc (powerpc-*): Allow native for with-cpu.
13495 2013-05-24 Jeff Law <law@redhat.com>
13497 PR tree-optimization/57124
13498 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
13499 conversion feeding a condition if the range has an overflow
13500 if -fstrict-overflow. Add warnings for when we do make the
13503 2013-05-24 Dehao Chen <dehao@google.com>
13505 * tree-cfg.c (locus_discrim_map): Fix the typo.
13506 (locus_discrim_hasher): Likewise.
13507 (locus_discrim_hasher::hash): Likewise.
13508 (locus_discrim_hasher::equal): Likewise.
13510 2013-05-24 Martin Jambor <mjambor@suse.cz>
13512 PR tree-optimization/57294
13513 * cgraph.h (ipa_record_stmt_references): Declare.
13514 * cgraphbuild.c (ipa_record_stmt_references): New function.
13515 (build_cgraph_edges): Use ipa_record_stmt_references.
13516 (rebuild_cgraph_edges): Likewise.
13517 (cgraph_rebuild_references): Likewise.
13518 * ipa-prop.c (ipa_modify_call_arguments): Discard references
13519 associated with the old statement and build references from the
13520 newly built statements.
13521 * ipa-ref.c (ipa_remove_stmt_references): New function.
13522 * ipa-ref.h (ipa_remove_stmt_references): Declare.
13524 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
13526 * lra-constraints.c (emit_spill_move): Use smaller mode for
13528 (check_and_process_move): Consider mem-reg moves for secondary
13530 (curr_insn_transform): Don't lose insns emitted before for
13531 secondary memory moves.
13532 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
13533 reg set up in the current insn.
13535 2013-05-24 Dehao Chen <dehao@google.com>
13537 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
13539 (locus_descrim_hasher::equal): Likewise.
13540 (build_gimple_cfg): New discriminator assignment algorithm.
13541 (make_edges): Likewise.
13542 (next_discriminator_for_locus): Likewise.
13543 (same_line_p): Likewise.
13544 (assign_discriminators): Likewise.
13545 (make_cond_expr_edges): Likewise.
13546 (make_gimple_switch_edges): Likewise.
13547 (make_goto_expr_edges): Likewise.
13548 (make_gimple_asm_edges): Likewise.
13550 2013-05-24 Ian Bolton <ian.bolton@arm.com>
13552 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
13553 X format specifier to only display bottom 16 bits.
13554 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
13555 immediate to match for operand 2, since it will be masked.
13557 2013-05-24 Richard Biener <rguenther@suse.de>
13559 PR tree-optimization/57287
13560 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
13561 all SSA names that occur in abnormal PHIs.
13563 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
13566 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
13567 that index is not negative.
13569 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
13571 PR rtl-optimization/55177
13572 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
13573 (simplify_byte_swapping_operation): New.
13574 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
13575 (simplify_relational_operation_1): Deal with BSWAP.
13577 2013-05-23 Richard Henderson <rth@redhat.com>
13580 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
13581 (ix86_reorg): Call it.
13583 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
13586 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
13587 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
13588 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
13590 2013-05-23 Christian Bruel <christian.bruel@st.com>
13593 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
13595 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13596 Marcus Shawcroft <marcus.shawcroft@arm.com>
13598 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
13599 * config/aarch64/constraints.md (Usa): Remove.
13600 * doc/md.texi (AArch64 Usa): Remove.
13602 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13603 Marcus Shawcroft <marcus.shawcroft@arm.com>
13605 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
13606 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
13607 * config/aarch64/predicates.md (aarch64_const_address): Remove.
13608 (aarch64_mov_operand): Use aarch64_mov_operand_p.
13610 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
13612 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
13613 instruction (AdvSIMD).
13614 * config/aarch64/aarch64-builtins.c
13615 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
13616 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
13618 2013-05-23 Martin Jambor <mjambor@suse.cz>
13620 PR middle-end/57347
13621 * tree.h (contains_bitfld_component_ref_p): Declare.
13622 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
13623 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
13625 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
13626 not access a bit-field. Assert all final offsets are byte-aligned.
13628 2013-05-23 Richard Biener <rguenther@suse.de>
13630 PR tree-optimization/57380
13631 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
13632 least one invariant or re-used load.
13633 * passes.c (init_optimization_passes): Move pass_phiprop before
13636 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
13638 * config/aarch64/aarch64-simd.md
13639 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
13641 2013-05-23 Richard Biener <rguenther@suse.de>
13643 PR middle-end/57381
13644 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13645 OEP_CONSTANT_ADDRESS_OF retained.
13647 2013-05-23 Jakub Jelinek <jakub@redhat.com>
13649 PR middle-end/57344
13650 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
13651 don't lower unit. Handle unit not being always BITS_PER_WORD.
13653 2013-05-23 Richard Biener <rguenther@suse.de>
13655 PR rtl-optimization/57341
13656 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
13657 instead of true_dependence.
13659 2013-05-22 David Malcolm <dmalcolm@redhat.com>
13661 * bb-reorder.c (branch_threshold): Make const.
13662 (exec_threshold): Ditto.
13664 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13665 Pat Haugen <pthaugen@us.ibm.com>
13666 Peter Bergner <bergner@vnet.ibm.com>
13668 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
13669 documentation for the power8 crypto builtins.
13671 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
13673 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
13674 macros for defining power8 builtin functions.
13675 (BU_P8V_AV_2): Likewise.
13676 (BU_P8V_AV_P): Likewise.
13677 (BU_P8V_VSX_1): Likewise.
13678 (BU_P8V_OVERLOAD_1): Likewise.
13679 (BU_P8V_OVERLOAD_2): Likewise.
13680 (BU_CRYPTO_1): Likewise.
13681 (BU_CRYPTO_2): Likewise.
13682 (BU_CRYPTO_3): Likewise.
13683 (BU_CRYPTO_OVERLOAD_1): Likewise.
13684 (BU_CRYPTO_OVERLOAD_2): Likewise.
13685 (XSCVSPDP): Fix typo, point to the correct instruction.
13686 (VCIPHER): Add power8 crypto builtins.
13687 (VCIPHERLAST): Likewise.
13688 (VNCIPHER): Likewise.
13689 (VNCIPHERLAST): Likewise.
13690 (VPMSUMB): Likewise.
13691 (VPMSUMH): Likewise.
13692 (VPMSUMW): Likewise.
13693 (VPERMXOR_V2DI): Likewise.
13694 (VPERMXOR_V4SI: Likewise.
13695 (VPERMXOR_V8HI: Likewise.
13696 (VPERMXOR_V16QI: Likewise.
13697 (VSHASIGMAW): Likewise.
13698 (VSHASIGMAD): Likewise.
13699 (VPMSUM): Likewise.
13700 (VPERMXOR): Likewise.
13701 (VSHASIGMA): Likewise.
13703 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13704 __CRYPTO__ if the crypto instructions are available.
13705 (altivec_overloaded_builtins): Add support for overloaded power8
13708 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
13709 support for power8 crypto builtins.
13710 (builtin_function_type): Likewise.
13711 (altivec_init_builtins): Add support for builtins that take vector
13712 long long (V2DI) arguments.
13714 * config/rs6000/crypto.md: New file, define power8 crypto
13717 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13718 Pat Haugen <pthaugen@us.ibm.com>
13719 Peter Bergner <bergner@vnet.ibm.com>
13721 * doc/invoke.texi (Option Summary): Add power8 options.
13722 (RS/6000 and PowerPC Options): Likewise.
13724 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
13725 constraints.md instead of rs6000.h. Reorder w* constraints. Add
13726 wm, wn, wr documentation.
13728 * config/rs6000/constraints.md (wm): New constraint for VSX
13729 registers if direct move instructions are enabled.
13730 (wn): New constraint for no registers.
13731 (wq): New constraint for quad word even GPR registers.
13732 (wr): New constraint if 64-bit instructions are enabled.
13733 (wv): New constraint if power8 vector instructions are enabled.
13734 (wQ): New constraint for quad word memory locations.
13736 * config/rs6000/predicates.md (const_0_to_15_operand): New
13737 constraint for 0..15 for crypto instructions.
13738 (gpc_reg_operand): If VSX allow registers in VSX registers as well
13739 as GPR and floating point registers.
13740 (int_reg_operand): New predicate to match only GPR registers.
13741 (base_reg_operand): New predicate to match base registers.
13742 (quad_int_reg_operand): New predicate to match even GPR registers
13743 for quad memory operations.
13744 (vsx_reg_or_cint_operand): New predicate to allow vector logical
13745 operations in both GPR and VSX registers.
13746 (quad_memory_operand): New predicate for quad memory operations.
13747 (reg_or_indexed_operand): New predicate for direct move support.
13749 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
13750 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
13751 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
13752 (POWERPC_MASKS): Add power8 options.
13753 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
13756 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
13757 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
13759 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
13760 (-mpower8-fusion): New power8 options.
13761 (-mpower8-fusion-sign): Likewise.
13762 (-mpower8-vector): Likewise.
13763 (-mcrypto): Likewise.
13764 (-mdirect-move): Likewise.
13765 (-mquad-memory): Likewise.
13767 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
13768 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
13770 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
13771 (rs6000_debug_vector_unit): Add p8_vector.
13772 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
13773 definitions. Also print fusion state.
13774 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
13775 (rs6000_builtin_mask_calculate): Add power8 builtin support.
13776 (rs6000_option_override_internal): Add support for power8.
13777 (rs6000_common_init_builtins): Add debugging for skipped builtins
13778 if -mdebug=builtin.
13779 (rs6000_adjust_cost): Add power8 support.
13780 (rs6000_issue_rate): Likewise.
13781 (insn_must_be_first_in_group): Likewise.
13782 (insn_must_be_last_in_group): Likewise.
13783 (force_new_group): Likewise.
13784 (rs6000_register_move_cost): Likewise.
13785 (rs6000_opt_masks): Likewise.
13787 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
13788 power8 capable assembler, default to power7 options.
13789 (TARGET_DIRECT_MOVE): Likewise.
13790 (TARGET_CRYPTO): Likewise.
13791 (TARGET_P8_VECTOR): Likewise.
13792 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
13793 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
13794 (VECTOR_MEM_P8_VECTOR_P): Likewise.
13795 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
13796 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
13797 (TARGET_XSCVDPSPN): Likewise.
13798 (TARGET_XSCVSPDPN): Likewsie.
13799 (TARGET_SYNC_HI_QI): Likewise.
13800 (TARGET_SYNC_TI): Likewise.
13801 (MASK_CRYPTO): Likewise.
13802 (MASK_DIRECT_MOVE): Likewise.
13803 (MASK_P8_FUSION): Likewise.
13804 (MASK_P8_VECTOR): Likewise.
13805 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
13806 temporary used by some of the direct move instructions to get two FP
13807 temporary registers does not force creation of a stack frame.
13808 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
13809 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
13810 that any VSX registers are tieable, even if they are also an
13811 Altivec vector mode.
13812 (r6000_reg_class_enum): Add wm, wr, wv constraints.
13813 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
13814 (RS6000_BTM_CRYPTO): Likewise.
13815 (RS6000_BTM_COMMON): Likewise.
13817 * config/rs6000/rs6000.md (cpu attribute): Add power8.
13818 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
13819 (enum rs6000_vector): Add power8 vector support.
13821 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13825 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
13826 (any_sibcall_could_use_r3): this and handle indirect calls.
13827 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
13829 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13831 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
13833 2013-05-22 Richard Biener <rguenther@suse.de>
13835 PR middle-end/57349
13836 * profile.c (branch_prob): Do not split blocks that are
13837 abnormally receiving from ECF_RETURNS_TWICE functions.
13839 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13841 * recog.c (offsettable_address_addr_space_p): Fix calculation of
13842 address mode. Move pointer mode initialization to the same place.
13844 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13846 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
13847 while it has any effect.
13849 2013-05-21 Easwaran Raman <eraman@google.com>
13851 PR tree-optimization/57322
13852 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
13853 UID of the statement added to the BB to be 1.
13855 2013-05-21 Jakub Jelinek <jakub@redhat.com>
13857 PR tree-optimization/57331
13858 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
13859 of conversion from pointer type to integral type with integer.
13861 2013-05-21 Martin Jambor <mjambor@suse.cz>
13864 * ipa-prop.c (ipa_read_node_info): Process param_used and
13865 controlled_uses in the same order as when writing.
13867 2013-05-21 Magnus Granberg <baldrick@free.fr>
13870 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
13872 2013-05-21 Richard Biener <rguenther@suse.de>
13874 PR tree-optimization/57318
13875 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
13876 estimate stmts with side-effects as likely eliminated.
13878 2013-05-21 Richard Biener <rguenther@suse.de>
13880 PR tree-optimization/57330
13881 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
13882 preserve the call stmts fntype.
13884 2013-05-21 Richard Biener <rguenther@suse.de>
13886 PR tree-optimization/57303
13887 * tree-ssa-sink.c (statement_sink_location): Improve killing
13888 stmt detection and properly handle self-assignments.
13890 2013-05-21 Christian Bruel <christian.bruel@st.com>
13892 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
13893 spanning registers. LEAF_REG_REMAP is supported only for contiguous
13894 registers. Set register size out of the PARALLEL loop.
13896 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
13899 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
13900 (*fmasf4, *fmasf4_media): New insns.
13902 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13904 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
13905 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
13906 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
13907 (mips_idiv_insns): Update the comments to say that the returned
13908 instruction counts are in units of BASE_INSN_LENGTH.
13909 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
13910 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
13911 using 2 rather than 4 as the length of indirect MIPS16 and
13912 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
13913 length of a NOP. Don't divide MIPS16 lengths by 2.
13914 (mips16_split_long_branches): Assume a branch is long if the
13915 length is greater than 4 rather than 8.
13916 * config/mips/mips.md (length): Give MIPS16 lengths directly,
13917 rather than multiplying them by 2. Multiply instruction counts
13918 by BASE_INSN_LENGTH rather than 4.
13919 (*jump_mips16, tls_get_tp_mips16_<mode>)
13920 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
13922 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13924 * config/mips/mips.md (extended_mips16): Remove branch case.
13925 (length): Remove duplicated extended_mips16 test.
13927 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13929 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
13931 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
13933 * recog.h (Recog_data): Rename to...
13934 (recog_data_d): ...this.
13935 (recog_data): Update accordingly.
13936 * recog.c (recog_data): Likewise.
13937 * reload.c (save_recog_data): Likewise.
13938 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
13939 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
13941 2013-05-17 Julian Brown <julian@codesourcery.com>
13943 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
13944 found in a REG_EQUAL note, invalidate it.
13946 2013-05-17 Easwaran Raman <eraman@google.com>
13948 * tree-ssa-reassoc.c (find_insert_point): New function.
13949 (insert_stmt_after): Likewise.
13950 (get_def_stmt): Likewise.
13951 (ensure_ops_are_available): Likewise.
13952 (not_dominated_by): Likewise.
13953 (rewrite_expr_tree): Do not move statements beyond what is
13954 necessary. Remove call to swap_ops_for_binary_stmt...
13955 (reassociate_bb): ... and move it here.
13956 (build_and_add_sum): Assign UIDs for new statements.
13957 (linearize_expr): Likewise.
13958 (do_reassoc): Renumber gimple statement UIDs.
13960 2013-05-17 Jan Hubicka <jh@suse.cz>
13962 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13964 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13965 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13967 (output_weakrefs): Update.
13969 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13970 Martin Jambor <mjambor@suse.cz>
13972 PR middle-end/57276
13973 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13974 value that corresponds to the given aggval is found in values vector.
13976 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13978 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13979 sse, sse2, sse3, ssse3 and sse4a flags to options.
13981 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13983 * gengtype-state.c: (s_expr_writer): New class, to handle
13984 prettifying of output layout of s-expressions.
13985 (state_writer): New class, to write out gtype.state.
13986 (state_written_type_count): Move this variable into member data of
13988 (s_expr_writer::s_expr_writer): New code: constructor for new class
13989 (state_writer::state_writer(): ditto
13990 (s_expr_writer::write_new_line): New function
13991 (s_expr_writer::write_any_indent): ditto
13992 (s_expr_writer::begin_s_expr): ditto
13993 (s_expr_writer::end_s_expr): ditto
13994 (write_state_fileloc): convert to method of state_writer...
13995 (state_writer:: write_state_fileloc): ...and use methods of
13996 s_expr_writer to write indentation into the gtype.state output file
13997 to visually represent the hierarchical structure of the list
13999 (write_state_fields): ditto, renaming to...
14000 (state_writer::write_state_fields)
14001 (write_state_a_string): ditto, renaming to...
14002 (state_writer::write_state_a_string)
14003 (write_state_string_option): ditto, renaming to...
14004 (state_writer::write_state_string_option)
14005 (write_state_type_option): ditto, renaming to...
14006 (state_writer::write_state_type_option)
14007 (write_state_nested_option): ditto, renaming to...
14008 (state_writer::write_state_nested_option)
14009 (write_state_option): ditto, renaming to...
14010 (state_writer::write_state_option)
14011 (write_state_options): ditto, renaming to...
14012 (state_writer::write_state_options)
14013 (write_state_lang_bitmap): ditto, renaming to...
14014 (state_writer::write_state_lang_bitmap)
14015 (write_state_version): ditto, renaming to...
14016 (state_writer::write_state_version)
14017 (write_state_scalar_type): ditto, renaming to...
14018 (state_writer::write_state_scalar_type)
14019 (write_state_string_type): ditto, renaming to...
14020 (state_writer::write_state_string_type)
14021 (write_state_undefined_type): ditto, renaming to...
14022 (state_writer::write_state_undefined_type)
14023 (write_state_struct_union_type): ditto, renaming to...
14024 (state_writer::write_state_struct_union_type)
14025 (write_state_struct_type): ditto, renaming to...
14026 (state_writer::write_state_struct_type)
14027 (write_state_user_struct_type): ditto, renaming to...
14028 (state_writer::write_state_user_struct_type)
14029 (write_state_lang_struct_type): ditto, renaming to...
14030 (state_writer::write_state_lang_struct_type)
14031 (write_state_param_struct_type): ditto, renaming to...
14032 (state_writer::write_state_param_struct_type)
14033 (write_state_pointer_type): ditto, renaming to...
14034 (state_writer::write_state_pointer_type)
14035 (write_state_array_type): ditto, renaming to...
14036 (state_writer::write_state_array_type)
14037 (write_state_gc_used): ditto, renaming to...
14038 (state_writer::write_state_gc_used)
14039 (write_state_common_type_content): ditto, renaming to...
14040 (state_writer::write_state_common_type_content)
14041 (write_state_type): ditto, renaming to...
14042 (state_writer::write_state_type)
14043 (write_state_pair_list): ditto, renaming to...
14044 (state_writer::write_state_pair_list)
14045 (write_state_pair): ditto, renaming to...
14046 (state_writer::write_state_pair)
14047 (write_state_typedefs): ditto, renaming to...
14048 (state_writer::write_state_typedefs)
14049 (write_state_structures): ditto, renaming to...
14050 (state_writer::write_state_structures)
14051 (write_state_param_structs): ditto, renaming to...
14052 (state_writer::write_state_param_structs)
14053 (write_state_variables): ditto, renaming to...
14054 (state_writer::write_state_variables)
14055 (write_state_srcdir): ditto, renaming to...
14056 (state_writer::write_state_srcdir)
14057 (write_state_files_list): ditto, renaming to...
14058 (state_writer::write_state_files_list)
14059 (write_state_languages): ditto, renaming to...
14060 (state_writer::write_state_languages)
14061 (write_state): create a state_writer instance and use it when
14062 writing out the state file
14064 2013-05-17 Mike Stump <mikestump@comcast.net>
14066 PR rtl-optimization/57304
14067 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
14068 accessing DF_REF_REAL_LOC.
14070 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14072 PR rtl-optimization/57281
14073 PR rtl-optimization/57300
14074 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
14075 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
14076 what the other splitter did if the registers are dead.
14078 2013-05-17 Richard Biener <rguenther@suse.de>
14080 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
14083 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14085 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
14088 2013-05-17 Marek Polacek <polacek@redhat.com>
14090 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
14091 length when doing non-zero store of storing '\0' to '\0'.
14093 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14095 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
14096 vect_external_def oprnd1 with loop_vinfo, try to emit
14097 optional cast, negation and and stmts on the loop preheader
14098 edge instead of into the pattern def seq.
14100 PR tree-optimization/57051
14101 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
14102 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
14104 2013-05-16 Nick Clifton <nickc@redhat.com>
14106 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
14107 (rl78_is_naked_func): New function.
14108 (rl78_expand_prologue): Skip prologue generation for naked functions.
14109 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
14110 * doc/extend.texi (naked): Add RL78 to the list of processors
14111 that supports this attribute.
14113 2013-05-16 Jeff Law <law@redhat.com>
14115 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
14117 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14119 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
14120 cache parameters using detect_caches_amd also for CYRIX,
14121 NSC and TM2 signatures.
14123 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14124 Dzianis Kahanovich <mahatma@eu.by>
14128 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
14129 VIA/Centaur processors and determine their cache parameters
14130 using detect_caches_amd.
14132 2013-05-16 Teresa Johnson <tejohnson@google.com>
14134 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
14135 (rtl_verify_edges): New function.
14136 (rtl_verify_bb_insns): Ditto.
14137 (rtl_verify_bb_pointers): Ditto.
14138 (rtl_verify_bb_insn_chain): Ditto.
14139 (rtl_verify_fallthru): Ditto.
14140 (rtl_verify_bb_layout): Ditto.
14141 (rtl_verify_flow_info_1): Outline checks into new functions.
14142 (rtl_verify_flow_info): Ditto.
14144 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
14146 * cfghooks.c (copy_bbs): Add update_dominance argument.
14147 * cfghooks.h (copy_bbs): Update prototype.
14148 * tree-cfg.c (gimple_duplicate_sese_region):
14149 Add update_dominance argument.
14150 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
14151 * tree-ssa-loop-ch.c (copy_loop_headers): Update
14152 gimple_duplicate_sese_region call.
14153 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
14154 Update copy_bbs call.
14155 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
14156 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
14158 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14160 * tree-vectorizer.h (NUM_PATTERNS): Increment.
14161 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
14162 vect_recog_rotate_pattern.
14163 (vect_recog_rotate_pattern): New function.
14165 2013-05-16 Jason Merrill <jason@redhat.com>
14167 * Makefile.in (LLINKER): New variable.
14168 (mostlyclean): Remove link mutex.
14169 * configure.ac: Handle --enable-link-mutex.
14170 * lock-and-run.sh: New script.
14172 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14175 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
14178 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14180 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
14182 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
14184 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
14185 * config/arm/arm.c (next_consecutive_mem): New function.
14186 (gen_movmem_ldrd_strd): Likewise.
14187 * config/arm/arm.md (movmemqi): Update condition and code.
14188 (unaligned_loaddi, unaligned_storedi): New patterns.
14190 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14192 * config.gcc: Obsolete *-*-solaris2.9*.
14193 * doc/install.texi (Specific, *-*-solaris2*): Document it.
14195 2013-05-16 Richard Biener <rguenther@suse.de>
14197 * passes.c (init_optimization_passes): Move pass_parallelize_loops
14198 earlier, after GRAPHITE transforms and IV canonicalization.
14200 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14202 * omp-low.c (extract_omp_for_data): For collapsed loops,
14203 if at least one of the loops is known at compile time to
14204 iterate zero times, set count to 0.
14205 (expand_omp_regimplify_p): New function.
14206 (expand_omp_for_generic): For collapsed loops, if at least
14207 one of the loops isn't known to iterate at least once,
14208 add runtime check with setting count to 0.
14209 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
14210 For unsigned types if it isn't known at compile time that
14211 the loop will iterate at least once, add runtime check to bypass
14212 the whole loop if initial condition isn't true.
14214 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
14216 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
14218 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
14220 PR middle-end/57286
14221 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
14222 transformations to avoid an infinite loop.
14224 2013-05-16 Marek Polacek <polacek@redhat.com>
14226 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
14228 2013-05-15 Leif Ekblad <leif@rdos.net>
14230 * config/i386/i386.c (ix86_decompose_address): Use
14231 DEFAULT_TLS_SEG_REG to access TLS segment register.
14232 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
14233 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
14234 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
14236 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
14239 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
14240 sibling calls to functions that would normally be lazily bound,
14241 unless $gp is call-clobbered.
14243 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
14245 * config/i386/i386.c (ix86_option_override_internal): Update
14246 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
14247 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
14248 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
14249 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
14251 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
14253 2013-05-15 Andreas Schwab <schwab@suse.de>
14255 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
14256 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
14259 2013-05-15 Teresa Johnson <tejohnson@google.com>
14261 * loop-unroll.c (report_unroll_peel): Check decision before
14262 emitting unroll/peel message.
14264 2013-05-15 Teresa Johnson <tejohnson@google.com>
14266 * function.h (has_bb_partition): New rtl_data flag.
14267 (bb_reorder_complete): Ditto.
14268 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
14269 instead of flag_reorder_blocks_and_partition.
14270 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
14271 with some enhancements.
14272 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
14273 * bb-reorder.c (connect_traces): Check for has_bb_partition
14274 instead of flag_reorder_blocks_and_partition.
14275 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
14276 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
14277 verify_hot_cold_block_grouping.
14278 (partition_hot_cold_basic_blocks): Set has_bb_partition.
14280 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14283 * config/arm/predicates.md (call_insn_operand): New predicate.
14284 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
14285 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
14286 if insn is not a tail call.
14287 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
14289 * config/arm/arm.h (enum reg_class): New caller save register class.
14290 (REG_CLASS_NAMES): Likewise.
14291 (REG_CLASS_CONTENTS): Likewise.
14292 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
14295 2013-05-15 Richard Biener <rguenther@suse.de>
14297 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
14298 of MSG_OPTIMIZED_LOCATIONS.
14299 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
14300 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
14302 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
14303 of MSG_OPTIMIZED_LOCATIONS.
14304 (execute_vect_slp): Likewise.
14305 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
14306 (vect_create_cond_for_alias_checks): Likewise.
14307 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
14308 (vect_recog_widen_mult_pattern): Likewise.
14309 (vect_recog_widen_sum_pattern): Likewise.
14310 (vect_recog_over_widening_pattern): Likewise.
14311 (vect_recog_widen_shift_pattern): Likewise.
14312 (vect_recog_vector_vector_shift_pattern): Likewise.
14313 (vect_recog_divmod_pattern): Likewise.
14314 (vect_recog_mixed_size_cond_pattern): Likewise.
14315 (vect_recog_bool_pattern): Likewise.
14316 (vect_pattern_recog_1): Likewise.
14318 2013-05-15 Martin Jambor <mjambor@suse.cz>
14320 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
14321 non-functions to builtin_unreachable.
14322 * ipa-inline-transform.c (inline_call): Do not assert estimates were
14323 correct when new direct edges were discovered.
14325 2013-05-15 Martin Jambor <mjambor@suse.cz>
14327 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
14328 header, print symbol order instead of node uid, print more information
14329 about indirect edge targets.
14330 (ipa_make_edge_direct_to_target): Print symbol order instead of node
14332 (ipa_make_edge_direct_to_target): Likewise.
14333 (remove_described_reference): Likewise.
14334 (propagate_controlled_uses): Likewise.
14335 (ipa_print_node_params): Also print symbol order.
14336 (ipcp_transform_function): Print symbol order instead of node uids.
14337 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
14338 (cgraph_get_create_real_symbol_node): Likewise.
14339 * ipa-cp.c (print_lattice): Likewise.
14340 (print_all_lattices): Likewise.
14341 (determine_versionability): Likewise.
14342 (initialize_node_lattices): Likewise.
14343 (estimate_local_effects): Likewise.
14344 (update_profiling_info): Likewise.
14345 (create_specialized_node): Likewise.
14346 (perhaps_add_new_callers): Likewise.
14347 (decide_about_value): Likewise.
14348 (decide_whether_version_node): Likewise.
14349 (identify_dead_nodes): Likewise.
14350 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
14351 (dump_inline_summary): Likewise.
14352 (estimate_node_size_and_time): Likewise.
14353 (inline_analyze_function): Likewise.
14354 * ipa-inline.c (report_inline_failed_reason): Likewise.
14355 (want_early_inline_function_p): Likewise.
14356 (edge_badness): Likewise.
14357 (update_edge_key): Likewise.
14358 (inline_small_functions): Likewise. Add dumping of order to two other
14360 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
14361 instead of node uids.
14362 (propagate_pure_const): Likewise.
14363 (propagate_pure_const): Likewise.
14364 * ipa-utils.c (dump_cgraph_node_set): Likewise.
14365 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
14366 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
14368 * tree-pretty-print.c (dump_function_header): Likewise.
14369 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
14370 Print symbol order instead of node uids.
14372 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14374 * config/s390/s390.c (s390_register_move_cost): Don't impose the
14375 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
14377 2013-05-15 Richard Biener <rguenther@suse.de>
14379 PR tree-optimization/57275
14380 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
14381 return value for fail to do runtime alias checks for gather loads.
14383 2013-05-15 Jan Hubicka <jh@suse.cz>
14387 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
14388 weakrefs are not external.
14389 (lto_symtab_merge_decls): Fix thinko when dealing with
14390 non-lto_symtab decls.
14391 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
14392 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
14393 * varpool.c (dump_varpool_node): Dump more flags.
14395 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14397 * config/i386/i386.c (processor_alias_table): Add instruction
14398 FSGSBASE for AMD bdver3 architecture.
14400 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14402 * tree.c (warn_deprecated_use): Print file:line using locus color.
14403 * diagnostic.c (diagnostic_report_current_module): Print file:line
14404 and file:line:column using locus color.
14406 2013-05-14 Mike Stump <mikestump@comcast.net>
14408 * gdbinit.in: Add __null.
14410 2013-05-14 Mike Stump <mikestump@comcast.net>
14412 * recog.h: Rename struct recog_data to Recog_data.
14413 * recog.c: Likewise.
14414 * reload.c (can_reload_into): Likewise.
14415 * config/picochip/picochip.c: Likewise.
14417 2013-05-14 Mike Stump <mikestump@comcast.net>
14419 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
14421 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
14423 * resource.h (struct resources): Remove unch_memory member.
14424 (CLEAR_RESOURCE): Don't clear unch_memory.
14425 * resource.c (mark_referenced_resources): Don't set it.
14426 (mark_set_resources): Likewise.
14427 (mark_target_live_regs): Don't clear it.
14428 (init_resource_info): Likewise.
14429 * reorg.c (resource_conflicts_p): Don't compare it.
14430 (redundant_insn): Don't set it.
14432 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
14434 * emit-rtl.c (next_label): Remove unused function.
14435 (skip_consecutive_labels, link_cc0_insns): Move to ...
14436 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
14437 only place where these functions are used, and make them static.
14439 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14441 * fold-const.c (fold_negate_expr): Handle vectors.
14442 (fold_truth_not_expr): Make it static.
14443 (fold_invert_truthvalue): New static function.
14444 (invert_truthvalue_loc): Handle vectors. Do not call
14445 fold_truth_not_expr directly.
14446 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
14447 <TRUTH_NOT_EXPR>: Do not cast to boolean.
14448 (fold_comparison): Handle vector constants.
14449 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
14450 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
14451 * tree.h (fold_truth_not_expr): Remove declaration.
14453 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
14455 * config/aarch64/aarch64-simd.md
14456 (aarch64_vcond_internal<mode>): Rename to...
14457 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
14458 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
14459 float modes. Clarify all iterator modes.
14460 (vcond<mode><mode>): Use new name for vcond expanders.
14461 (vcond<v_cmp_result><mode>): Likewise.
14462 (vcondu<mode><mode>: Likewise.
14463 * config/aarch64/iterators.md (VDQF_COND): New.
14465 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14468 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
14469 variable for the shift amount. Check that we shift by non-negative
14472 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
14475 * config/arm/arm.h (EPILOGUE_USES): Only return true
14476 for LR_REGNUM after epilogue_completed.
14478 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
14480 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
14481 is error_mark_node.
14483 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14486 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
14487 and Solaris 11+/x86 with gld.
14488 * configure: Regenerate.
14490 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14492 * expmed.c (expand_shift_1): Canonicalize rotates by
14493 constant bitsize / 2 to bitsize - 1.
14494 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
14495 case ROTATERT>: Likewise.
14498 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14500 * config/i386/i386.md (rotateinv): New code attr.
14501 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14502 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14503 roll $31, %eax, etc.
14505 2013-05-14 Richard Biener <rguenther@suse.de>
14507 PR middle-end/57235
14508 * tree-eh.c (sink_clobbers): Give up for successors with
14509 multiple predecessors and no virtual uses.
14511 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
14513 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
14514 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
14516 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14518 PR middle-end/57251
14519 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
14520 the case when both op0 and op1 have VOIDmode.
14522 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14524 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
14525 in multiply-accumulate mode.
14527 2013-05-13 Guozhi Wei <carrot@google.com>
14529 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
14531 2013-05-13 Kai Tietz <ktietz@redhat.com>
14534 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
14535 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
14536 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
14537 * config/i386/i386.c (ix86_option_override_internal): Likewise.
14538 (ix86_expand_prologue): Likewise.
14539 (ix86_expand_split_stack_prologue): Likewise.
14540 (legitimate_pic_address_disp_p): Likewise.
14541 (legitimize_pic_address): Likewise.
14542 (legitimize_tls_address): Likewise.
14543 (legitimize_pe_coff_symbol): Likewise.
14544 (output_pic_addr_const): Likewise.
14545 (construct_plt_address): Likewise.
14546 (ix86_expand_call): Likewise.
14547 (x86_output_mi_thunk): Likewise.
14548 (x86_function_profiler): Likewise.
14550 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
14552 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
14553 similar switch cases.
14554 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
14555 (aarch64_simd_mov_to_<mode>low): Delete.
14556 (aarch64_simd_mov_to_<mode>high): Delete.
14557 (move_lo_quad_<mode>): Add w<-r alternative.
14558 (aarch64_simd_move_hi_quad_<mode>): Likewise.
14559 (aarch64_simd_mov_from_*): Update type attribute.
14560 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
14563 2013-05-13 Jan Hubicka <jh@suse.cz>
14565 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
14566 * config/i386/i386.c (ix86_compute_frame_layout,
14567 ix86_expand_epilogue, emit_i387_cw_initialization,
14568 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
14569 ix86_local_alignment): Fix use of size/speed predicates.
14571 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14573 PR tree-optimization/45216
14574 PR tree-optimization/57157
14575 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
14576 the (-Y) & (B - 1) variant if OP is |.
14577 * expmed.c (expand_shift_1): For rotations by const0_rtx just
14578 return shifted. Use (-op1) & (prec - 1) as other_amount
14579 instead of prec - op1.
14581 2013-05-13 Martin Jambor <mjambor@suse.cz>
14583 PR middle-end/42371
14584 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
14585 (ipa_constant_data): New type.
14586 (ipa_jump_func): Use ipa_constant_data to hold information about
14587 constant jump functions.
14588 (ipa_get_jf_constant): Adjust to jump function type changes.
14589 (ipa_get_jf_constant_rdesc): New function.
14590 (ipa_param_descriptor): New field controlled_uses.
14591 (ipa_get_controlled_uses): New function.
14592 (ipa_set_controlled_uses): Likewise.
14593 * ipa-ref.h (ipa_find_reference): Declare.
14594 * ipa-prop.c (ipa_cst_ref_desc): New type.
14595 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
14597 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
14598 New parameter cs. Adjust all callers.
14599 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
14600 (remove_described_reference): New function.
14601 (jfunc_rdesc_usable): Likewise.
14602 (try_make_edge_direct_simple_call): Decrement controlled use count,
14603 attempt to remove reference if it hits zero.
14604 (combine_controlled_uses_counters): New function.
14605 (propagate_controlled_uses): Likewise.
14606 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
14607 (ipa_edge_duplication_hook): Duplicate reference descriptions.
14608 (ipa_print_node_params): Print described use counter.
14609 (ipa_write_jump_function): Adjust to jump function type changes.
14610 (ipa_read_jump_function): New parameter CS, pass it to
14611 ipa_set_jf_constant. Adjust caller.
14612 (ipa_write_node_info): Stream controlled use count
14613 (ipa_read_node_info): Likewise.
14614 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
14616 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
14617 count. Remove cloning-added reference if it reaches zero.
14618 * ipa-ref.c (ipa_find_reference): New function.
14620 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14622 * config/i386/i386.c (processor_target_table): Modified default
14623 alignment values for AMD BD and BT architectures.
14625 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
14627 * tree-vect-generic.c (uniform_vector_p): Move ...
14628 * tree.c (uniform_vector_p): ... here.
14629 * tree.h (uniform_vector_p): Declare it.
14630 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
14633 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14635 PR tree-optimization/57230
14636 * tree-ssa-strlen.c (handle_char_store): Record length for
14637 array store from STRING_CST.
14639 PR tree-optimization/57230
14640 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
14643 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
14645 * config/epiphany/epiphany.c (epiphany_init): Check size of
14646 NUM_MODES_FOR_MODE_SWITCHING.
14647 (epiphany_expand_prologue):
14648 Remove CONFIG_REGNUM initial value handling code.
14649 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
14650 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
14651 (emit_set_fp_mode, epiphany_mode_after): Likewise.
14652 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
14653 Don't return 1 for FP_MODE_NONE.
14654 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
14655 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
14656 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
14657 * config/epiphany/epiphany.md (save_config): New pattern.
14659 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
14661 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
14663 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
14665 * config/i386/i386.md (memory): Handle sseishft1.
14666 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
14667 (*vec_extractv2di_1): Ditto.
14669 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
14671 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
14674 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
14676 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
14677 Add mthumb/march=armv7-a multilib.
14678 Add mthumb/march=armv7-r multilib.
14679 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
14681 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
14683 * config/v850/t-rtems: Add more multilibs.
14685 2013-05-10 Richard Biener <rguenther@suse.de>
14687 PR tree-optimization/57214
14688 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
14689 not propagate from SSA names that occur in abnormal PHI nodes.
14691 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
14693 * stor-layout.c (element_precision): New function.
14694 * machmode.h (element_precision): Declare it.
14695 * tree.c (build_minus_one_cst): New function.
14696 (element_precision): Likewise.
14697 * tree.h (build_minus_one_cst): Declare new function.
14698 (element_precision): Likewise.
14699 * fold-const.c (operand_equal_p): Use element_precision.
14700 (fold_binary_loc): Handle vector types.
14701 * convert.c (convert_to_integer): Use element_precision.
14702 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
14705 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
14707 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
14708 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
14709 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
14710 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14711 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
14712 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
14713 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
14714 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14715 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
14716 (Uuw8): New constraints.
14717 (Usb4): Move into alphabetical order.
14718 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
14719 (sd8_operand, ub8_operand, uw8_operand): New predicates.
14720 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
14721 previously unnamed patterns.
14722 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
14723 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
14724 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
14725 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
14726 of set_attr_alternative/if_then_else. Use extended_mips16 instead
14727 of specific lengths.
14729 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14731 * config/i386/i386.md (rotateinv): New code attr.
14732 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14733 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14734 roll $31, %eax, etc.
14736 PR tree-optimization/45216
14737 PR tree-optimization/57157
14738 * tree-ssa-forwprop.c (simplify_rotate): New function.
14739 (ssa_forward_propagate_and_combine): Call it.
14741 2013-05-10 Richard Biener <rguenther@suse.de>
14743 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
14744 disable peeling when we version for aliasing.
14745 (vector_alignment_reachable_p): Honor explicit user alignment.
14746 (vect_supportable_dr_alignment): Likewise.
14747 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
14748 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
14749 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
14750 then peeling to arrange for the cost-model check to come first.
14752 2013-05-10 Alan Modra <amodra@gmail.com>
14754 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
14755 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
14756 * configure: Regenerate.
14758 2013-05-10 Alan Modra <amodra@gmail.com>
14761 * varasm.c (default_elf_select_section): Move !DECL_P check..
14762 (get_named_section): ..to here before calling get_section_name.
14764 (default_section_type_flags): Add DECL_P check.
14765 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
14766 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
14768 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
14770 * config/epiphany/epiphany.c (epiphany_expand_prologue):
14771 When using gen_stack_adjust_str with a register offset, add a
14772 REG_FRAME_RELATED_EXPR note.
14774 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
14776 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
14777 (*vec_extractv4si_zext_mem): Ditto.
14778 (*vec_extractv2di): Add 0->x and x->x alternatives.
14779 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
14780 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
14782 2013-05-09 Jason Merrill <jason@redhat.com>
14784 N3639 C++1y VLA support
14785 * gimplify.c (gimplify_vla_decl): Don't touch an existing
14788 * tree.c (build_constructor_va): New.
14789 * tree.h: Declare it.
14791 2013-05-09 Martin Jambor <mjambor@suse.cz>
14794 * gimple-fold.c (canonicalize_constructor_val): Call
14795 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
14797 2013-05-09 Jan Hubicka <jh@suse.cz>
14798 Richard Biener <rguenther@suse.de>
14801 * symtab.c (symtab_make_decl_local): Do not add private names.
14803 2013-05-09 Jan Hubicka <jh@suse.cz>
14806 * symtab.c (insert_to_assembler_name_hash): Handle clones.
14807 (unlink_from_assembler_name_hash): Likewise.
14808 (symtab_prevail_in_asm_name_hash, symtab_register_node,
14809 symtab_unregister_node, symtab_initialize_asm_name_hash,
14810 change_decl_assembler_name): Update.
14812 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
14814 * config/aarch64/aarch64.md: New movtf split.
14815 (*movtf_aarch64): Update.
14816 (aarch64_movdi_tilow): Handle TF modes and rename to
14817 aarch64_movdi_<mode>low.
14818 (aarch64_movdi_tihigh): Handle TF modes and rename to
14819 aarch64_movdi_<mode>high
14820 (aarch64_movtihigh_di): Handle TF modes and rename to
14821 aarch64_mov<mode>high_di
14822 (aarch64_movtilow_di): Handle TF modes and rename to
14823 aarch64_mov<mode>low_di
14824 (aarch64_movtilow_tilow): Remove spurious whitespace.
14825 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
14827 (aarch64_print_operand): Update.
14829 2013-05-09 Alan Modra <amodra@gmail.com>
14831 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
14833 * configure: Regenerate.
14835 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14837 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
14838 splitter preparation statements.
14839 * config/i386/sse.md (*vec_extract* splitters): Ditto.
14840 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
14843 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14845 * gimple-ssa-strength-reduction.c (count_candidates): Change
14846 return value to int.
14847 (analyze_candidates_and_replace): Change type of length to int.
14849 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14851 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
14852 (*vec_extract<mode>): Use VI12_128 mode iterator.
14853 (*vec_extract<mode>_mem): Ditto.
14854 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
14857 2013-05-08 Diego Novillo <dnovillo@google.com>
14862 2012-08-17 Diego Novillo <dnovillo@google.com>
14865 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
14866 * config.in: Regenerate.
14867 * configure: Regenerate.
14868 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
14870 2013-05-08 Jan Hubicka <jh@suse.cz>
14873 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
14874 * cgraph.h (symtab_node_base): Add unique_name.
14875 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14876 input_overwrite_node, input_varpool_node): Stream unique_name.
14877 * cgraphclones.c (cgraph_create_virtual_clone,
14878 cgraph_function_versioning): Set unique_name.
14879 * ipa.c (function_and_variable_visibility): Set unique_name.
14881 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14883 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
14884 (alloc_cand_and_find_basis): Restrict conditional candidate
14885 processing to CAND_MULTs.
14887 2013-05-08 Jan Hubicka <jh@suse.cz>
14890 lto-symtab.c (lto_symtab_symbol_p): New function.
14891 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
14892 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
14893 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
14894 Skip static symbols.
14896 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
14898 PR tree-optimization/57200
14899 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
14900 Only call inform if the preceding warning_at returns true.
14902 2013-05-07 Han Shen <shenhan@google.com>
14904 * cfgexpand.c (record_or_union_type_has_array_p): New function.
14905 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
14906 * common.opt (fstack-protector-strong): New option.
14907 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
14908 * doc/invoke.texi (Optimization Options): Document
14909 "-fstack-protector-strong".
14910 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
14912 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14914 * config/mips/mips.c (mips_machine_reorg2): Return 0.
14916 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
14918 * ira.c (update_equiv_regs): Add insn having equiv memory even if
14919 it is not lhs of the insn.
14920 (setup_reg_equiv): Remove insn having equiv memory which it is not
14922 * lra-constraints.c (process_address): Try to improve generation
14923 code for address base + disp.
14924 (lra_constraints): Make correct the code for checking insn setting
14925 up backward equivalence. Remove insn only if it is in the init
14927 * lra-eliminations.c (update_reg_eliminate): Change return value.
14928 (lra_eliminate): Use the result.
14930 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
14932 * config/i386/sse.md (ssescalarnummask): New mode attribute.
14933 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
14934 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
14935 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
14936 register target operands.
14937 (*vec_extractv8hi_sse2): New pattern.
14938 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
14939 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
14940 (*vec_extract<mode>_mem): New insn and split pattern.
14942 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
14944 * config/arm/arm.c (arm_asan_shadow_offset): New function.
14945 (TARGET_ASAN_SHADOW_OFFSET): Define.
14946 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
14947 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
14949 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14951 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
14952 (incr_vec_index): Return -1 if increment not found.
14953 (create_add_on_incoming_edge): Assert if increment not found.
14954 (record_increment): Limit number of increments recorded.
14955 (all_phi_incrs_profitable): Return false if an increment not found.
14956 (replace_profitable_candidates): Don't process increments that were
14958 (analyze_candidates_and_replace): Limit size of incr_vec.
14960 2013-05-07 Richard Biener <rguenther@suse.de>
14962 * calls.c (special_function_p): setjmp-like functions are leaf.
14963 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14964 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14966 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14968 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14969 (aarch64_simd_mov<mode>): New expander.
14970 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14971 (aarch64_simd_mov_to_<mode>high): Likewise.
14972 (aarch64_simd_mov_from_<mode>low): Likewise.
14973 (aarch64_simd_mov_from_<mode>high): Likewise.
14974 (aarch64_dup_lane<mode>): Update.
14975 (aarch64_dup_lanedi): New instruction pattern.
14976 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14977 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14979 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14981 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14982 (replace_mult_candidate): Remove unnecessary argument; remove
14983 unnecessary parameter from call to introduce_cast_before_cand.
14984 (replace_unconditional_candidate): Remove unnecessary parameter
14985 from call to replace_mult_candidate.
14986 (replace_conditional_candidate): Likewise.
14987 (insert_initializers): Use make_temp_ssa_name.
14988 (introduce_cast_before_cand): Remove unnecessary argument; use
14989 make_temp_ssa_name.
14990 (replace_one_candidate): Remove unnecessary argument; remove
14991 unnecessary parameter from calls to introduce_cast_before_cand.
14992 (replace_profitable_candidates): Remove unnecessary parameters
14993 from calls to replace_one_candidate.
14995 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14997 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14998 phi def as possibly hiding a basis for a CAND_ADD whose operands
14999 have been commuted in the analysis.
15000 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
15002 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15004 * config/aarch64/aarch64.md
15005 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
15006 shift value between 0-4.
15008 2013-05-07 Richard Biener <rguenther@suse.de>
15010 * double-int.h (rshift): New overload.
15011 * double-int.c (rshift): New function.
15012 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
15013 (create_reference_ops_from_ref): Remove.
15014 (vn_reference_insert): Use shared ops for constructing the
15015 reference and copy it.
15017 2013-05-07 Richard Biener <rguenther@suse.de>
15019 PR middle-end/57190
15020 * tree-eh.c (sink_clobbers): Properly propagate
15021 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
15023 2013-05-07 Jakub Jelinek <jakub@redhat.com>
15025 PR tree-optimization/57149
15026 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
15027 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
15028 collect_phi_def_edges, execute_late_warn_uninitialized): Use
15029 uninit_undefined_value_p instead of ssa_undefined_value_p.
15032 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
15033 for modifier == EXPAND_INITIALIZER.
15035 2013-05-07 Anton Blanchard <anton@samba.org>
15037 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
15038 for powerpc64 little endian.
15039 * configure: Regenerate.
15041 2013-05-06 Graham Stott <grahams@btinternet.com>
15043 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
15044 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
15045 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
15046 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
15048 2013-05-06 Graham Stott <grahams@btinternet.com>
15050 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
15051 codes which allow non-lvalues.
15053 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
15055 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
15056 components are all 1s.
15057 (integer_minus_onep): New function.
15058 * tree.h (integer_minus_onep): Declare it.
15059 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
15060 integer_minus_onep instead of integer_all_onesp.
15062 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15065 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
15066 variations of these patterns.
15068 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15070 * config/i386/i386.md (isa): Add x64_sse4 member.
15071 (enabled): Handle x64_sse4.
15072 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
15073 instruction for 64bit SSE4_1 targets. Update insn attributes.
15074 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
15075 instruction for SSE4_1 targets. Update insn attributes.
15076 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15077 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
15078 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
15080 (*vec_extractv4si): Rename from *sse4_1_pextrd.
15081 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
15082 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
15084 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15087 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
15089 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
15091 * final.c (do_assembler_dialects): Don't handle curly braces and
15092 vertical bar escaped by % as dialect delimiters.
15093 (output_asm_insn): Print curly braces and vertical bar if escaped
15094 by % and ASSEMBLER_DIALECT defined.
15095 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
15096 * doc/tm.texi: Regenerated.
15098 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15100 * config/mips/mips.c: Include tree-pass.h.
15101 (mips_reorg): Split in pre- and post-dbr_schedule parts.
15102 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
15103 (pass_mips_machine_reorg2): New machine specific pass.
15104 (insert_pass_mips_machine_reorg2): New pass plugin definition.
15105 (mips_option_override): Register the new pass.
15106 * rtl.h (cleanup_barriers): Remove prototype.
15107 (dbr_schedule): Likewise.
15108 * jump.c (cleanup_barriers): Make static.
15109 * reorg.c (dbr_schedule): Likewise.
15111 2013-05-06 Richard Biener <rguenther@suse.de>
15113 PR tree-optimization/57185
15114 * tree-parloops.c (add_field_for_reduction): Handle anonymous
15115 SSA names properly.
15117 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15120 * config/i386/i386.c (add_parameter_dependencies): Add dependence
15121 between "first_arg" and "insn", not "last" and "insn".
15123 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15125 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
15126 (find_candidates_in_block): Re-enable slsr_process_phi.
15127 (create_phi_basis): Fix double counting of candidate adjustment.
15129 2013-05-06 Richard Biener <rguenther@suse.de>
15131 PR middle-end/57147
15132 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
15133 the edge is also fallthru, preserve it and just clear the
15135 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
15136 also complex, preserve that and just clear the fallthru flag.
15137 * tree-inline.c (update_ssa_across_abnormal_edges): Also
15138 update virtual operands.
15140 2013-05-06 Alan Modra <amodra@gmail.com>
15142 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
15143 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
15144 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
15145 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
15146 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15147 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
15149 2013-05-06 Alan Modra <amodra@gmail.com>
15151 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
15152 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
15153 (DEFAULT_ASM_ENDIAN): Define.
15154 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15155 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
15156 Update -K PIC clause from sysv4.h.
15157 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
15158 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
15160 2013-05-06 Alan Modra <amodra@gmail.com>
15162 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
15163 twice for little-endian.
15164 (ashrdi3_no_power, ashrdi3): Support little-endian.
15166 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15169 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
15170 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
15172 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
15173 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
15174 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
15177 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
15178 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15180 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
15181 * config/pa/pa.opt: Make mbig-switch a no-op.
15182 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
15183 (CASE_VECTOR_MODE): Always return SImode.
15184 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
15185 for the !TARGET_BIG_SWITCH case.
15186 * config/pa/pa-linux.h: Likewise.
15187 * config/pa/pa-openbsd.h: Likewise.
15188 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
15189 * config/pa/pa.md (short_jump): Remove define_insn.
15190 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
15191 (casesi0): Remove define_insn.
15192 (type): Remove btable_branch.
15193 (pa_combine_type): Likewise.
15194 (in_nullified_branch_delay): Likewise.
15195 (in_call_delay): Likewise.
15196 (define_delay): Likewise.
15197 (define_insn_reservation "Z3"): Likewise.
15198 (define_insn_reservation "Z4"): Likewise.
15199 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
15200 (pa_adjust_insn_length): Remove adjustment for btable branches.
15201 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
15204 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
15206 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15207 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
15208 Add m->r,x alternatives.
15209 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
15210 splitters using SWI48x mode iterator.
15211 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
15212 TARGET_64BIT. Add m->x alternative.
15213 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
15214 Add o->x alternative. Enable for TARGET_SSE.
15215 (sse_storeq): Remove expander.
15216 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
15217 with memory input operand.
15218 (*vec_extractv2di_1 splitter): New.
15219 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
15220 * config/i386/i386.md (ssevecmodelower): New mode attribute.
15222 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15224 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
15225 (INT_LOWPART): Delete.
15226 (extract_MB): Adjust.
15227 (extract_ME): Adjust.
15228 (print_operand): Adjust.
15230 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15232 * config/rs6000/predicates.md (reg_or_add_cint_operand,
15233 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
15234 (reg_or_logical_cint_operand, easy_fp_constant,
15235 logical_const_operand): Delete "CONST_DOUBLE" case.
15236 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
15237 "HOST_BITS_PER_WIDE_INT == 64" test.
15238 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
15239 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
15240 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
15241 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
15243 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
15244 CONST_DOUBLE DImode/VOIDmode case.
15245 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
15246 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
15247 CONST_DOUBLE VOIDmode case.
15248 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
15249 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
15250 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
15251 Delete CONST_DOUBLE case.
15252 (splitters for mov FMOVE64 const_double): Delete
15253 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
15254 "HOST_BITS_PER_WIDE_INT >= 64" test.
15255 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
15257 (mov DI const_double): Delete.
15259 2013-05-04 Jakub Jelinek <jakub@redhat.com>
15261 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
15262 on op shows all bits zero in mode of a lowpart subreg, return zero.
15264 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
15267 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
15268 to save TFmode registers and DImode to save TImode registers for
15269 caller save operations.
15270 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
15271 mark being partially clobbered since they only use the first
15274 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
15275 and TDmode only use the upper 64-bits of each VSX register.
15277 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15279 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
15280 (find_candidates_in_block): Disable slsr_process_phi.
15282 2013-05-03 Guozhi Wei <carrot@google.com>
15284 * coverage.c (coverage_obj_init): Move the construction of gcov
15286 (build_init_ctor): ... here.
15288 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15290 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
15291 (slsr_cand_d): Redefine def_phi.
15292 (stride_status, phi_adjust_status, count_phis_status): New enums.
15293 (find_phi_def): New.
15294 (find_basis_for_base_expr): New.
15295 (find_basis_for_candidate): Handle hidden bases.
15296 (alloc_cand_and_find_basis): Handle phi candidates.
15297 (slsr_process_phi): New.
15298 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
15299 (create_mul_imm_cand): Likewise.
15300 (create_add_ssa_cand): Exclude phi base candidates.
15301 (create_add_imm_cand): Likewise.
15302 (slsr_process_cast): Likewise.
15303 (slsr_process_copy): Likewise.
15304 (find_candidates_in_block): Handle phi candidates.
15305 (dump_candidate): Likewise.
15306 (unconditional_cands): Delete.
15307 (unconditional_cands_with_known_stride_p): Delete.
15308 (phi_dependent_cand_p): New.
15309 (cand_increment): Handle phi-dependent candidates.
15310 (replace_dependent): Delete.
15311 (replace_mult_candidate): New.
15312 (replace_unconditional_candidate): New.
15313 (incr_vec_index): Move to avoid forward reference.
15314 (create_add_on_incoming_edge): New.
15315 (create_phi_basis): New.
15316 (replace_dependents): Delete.
15317 (replace_conditional_candidate): New.
15318 (phi_add_costs): New.
15319 (replace_uncond_cands_and_profitable_phis): New.
15320 (record_increment): Handle phi adjustments.
15321 (record_phi_increments): New.
15322 (record_increments): Handle phi adjustments.
15323 (phi_incr_cost): New.
15324 (lowest_cost_path): Handle phis.
15325 (total_savings): Likewise.
15326 (analyze_increments): Likewise.
15327 (ncd_with_phi): New.
15328 (ncd_of_cand_and_phis): New.
15329 (nearest_common_dominator_for_cands): Handle phi increments.
15330 (all_phi_incrs_profitable): New.
15331 (replace_profitable_candidates): Handle phi-dependent candidates.
15332 (analyze_candidates_and_replace): Likewise.
15334 2013-05-03 Teresa Johnson <tejohnson@google.com>
15337 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
15338 do not exceed REG_BR_PROB_BASE.
15340 2013-05-03 Jeff Law <law@redhat.com>
15342 PR tree-optimization/57144
15343 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
15344 operand of the condition will bit into the new type when eliminating
15345 a cast feeding a condition.
15347 2013-05-03 Jakub Jelinek <jakub@redhat.com>
15349 PR rtl-optimization/57130
15350 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
15351 of COMPARE as in_code to the recursive call if needed.
15353 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
15355 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
15356 (enabled): Handle new members.
15357 * config/i386/sse.md (*vec_concatv2si): Merge from
15358 *vec_concatv2si_sse2 and vec_concatv2si_sse.
15359 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
15361 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
15363 PR tree-optimization/57027
15364 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
15365 for fnms opportunity, check we got the prerequisite kind
15366 of tree / gimple before using accessor functions.
15368 2013-05-03 Richard Biener <rguenther@suse.de>
15370 * double-int.h (lshift): New overload without precision
15371 and arith argument.
15372 (operator *=, operator +=, operator -=): Move ...
15373 * double-int.c (operator *=, operator +=, operator -=): ... here
15374 and implement more efficiently.
15375 (mul_double_with_sign): Remove.
15376 (lshift_double): Adjust to take unsinged shift argument, push
15377 dispatching code to callers.
15378 (mul_double_wide_with_sign): Add early out for callers that
15379 are not interested in high parts or overflow.
15380 (lshift): New function.
15381 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
15382 dispatch code here.
15383 (lrotate, rrotate): Use logical shifts.
15384 * expr.c (get_inner_reference): Use lshift.
15385 * fixed-value.c (do_fixed_divide): Likewise.
15386 * tree-dfa.c (get_ref_base_and_extent): Likewise.
15387 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
15388 (indirect_refs_may_alias_p): Likewise.
15389 (stmt_kills_ref_p_1): Likewise.
15391 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15393 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
15395 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15397 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
15398 scalar form of FABD instruction.
15400 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
15402 * lra-constraints.c (process_alt_operands): Add checking alt
15403 number to choose the best alternative.
15405 2013-05-02 Richard Biener <rguenther@suse.de>
15407 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
15408 bitmap and its handling.
15409 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
15411 2013-05-02 Richard Biener <rguenther@suse.de>
15413 PR middle-end/57140
15414 * tree-inline.c (copy_loops): Properly handle removed loops.
15415 (copy_cfg_body): Mark destination loops for fixup if source
15416 loops needed fixup.
15418 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
15421 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
15422 generating simple_return for naked functions.
15424 2013-05-02 Martin Jambor <mjambor@suse.cz>
15426 PR middle-end/56988
15427 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
15428 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
15430 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
15431 ipa_agg_replacement_value structures.
15432 (known_aggs_to_agg_replacement_list): Likewise.
15433 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
15434 (read_agg_replacement_chain): Likewise.
15435 (ipcp_transform_function): Also check that by_ref flags match.
15437 2013-05-02 Richard Biener <rguenther@suse.de>
15439 * graphds.h (struct graph): Add obstack member.
15440 * graphds.c (new_graph): Initialize obstack and allocate
15442 (add_edge): Allocate edge from the obstack.
15443 (free_graph): Free the obstack instead of all edges and vertices.
15445 2013-05-02 Teresa Johnson <tejohnson@google.com>
15447 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
15449 * cfg.c (update_bb_profile_for_threading): Ditto.
15450 * tree-inline.c (copy_bb): Ditto.
15451 (copy_edges_for_bb): Ditto.
15452 (initialize_cfun): Ditto.
15453 (copy_cfg_body): Ditto.
15454 (expand_call_inline): Ditto.
15455 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
15456 (estimate_node_size_and_time): Ditto.
15457 (inline_merge_summary): Ditto.
15458 * cgraphclones.c (cgraph_clone_edge): Ditto.
15459 (cgraph_clone_node): Ditto.
15460 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15461 (compute_trg_info): Ditto.
15463 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15465 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
15466 S reg when fp attribute set.
15467 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
15469 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15471 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
15473 (*and_one_cmplsi3_compare0_uxtw): Likewise.
15474 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
15475 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
15477 2013-05-02 Richard Biener <rguenther@suse.de>
15479 * tree-scalar-evolution.c (scev_info_hasher): Remove.
15480 (struct instantiate_cache_entry): New type.
15481 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
15482 (struct instantiate_cache_type): New type.
15483 (set_instantiated_value, get_instantiated_value): Remove.
15484 (get_instantiated_value_entry): New function.
15485 (instantiate_scev_name): Use the new cache and adjust.
15486 (instantiate_scev_poly): Adjust.
15487 (instantiate_scev_binary): Likewise.
15488 (instantiate_array_ref): Likewise.
15489 (instantiate_scev_convert): Likewise.
15490 (instantiate_scev_not): Likewise.
15491 (instantiate_scev_3): Likewise.
15492 (instantiate_scev_2): Likewise.
15493 (instantiate_scev_r): Likewise.
15494 (instantiate_scev): Likewise.
15495 (resolve_mixers): Likewise.
15497 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
15500 * lra-constraints.c (best_small_class_operands_num): Remove.
15501 (process_alt_operands): Remove small_class_operands_num. Take
15502 small classes operands into losers and only if the operand is not
15503 matched. Modify debugging output.
15504 (curr_insn_transform): Remove best_small_class_operands_num.
15507 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15509 * config/aarch64/aarch64-builtins.c
15510 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
15511 * config/aarch64/aarch64-simd-builtins.def
15512 (reduc_splus_): Add new modes.
15513 (reduc_uplus_): New.
15514 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
15515 (reduc_uplus_v4sf): Likewise.
15516 (reduc_splus_v4sf): Likewise.
15517 (aarch64_addv<mode>): Likewise.
15518 (reduc_uplus_<mode>): Likewise.
15519 (reduc_splus_<mode>): Likewise.
15520 (aarch64_addvv2di): Likewise.
15521 (reduc_uplus_v2di): Likewise.
15522 (reduc_splus_v2di): Likewise.
15523 (aarch64_addvv2si): Likewise.
15524 (reduc_uplus_v2si): Likewise.
15525 (reduc_splus_v2si): Likewise.
15526 (reduc_<sur>plus_<mode>): New.
15527 (reduc_<sur>plus_v2di): Likewise.
15528 (reduc_<sur>plus_v2si): Likewise.
15529 (reduc_<sur>plus_v4sf): Likewise.
15530 (aarch64_addpv4sf): Likewise.
15531 * config/aarch64/arm_neon.h
15532 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
15533 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
15534 add UNSPEC_SADDV, UNSPEC_UADDV.
15536 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
15538 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15540 * config/aarch64/arm_neon.h
15541 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
15543 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15545 * config/aarch64/aarch64-builtins
15546 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
15548 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15550 * config/aarch64/aarch64-simd-builtins.def
15551 (reduc_smax_): New.
15552 (reduc_smin_): Likewise.
15553 (reduc_umax_): Likewise.
15554 (reduc_umin_): Likewise.
15555 (reduc_smax_nan_): Likewise.
15556 (reduc_smin_nan_): Likewise.
15559 (smax): Update for V2SF, V4SF and V2DF modes.
15562 (smin_nan): Likewise.
15563 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
15564 (<su><maxmin><mode>3): ...This, refactor.
15565 (s<maxmin><mode>3): New.
15566 (<maxmin_uns><mode>3): Likewise.
15567 (reduc_<maxmin_uns>_<mode>): Refactor.
15568 (reduc_<maxmin_uns>_v4sf): Likewise.
15569 (reduc_<maxmin_uns>_v2si): Likewise.
15570 (aarch64_<fmaxmin><mode>: Remove.
15571 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
15573 (vmin<q>_f<32,64>): Likewise.
15574 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15576 (su): Add mappings for smax, smin, umax, umin.
15578 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15579 (FMAXMIN): Rename as...
15580 (FMAXMIN_UNS): ...This.
15582 (fmaxminv): Likewise.
15583 (fmaxmin): Likewise.
15585 (maxmin_uns_op): Likewise.
15587 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15589 * config/aarch64/arm_neon.h
15590 (vac<ge, gt><sd>_f<32, 64>): Rename to...
15591 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
15592 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
15594 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15596 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
15597 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
15599 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15601 * config/aarch64/aarch64-simd.md
15602 (vcond<mode>_internal): Handle special cases for constant masks.
15603 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
15604 (vcondu<mode><mode>): Likewise.
15605 (vcond<v_cmp_result><mode>): New.
15607 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15609 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
15610 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
15611 * config/aarch64/aarch64-simd-builtins.def
15612 (cmeq): Update to BUILTIN_VALLDI.
15617 * config/aarch64/arm_neon.h
15618 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
15619 to builtins or C as appropriate.
15621 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15623 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
15625 (cmhi): Rename to...
15627 * config/aarch64/aarch64-simd.md
15628 (simd_mode): Add SF.
15629 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
15630 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
15631 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
15632 (cstore<mode>_neg): ...This.
15633 * config/aarch64/iterators.md
15635 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
15636 (COMPARISONS): New.
15637 (UCOMPARISONS): Likewise.
15638 (optab): Add missing comparisons.
15644 (VCMP_S): Likewise.
15645 (VCMP_U): Likewise.
15646 (V_cmp_result): Add DF, SF modes.
15647 (v_cmp_result): Likewise.
15649 (vmtype): Likewise.
15650 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
15652 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
15654 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
15655 define_insn to define_insn_and_split.
15656 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
15657 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
15658 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
15659 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
15660 (thumb2_negscc): Likewise.
15662 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15664 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
15666 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15668 * config/arm/thumb2.md: Remove trailing whitespaces.
15670 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15672 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
15673 Use gen_int_mode rather than GEN_INT.
15675 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
15677 * value-prof.c (stream_in_histogram_value): Remove the strayed
15680 2013-04-30 Richard Biener <rguenther@suse.de>
15682 PR middle-end/57122
15683 * cfghooks.c (split_edge): Properly check for the loop latch edge.
15685 2013-04-30 Richard Biener <rguenther@suse.de>
15687 PR middle-end/57107
15688 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
15690 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
15692 PR rtl-optimization/56957
15693 PR rtl-optimization/57105
15694 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
15695 variable. Use just INSN_UID for determining whether an insn
15696 should be only disconnected from the insn stream.
15697 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
15699 2013-04-30 Jakub Jelinek <jakub@redhat.com>
15701 PR tree-optimization/57104
15702 * tsan.c (instrument_expr): Don't instrument accesses to
15703 DECL_HARD_REGISTER VAR_DECLs.
15705 2013-04-30 Richard Biener <rguenther@suse.de>
15707 * function.h (loops_for_fn): New inline function.
15708 (set_loops_for_fn): Likewise.
15709 * cfgloop.h (place_new_loop): Add struct function parameter.
15710 (get_loop): Likewise.
15711 (get_loops): Likewise.
15712 (number_of_loops): Likewise.
15713 (fel_next): Adjust.
15714 (fel_init): Likewise.
15715 * cfg.c (get_loop_copy): Adjust.
15716 * cfgloop.c (flow_loops_dump): Likewise.
15717 (record_loop_exits): Likewise.
15718 (verify_loop_structure): Likewise.
15719 * cfgloopanal.c (mark_irreducible_loops): Likewise.
15720 (estimate_reg_pressure_cost): Likewise.
15721 (mark_loop_exit_edges): Likewise.
15722 * cfgloopmanip.c (place_new_loop): Likewise.
15723 (add_loop): Likewise.
15724 (duplicate_loop): Likewise.
15725 * graph.c (draw_cfg_nodes): Likewise.
15726 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
15727 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
15728 (extract_affine_chrec): Likewise.
15729 (build_scop_iteration_domain): Likewise.
15730 * graphite.c (graphite_initialize): Likewise.
15731 * ira-build.c (create_loop_tree_nodes): Likewise.
15732 (more_one_region_p): Likewise.
15733 (rebuild_regno_allocno_maps): Likewise.
15734 (mark_loops_for_removal): Likewise.
15735 (mark_all_loops_for_removal): Likewise.
15736 (remove_unnecessary_regions): Likewise.
15737 (ira_build): Likewise.
15738 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15739 * loop-init.c (fix_loop_structure): Likewise.
15740 (gate_rtl_move_loop_invariants): Likewise.
15741 (gate_rtl_unswitch): Likewise.
15742 (gate_rtl_unroll_and_peel_loops): Likewise.
15743 (rtl_doloop): Likewise.
15744 * lto-streamer-in.c (input_cfg): Likewise.
15745 * lto-streamer-out.c (output_cfg): Likewise.
15746 * modulo-sched.c (sms_schedule): Likewise.
15747 * predict.c (tree_estimate_probability): Likewise.
15748 (tree_estimate_probability_driver): Likewise.
15749 (estimate_loops): Likewise.
15750 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
15751 (move_sese_region_to_fn): Likewise.
15752 (debug_loop_num): Likewise.
15753 * tree-chrec.c (chrec_evaluate): Likewise.
15754 (hide_evolution_in_other_loops_than_loop): Likewise.
15755 (chrec_component_in_loop_num): Likewise.
15756 (reset_evolution_in_loop): Likewise.
15757 (evolution_function_is_invariant_rec_p): Likewise.
15758 * tree-if-conv.c (main_tree_if_conversion): Likewise.
15759 * tree-inline.c (copy_loops): Likewise.
15760 (copy_cfg_body): Likewise.
15761 (tree_function_versioning): Likewise.
15762 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
15763 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
15765 (add_to_evolution_1): Likewise.
15766 (scev_const_prop): Likewise.
15767 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
15768 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
15769 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
15770 (tree_ssa_lim_initialize): Likewise.
15771 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
15772 (verify_loop_closed_ssa): Likewise.
15773 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
15774 (tree_ssa_loop_im): Likewise.
15775 (tree_ssa_loop_unswitch): Likewise.
15776 (tree_vectorize): Likewise.
15777 (check_data_deps): Likewise.
15778 (tree_ssa_loop_ivcanon): Likewise.
15779 (tree_ssa_loop_bounds): Likewise.
15780 (tree_complete_unroll): Likewise.
15781 (tree_complete_unroll_inner): Likewise.
15782 (tree_parallelize_loops): Likewise.
15783 (tree_ssa_loop_prefetch): Likewise.
15784 (tree_ssa_loop_ivopts): Likewise.
15785 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
15786 * tree-vectorizer.c (vectorize_loops): Likewise.
15788 2013-04-29 Mike Frysinger <vapier@gentoo.org>
15790 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
15791 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
15792 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
15793 with EABI_LINK_SPEC.
15795 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15798 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
15801 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
15804 * lra-constraints.c (process_alt_operands): Discourage a bit more
15805 using memory for pseudos. Print cost dump for alternatives.
15806 Modify cost values for conflicts with early clobbers.
15807 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
15809 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15812 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
15814 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15816 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
15817 from/to S register.
15818 (movdi_aarch64): Support LDR/STR from/to D register.
15820 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15822 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
15823 or higher by default.
15825 2013-04-29 Richard Biener <rguenther@suse.de>
15827 PR middle-end/57075
15828 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
15829 even if not adding abnormal edges for calls that can make
15832 2013-04-29 Richard Biener <rguenther@suse.de>
15834 PR middle-end/57103
15835 * tree-cfg.c (move_stmt_op): Fix condition under which to update
15837 (move_stmt_r): Remove redundant checking.
15839 2013-04-29 Teresa Johnson <tejohnson@google.com>
15842 * basic-block.h (apply_scale): New function.
15843 (apply_probability): Use apply_scale.
15844 * gimple-streamer-in.c (input_bb): Ditto.
15845 * lto-streamer-in.c (input_cfg): Ditto.
15846 * lto-cgraph.c (merge_profile_summaries): Ditto.
15847 * tree-optimize.c (execute_fixup_cfg): Ditto.
15848 * tree-inline.c (copy_bb): Update comment to use apply_scale.
15849 (copy_edges_for_bb): Ditto.
15850 (copy_cfg_body): Ditto.
15852 2013-04-29 Tom de Vries <tom@codesourcery.com>
15854 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
15855 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
15856 (tail_merge_optimize): Handle current_loops == NULL.
15858 2013-04-26 Jeff Law <law@redhat.com>
15860 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
15861 (simplify_cond_using_ranges): Generalize code to simplify
15862 COND_EXPRs where one argument is a constant and the other
15863 is an SSA_NAME created by an integral type conversion.
15865 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15867 * config/arm/arm.md (store_minmaxsi): Use only when
15868 optimize_insn_for_size_p.
15870 2013-04-29 Christian Bruel <christian.bruel@st.com>
15873 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
15875 2013-04-29 Richard Biener <rguenther@suse.de>
15877 PR middle-end/57089
15878 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
15879 loop tree make sure to schedule a fixup for the child as well.
15880 (expand_omp_for_generic): Properly add loops.
15881 (expand_omp_for_static_nochunk): Likewise.
15882 (expand_omp_for_static_chunk): Likewise.
15883 (expand_omp_for): For the degenerate case fixup loops.
15884 (expand_omp_sections): Fix default bb placement in loops.
15885 (expand_omp_atomic_pipeline): Properly add loops.
15887 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15889 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
15891 2013-04-29 Tom de Vries <tom@codesourcery.com>
15893 * tree-ssa-tail-merge.c: Update header comment.
15895 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15897 * config/aarch64/arm_neon.h
15898 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
15899 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
15900 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
15901 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
15902 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
15903 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
15904 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
15905 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
15907 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15909 * config/aarch64/aarch64-simd.md
15910 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
15911 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
15912 fix_trunc, fixuns_trunc.
15913 (ftrunc<VDQF:mode>2): New.
15914 * config/aarch64/iterators.md (optab): Add fix, fixuns.
15915 (fix_trunc_optab): New.
15917 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15919 * config/aarch64/aarch64-builtins.c
15920 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
15921 iceilf, lround, iroundf.
15923 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15926 * config/i386/i386.h (enum ix86_tune_indices)
15927 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
15928 New, split from X86_TUNE_INTER_UNIT_MOVES.
15929 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
15930 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
15931 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
15932 (TARGET_INTER_UNIT_MOVES): Remove.
15933 * config/i386/i386.c (initial_ix86_tune_features): Update.
15934 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
15935 (ix86_expand_convert_uns_didf_sse): Use
15936 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15937 (ix86_expand_vector_init_one_nonzero): Ditto.
15938 (ix86_expand_vector_init_interleave): Ditto.
15939 (inline_secondary_memory_needed): Return true for moves from SSE class
15940 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
15941 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
15942 * config/i386/constraints.md (Yi, Ym): Depend on
15943 TARGET_INTER_UNIT_MOVES_TO_VEC.
15944 (Yj, Yn): New constraints.
15945 * config/i386/i386.md (*movdi_internal): Change constraints of
15946 operand 1 from Yi to Yj and from Ym to Yn.
15947 (*movsi_internal): Ditto.
15948 (*movdf_internal): Ditto.
15949 (*movsf_internal): Ditto.
15950 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
15951 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15952 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
15953 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
15954 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
15955 * config/i386/sse.md (movdi_to_sse): Ditto.
15956 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
15957 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15958 TARGET_INTER_UNIT_MOVES.
15959 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15960 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15961 instead of TARGET_INTER_UNIT_MOVES.
15962 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15963 operand 1 from Yi to Yj and from Ym to Yn.
15965 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15967 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15968 (float_truncate_hi_): Likewise.
15969 (float_extend_lo_): Likewise.
15970 (float_truncate_lo_): Likewise.
15971 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15972 (aarch64_float_extend_lo_v2df): Likewise.
15973 (vec_unpacks_hi_v4sf): Likewise.
15974 (aarch64_float_truncate_lo_v2sf): Likewise.
15975 (aarch64_float_truncate_hi_v4sf): Likewise.
15976 (vec_pack_trunc_v2df): Likewise.
15977 (vec_pack_trunc_df): Likewise.
15979 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15981 * config/aarch64/aarch64-builtins.c
15982 (aarch64_fold_builtin): Fold float conversions.
15983 * config/aarch64/aarch64-simd-builtins.def
15984 (floatv2si, floatv4si, floatv2di): New.
15985 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15986 * config/aarch64/aarch64-simd.md
15987 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15988 * config/aarch64/iterators.md (FLOATUORS): New.
15989 (optab): Add float, floatuns.
15990 (su_optab): Likewise.
15992 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15994 * config/aarch64/aarch64-builtins.c
15995 (aarch64_builtin_vectorized_function): Use new names for
15997 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15998 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15999 (fcvtzu): Split as...
16000 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
16001 (fcvtas): Split as...
16002 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
16003 (fcvtau): Split as...
16004 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
16005 (fcvtps): Split as...
16006 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
16007 (fcvtpu): Split as...
16008 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
16009 (fcvtms): Split as...
16010 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
16011 (fcvtmu): Split as...
16012 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
16013 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
16014 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
16015 (lfrintnusf, lfrintnudf): Likewise.
16016 * config/aarch64/aarch64-simd.md
16017 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
16019 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
16020 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
16021 (fcvt_pattern): Likewise.
16023 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16025 * config/aarch64/aarch64-simd.md
16026 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
16027 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
16029 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16031 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
16032 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
16033 (vrnd<a,m,n,p>_f32): Implement using builtins.
16034 (vrnd<i,x><q>_f<32, 64>): New.
16036 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16038 * config/aarch64/aarch64-builtins.c
16039 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
16040 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
16041 (frintz): Rename to...
16043 (frintp): Rename to...
16045 (frintm): Rename to...
16047 (frinti): Rename to...
16048 (nearbyint): ...this.
16049 (frintx): Rename to...
16051 (frinta): Rename to...
16053 * config/aarch64/aarch64-simd.md
16054 (aarch64_frint<frint_suffix><mode>): Delete.
16055 (<frint_pattern><mode>2): Convert to insn.
16056 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
16057 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
16058 (frint_pattern): Likewise.
16059 (frint_suffix): Likewise.
16061 2013-04-29 Richard Biener <rguenther@suse.de>
16063 PR tree-optimization/57081
16064 * loop-init.c: Include tree-flow.h.
16065 (loop_optimizer_finalize): Free number of iteration estimates.
16066 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
16068 2013-04-29 Jakub Jelinek <jakub@redhat.com>
16070 PR tree-optimization/57083
16071 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
16072 non-singleton shift count range, zero extend low_bound for uns case.
16074 * config/i386/predicates.md (general_vector_operand): New predicate.
16075 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
16076 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
16077 if they aren't nonimmediate operands. If their original values
16078 satisfy const_vector_equal_evenodd_p, don't shift them.
16079 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
16080 predicates. For the SSE4.1 case force operands[{1,2}] into registers
16081 if not nonimmediate_operand.
16082 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
16083 instead of register_operand.
16084 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
16086 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
16088 * stor-layout.c (finalize_size_functions): Allocate a structure and
16089 reset cfun before dumping the functions.
16091 2013-04-27 Jakub Jelinek <jakub@redhat.com>
16093 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
16096 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
16097 use xop_pmacsdqh if uns_p.
16098 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
16099 the immediate rotate count.
16101 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
16103 * rtl.h (struct rtx_def): Add comment for field jump.
16104 (LRA_SUBREG_P): New macro.
16105 * recog.c (register_operand): Check LRA_SUBREG_P.
16106 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16108 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16109 (lra_final_code_change): Skip subreg change for operators.
16110 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16111 if there are no operand changes.
16112 * lra-constraints.c (curr_insn_set): New.
16113 (match_reload): Set LRA_SUBREG_P.
16114 (emit_spill_move): Ditto.
16115 (check_and_process_move): Use curr_insn_set. Process only single
16116 set insns. Don't initialize sec_mem_p and change_p.
16117 (simplify_operand_subreg): Use LRA_SUBREG_P.
16118 (reg_in_class_p): New function.
16119 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16120 of #ifdef. Add code to remove cycling.
16121 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16122 non-null disp. Reload inner instead of disp when base and index
16123 are null. Try to put lo_sum into register.
16124 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16125 (check_and_process_move): Move code for move cost check to
16126 simple_move_p. Remove equiv_substitution.
16127 (simple_move_p): New function.
16128 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16129 curr_insn_set. Call check_and_process_move only for single set
16130 insns. Use the new function. Move call of check_and_process_move
16131 after operand equiv substitution and address process.
16133 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16136 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
16137 with nonlocal goto receivers or returns twice calls, ignore
16138 unininitialized values from abnormal edges to nl goto receiver
16139 or returns twice call.
16141 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16143 PR tree-optimization/57051
16144 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
16145 and VEC_RSHIFT_EXPR if shift count is a multiple of element
16148 2013-04-26 Richard Biener <rguenther@suse.de>
16150 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
16151 (expand_omp_taskreg): Likewise. Mark loops for fixup.
16152 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
16153 (fixup_loop_arrays_after_move): New function.
16154 (move_sese_region_to_fn): Properly outline the loop tree parts
16155 of the SESE region.
16157 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
16159 * config/i386/i386.md (type, unit): Fix long lines.
16161 2013-04-26 Richard Biener <rguenther@suse.de>
16163 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
16164 (lto-streamer-out.o): Likewise.
16165 * cfgloop.c (init_loops_structure): Export, add struct function
16166 argument and adjust.
16167 (flow_loops_find): Adjust.
16168 * cfgloop.h (enum loop_estimation): Add EST_LAST.
16169 (init_loops_structure): Declare.
16170 * lto-streamer-in.c: Include cfgloop.h.
16171 (input_cfg): Input the loop tree.
16172 * lto-streamer-out.c: Include cfgloop.h.
16173 (output_cfg): Output the loop tree.
16174 (output_struct_function_base): Do not drop PROP_loops.
16176 2013-03-26 Richard Biener <rguenther@suse.de>
16178 * tree-cfg.c (execute_build_cfg): Build the loop tree.
16179 (pass_build_cfg): Provide PROP_loops.
16180 (move_sese_region_to_fn): Remove loops that are outlined into fn
16182 * tree-inline.c: Include cfgloop.h.
16183 (initialize_cfun): Do not drop PROP_loops.
16184 (copy_loops): New function.
16185 (copy_cfg_body): Copy loop structure.
16186 (tree_function_versioning): Initialize destination loop tree.
16187 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
16188 (pass_parallelize_loops): Do IL verification.
16189 * loop-init.c (loop_optimizer_init): Fixup loops if required.
16190 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
16191 the CFG make sure we fixup loops as well.
16192 * tree-ssa-tail-merge.c: Include cfgloop.h.
16193 (replace_block_by): When merging loop latches mark loops for fixup.
16194 * lto-streamer-out.c (output_struct_function_base): Drop
16195 PROP_loops for now.
16196 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
16197 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
16198 * ipa-split.c: Include cfgloop.h.
16199 (split_function): Add the new return block to the loop tree root.
16200 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
16201 whether we have removed the forwarder block.
16202 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
16203 * cfgloop.h (place_new_loop): Declare.
16204 * cfgloopmanip.c (place_new_loop): Export.
16205 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
16206 (tree-switch-conversion.o): Likewise.
16207 (tree-complex.o): Likewise.
16208 (tree-inline.o): Likewise.
16209 (tree-ssa-tailmerge.o): Likewise.
16210 (ipa-split.o): Likewise.
16211 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
16212 (tree-ssa-copy.o): Likewise.
16213 * tree-switch-conversion.c: Include cfgloop.h
16214 (process_switch): If we emit a bit-test cascade, schedule loops
16216 * tree-complex.c: Include cfgloop.h.
16217 (expand_complex_div_wide): Properly add new basic-blocks to loops.
16218 * asan.c: Include cfgloop.h.
16219 (create_cond_insert_point): Properly add new basic-blocks to
16220 loops, schedule loop fixup.
16221 * cfgloop.c (verify_loop_structure): Check that looks are not
16223 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
16225 (expand_omp_for_generic): Likewise.
16226 (expand_omp_sections): Likewise.
16227 (expand_omp_atomic_pipeline): Schedule loops for fixup.
16228 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
16229 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
16230 is initialized, not when loops are present.
16231 * tree-parloops.c (parallelize_loops): Remove checking here.
16232 * passes.c (init_optimization_passes): Schedule a copy-propagation
16233 pass before complete unrolling of inner loops.
16235 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16237 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
16238 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
16239 (colorize_init): Add argument to _WIN32 version.
16240 * toplev.c: Include diagnostic-color.h.
16241 (process_options): Default to -fdiagnostics-color=auto if
16242 GCC_COLORS env var is in the environment.
16243 * common.opt (fdiagnostics-color=): Add Var and Init.
16244 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
16245 env var is in the environment, the default is auto rather than never.
16247 * diagnostic.h (file_name_as_prefix): Add context argument.
16248 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
16249 the string as locus.
16250 * langhooks.c (lhd_print_error_function): Adjust caller.
16252 2013-04-25 Lawrence Crowl <crowl@google.com>
16254 * var-tracking.c (shared_hash_def::htab):
16255 Change type to hash_table. Update dependent calls and types.
16257 2013-04-25 Lawrence Crowl <crowl@google.com>
16259 * Makefile.in: Update as needed below.
16261 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
16262 Move declaration to after the type's method definitons.
16264 * attribs.c (htab_t scoped_attributes::attribute_hash):
16265 Change type to hash_table. Update dependent calls and types.
16267 * bitmap.c (htab_t bitmap_desc_hash):
16268 Change type to hash_table. Update dependent calls and types.
16270 * cselib.c (htab_t cselib_hash_table):
16271 Change type to hash_table. Update dependent calls and types.
16273 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
16274 (hash_string_slot_node): Move implementation into lto-streamer.h
16275 struct string_slot_hasher.
16276 (eq_string_slot_node): Likewise.
16278 * data-streamer-out.c: Update output_block::string_hash_table
16279 dependent calls and types.
16281 * dwarf2cfi.c (htab_t trace_index):
16282 Change type to hash_table. Update dependent calls and types.
16284 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
16285 Change type to hash_table. Update dependent calls and types.
16286 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
16287 (htab_t optimize_external_refs::map): Likewise.
16288 (htab_t output_comp_unit::extern_map): Likewise.
16289 (htab_t output_comdat_type_unit::extern_map): Likewise.
16290 (htab_t output_macinfo::macinfo_htab): Likewise.
16291 (htab_t optimize_location_lists::htab): Likewise.
16292 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
16294 * except.c (htab_t ehspec_hash_type):
16295 Change type to hash_table. Update dependent calls and types.
16296 (assign_filter_values::ttypes): Likewise.
16297 (assign_filter_values::ehspec): Likewise.
16298 (sjlj_assign_call_site_values::ar_hash): Likewise.
16299 (convert_to_eh_region_ranges::ar_hash): Likewise.
16301 * gcse.c (htab_t pre_ldst_table):
16302 Change type to hash_table. Update dependent calls and types.
16304 * ggc-common.c (htab_t saving_htab):
16305 Change type to hash_table. Update dependent calls and types.
16306 (htab_t loc_hash): Likewise.
16307 (htab_t ptr_hash): Likewise.
16308 (call_count): Rename ggc_call_count.
16309 (call_alloc): Rename ggc_call_alloc.
16310 (loc_descriptor): Rename make_loc_descriptor.
16311 (add_statistics): Rename ggc_add_statistics.
16313 * ggc-common.c (saving_htab):
16314 Change type to hash_table. Update dependent calls and types.
16316 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
16317 (push_gimplify_context): Likewise.
16318 (pop_gimplify_context): Likewise.
16319 (struct gimple_temp_hash_elt): Added.
16320 (struct gimplify_hasher): Likewise.
16321 (struct gimplify_ctx.temp_htab):
16322 Change type to hash_table. Update dependent calls and types.
16324 * gimple-fold.c: Include gimplify-ctx.h.
16326 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
16327 Change type to hash_table. Update dependent calls and types.
16328 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
16329 avoid potential global name collision.
16331 * gimplify.c: Include gimplify-ctx.h.
16332 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
16333 (htab_t gimplify_ctx::temp_htab):
16334 Update dependent calls and types for new type hash_table.
16335 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
16336 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
16338 * gimplify-ctx.h: New.
16339 (struct gimple_temp_hash_elt): Move from gimplify.c.
16340 (class gimplify_hasher): New.
16341 (struct gimplify_ctx): Move from gimple.h.
16342 (htab_t gimplify_ctx::temp_htab):
16343 Change type to hash_table. Update dependent calls and types.
16345 * graphite-clast-to-gimple.c: Include graphite-htab.h.
16346 (htab_t ivs_params::newivs_index):
16347 Change type to hash_table. Update dependent calls and types.
16348 (htab_t ivs_params::params_index): Likewise.
16349 (htab_t print_generated_program::params_index): Likewise.
16350 (htab_t gloog::newivs_index): Likewise.
16351 (htab_t gloog::params_index): Likewise.
16353 * graphite.c: Include graphite-htab.h.
16354 4htab_t graphite_transform_loops::bb_pbb_mapping):
16355 Change type to hash_table. Update dependent calls and types.
16357 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
16358 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
16359 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
16361 * graphite-dependences.c: Include graphite-htab.h.
16362 (loop_is_parallel_p): Change hash table type of parameter.
16364 * graphite-htab.h: New.
16365 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
16366 (extern find_pbb_via_hash): Move from graphite-poly.h.
16367 (extern loop_is_parallel_p): Move from graphite-poly.h.
16368 (extern get_loop_body_pbbs): Move from graphite-poly.h.
16370 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
16371 (extern loop_is_parallel_p): Move to graphite-htab.h.
16372 (extern get_loop_body_pbbs): Move to graphite-htab.h.
16374 * haifa-sched.c (htab_t delay_htab):
16375 Change type to hash_table. Update dependent calls and types.
16376 (htab_t delay_htab_i2): Likewise.
16378 * ira-color.c (htab_t allocno_hard_regs_htab):
16379 Change type to hash_table. Update dependent calls and types.
16381 * ira-costs.c (htab_t cost_classes_htab):
16382 Change type to hash_table. Update dependent calls and types.
16384 * loop-invariant.c (htab_t merge_identical_invariants::eq):
16385 Change type to hash_table. Update dependent calls and types.
16387 * loop-iv.c (htab_t bivs):
16388 Change type to hash_table. Update dependent calls and types.
16390 * loop-unroll.c (htab_t opt_info::insns_to_split):
16391 Change type to hash_table. Update dependent calls and types.
16392 (htab_t opt_info::insns_with_var_to_expand): Likewise.
16394 * lto-streamer.h (struct string_slot): Move from data-streamer.h
16395 (struct string_slot_hasher): New.
16396 (htab_t output_block::string_hash_table):
16397 Change type to hash_table. Update dependent calls and types.
16399 * lto-streamer-in.c (freeing_string_slot_hasher): New.
16400 (htab_t file_name_hash_table):
16401 Change type to hash_table. Update dependent calls and types.
16403 * lto-streamer-out.c: Update output_block::string_hash_table dependent
16406 * lto-streamer.c (htab_t tree_htab):
16407 Change type to hash_table. Update dependent calls and types.
16409 * omp-low.c: Include gimplify-ctx.h.
16411 * passes.c (htab_t name_to_pass_map):
16412 Change type to hash_table. Update dependent calls and types.
16413 (pass_traverse): Rename to passes_pass_traverse.
16415 * plugin.c (htab_t event_tab):
16416 Change type to hash_table. Update dependent calls and types.
16418 * postreload-gcse.c (htab_t expr_table):
16419 Change type to hash_table. Update dependent calls and types.
16420 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
16422 * sese.c (debug_rename_map_1): Make extern.
16423 (htab_t copy_bb_and_scalar_dependences::rename_map):
16424 Change type to hash_table. Update dependent calls and types.
16426 * sese.h (extern debug_rename_map): Move to .c file.
16428 * store-motion.c (htab_t store_motion_mems_table):
16429 Change type to hash_table. Update dependent calls and types.
16431 * trans-mem.c (htab_t tm_new_mem_hash):
16432 Change type to hash_table. Update dependent calls and types.
16434 * tree-browser.c (htab_t TB_up_ht):
16435 Change type to hash_table. Update dependent calls and types.
16437 * tree-cfg.c (htab_t discriminator_per_locus):
16438 Change type to hash_table. Update dependent calls and types.
16440 * tree-complex.c: Include tree-hasher.h
16441 (htab_t complex_variable_components):
16442 Change type to hash_table. Update dependent calls and types.
16444 * tree-eh.c (htab_t finally_tree):
16445 Change type to hash_table. Update dependent calls and types.
16447 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
16448 struct int_tree_hasher.
16449 (extern int_tree_map_eq): Likewise.
16450 (uid_decl_map_hash): Removed.
16451 (extern decl_tree_map_eq): Likewise.
16453 * tree-hasher.h: New.
16454 (struct int_tree_hasher): New.
16455 (typedef int_tree_htab_type): New.
16457 * tree-inline.c: Include gimplify-ctx.h.
16459 * tree-mudflap.c: Include gimplify-ctx.h.
16461 * tree-parloops.c: Include tree-hasher.h.
16462 (htab_t eliminate_local_variables_stmt::decl_address):
16463 Change type to hash_table. Update dependent calls and types.
16464 (htab_t separate_decls_in_region::decl_copies): Likewise.
16466 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
16467 Change type to hash_table. Update dependent calls and types.
16469 * tree-sra.c (candidates):
16470 Change type to hash_table. Update dependent calls and types.
16472 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
16474 (int_tree_map_hash): Likewise.
16476 * tree-ssa-dom.c (htab_t avail_exprs):
16477 Change type to hash_table. Update dependent calls and types.
16479 * tree-ssa-live.c (var_map_base_init::tree_to_index):
16480 Change type to hash_table. Update dependent calls and types.
16482 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
16483 Change type to hash_table. Update dependent calls and types.
16485 * tree-ssa-phiopt.c (seen_ssa_names):
16486 Change type to hash_table. Update dependent calls and types.
16488 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
16489 Change type to hash_table. Update dependent calls and types.
16491 * tree-ssa-uncprop.c (equiv):
16492 Change type to hash_table. Update dependent calls and types.
16494 2013-04-25 Jakub Jelinek <jakub@redhat.com>
16496 PR rtl-optimization/57003
16497 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
16498 call note_stores with kill_clobbered_value callback again after
16499 killing regs_invalidated_by_call.
16501 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16503 * config/aarch64/aarch64-simd.md
16504 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
16505 (aarch64_simd_bsl<mode>): Likewise.
16506 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
16508 2013-04-25 Marek Polacek <polacek@redhat.com>
16510 PR tree-optimization/57066
16511 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
16513 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
16515 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
16517 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16519 * config/aarch64/aarch64-builtins.c
16520 (aarch64_fold_builtin): New.
16521 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
16522 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
16523 * config/aarch64/aarch64-simd-builtins.def (abs): New.
16524 * config/aarch64/arm_neon.h
16525 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
16527 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16528 Tejas Belagod <tejas.belagod@arm.com>
16530 * config/aarch64/aarch64-builtins.c
16531 (aarch64_gimple_fold_builtin): New.
16532 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
16533 * config/aarch64/aarch64-simd-builtins.def (addv): New.
16534 * config/aarch64/aarch64-simd.md (addpv4sf): New.
16535 (addvv4sf): Update.
16536 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
16538 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16540 * config/aarch64/aarch64.md
16541 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
16543 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16545 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
16546 (*ngcsi_uxtw): New pattern.
16548 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16549 Julian Brown <julian@codesourcery.com>
16551 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
16552 (TB_DREG): Add T_V4HF.
16553 (v4hf_UP): New macro.
16554 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16555 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16556 Handle initialisation of V4HF. Adjust initialisation of reinterpret
16558 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16559 (arm_vector_mode_supported_p): Handle V4HF.
16560 (arm_mangle_map): Handle V4HFmode.
16561 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
16562 * config/arm/arm_neon_builtins.def: Add entries for
16563 vcvtv4hfv4sf, vcvtv4sfv4hf.
16564 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
16565 (neon_vcvtv4hfv4sf): Likewise.
16566 * config/arm/neon-gen.ml: Handle half-precision floating point
16568 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
16569 * config/arm/arm_neon.h: Regenerate.
16570 * config/arm/neon.ml (type elts): Add F16.
16571 (type vectype): Add T_float16x4, T_floatHF.
16572 (type vecmode): Add V4HF.
16573 (type features): Add Requires_FP_bit feature.
16574 (elt_width): Handle F16.
16575 (elt_class): Likewise.
16576 (elt_of_class_width): Likewise.
16577 (mode_of_elt): Refactor.
16578 (type_for_elt): Handle F16, fix error messages.
16579 (vectype_size): Handle T_float16x4.
16580 (vcvt_sh): New function.
16581 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
16582 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
16583 (string_of_mode): Handle V4HF.
16584 * doc/arm-neon-intrinsics.texi: Regenerate.
16586 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16588 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
16589 format specifier in 'X' case.
16591 2013-04-25 Alan Modra <amodra@gmail.com>
16594 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
16595 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
16596 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
16597 Repeat for many other rotate/shift and mask patterns using subregs.
16598 Name lshiftrt insns.
16599 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
16600 on WORDS_BIG_ENDIAN.
16602 2013-04-25 Alan Modra <amodra@gmail.com>
16604 * config.gcc: Support little-endian powerpc-linux targets.
16605 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
16606 (LINK_OS_LINUX_SPEC): Define.
16607 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
16608 Preserve MASK_LITTLE_ENDIAN.
16609 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
16610 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
16611 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
16612 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
16613 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
16614 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
16615 Correct fp word order for little-endian. Don't shift toc entries
16616 smaller than a word for little-endian.
16617 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
16618 (bswapdi2 splits): Correct low-part subreg for little-endian.
16619 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
16620 low/high where such is correct only for be.
16621 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
16622 little-endian for -mcall-aixdesc.
16624 2013-04-25 Alan Modra <amodra@gmail.com>
16626 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
16627 replace_equiv_address_nv.
16629 2013-04-25 Alan Modra <amodra@gmail.com>
16631 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
16633 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16636 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16637 * rtl.h (struct rtx_def): ...
16639 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16641 PR rtl-optimizations/57046
16642 * lra-constraints (split_reg): Set up lra_risky_transformations_p
16643 for multi-reg splits.
16645 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
16647 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
16649 2013-04-24 Sterling Augustine <saugustine@google.com>
16651 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
16652 (comp_dir_string, debug_str_dwo_section): New.
16653 (DEBUG_STR_DWO_SECTION): Rename to ...
16654 (DEBUG_DWO_STR_SECTION): ... this.
16655 (DEBUG_NORM_STR_SECTION): Delete.
16656 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
16657 (DEBUG_STR_DWO_SECTION_FLAGS): New.
16658 (find_AT_string): Move most logic to ...
16659 (find_AT_string_in_table): ... here. New.
16660 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
16661 add_skeleton_AT_string. Delete logic.
16662 (output_skeleton_debug_sections): Remove call to
16663 add_top_level_skeleton_die_attrs.
16664 (add_comp_dir_attribute): Move logic to comp_dir_string.
16665 (dwarf2out_init): Initialize debug_str_dwo_section.
16666 (output_indirect_string): Call find_string_form.
16667 (output_indirect_strings): Rewrite.
16668 (prune_unused_types): Empty skeleton_debug_str_hash.
16669 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
16670 (dwarf2out_finish): Call output_indirect_strings.
16672 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16674 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
16676 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16678 * rtl.h (struct rtx_def): Add comment for field jump.
16679 (LRA_SUBREG_P): New macro.
16680 * recog.c (register_operand): Check LRA_SUBREG_P.
16681 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16683 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16684 (lra_final_code_change): Skip subreg change for operators.
16685 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16686 if there are no operand changes.
16687 * lra-constraints.c (curr_insn_set): New.
16688 (match_reload): Set LRA_SUBREG_P.
16689 (emit_spill_move): Ditto.
16690 (check_and_process_move): Use curr_insn_set. Process only single
16691 set insns. Don't initialize sec_mem_p and change_p.
16692 (simplify_operand_subreg): Use LRA_SUBREG_P.
16693 (reg_in_class_p): New function.
16694 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16695 of #ifdef. Add code to remove cycling.
16696 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16697 non-null disp. Reload inner instead of disp when base and index
16698 are null. Try to put lo_sum into register.
16699 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16700 (check_and_process_move): Move code for move cost check to
16701 simple_move_p. Remove equiv_substitution.
16702 (simple_move_p): New function.
16703 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16704 curr_insn_set. Call check_and_process_move only for single set
16705 insns. Use the new function. Move call of check_and_process_move
16706 after operand equiv substitution and address process.
16708 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
16710 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
16711 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
16712 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
16714 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16716 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
16718 2013-04-24 Marek Polacek <polacek@redhat.com>
16720 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
16721 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
16722 (select_loops_exit_conditions): Likewise.
16723 (number_of_iterations_for_all_loops): Likewise.
16724 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
16725 (scev_analysis): Likewise.
16727 2013-04-02 Catherine Moore <clm@codesourcery.com>
16728 Chao-ying Fu <fu@mips.com>
16730 * config/mips/micromips.md (jraddiusp): New pattern.
16731 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
16732 instruction if possible.
16734 2013-04-24 Alan Modra <amodra@gmail.com>
16736 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
16738 2013-04-24 Julian Brown <julian@codesourcery.com>
16739 Chung-Lin Tang <cltang@codesourcery.com>
16741 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
16742 dependency behavior in enumeration type DIE generation. Add TODO note
16743 to comments about future DW_FORM_sdata/udata re-work of related code.
16745 2013-04-23 Lawrence Crowl <crowl@google.com>
16747 * Makefile.in: Update as needed below.
16749 * hash-table.h (class hash_table):
16750 Correct many methods with parameter types compare_type to the correct
16751 value_type. (Correct code was unlikely to notice the change.)
16752 (hash_table::elements_with_deleted) New.
16753 (class hashtable::iterator): New.
16754 (hashtable::begin()): New.
16755 (hashtable::end()): New.
16756 (FOR_EACH_HASH_TABLE_ELEMENT): New.
16758 * statistics.c (statistics_hashes):
16759 Change type to hash_table. Update dependent calls and types.
16761 * tree-into-ssa.c (var_infos):
16762 Change type to hash_table. Update dependent calls and types.
16764 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
16765 Change type to hash_table. Update dependent calls and types.
16767 * tree-ssa-loop-im.c (struct mem_ref.refs):
16768 Change type to hash_table. Update dependent calls and types.
16770 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
16771 Change type to hash_table. Update dependent calls and types.
16773 * tree-ssa-sccvn.c (vn_tables_s::nary):
16774 Change type to hash_table. Update dependent calls and types.
16775 (vn_tables_s::phis): Likewise.
16776 (vn_tables_s::references): Likewise.
16778 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
16779 (vn_reference_eq): Update parameter and return types.
16781 * tree-ssa-structalias.c (pointer_equiv_class_table):
16782 Change type to hash_table. Update dependent calls and types.
16783 (location_equiv_class_table): Likewise.
16785 * tree-vect-data-refs.c: Consequential changes for making
16786 peeling a hash_table.
16788 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
16789 (destroy_loop_vec_info): Dependent hash_table update.
16791 * tree-vectorizer.h (peeling_htab):
16792 Change type to hash_table. Update dependent calls and types.
16794 2013-04-23 Shiva Chen <shiva0217@gmail.com>
16796 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
16797 to check the register content is equal or not.
16798 * lra-constraints.c (match_reload): Use lra_assign_reg_val
16799 to assign register content record.
16800 * lra-eliminations.c (update_reg_eliminate): Use
16801 lra_update_reg_val_offset to update register content offset.
16802 * lra-int.h (struct lra_reg): Add offset member.
16803 (lra_reg_val_equal_p): New static inline function.
16804 (lra_update_reg_val_offset): New static inline function.
16805 (lra_assign_reg_val): New static inline function.
16806 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
16807 to assign register content record.
16808 (initialize_lra_reg_info_element): Initial offset to zero.
16810 2013-04-23 Catherine Moore <clm@codesourcery.com>
16812 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
16813 operands. Record compression.
16815 2013-04-23 Xinliang David Li <davidxl@google.com>
16817 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
16819 2013-04-23 Richard Biener <rguenther@suse.de>
16821 PR middle-end/57036
16822 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
16823 parameter, only add abnormal goto edges from the copied body
16824 if the call could perform abnormal gotos.
16825 (copy_cfg_body): Adjust.
16827 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
16829 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
16831 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
16833 * coretypes.h (gimple_stmt_iterator): Add struct to make
16836 2013-04-23 Richard Biener <rguenther@suse.de>
16838 PR tree-optimization/57026
16839 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
16840 from SSA names occuring in abnormal PHI nodes.
16842 2013-04-22 Andi Kleen <ak@linux.intel.com>
16844 * lto/lto.c (print_lto_report_1): Fix LTO report names.
16846 2013-04-22 Andi Kleen <ak@linux.intel.com>
16848 * lto/lto.c (print_lto_report_1): Declare early.
16849 (read_cgraph_and_symbols): Call print_lto_report_1 early.
16851 2013-04-22 Andi Kleen <ak@linux.intel.com>
16853 * common.opt (-flto-report-wpa): Add.
16854 * doc/invoke.texi (-flto-report-wpa): Add.
16855 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
16858 2013-04-22 Xinliang David Li <davidxl@google.com>
16860 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
16861 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
16862 * Makefile.in: New dependency
16864 David Daney <ddaney.cavm@gmail.com>
16866 * configure.ac (gcc_cv_as_micromips_support): Use the
16867 --fatal-warnings option.
16868 * configure: Regenerate.
16870 2013-04-22 Marek Polacek <polacek@redhat.com>
16873 * tsan.c (instrument_expr): Don't instrument expression
16874 in case its size is zero.
16876 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
16880 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16882 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16884 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16886 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
16887 (gimple_stmt_iterator): New typedef.
16888 * gimple.h (gimple_stmt_iterator): Rename to...
16889 (gimple_stmt_iterator_d): ... This.
16890 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
16891 trees be valid for GIMPLE and GENERIC.
16892 (TARGET_GIMPLE_FOLD_BUILTIN): New.
16893 * gimple-fold.c (gimple_fold_call): Call target hook
16894 gimple_fold_builtin.
16895 * hooks.c (hook_bool_gsiptr_false): New.
16896 * hooks.h (hook_bool_gsiptr_false): New.
16897 * target.def (fold_stmt): New.
16898 * doc/tm.texi: Regenerate.
16900 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
16903 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
16904 a set sp if no stack realignment.
16906 2013-04-22 Nick Clifton <nickc@redhat.com>
16908 * config.gcc (tilegx-linux): Extend extra_objs rather than
16910 (tilepro-linux): Likewise.
16912 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16914 * config/aarch64/aarch64-builtins.c
16916 (CF0, CF1, CF2, CF3, CF4, CF10): New.
16917 (VAR<1-12>): Add MAP parameter.
16918 (BUILTIN_*): Likewise.
16919 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
16920 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
16921 (aarch64_ushl_n<mode>): Likewise.
16922 (aarch64_sshr_n<mode>): Likewise.
16923 (aarch64_ushr_n<mode>): Likewise.
16924 (aarch64_<maxmin><mode>): Likewise.
16925 (aarch64_sqrt<mode>): Likewise.
16926 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
16927 (vshr<q>_n_*): Likewise.
16929 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16931 * config/aarch64/aarch64-builtins.c
16932 (aarch64_simd_builtin_type_mode): Handle SF types.
16934 (BUILTIN_GPF): Define.
16935 (aarch64_init_simd_builtins): Handle SF types.
16936 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
16937 (frecps): Likewise.
16938 (frecpx): Likewise.
16939 * config/aarch64/aarch64-simd.md
16940 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
16941 (aarch64_frecpe<mode>): New.
16942 (aarch64_frecps<mode>): Likewise.
16943 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
16944 (v8type): Add frecp<esx>.
16945 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
16946 (aarch64_frecps<mode>): Likewise.
16947 * config/aarch64/iterators.md (FRECP): New.
16948 (frecp_suffix): Likewise.
16949 * config/aarch64/arm_neon.h
16950 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
16952 2013-04-22 Christian Bruel <christian.bruel@st.com>
16955 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
16956 (REG_CLASS_NAMES): Idem.
16957 (REG_CLASS_CONTENTS): Idem.
16958 (REGCLASS_HAS_FP_REG): Idem.
16959 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16960 (sh_conditional_register_usage): Idem.
16962 2013-04-21 Jeff Law <law@redhat.com>
16964 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16965 (ssa_forward_propagate_and_combine): Use it.
16967 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16969 * lra.c: Update the flow chart diagram.
16971 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16973 PR rtl-optimization/56847
16974 * lra-constraints.c (process_alt_operands): Discourage alternative
16975 with non-matche doffsettable memory constraint fro memory with
16978 2013-04-19 Richard Biener <rguenther@suse.de>
16980 PR tree-optimization/56982
16981 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16983 * gimplify.c (gimplify_call_expr): Notice special calls.
16984 (gimplify_modify_expr): Likewise.
16985 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16986 abnormal control flow receivers.
16987 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16988 in the same way as cfun->has_nonlocal_labels.
16989 (gimple_purge_dead_abnormal_call_edges): Likewise.
16990 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16991 receivers start a basic-block.
16993 2013-04-19 Richard Biener <rguenther@suse.de>
16995 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16997 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16998 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16999 (SLP_TREE_LOAD_PERMUTATION): Add.
17000 (vect_transform_slp_perm_load): Adjust prototype.
17001 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
17002 (vect_free_slp_instance): Likewise.
17003 (vect_create_new_slp_node): Likewise.
17004 (vect_supported_slp_permutation_p): Remove.
17005 (vect_slp_rearrange_stmts): Adjust.
17006 (vect_supported_load_permutation_p): Likewise. Inline
17007 vect_supported_slp_permutation_p here.
17008 (vect_analyze_slp_instance): Compute load permutations per
17009 slp node instead of per instance.
17010 (vect_get_slp_defs): Adjust.
17011 (vect_transform_slp_perm_load): Likewise.
17012 (vect_schedule_slp_instance): Remove redundant code.
17013 (vect_schedule_slp): Remove hack for PR56270, add it ...
17014 * tree-vect-stmts.c (vectorizable_load): ... here, do not
17015 CSE loads for SLP. Adjust.
17017 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17019 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
17020 spelling in two comments.
17022 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17025 * config/arm/arm.c (load_multiple_sequence): Require SP
17026 as base register for loads if SP is in the register list.
17028 2013-04-19 Martin Jambor <mjambor@suse.cz>
17030 PR tree-optimization/56718
17031 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
17032 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
17033 and made public. Adjusted all callers.
17034 (ipa_intraprocedural_devirtualization): New function.
17035 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
17036 (ipa_intraprocedural_devirtualization): Likewise.
17037 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
17039 2013-04-19 Richard Biener <rguenther@suse.de>
17041 PR tree-optimization/57000
17042 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
17044 2013-04-19 Terry Guo <terry.guo@arm.com>
17046 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
17048 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
17049 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
17050 (cortex_m4_fmacs): Use new reservations.
17051 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
17053 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
17055 PR rtl-optimization/56999
17056 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
17058 (lra_coalesce): Remove split_origin_bitmap and related code.
17059 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
17060 ranges if necessary.
17062 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
17064 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
17066 (ix86_expand_call): Remove clobbered_registers array and use
17067 x86_64_ms_sysv_extra_clobbered_registers instead.
17068 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
17070 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
17072 * config/i386/i386.md (*call_rex64_ms_sysv): Use
17073 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
17074 (*call_value_rex64_ms_sysv): Ditto.
17076 2013-04-18 Cary Coutant <ccoutant@google.com>
17078 * dwarf2out.c (output_pubnames): Check die_perennial_p of
17079 parent instead of die_mark.
17081 2013-04-18 Diego Novillo <dnovillo@google.com>
17083 * gimple.c (create_gimple_tmp): New.
17084 (get_expr_type): New.
17085 (build_assign): New.
17086 (build_type_cast): New.
17087 * gimple.h (enum ssa_mode): Define.
17088 (gimple_seq_set_location): New.
17089 * asan.c (build_check_stmt): Change some gimple_build_* calls
17090 to use build_assign and build_type_cast.
17092 2013-04-18 Richard Biener <rguenther@suse.de>
17094 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
17095 handle negative step. Remove redundant checks.
17096 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
17097 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
17098 for negative step and grouped loads fail to vectorize.
17100 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17102 * emit-rtl.c (reset_insn_used_flags): New function.
17103 (reset_all_used_flags): Use it.
17104 (verify_insn_sharing): New function.
17105 (verify_rtl_sharing): Fix verification for SEQUENCEs.
17107 2013-04-18 Jakub Jelinek <jakub@redhat.com>
17109 PR tree-optimization/56984
17110 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
17111 and (x >> M) >= N don't register any assertion if N << M is the
17114 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17116 * lower-subreg.c (resolve_simple_move): If called self-recursive,
17117 do not delete_insn insns that have not yet been emitted, only
17118 unlink them with remove_insn.
17119 * df-scan.c (df_insn_delete): Revert r197492.
17121 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
17123 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
17124 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
17126 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17128 * config/arm/arm.md (movsicc_insn): Convert define_insn into
17129 define_insn_and_split.
17130 (and_scc,ior_scc,negscc): Likewise.
17131 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
17133 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17135 * config/arm/arm.c (use_return_insn): Return 0 for targets that
17136 can benefit from using a sequence of LDRD instructions in epilogue
17137 instead of a single LDM instruction.
17139 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
17142 * doc/extend.texi: Fix typo.
17144 2013-04-17 Richard Biener <rguenther@suse.de>
17146 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
17147 (vect_build_slp_tree): ... here.
17148 (vect_build_slp_tree_1): Compute which stmts of the SLP group
17149 match. Remove special-casing of mismatched complex loads.
17150 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
17151 re-try the match with swapped commutative operands.
17152 (vect_supported_load_permutation_p): Remove special-casing of
17153 mismatched complex loads.
17154 (vect_analyze_slp_instance): Adjust.
17156 2013-04-17 Richard Biener <rguenther@suse.de>
17158 PR rtl-optimization/56921
17159 * cfgloop.h (struct loop): Add simple_loop_desc member.
17160 (struct niter_desc): Mark with GTY(()).
17161 (simple_loop_desc): Do not use aux field but simple_loop_desc.
17162 * loop-iv.c (get_simple_loop_desc): Likewise.
17163 (free_simple_loop_desc): Likewise.
17166 2013-04-16 Richard Biener <rguenther@suse.de>
17168 PR rtl-optimization/56921
17169 * loop-init.c (pass_rtl_move_loop_invariants): Add
17170 TODO_do_not_ggc_collect to todo_flags_finish.
17171 (pass_rtl_unswitch): Same.
17172 (pass_rtl_unroll_and_peel_loops): Same.
17173 (pass_rtl_doloop): Same.
17175 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
17177 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
17178 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
17179 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
17181 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
17182 * tree-streamer.c (record_common_node): Adjust reference in comment.
17184 2013-04-17 Terry Guo <terry.guo@arm.com>
17186 * config/arm/cortex-m4.md: Add a new bypass.
17188 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17190 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
17192 (*subs_<optab><mode>_multp2): New pattern.
17193 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17194 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17196 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17198 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
17199 (*subs_mul_imm_<mode>): New pattern.
17201 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
17204 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
17205 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
17206 (vsx_movti_32bit): Change j->wa to O->wa.
17208 2013-04-16 Richard Biener <rguenther@suse.de>
17210 PR rtl-optimization/56921
17211 * loop-init.c (pass_rtl_move_loop_invariants): Add
17212 TODO_do_not_ggc_collect to todo_flags_finish.
17213 (pass_rtl_unswitch): Same.
17214 (pass_rtl_unroll_and_peel_loops): Same.
17215 (pass_rtl_doloop): Same.
17217 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
17219 * config/arm/arm.c (emit_multi_reg_push): New declaration
17220 for an existing function.
17221 (arm_emit_strd_push): New function.
17222 (arm_expand_prologue): Used here.
17223 (arm_emit_ldrd_pop): New function.
17224 (arm_expand_epilogue): Used here.
17225 (arm_get_frame_offsets): Update condition.
17226 (arm_emit_multi_reg_pop): Add a special case for load of a single
17227 register with writeback.
17229 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
17231 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
17234 2013-04-16 Richard Biener <rguenther@suse.de>
17236 PR tree-optimization/56756
17237 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
17238 (first_mem_ref_loc): New.
17239 (execute_sm): Place the load temporarily before a previous
17240 access instead of in the latch edge to ensure its SSA dependencies
17241 are defined at points dominating the load.
17243 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
17245 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
17246 correct fix by moving header and footer insn to the footer of
17247 the merged basic block. Clear BB_END of the merged-away block.
17249 PR middle-end/43631
17250 * emit-rtl.c (make_note_raw): New function.
17251 (link_insn_into_chain): New static inline function.
17252 (add_insn): Use it.
17253 (add_insn_before, add_insn_after): Factor insn chain linking code...
17254 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
17255 using link_insn_into_chain.
17256 (note_outside_basic_block_p): New helper function for emit_note_after
17257 and emit_note_before.
17258 (emit_note_after): Use nobb variant of add_insn_after if the note
17259 should not be contained in a basic block.
17260 (emit_note_before): Use nobb variant of add_insn_before if the note
17261 should not be contained in a basic block.
17262 (emit_note_copy): Use make_note_raw.
17263 (emit_note): Likewise.
17264 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
17265 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
17266 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
17267 the moved barrier the tail of the basic block it follows.
17268 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
17270 2013-04-15 Jakub Jelinek <jakub@redhat.com>
17272 PR tree-optimization/56962
17273 * gimple-ssa-strength-reduction.c (record_increment): Only set
17274 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
17275 either rhs1 or rhs2 is equal to c->base_expr.
17277 2013-04-15 Richard Biener <rguenther@suse.de>
17279 PR tree-optimization/56933
17280 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
17282 (GROUP_READ_WRITE_DEPENDENCE): Remove.
17283 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
17284 * tree-vect-data-refs.c (vect_analyze_group_access): Move
17285 dependence check ...
17286 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
17288 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
17289 GROUP_READ_WRITE_DEPENDENCE.
17291 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17293 * emit-rtl.c (reset_all_used_flags): New function.
17294 (verify_rtl_sharing): Call reset_all_used_flags before and after
17295 performing the checks.
17297 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17299 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
17300 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
17301 * config/arm/constraints.md (De): New constraint.
17302 * config/arm/neon.md (anddi3_neon): Delete.
17303 (neon_vand<mode>): Expand to standard anddi3 pattern.
17304 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
17305 Move earlier in the file.
17306 (neon_inv_logic_op2): Likewise.
17307 (arm_anddi_operand_neon): New predicate.
17309 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17311 * configure.ac (gcc_cv_ld_as_needed): Set
17312 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
17313 Use -z ignore, -z record on *-*-solaris2*.
17314 (HAVE_LD_AS_NEEDED): Update comment.
17315 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
17316 * configure: Regenerate.
17317 * config.in: Regenerate.
17318 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
17319 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
17320 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
17321 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
17322 equivalents. Fix markup.
17323 * doc/tm.texi: Regenerate.
17325 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
17327 * config/i386/i386.opt: New option mstack-protector-guard=.
17328 * config/i386/i386-opts.h: Add enum stack_protector_guard.
17329 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
17330 TARGET_SSP_TLS_GUARD.
17331 * config/i386/i386.c (ix86_option_override_internal): Set
17332 ix86_stack_protector_guard.
17333 * config/i386/i386.md (stack_protect_set): Enable for
17334 TARGET_SSP_TLS_GUARD only.
17335 (stack_protect_set_<mode>): Ditto.
17336 (stack_protect_test): Ditto.
17337 (stack_protect_test_<mode>): Ditto.
17338 * doc/invoke.texi (i386 Option): Document.
17340 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
17343 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
17344 (S_MODES): Set H_MODE bit.
17345 (SF_MODES): Set only S_MODE and SF_MODE bits.
17346 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
17347 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
17348 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
17349 <MODE_FLOAT>: Likewise.
17351 2013-04-15 Joey Ye <joey.ye@arm.com>
17353 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
17355 2013-04-15 Joey Ye <joey.ye@arm.com>
17357 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
17359 (thumb_far_jump_used_p): Count instruction size and set
17362 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
17364 * reorg.c (fill_simple_delay_slots): Reindent block of code.
17365 * resource.c (mark_target_live_regs): Reformat conditional block.
17367 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17369 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
17370 notes, they are emitted only just before final.
17371 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
17373 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17375 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
17376 * cfgrtl.c (delete_insn): Call it here instead.
17377 * lra-spills.c (lra_final_code_change): Use delete_insn.
17378 * haifa-sched.c (sched_remove_insn): Likewise.
17379 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
17380 returning to the nop pool.
17381 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
17382 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
17384 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17386 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
17387 * doc/tm.texi: Regenerated.
17389 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
17391 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
17394 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17396 * df-core.c (df_find_def): Compare register numbers.
17397 (df_find_use): Likewise.
17399 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
17402 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
17403 lra_in_progress for return.
17405 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17407 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
17408 define_insn into define_insn_and_split and emit movsicc patterns.
17410 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17412 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
17414 2013-04-12 Richard Biener <rguenther@suse.de>
17416 * tree-pass.h (TODO_do_not_ggc_collect): New.
17417 * passes.c (execute_one_ipa_transform_pass): Honor
17418 TODO_do_not_ggc_collect.
17419 (execute_one_pass): Likewise.
17422 2013-04-10 Richard Biener <rguenther@suse.de>
17424 * passes.c (init_optimization_passes): Remove reload pass.
17425 * ira.c (do_reload): Merge into ...
17427 (rest_of_handle_reload): Remove.
17428 (pass_reload): Likewise.
17429 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17430 of reload for vzeroupper pass placement.
17432 2013-04-12 Jakub Jelinek <jakub@redhat.com>
17434 PR tree-optimization/56918
17435 PR tree-optimization/56920
17436 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
17437 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
17438 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
17439 use wide_mul_with_sign method.
17441 2013-04-12 Richard Biener <rguenther@suse.de>
17443 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
17444 not be considered a gimple constant.
17446 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
17448 * fold-const.c (const_binop): Handle vector shifts by a scalar.
17449 (fold_binary_loc): Call const_binop also for mixed vector-scalar
17452 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
17453 Jakub Jelinek <jakub@redhat.com>
17455 * opts.c: Include diagnostic-color.h.
17456 (common_handle_option): Handle OPT_fdiagnostics_color_.
17457 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
17458 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
17459 (diagnostic-color.o): New.
17460 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
17461 (diagnostic_color_rule): New enum.
17462 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
17463 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
17464 the location string.
17465 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
17466 either NULL, or color kind.
17467 * diagnostic-color.c: New file.
17468 * diagnostic-color.h: New file.
17469 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
17471 * doc/invoke.texi (-fdiagnostics-color): Document.
17472 * pretty-print.h (pp_show_color): Define.
17473 (struct pretty_print_info): Add show_color field.
17474 * diagnostic.c: Include diagnostic-color.h.
17475 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
17476 macros. Colorize error:, warning: etc. strings and also the location
17478 (diagnostic_show_locus): Colorize the caret line.
17479 * pretty-print.c: Include diagnostic-color.h.
17480 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
17481 inside of %< %> quotes or quoted through q format modifier.
17483 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17485 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
17487 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17489 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
17490 code in CC_NZ mode.
17491 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
17494 2013-04-11 Marek Polacek <polacek@redhat.com>
17496 PR tree-optimization/48184
17497 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
17499 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
17501 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
17502 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
17503 (skip_simple_arithmetic): Tidy up.
17504 * tree.h (skip_simple_constant_arithmetic): Declare.
17506 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17508 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
17510 2013-04-11 Richard Biener <rguenther@suse.de>
17512 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17513 generate vector constants.
17515 2013-04-11 Richard Biener <rguenther@suse.de>
17517 PR tree-optimization/56878
17518 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
17519 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
17521 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
17522 Prefer to align the DR with the most invariant base address.
17524 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17526 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
17529 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
17531 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
17532 floating-point vector comparisons against 0.
17534 2013-04-11 Jakub Jelinek <jakub@redhat.com>
17536 PR tree-optimization/56899
17537 * fold-const.c (extract_muldiv_1): Apply distributive law
17538 only if TYPE_OVERFLOW_WRAPS (ctype).
17540 2013-04-11 Bin Cheng <bin.cheng@arm.com>
17543 * ira-costs.c (scan_one_insn): Check whether the source rtx of
17544 loading has side effect.
17546 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
17548 * config/sparc/sparc.c: Include tree-pass.h.
17549 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
17550 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
17551 head of file. Change return type. Split off gate function.
17552 (sparc_gate_work_around_errata): New function.
17553 (pass_work_around_errata): New pass definition.
17554 (insert_pass_work_around_errata) New pass insert definition to
17555 insert pass_work_around_errata just after delayed-branch scheduling.
17556 (sparc_option_override): Insert the pass.
17557 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
17559 2013-04-10 David S. Miller <davem@davemloft.net>
17561 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
17562 or -mcpu=hypersparc.
17564 * target.def (cstore_mode): New hook.
17565 * target.h: Include insn-codes.h
17566 * targhooks.c: Likewise.
17567 (default_cstore_mode): New function.
17568 * targhooks.h: Declare it.
17569 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
17570 * doc/tm.texi: Rebuild.
17571 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
17572 target hook, rather than inspecting the insn_data.
17573 * config/sparc/sparc.c (sparc_cstore_mode): New function.
17574 (TARGET_CSTORE_MODE): Redefine.
17575 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
17577 * config/sparc/predicates.md (cstore_result_operand): New special
17579 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
17580 Use it for operand 0.
17581 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
17582 (*snesi_special): Likewise.
17583 (*snesi_zero): Likewise.
17584 (*seqsi_zero): Likewise.
17585 (*sltu_insn): Likewise.
17586 (*sgeu_insn): Likewise.
17587 (*seqdi_special): Make operand 0 and comparison operation be of
17589 (*snedi_special): Likewise.
17590 (*snedi_special_vis3): Likewise.
17591 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
17592 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
17593 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
17594 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
17595 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
17596 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
17597 (*sltu_extend_sp64): Likewise.
17598 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
17599 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
17600 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
17601 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
17602 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
17604 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
17606 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
17607 (aarch64_start_file): Use the new function.
17609 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17610 Jason Merrill <jason@redhat.com>
17612 * common.opt: Add -gdwarf.
17613 * opts.c (common_handle_option): Handle it.
17614 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
17616 2013-04-10 Richard Biener <rguenther@suse.de>
17618 * passes.c (execute_todo): Do not call ggc_collect conditional here.
17619 (execute_one_ipa_transform_pass): But unconditionally here.
17620 (execute_one_pass): And here.
17621 (init_optimization_passes): Remove reload pass.
17622 * tree-pass.h (TODO_ggc_collect): Remove.
17623 (pass_reload): Likewise.
17624 * ira.c (do_reload): Merge into ...
17626 (rest_of_handle_reload): Remove.
17627 (pass_reload): Likewise.
17628 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17629 of reload for vzeroupper pass placement.
17630 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
17631 and todo_flags_finish of all passes.
17633 2013-04-10 Richard Biener <rguenther@suse.de>
17635 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
17636 first_const_oprnd field, rename first_def_type to first_op_type.
17637 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
17638 (vect_get_and_check_slp_defs): Always use the type of the
17639 operand. Allow mixed vect_external_def, vect_constant_def types.
17640 (vect_get_constant_vectors): Handle mixed vect_external_def,
17641 vect_constant_def types.
17643 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
17645 PR tree-optimization/55524
17646 * tree-ssa-math-opts.c
17647 (convert_mult_to_fma): Don't use an fms construct
17648 when we don't have an fms operation, but fnma, and it looks
17649 likely that we'll be able to use the latter.
17651 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
17653 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
17655 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
17656 inline fail caused by overwritable functions.
17658 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
17660 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
17661 unnecessary bits in the constant power of two case.
17663 2013-04-10 Richard Biener <rguenther@suse.de>
17665 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
17666 broken code swapping operands.
17667 (vect_build_slp_tree): Do not compute load permutations here.
17668 (vect_analyze_slp_instance): Compute load permutations here,
17669 after building the SLP tree.
17671 2013-04-09 Christian Bruel <christian.bruel@st.com>
17673 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
17674 of next/prev_real_insn.
17676 2013-04-09 Jan Hubicka <jh@suse.cz>
17678 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17679 Drop aliased parameter.
17680 (function_and_variable_visibility): Do not handle alias pairs.
17681 * cgraph.c (varpool_externally_visible_p): Update prototype.
17682 * varpool.c (varpool_add_new_variable): Update.
17684 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17686 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
17688 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
17690 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
17692 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
17694 2013-04-09 Marek Polacek <polacek@redhat.com>
17696 PR tree-optimization/48762
17697 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
17699 2013-04-09 Richard Biener <rguenther@suse.de>
17701 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
17703 (vect_build_slp_tree): Likewise.
17704 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
17705 calculating the cost of a SLP instance.
17706 (vect_analyze_slp_instance): Use it from here, after building
17709 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17711 PR middle-end/56883
17712 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
17713 expand_omp_for_static_chunk): Use simple_p = true in
17714 force_gimple_operand_gsi calls when assigning to addressable decls.
17716 2013-04-09 Jeff Law <law@redhat.com>
17718 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
17719 when the boolean was created by converting a wider object which
17720 had a boolean range.
17722 2013-04-09 Richard Biener <rguenther@suse.de>
17724 * tree-vectorizer.h (slp_void_p): Remove.
17725 (slp_tree): Typedef before _slp_tree declaration.
17726 (struct _slp_tree): Use a vector of slp_tree as children.
17727 (vect_get_place_in_interleaving_chain): Remove.
17728 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
17730 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
17732 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
17733 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
17734 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
17735 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
17736 Use slp_node instead of slp_void_p and adjust.
17738 2013-04-09 Richard Biener <rguenther@suse.de>
17740 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
17741 work that is not necessary.
17743 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17745 PR tree-optimization/56854
17746 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
17747 forward into clobber stmts if it would change MEM_REF lhs into
17750 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
17752 * tree.c (type_hash_lookup, type_hash_add): Make static.
17753 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
17755 2013-04-09 Richard Biener <rguenther@suse.de>
17757 * tree.h (unsave_expr_now): Remove.
17758 * tree-inline.c (mark_local_for_remap_r): Remove.
17759 (unsave_expr_1): Likewise.
17760 (unsave_r): Likewise.
17761 (unsave_expr_now): Likewise.
17762 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
17763 (propagate_tree_value): Likewise.
17765 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
17767 * doc/rtl.texi (sequence): Rewrite documentation to match the
17768 current use of SEQUENCE rtl objects.
17769 * rtl.def (SEQUENCE): Likewise.
17771 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
17772 Update documentation.
17773 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
17774 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
17776 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
17778 2013-04-08 Teresa Johnson <tejohnson@google.com>
17780 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
17781 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
17783 (estimate_edge_size_and_time): Add comment to suggest using rounding
17785 (estimate_node_size_and_time): Ditto.
17786 (remap_edge_change_prob): Use helper rounding divide methods.
17787 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
17788 (gimple_mod_pow2_value_transform): Ditto.
17789 (gimple_mod_subtract_transform): Ditto.
17790 (gimple_ic_transform): Ditto.
17791 (gimple_stringops_transform): Ditto.
17792 * stmt.c (conditional_probability): Ditto.
17793 (emit_case_dispatch_table): Ditto.
17794 * lto-cgraph.c (merge_profile_summaries): Ditto.
17795 * tree-optimize.c (execute_fixup_cfg): Ditto.
17796 * cfgcleanup.c (try_forward_edges): Ditto.
17797 * cfgloopmanip.c (scale_loop_profile): Ditto.
17799 (duplicate_loop_to_header_edge): Ditto.
17800 (lv_adjust_loop_entry_edge): Ditto.
17801 * tree-vect-loop.c (vect_transform_loop): Ditto.
17802 * profile.c (compute_branch_probabilities): Ditto.
17803 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
17804 * lto-streamer-in.c (input_cfg): Ditto.
17805 * gimple-streamer-in.c (input_bb): Ditto.
17806 * ipa-cp.c (update_profiling_info): Ditto.
17807 (update_specialized_profile): Ditto.
17808 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
17809 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
17811 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17812 (compute_trg_info): Ditto.
17813 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
17814 (purge_dead_edges): Ditto.
17815 * loop-unswitch.c (unswitch_loop): Ditto.
17816 * cgraphclones.c (cgraph_clone_edge): Ditto.
17817 (cgraph_clone_node): Ditto.
17818 * tree-inline.c (copy_bb): Ditto.
17819 (copy_edges_for_bb): Ditto.
17820 (initialize_cfun): Ditto.
17821 (copy_cfg_body): Ditto.
17822 (expand_call_inline): Ditto.
17824 2013-04-08 Kai Tietz <ktietz@redhat.com>
17826 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
17827 TARGET_CYGWIN64 by TARGET_64BIT.
17829 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17831 * config/epiphany/epiphany.md (GPR_1): New constant.
17832 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
17833 * config/epiphany/epiphany.c (gen_compare_reg):
17834 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
17835 is already in place.
17836 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
17837 Don't require being called during rtl expansion; If y operlaps r0,
17839 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
17840 (epiphany_expand_epilogue): Likewise.
17842 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
17843 Don't use CC_FPmode for ORDERED / UNORDERED.
17844 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
17846 * config/epiphany/constraints.md (CnL): New constraint.
17847 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
17848 * config/epiphany/predicates.md (add_operand): Allow 1024.
17850 * config/epiphany/epiphany.md (logical_op): New code iterator.
17851 (op_mnc): New code attribute.
17852 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
17853 (mov_f+1, mov_f+2): New peephole2 patterns.
17855 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
17856 (cstoresi4): Also allow re-use of zero result when doing a NE
17857 comparison to a non-zero operand.
17858 Use (clobber (scratch)) for first insn if the gpr output is not needed.
17860 * config/epiphany/epiphany.md (<insn_opname>v2si3):
17861 Use gen_addsi3_i / gen_subsi3_i.
17863 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17867 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
17868 contain anything but clobbers, at most one __builtin_stack_restore,
17869 optionally debug stmts and final resx, and if it has at least one
17870 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
17871 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
17872 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
17873 which isn't defaut definition, remove them.
17874 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
17875 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
17876 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
17877 with MEM_REF LHS with SSA_NAME address.
17879 2013-04-08 Jeff Law <law@redhat.com>
17881 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
17883 2013-04-08 Richard Biener <rguenther@suse.de>
17885 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
17887 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
17888 determined vector type.
17889 (vect_analyze_data_refs): Likewise.
17890 (vect_get_new_vect_var): Adjust.
17891 (vect_create_destination_var): Preserve SSA name versions.
17892 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
17893 not dump anything here.
17895 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17897 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
17898 Add member lr_slot_known.
17899 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
17901 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
17902 Remove code that sets lr_slot_offset according to what a previous
17903 version of epiphany_emit_save_restore used to do.
17904 (epiphany_emit_save_restore): When doing an lr save or restore,
17905 set/verify lr_slot_known and lr_slot_offset.
17907 2013-04-08 Xinyu Qi <xyqi@marvell.com>
17910 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
17913 2013-04-08 Richard Biener <rguenther@suse.de>
17915 * alias.c (find_base_term): Fix thinko in previous change.
17917 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17919 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
17920 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
17921 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
17922 if possible to compute val.
17923 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
17924 For QImode integers don't require anything about precision. Use
17925 const_with_all_bytes_same to find out if the constant doesn't have
17926 repeated bytes in it.
17928 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17930 * config/s390/s390.c (s390_expand_insv): Only accept insertions
17933 2013-04-08 Marek Polacek <polacek@redhat.com>
17935 PR rtl-optimization/48182
17936 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
17939 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17942 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
17943 nuses, make sure we have a label.
17945 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17948 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
17949 (rs6000_emit_swdiv_low_precision): Remove.
17950 (rs6000_emit_swdiv): Rewrite to handle between one and four
17951 iterations of Newton-Raphson generally; modify required number of
17952 iterations for some cases.
17953 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
17955 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
17957 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17958 set-but-unused variable.
17960 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17961 basic blocks of released function bodies garbage-collectable.
17963 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17964 (struct rtl_opt_pass): Add TODO_df_finish.
17966 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17968 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17970 * config/arm/constraints.md (q): New constraint.
17971 * config/arm/ldrdstrd.md: New file.
17972 * config/arm/arm.md (ldrdstrd.md) New include.
17973 (arm_movdi): Use "q" instead of "r" constraint
17974 for double-word memory access.
17975 (movdf_soft_insn): Likewise.
17976 * config/arm/vfp.md (movdi_vfp): Likewise.
17977 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17978 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17979 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17980 (mem_ok_for_ldrd_strd): Likewise.
17981 (output_move_double): Update assertion.
17983 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17985 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17987 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17989 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17990 define_insn_and_split.
17991 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17993 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17995 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17996 define_insn_and_split.
17997 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17998 (shiftsi3_compare): New pattern.
17999 (rrx): New pattern.
18000 * config/arm/unspecs.md (UNSPEC_RRX): New.
18002 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18004 * config/arm/arm.md (negdi_extendsidi): New pattern.
18005 (negdi_zero_extendsidi): Likewise.
18007 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18009 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
18010 define_insn_and_split.
18011 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
18012 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
18014 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18016 * config/arm/arm.md (arm_subdi3): Convert define_insn into
18017 define_insn_and_split.
18018 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
18019 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
18021 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18023 * config/arm/arm.md (subsi3_carryin): New pattern.
18024 (subsi3_carryin_const): Likewise.
18025 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
18026 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
18028 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18030 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
18032 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18034 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
18035 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
18037 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18039 * config/arm/arm.c (arm_expand_builtin): Change fcode
18040 type to unsigned int.
18042 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18044 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
18046 2013-04-04 Ian Lance Taylor <iant@google.com>
18048 * doc/standards.texi (Standards): The Go frontend supports the Go 1
18051 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
18053 PR middle-end/56729
18054 * df-scan.c (df_insn_delete): Disable failing assert.
18056 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18058 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
18059 New function prototype.
18060 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
18061 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
18062 (arm_builtin_vectorized_function): New function.
18064 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18066 * config/arm/arm_neon_builtins.def: New file.
18067 * config/arm/arm.c (neon_builtin_data): Move contents to
18068 arm_neon_builtins.def.
18069 (enum arm_builtins): Include neon builtin definitions.
18070 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
18071 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
18073 2013-04-04 Marek Polacek <polacek@redhat.com>
18075 PR tree-optimization/48186
18076 * predict.c (maybe_hot_frequency_p): Return false if
18077 HOT_BB_FREQUENCY_FRACTION is 0.
18078 (cgraph_maybe_hot_edge_p): Likewise.
18080 2013-04-04 Richard Biener <rguenther@suse.de>
18082 PR tree-optimization/56826
18083 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
18086 2013-04-04 Richard Biener <rguenther@suse.de>
18088 PR tree-optimization/56213
18089 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
18090 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
18092 2013-04-04 Richard Biener <rguenther@suse.de>
18094 PR tree-optimization/56837
18095 * tree-loop-distribution.c (classify_partition): For non-zero
18096 values require that the value has the same precision as its
18097 mode to be useful as memset value.
18099 2013-04-03 Nick Clifton <nickc@redhat.com>
18101 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
18102 (fmssf4): Use fmsf.s on E3V5 architectures.
18103 (fnmasf4): Use fnmaf.s on E3V5 architectures.
18104 (fnmssf4): Use fnmsf.s on E3V5 architectures.
18106 2013-04-03 Jeff Law <law@redhat.com>
18108 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
18109 (lra-eliminations.o): Likewise.
18111 2013-04-03 Teresa Johnson <tejohnson@google.com>
18113 * gcov-io.c (compute_working_sets): Moved most of body of old
18114 compute_working_sets here from profile.c.
18115 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
18116 (gcov_working_set_t): Moved typedef here from basic-block.h
18117 (compute_working_set): Declare.
18118 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
18119 (get_working_sets): Renamed from compute_working_set,
18120 replace most of body with call to new compute_working_sets.
18121 (get_exec_counts): Replace call to compute_working_sets
18122 to get_working_sets.
18123 * profile.h (get_working_sets): Renamed from compute_working_set.
18124 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
18125 to get_working_sets.
18126 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
18127 * gcov-dump.c (dump_working_sets): New function.
18129 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
18131 * hwint.c (sext_hwi, zext_hwi): New functions.
18132 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
18133 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
18134 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
18135 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
18136 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
18137 (sext_hwi, zext_hwi): New functions.
18139 2013-04-03 Jeff Law <law@redhat.com>
18141 PR tree-optimization/56799
18142 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
18143 back test for widening conversion erroneously dropped in prior change.
18145 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18148 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
18149 instead of next_real_insn.
18151 2013-04-03 Marek Polacek <polacek@redhat.com>
18154 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
18156 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18159 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
18161 (thumb1_output_casesi): Likewise.
18162 (thumb2_output_casesi): Likewise.
18164 2013-04-03 Richard Biener <rguenther@suse.de>
18166 PR tree-optimization/56817
18167 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
18169 (tree_unroll_loops_completely_1): ... new function to manually
18170 walk the loop tree, properly defering outer loops of unrolled
18171 loops to later iterations.
18173 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18175 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
18176 (vectorizable_load): Likewise.
18177 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
18178 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
18180 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18182 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
18185 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
18187 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
18189 2013-04-03 Bin Cheng <bin.cheng@arm.com>
18191 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
18193 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18195 PR tree-optimization/56790
18196 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
18199 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18201 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
18203 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
18204 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
18207 2013-04-03 Jakub Jelinek <jakub@redhat.com>
18210 * tree.h (force_folding_builtin_constant_p): New decl.
18211 * builtins.c (force_folding_builtin_constant_p): New variable.
18212 (fold_builtin_constant_p): Fold immediately also if
18213 force_folding_builtin_constant_p.
18215 2013-04-03 Richard Biener <rguenther@suse.de>
18217 PR tree-optimization/56812
18218 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
18219 DRs of the same interleaving chain are independent.
18221 2013-04-02 Jason Merrill <jason@redhat.com>
18223 * gdbinit.in (pbb): Use debug fn.
18225 2013-04-02 Lawrence Crowl <crowl@google.com>
18227 * sese.h (struct ivtype_map_elt_s): Remove unused.
18228 (extern debug_ivtype_map): Remove unused.
18229 (extern eq_ivtype_map_elts): Remove unused.
18230 * sese.c (debug_ivtype_map): Removed unused.
18231 (debug_ivtype_map_1): Removed unused.
18232 (debug_ivtype_elt): Remove unused.
18233 (eq_ivtype_map_elts): Remove unused.
18236 2013-04-02 Kai Tietz <ktietz@redhat.com>
18239 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
18240 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
18241 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
18243 (legitimize_pe_coff_symbol): Likewise.
18244 (is_imported_p): New helper-function.
18245 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
18246 for Windows x64 targets.
18247 (ix86_expand_prologue): Optimize for pe-coff targets.
18248 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
18249 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
18250 medium/large code-model.
18251 (legitimize_pic_address): Likewise.
18252 (legitimize_tls_address): Likewise.
18253 (ix86_expand_call): Likewise.
18254 (x86_output_mi_thunk): Likewise.
18255 (get_dllimport_decl): Add new beimport argument.
18256 (construct_plt_address): Don't assert for x64 pe-coff targets.
18257 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
18259 (SYMBOL_FLAG_STUBVAR): New macro.
18260 (SYMBOL_REF_STUBVAR_P): Likewise.
18261 * config/i386/winnt.c (stub_list): New structure.
18262 (stub_head): New local variable.
18263 (i386_pe_record_stub): New function.
18264 (i386_pe_file_end): Emit refptr-stubs.
18266 2013-04-02 Jakub Jelinek <jakub@redhat.com>
18268 PR rtl-optimization/56745
18269 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
18270 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
18273 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
18274 and both of them are MEM_REFs, just compare first argument for
18275 equality and attempt to deal even with differing offsets.
18278 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
18279 of gimple_clobber_p to be MEM_REF.
18280 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
18281 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
18282 after gimplification.
18283 * asan.c (get_mem_ref_of_assignment): Don't instrument
18284 gimple_clobber_p stmts.
18285 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
18286 gimple_clobber_p stmt if they have MEM_REF lhs and
18287 are dead because of another gimple_clobber_p stmt.
18288 * tree-ssa-live.c (clear_unused_block_pointer): Treat
18289 gimple_clobber_p stmts like debug stmts.
18290 (remove_unused_locals): Remove clobbers with MEM_REF lhs
18291 that refer to unused VAR_DECLs or uninitialized values.
18292 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
18293 gimple_clobber_p stmts if they refer to removed parameters.
18294 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
18297 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
18299 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
18300 using SWI48 mode attribute.
18302 2013-04-02 Wei Mi <wmi@google.com>
18304 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
18305 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
18306 *<rotate_insn><mode>3_mask in i386.md.
18308 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
18310 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
18312 2013-04-02 Richard Biener <rguenther@suse.de>
18314 PR tree-optimization/56778
18315 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
18316 Runtime alias tests are not supported for gather loads.
18317 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
18318 stmts referenced from SSA operands before updating SSA form.
18320 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
18321 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18323 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
18324 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
18325 * config/arm/cortex-a53.md: New file.
18326 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
18327 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
18328 * config/arm/arm.c (arm_issue_rate): Likewise.
18329 * config/arm/arm-tune.md: Regenerate
18330 * config/arm/arm-tables.opt: Regenerate.
18331 * config/arm/arm-cores.def: Add cortex-a53.
18333 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
18335 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
18338 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18340 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
18341 scalar load/store operations using B/H registers.
18342 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
18344 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18346 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
18348 * config/aarch64/aarch64.c
18349 (aarch64_simd_scalar_immediate_valid_for_move): New.
18350 * config/aarch64/aarch64-protos.h
18351 (aarch64_simd_scalar_immediate_valid_for_move): New.
18352 * config/aarch64/constraints.md (Dh, Dq): New.
18353 * config/aarch64/iterators.md (hq): New.
18355 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
18357 * reorg.c (get_branch_condition): Deal with conditional returns.
18358 (fill_simple_delay_slots): Remove dead code dealing with jumps.
18360 2013-04-01 Wei Mi <wmi@google.com>
18362 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
18363 Truncate operand 2 using %b asm operand modifier.
18364 (*<shift_insn><mode>3_mask): Ditto.
18365 (*<rotate_insn><mode>3_mask): Ditto.
18367 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
18369 PR middle-end/56798
18370 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
18372 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
18374 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
18376 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
18378 2013-03-30 Lawrence Crowl <crowl@google.com>
18380 * dse.c (clear_alias_sets): Remove never set.
18381 (disqualified_clear_alias_sets): Remove never set.
18382 (clear_alias_mode_pool): Remove never set.
18383 (dse_step0): Remove condition that is never true.
18384 (canon_address): Remove condition that is never true.
18385 (dse_step7): Remove condition that is never true.
18386 (rest_of_handle_dse): Remove condition that is never true.
18387 (rest_of_handle_dse::did_global): Remove never read from above.
18388 (dse_step2_spill): Remove never called from above.
18389 (dse_step5_spill): Remove never called from above.
18391 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
18393 * doc/md.texi (Standard Names) <casesi>: Update documentation for
18394 JUMP_TABLE_DATA changes.
18395 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
18396 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
18397 (Insns) <jump_table_data>: New entry.
18398 * doc/tm.texi: Regenerate.
18400 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
18402 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
18403 for table jump at the end of a basic block using tablejump_p.
18404 * targhooks.c (default_invalid_within_doloop): Likewise.
18405 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
18406 target hook implementation that is identical to the default hook.
18407 (rs6000_invalid_within_doloop): Remove.
18409 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
18410 unused variable from tablejump_p call.
18412 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
18413 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
18414 (INSN_DELETED_P): Likewise.
18415 (emit_jump_table_data): New prototype.
18416 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
18417 after 4th as unused.
18418 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
18419 * sched-vis.c (print_insn): Likewise.
18420 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
18421 insn for compatibility with back ends that use next_active_insn to
18422 identify jump table data.
18423 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
18424 (remove_insn): Likewise.
18425 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
18427 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
18428 (emit_jump_table_data): New function.
18430 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
18431 basic block, a JUMP_TABLE_DATA never is.
18432 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
18433 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
18434 off from code handling real insns.
18435 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
18436 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
18437 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
18438 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
18439 is not a NONDEBUG_INSN_P.
18440 * ira-costs.c (scan_one_insn): Likewise.
18441 * jump.c (mark_all_labels): Likewise.
18442 (mark_jump_label_1): Likewise.
18443 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18444 * lra.c (get_insn_freq): Expect all insns reaching here to be in
18446 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
18447 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
18448 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
18449 JUMP_TABLE_DATA_P insns.
18450 (calculate_elim_costs_all_insns): Likewise.
18451 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
18452 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
18453 (delete_output_reload): Code style fixups.
18454 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
18455 insn flags on this non-insn.
18456 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
18457 as scheduling barriers, for pre-change compatibility.
18458 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
18459 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
18461 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
18462 redundant JUMP_TABLE_DATA_P test.
18463 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
18464 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18465 (frv_for_each_packet): Likewise.
18466 * config/i386/i386.c (min_insn_size): Likewise.
18467 (ix86_avoid_jump_mispredicts): Likewise.
18468 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18469 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
18470 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18471 (mips16_insn_length): Robustify.
18472 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
18473 (mips16_split_long_branches): Likewise.
18474 * config/pa/pa.c (pa_combine_instructions): Likewise.
18475 * config/rs6000/rs6000.c (get_next_active_insn): Treat
18476 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
18477 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
18478 as contributing to pool range lengths.
18479 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
18480 Remove redundant JUMP_TABLE_DATA_P test.
18481 (sh_loop_align): Likewise.
18482 (split_branches): Likewise.
18483 (sh_insn_length_adjustment): Likewise.
18484 * config/spu/spu.c (get_branch_target): Likewise.
18486 2013-03-29 Jan Hubicka <jh@suse.cz>
18488 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
18489 gcov streaming; stream hot bb threshold to ltrans.
18490 * predict.c (get_hot_bb_threshold): Break out from ....
18491 (maybe_hot_count_p): ... here.
18492 (set_hot_bb_threshold): New function.
18493 * lto-section-in.c (lto_section_name): Add profile.
18494 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
18495 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
18496 and data-streamer.h
18497 (histogram_entry): New structure.
18498 (histogram, histogram_pool): New global vars.
18499 (histogram_hash): New structure.
18500 (histogram_hash::hash): New method.
18501 (histogram_hash::equal): Likewise.
18502 (account_time_size): New function.
18503 (cmp_counts): New function.
18504 (dump_histogram): New function.
18505 (ipa_profile_generate_summary): New function.
18506 (ipa_profile_write_summary): New function.
18507 (ipa_profile_read_summary): New function.
18508 (ipa_profile): Decide on threshold.
18509 (pass_ipa_profile): Add ipa_profile_write_summary and
18510 ipa_profile_read_summary.
18511 * Makefile.in (ipa.o): Update dependencies.
18512 * lto-streamer.h (LTO_section_ipa_profile): New section.
18514 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
18516 * tree.h (VAR_P): New.
18518 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
18521 * doc/invoke.texi ([-fwhole-program]): Fix typo.
18523 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
18525 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
18526 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18527 (control_flow_insn_p): Likewise.
18528 * cfgrtl.c (duplicate_insn_chain): Likewise.
18529 * final.c (get_attr_length_1): Likewise.
18530 (shorten_branches): Likewise.
18531 (final_scan_insn): Likewise.
18532 * function.c (instantiate_virtual_regs): Likewise.
18533 * gcse.c (insert_insn_end_basic_block): Likewise.
18534 * ira-costs.c (scan_one_insn): Likewise.
18535 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18536 * lra.c (check_rtl): Likewise.
18537 * reload1.c (elimination_costs_in_insn): Likewise.
18538 * reorg.c (follow_jumps): Likewise.
18540 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
18541 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18542 (thumb_far_jump_used_p): Likewise.
18543 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
18544 (workaround_speculation): Likewise.
18545 (add_sched_insns_for_speculation): Likewise.
18546 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
18547 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18548 (frv_for_each_packet): Likewise.
18549 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
18550 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
18551 (final_emit_insn_group_barriers): Likewise.
18552 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18553 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18554 (mips16_insn_length): Likewise.
18555 * config/pa/pa.c (pa_reorg): Likewise.
18556 (pa_combine_instructions): Likewise.
18557 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
18558 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
18559 (sh_reorg): Likewise.
18560 (split_branches): Likewise.
18561 * config/spu/spu.c (get_branch_target): Likewise.
18563 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
18566 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
18568 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
18569 Fix declaration name.
18571 2013-03-28 Lawrence Crowl <crowl@google.com>
18573 * graphds.h (struct graph.indicies): Remove unused.
18574 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
18575 (SCOP_ORIGINAL_PDDRS): Remove unused.
18576 * sese.h (extern insert_loop_close_phis): Removed unused.
18577 (extern insert_guard_phis): Removed unused.
18578 (extern ivtype_map_elt_info): Removed unused.
18579 (new_ivtype_map_elt): Removed unused.
18580 * sese.c (ivtype_map_elt_info): Removed unused.
18582 2013-03-28 Lawrence Crowl <crowl@google.com>
18584 * Makefile.in: Add several missing include dependences.
18586 (test-dump.o): New. This object is not added to any executable,
18587 but is present for ad-hoc testing.
18589 (debug (const bitmap_head_def &)): New.
18590 (debug (const bitmap_head_def *)): New.
18592 (extern debug (const bitmap_head_def &)): New.
18593 (extern debug (const bitmap_head_def *)): New.
18595 (debug (edge_def &)): New.
18596 (debug (edge_def *)): New.
18598 (debug (basic_block_def &)): New.
18599 (debug (basic_block_def *)): New.
18601 (dump_node (const_tree, int, FILE *)): Correct source file.
18603 (debug (die_struct &)): New.
18604 (debug (die_struct *)): New.
18606 (extern debug (die_struct &)): New.
18607 (extern debug (die_struct *)): New.
18608 * gimple-pretty-print.c
18609 (debug (gimple_statement_d &)): New.
18610 (debug (gimple_statement_d *)): New.
18611 * gimple-pretty-print.h
18612 (extern debug (gimple_statement_d &)): New.
18613 (extern debug (gimple_statement_d *)): New.
18615 (debug (ira_allocno_copy &)): New.
18616 (debug (ira_allocno_copy *)): New.
18617 (debug (ira_allocno &)): New.
18618 (debug (ira_allocno *)): New.
18620 (extern debug (ira_allocno_copy &)): New.
18621 (extern debug (ira_allocno_copy *)): New.
18622 (extern debug (ira_allocno &)): New.
18623 (extern debug (ira_allocno *)): New.
18625 (debug (live_range &)): New.
18626 (debug (live_range *)): New.
18628 (debug (lra_live_range &)): New.
18629 (debug (lra_live_range *)): New.
18631 (debug (lra_live_range &)): New.
18632 (debug (lra_live_range *)): New.
18634 (debug (omega_pb_d &)): New.
18635 (debug (omega_pb_d *)): New.
18637 (extern debug (omega_pb_d &)): New.
18638 (extern debug (omega_pb_d *)): New.
18640 (debug (const rtx_def &)): New.
18641 (debug (const rtx_def *)): New.
18643 (debug_tree (tree): Move within file.
18644 (debug_raw (const tree_node &)): New.
18645 (debug_raw (const tree_node *)): New.
18646 (dump_tree_via_hooks (const tree_node *, int)): New.
18647 (debug (const tree_node &)): New.
18648 (debug (const tree_node *)): New.
18649 (debug_verbose (const tree_node &)): New.
18650 (debug_verbose (const tree_node *)): New.
18651 (debug_head (const tree_node &)): New.
18652 (debug_head (const tree_node *)): New.
18653 (debug_body (const tree_node &)): New.
18654 (debug_body (const tree_node *)): New.
18655 (debug_vec_tree (tree): Move and reimplement in terms of dump.
18656 (debug (vec<tree, va_gc> &)): New.
18657 (debug (vec<tree, va_gc> *)): New.
18659 (extern debug (const rtx_def &)): New.
18660 (extern debug (const rtx_def *)): New.
18662 (debug_raw (simple_bitmap_def &)): New.
18663 (debug_raw (simple_bitmap_def *)): New.
18664 (debug (simple_bitmap_def &)): New.
18665 (debug (simple_bitmap_def *)): New.
18667 (extern debug (simple_bitmap_def &)): New.
18668 (extern debug (simple_bitmap_def *)): New.
18669 (extern debug_raw (simple_bitmap_def &)): New.
18670 (extern debug_raw (simple_bitmap_def *)): New.
18672 (debug (vinsn_def &)): New.
18673 (debug (vinsn_def *)): New.
18674 (debug_verbose (vinsn_def &)): New.
18675 (debug_verbose (vinsn_def *)): New.
18676 (debug (expr_def &)): New.
18677 (debug (expr_def *)): New.
18678 (debug_verbose (expr_def &)): New.
18679 (debug_verbose (expr_def *)): New.
18680 (debug (vec<rtx> &)): New.
18681 (debug (vec<rtx> *)): New.
18683 (extern debug (vinsn_def &)): New.
18684 (extern debug (vinsn_def *)): New.
18685 (extern debug_verbose (vinsn_def &)): New.
18686 (extern debug_verbose (vinsn_def *)): New.
18687 (extern debug (expr_def &)): New.
18688 (extern debug (expr_def *)): New.
18689 (extern debug_verbose (expr_def &)): New.
18690 (extern debug_verbose (expr_def *)): New.
18691 (extern debug (vec<rtx> &)): New.
18692 (extern debug (vec<rtx> *)): New.
18694 (_list_iter_cond_expr): Make inline instead of static.
18696 (debug (sreal &)): New.
18697 (debug (sreal *)): New.
18699 (extern debug (sreal &)): New.
18700 (extern debug (sreal *)): New.
18702 (extern debug_raw (const tree_node &)): New.
18703 (extern debug_raw (const tree_node *)): New.
18704 (extern debug (const tree_node &)): New.
18705 (extern debug (const tree_node *)): New.
18706 (extern debug_verbose (const tree_node &)): New.
18707 (extern debug_verbose (const tree_node *)): New.
18708 (extern debug_head (const tree_node &)): New.
18709 (extern debug_head (const tree_node *)): New.
18710 (extern debug_body (const tree_node &)): New.
18711 (extern debug_body (const tree_node *)): New.
18712 (extern debug (vec<tree, va_gc> &)): New.
18713 (extern debug (vec<tree, va_gc> *)): New.
18715 (debug (struct loop &)): New.
18716 (debug (struct loop *)): New.
18717 (debug_verbose (struct loop &)): New.
18718 (debug_verbose (struct loop *)): New.
18719 * tree-dump.c: Add header dependence.
18721 (extern debug (struct loop &)): New.
18722 (extern debug (struct loop *)): New.
18723 (extern debug_verbose (struct loop &)): New.
18724 (extern debug_verbose (struct loop *)): New.
18726 (debug (data_reference &)): New.
18727 (debug (data_reference *)): New.
18728 (debug (vec<data_reference_p> &)): New.
18729 (debug (vec<data_reference_p> *)): New.
18730 (debug (vec<ddr_p> &)): New.
18731 (debug (vec<ddr_p> *)): New.
18733 (extern debug (data_reference &)): New.
18734 (extern debug (data_reference *)): New.
18735 (extern debug (vec<data_reference_p> &)): New.
18736 (extern debug (vec<data_reference_p> *)): New.
18737 (extern debug (vec<ddr_p> &)): New.
18738 (extern debug (vec<ddr_p> *)): New.
18740 (debug (pt_solution &)): New.
18741 (debug (pt_solution *)): New.
18743 (extern debug (pt_solution &)): New.
18744 (extern debug (pt_solution *)): New.
18746 (debug (_var_map &)): New.
18747 (debug (_var_map *)): New.
18748 (debug (tree_live_info_d &)): New.
18749 (debug (tree_live_info_d *)): New.
18751 (extern debug (_var_map &)): New.
18752 (extern debug (_var_map *)): New.
18753 (extern debug (tree_live_info_d &)): New.
18754 (extern debug (tree_live_info_d *)): New.
18756 2013-03-28 Jan Hubicka <jh@suse.cz>
18758 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
18760 2013-03-28 Ian Bolton <ian.bolton@arm.com>
18762 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
18763 record only when desired or required.
18765 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
18767 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
18768 *vec_extractv2di_1_rex64. Use x64 isa attribute.
18770 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18772 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
18773 (*andsi3_compare0_uxtw): New pattern.
18774 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
18775 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
18777 2013-03-28 Jan Hubicka <jh@suse.cz>
18779 * data-streamer-in.c (streamer_read_gcov_count): New function.
18780 * gimple-streamer-out.c: Include value-prof.h.
18781 (output_gimple_stmt): Output histogram.
18782 (output_bb): Use streamer_write_gcov_count.
18783 * value-prof.c: Include data-streamer.h
18784 (dump_histogram_value): Add HIST_TYPE_MAX.
18785 (stream_out_histogram_value): New function.
18786 (stream_in_histogram_value): New function.
18787 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
18788 (stream_out_histogram_value, stream_in_histogram_value): Declare.
18789 * data-streamer-out.c (streamer_write_gcov_count): New function.
18790 (streamer_write_gcov_count_stream): New function.
18791 * lto-cgraph.c (lto_output_edge): Update counter streaming.
18792 (lto_output_node): Likewise.
18793 (input_node, input_edge): Likewise.
18794 * lto-streamer-out.c (output_cfg): Update streaming.
18795 * lto-streamer-in.c (input_cfg): Likewise.
18796 * data-streamer.h (streamer_write_gcov_count,
18797 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
18798 * gimple-streamer-in.c: Include value-prof.h
18799 (input_gimple_stmt): Input histograms.
18800 (input_bb): Update profile streaming.
18802 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18804 * genmodes.c (emit_max_int): New function.
18805 (emit_insn_modes_h): Added call to emit_max_function.
18806 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
18808 * machmode.def: Fixed comment.
18810 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18812 * combine.c (try_combine): Removed useless assert.
18813 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
18815 2013-03-28 Marek Polacek <polacek@redhat.com>
18816 Richard Biener <rguenther@suse.de>
18818 PR tree-optimization/56695
18819 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
18820 build signed result of a vector comparison.
18821 * tree-cfg.c (verify_gimple_comparison): Check that a result
18822 of a vector comparison has signed type.
18824 2013-03-28 Richard Biener <rguenther@suse.de>
18826 PR tree-optimization/37021
18827 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
18828 do not restrict gaps between groups.
18829 * tree-vect-stmts.c (vectorizable_load): Properly account for
18830 a gap between groups.
18832 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
18834 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
18835 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
18838 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
18840 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
18841 * doc/extend.texi (Named Address Spaces): Ditto.
18842 (Variable Attributes): Ditto.
18844 2013-03-27 Kai Tietz <ktietz@redhat.com>
18846 * config.build: Add support for cygwin x64 target.
18847 * config.gcc: Likewise.
18848 * config.host: Likewise.
18849 * configure.ac: Likewise
18850 * configure: Regenerated.
18852 2013-03-27 Kai Tietz <ktietz@redhat.com>
18854 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
18855 * config/i386/t-cygwin-w64: New file.
18856 * config/i386/cygwin-w64.h: New file.
18857 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
18858 and add support for x64-cygwin target.
18859 (CPP_SPEC): Likewise.
18860 (CXX_WRAP_SPEC_LIST): Undefine before define.
18861 (LIBGCJ_SONAME): Use 15 as version.
18863 2013-03-27 Richard Biener <rguenther@suse.de>
18865 PR tree-optimization/56716
18866 * tree-ssa-structalias.c (perform_var_substitution): Adjust
18867 dumping for ref nodes.
18869 2013-03-27 Martin Jambor <mjambor@suse.cz>
18871 PR tree-optimization/55334
18872 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
18873 restricted pointers to arrays.
18875 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
18877 * Makefile.in (.SUFFIXES): Add .cc.
18878 (.c.o): Apply same recipe for implicit rule .cc.o.
18880 2013-03-27 Richard Biener <rguenther@suse.de>
18882 PR tree-optimization/37021
18883 * tree-vect-data-refs.c (vect_check_strided_load): Allow
18884 REALPART/IMAGPART_EXPRs around the supported refs.
18885 * tree-ssa-structalias.c (find_func_aliases): Assume that
18886 floating-point values are not used to transfer pointers.
18888 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
18890 * target.def (TARGET_HAS_IFUNC_P): New target hook.
18891 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
18892 * doc/tm.texi: Regenerate.
18893 * targhooks.h (default_has_ifunc_p): New.
18894 * targhooks.c (default_has_ifunc_p): Ditto.
18895 * config/linux-protos.h: New file.
18896 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
18897 hook for linux which disables support of indirect functions in android.
18898 * config/linux-android.c: New file.
18899 * config/t-linux-android.c: Ditto.
18900 * config.gcc: Added new object file linux-android.o.
18901 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
18902 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
18903 * varasm.c (do_assemble_alias): Likewise.
18904 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
18905 doesn't support indirect functions.
18906 * configure: Regenerate.
18908 2013-03-27 Bin Cheng <bin.cheng@arm.com>
18911 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
18912 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
18915 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18917 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
18919 2013-03-27 Terry Guo <terry.guo@arm.com>
18921 * config/arm/arm-cores.def: Added core cortex-r7.
18922 * config/arm/arm-tune.md: Regenerated.
18923 * config/arm/arm-tables.opt: Regenerated.
18924 * doc/invoke.texi: Added entry for core cortex-r7.
18926 2013-03-27 Walter Lee <walt@tilera.com>
18928 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
18929 double-decrement of next_scratch_regno.
18931 2013-03-27 Walter Lee <walt@tilera.com>
18933 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
18935 (insn_v1mulus): Ditto.
18936 (insn_v2muls): Ditto.
18938 2013-03-27 Walter Lee <walt@tilera.com>
18940 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
18941 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
18943 2013-03-27 Walter Lee <walt@tilera.com>
18945 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
18946 (*sibcall_value): Ditto.
18948 2013-03-27 Walter Lee <walt@tilera.com>
18950 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
18951 (insn_mnz_v8qi): ... this ...
18952 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
18954 (insn_v<n>mnz): Replaced by ...
18955 (insn_v1mnz): ... this ...
18956 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18958 (insn_mz_<mode>): Replaced by ...
18959 (insn_mz_v8qi): ... this ...
18960 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18962 (insn_v<n>mz): Replaced by ...
18963 (insn_v1mz): ... this ...
18964 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18967 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18969 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18971 2013-03-26 Roland McGrath <mcgrathr@google.com>
18973 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18974 than fprintf with a non-constant, non-format string.
18976 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18978 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18979 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18980 operand 0 predicate.
18981 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18982 attribute. Use general_x64nomem_operand as operand 1 predicate.
18983 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18984 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18985 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18986 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18987 (mov<mode>_insv_1): Remove expander. Merge insn with
18988 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18989 Use general_x64nomem_operand as operand 1 predicate.
18990 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18991 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18992 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18993 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18994 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18995 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18996 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18997 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18998 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18999 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
19000 (general_x64nomem_operand): Ditto.
19002 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
19004 * config/rtems.opt: Add -pthread option.
19006 2013-03-26 Richard Biener <rguenther@suse.de>
19008 * alias.c (find_base_term): Avoid redundant and not used recursion.
19009 (base_alias_check): Get the initial base term from the caller.
19010 (true_dependence_1): Compute and pass base terms to base_alias_check.
19011 (write_dependence_p): Likewise.
19012 (may_alias_p): Likewise.
19014 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
19016 * config/aarch64/aarch64.c (aarch64_classify_address): Support
19017 PC-relative load in SI modes and above only.
19019 2013-03-26 Xinyu Qi <xyqi@marvell.com>
19021 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
19022 * config/arm/iwmmxt.md (WCGR0): Update.
19023 (WCGR1, WCGR2, WCGR3): Likewise.
19025 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19027 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
19028 Use x64 and nox64 isa attributes.
19030 2013-03-26 Richard Biener <rguenther@suse.de>
19032 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
19033 alignment computations and rely on get_object_alignment_1
19034 for the !TYPE_P case.
19035 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
19037 2013-03-26 Walter Lee <walt@tilera.com>
19039 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
19040 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
19042 2013-03-25 Jeff Law <law@redhat.com>
19044 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
19045 check for INTEGRAL_TYPE_P that was missing due to checking in
19046 wrong version of prior patch.
19048 2013-03-25 Walter Lee <walt@tilera.com>
19050 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
19051 TILEGX_INSN_SHUFFLEBYTES1.
19052 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
19054 (tilegx_builtins): Ditto.
19055 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
19057 2013-03-25 Walter Lee <walt@tilera.com>
19059 * config/tilegx/tilegx.md (floatsisf2): New pattern.
19060 (floatunssisf2): New pattern.
19061 (floatsidf2): New pattern.
19062 (floatunssidf2): New pattern.
19064 2013-03-25 Walter Lee <walt@tilera.com>
19066 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
19067 tests for constraint J, K, N, P.
19069 2013-03-25 Walter Lee <walt@tilera.com>
19071 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
19072 Use indirect/pcrel encoding.
19073 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
19076 2013-03-25 Steve Ellcey <sellcey@mips.com>
19078 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
19079 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
19080 * config/mips/mips.c (mips_option_override): Set IMADD default.
19081 * config/mips/mips.h (PTF_AVOID_IMADD): New.
19082 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
19083 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
19084 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
19085 * doc/invoke.texi (-mimadd/-mno-imadd): New.
19087 2013-03-25 Jeff Law <law@redhat.com>
19089 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
19090 slightly to avoid creating and folding useless trees. Simplify
19091 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
19093 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
19095 * config/i386/i386.md (*zero_extendsidi2): Merge with
19096 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
19097 * config/i386/predicates.md (x86_64_zext_operand): Rename from
19098 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
19099 targets. Clarify comment.
19101 2013-03-25 Martin Jambor <mjambor@suse.cz>
19103 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
19104 pass-through jump functions differently.
19105 (ipa_read_jump_function): Likewise. Also use setter functions to set
19108 2013-03-25 Martin Jambor <mjambor@suse.cz>
19110 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
19111 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
19113 (ipa_get_indirect_edge_target): New function.
19114 (devirtualization_time_bonus): New parameter known_aggs, pass it to
19115 ipa_get_indirect_edge_target. Update all callers.
19116 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
19117 ipa_get_indirect_edge_target_1 instead of calling
19118 ipa_get_indirect_edge_target.
19119 (create_specialized_node): Pass aggvlas to
19120 ipcp_discover_new_direct_edges.
19122 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19124 * config/arm/arm.md (f_sels, f_seld): New types.
19125 (*cmov<mode>): New pattern.
19126 * config/arm/predicates.md (arm_vsel_comparison_operator): New
19129 2013-03-25 Kai Tietz <ktietz@redhat.com>
19131 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
19132 POSIX-printf for mingw-hosted builds.
19134 2013-03-25 Richard Biener <rguenther@suse.de>
19136 PR middle-end/56694
19137 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
19138 must-not-throw stmt location.
19140 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19142 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
19143 Emit load-acquire versions when acq is true.
19144 (arm_emit_store_exclusive): Add rel parameter.
19145 Emit store-release versions when rel is true.
19146 (arm_split_compare_and_swap): Use acquire-release instructions
19148 of barriers when appropriate.
19149 (arm_split_atomic_op): Likewise.
19150 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
19151 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
19152 (VUNSPEC_SLX): Likewise.
19153 (VUNSPEC_LDA): Likewise.
19154 (VUNSPEC_STL): Likewise.
19155 * config/arm/sync.md (atomic_load<mode>): New pattern.
19156 (atomic_store<mode>): Likewise.
19157 (arm_load_acquire_exclusive<mode>): Likewise.
19158 (arm_load_acquire_exclusivesi): Likewise.
19159 (arm_load_acquire_exclusivedi): Likewise.
19160 (arm_store_release_exclusive<mode>): Likewise.
19162 2013-03-25 Catherine Moore <clm@codesourcery.com>
19164 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
19165 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
19166 * config/mip/predicates.md (lwsp_swsp_operand,
19167 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
19168 sb16_operand, db4_operand, db7_operand, ib3_operand,
19169 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
19170 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
19171 andi16_operand): New predicates.
19172 * config/mips/mips.md (compression): New attribute.
19173 (enabled): New attribute.
19174 (length): Consider compression in computing length.
19175 (shift_compression): New code attribute.
19176 (*add<mode>3): New operands. Record compression.
19177 (sub<mode>3): Likewise.
19178 (one_cmpl<mode>2): Likewise.
19179 (*and<mode>3): Likewise.
19180 (*ior<mode>3): Likewise.
19181 (unnamed pattern for xor): Likewise.
19182 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
19183 (*<optab><mode>3): Likewise.
19184 (*mov<mode>_internal: Likewise.
19185 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
19186 (mips_unsigned_immediate_p): New.
19187 (umips_lwsp_swsp_address_p): New.
19188 (m16_based_address_p): New.
19189 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
19190 (mips_unsigned_immediate_p): New prototype.
19191 (lwsp_swsp_address_p): New prototype.
19192 (m16_based_address_p): New prototype.
19193 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
19194 (mips_signed_immediate_p): New function.
19195 (m16_based_address_p): New function.
19196 (lwsp_swsp_address_p): New function.
19197 (mips_print_operand_punctuation): Recognize short delay slot insns
19198 for microMIPS.add<mode>3"
19200 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19203 * config/arm/iterators.md (v_cmp_result): New mode attribute.
19204 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
19206 2013-03-25 Richard Biener <rguenther@suse.de>
19208 PR tree-optimization/56689
19209 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
19212 2013-03-25 Richard Biener <rguenther@suse.de>
19214 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
19216 (memory_references): Likewise.
19217 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
19218 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
19219 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
19220 (gather_mem_refs_in_loops): Fold into ...
19221 (analyze_memory_references): ... this. Move initialization
19222 to tree_ssa_lim_initialize.
19223 (fill_always_executed_in): Rename to ...
19224 (fill_always_executed_in_1): ... this.
19225 (fill_always_executed_in): Move contains_call computation to
19226 this new function from ...
19227 (tree_ssa_lim_initialize): ... here.
19228 (tree_ssa_lim): Call fill_always_executed_in.
19230 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19232 * postreload.c (reload_combine): Fix code detecting returns.
19234 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19236 * function.c (emit_use_return_register_into_block): On cc0 targets,
19237 do not emit the sequence between cc0 setter and user.
19239 2013-03-25 Kai Tietz <ktietz@redhat.com>
19241 * config/i386/predicates.md (local_symbolic_operand): Interpret
19242 dll-imported symbols as none-local.
19244 2013-03-25 Richard Biener <rguenther@suse.de>
19246 * tree-ssa-loop-im.c (struct depend): Remove.
19247 (struct lim_aux_data): Make depends a vec of gimples.
19248 (free_lim_aux_data): Adjust.
19249 (add_dependency): Likewise.
19250 (set_level): Likewise.
19252 2013-03-25 Richard Biener <rguenther@suse.de>
19254 PR middle-end/56434
19255 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
19256 the pointer returned by calls with ECF_MALLOC set.
19258 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
19260 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
19262 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
19264 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
19265 using MMXMODE mode iterator.
19266 (*move<mode>_internal): Merge with *movv2sf_internal and
19267 *movv2sf_internal_rex64 using MMXMODE mode iterator.
19269 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
19271 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
19272 (record_last_mem_set_info): Likewise.
19274 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
19275 of XNEWVEC followed by memset.
19276 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
19278 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
19280 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
19281 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
19282 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
19283 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
19284 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
19285 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
19286 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
19287 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
19288 BARRIER_P instead of GET_CODE.
19290 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19292 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
19293 inaccuracy in the probing code.
19295 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
19296 (ctrapdi4): Likewise.
19298 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19300 * calls.c (expand_call): Add missing guard to code handling return
19301 of non-BLKmode structures in MSB.
19302 * function.c (expand_function_end): Likewise.
19304 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19306 * combine.c (try_combine): Adjust comment. Do not add the set of
19307 insn #0 if the destination indirectly is set or dies in insn #2.
19308 Tidy up code to distribute a new note.
19310 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19312 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
19313 also for alternatives 16 and 17.
19315 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19317 * config/i386/sse.md (*mov<mode>_internal): Merge with
19318 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
19319 Emit insn template depending on type attribute. Use
19320 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19321 movd instead of movq mnemonic for interunit moves. Rewrite mode
19322 attribute calculation. Remove unit attribute calculation.
19323 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
19324 Set prefix_data16 attribute for DImode ssemov types.
19325 Use Ym instead of y for SSE-MMX conversion alternatives.
19326 Reorder operand constraints.
19328 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
19330 * df.h (df_insn_delete): Adjust prototype.
19331 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
19332 and let it decide whether mark the basic block dirty.
19333 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
19334 * df-scan.c (df_insn_info_delete): New helper function, split
19335 off from df_insn_delete.
19336 (df_scan_free_bb_info): Use it.
19337 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
19339 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
19340 that the insn is actually an insn and it has a non-NULL basic block.
19341 Do not mark basic block dirty if only deleting a DEBUG_INSN.
19343 2013-03-22 Richard Biener <rguenther@suse.de>
19345 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
19347 (mem_ref_alloc): Do not allocate them.
19348 (refs_independent_p): Do not query or maintain a cache.
19350 2013-03-22 Richard Biener <rguenther@suse.de>
19352 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
19353 (gather_mem_refs_in_loops): Do not compute it.
19354 (analyze_memory_references): Do not allocate it.
19355 (tree_ssa_lim_finalize): Do not free it.
19356 (for_all_locs_in_loop): Do not query all_refs_in_loop.
19358 2013-03-22 Richard Biener <rguenther@suse.de>
19360 * is-a.h (as_a): Use gcc_checking_assert.
19362 2013-03-22 Ian Bolton <ian.bolton@arm.com>
19364 * config/aarch64/aarch64.c (aarch64_print_operand): New
19365 format specifier for printing a constant in hex.
19366 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
19367 format specifier for printing second operand.
19369 2013-03-22 Richard Biener <rguenther@suse.de>
19371 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
19373 (gather_mem_refs_in_loops): Perform store accumulation here.
19374 (create_vop_ref_mapping_loop): Remove.
19375 (create_vop_ref_mapping): Likewise.
19376 (analyze_memory_references): Initialize refs_stored_in_loop.
19377 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
19378 (record_indep_loop): Remove.
19379 (record_dep_loop): New function.
19380 (ref_indep_loop_p_1): Adjust to only walk over references
19381 in the loop, not its subloops.
19382 (ref_indep_loop_p): Rename to ...
19383 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
19384 maintaining a more fine-grained cache.
19385 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
19386 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
19388 2013-03-22 Richard Biener <rguenther@suse.de>
19390 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
19391 (struct mem_ref): Make accesses_in_loop a vec of a vec of
19392 aggregate mem_ref_loc.
19393 (free_mem_ref_locs): Inline into ...
19394 (memref_free): ... this and adjust.
19395 (mem_ref_alloc): Adjust.
19396 (mem_ref_locs_alloc): Remove.
19397 (record_mem_ref_loc): Adjust.
19398 (get_all_locs_in_loop): Rewrite into ...
19399 (for_all_locs_in_loop): ... this iterator.
19400 (rewrite_mem_ref_loc): New functor.
19401 (rewrite_mem_refs): Use for_all_locs_in_loop.
19402 (sm_set_flag_if_changed): New functor.
19403 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
19404 (ref_always_accessed): New functor.
19405 (ref_always_accessed_p): Use for_all_locs_in_loop.
19407 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19409 * tree-pass.h (PROP_gimple_lvec): New.
19410 * passes.c (dump_properties): Handle PROP_gimple_lvec.
19411 (init_optimization_passes): Move pass_lower_vector.
19412 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
19414 (pass_lower_vector): Provide PROP_gimple_lvec.
19415 (pass_lower_vector_ssa): Likewise.
19416 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
19418 2013-03-21 Mark Wielaard <mjw@redhat.com>
19420 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
19422 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19424 * config/i386/i386.md (*movdi_internal): Disparage slightly
19425 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
19426 conversion alternatives.
19428 2013-03-21 Jakub Jelinek <jakub@redhat.com>
19430 PR middle-end/48087
19431 * diagnostic.def (DK_WERROR): New kind.
19432 * diagnostic.h (werrorcount): Define.
19433 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
19434 promoted to DK_ERROR, increment DK_WERROR counter instead of
19436 * toplev.c (toplev_main): Call print_ignored_options even if
19437 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
19438 even if just werrorcount is non-zero.
19441 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
19443 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
19444 (string_cst_pool_decl): New function.
19445 (optimize_one_addr_into_implicit_ptr): New function.
19446 (resolve_addr_in_expr): Optimize DWARF location expression
19447 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
19448 which doesn't live in memory, but has DW_AT_location or
19449 DW_AT_const_value, or refers to a string literal, into
19450 DW_OP_GNU_implicit_pointer.
19451 (optimize_location_into_implicit_ptr): New function.
19452 (resolve_addr): If removing DW_AT_location of a variable because
19453 it was DW_OP_addr of address of the variable, but the variable doesn't
19454 live in memory, try to emit const value attribute for the initializer.
19456 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19458 * tree.h (VECTOR_TYPE_P): New macro.
19459 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
19460 TYPE_MODE): Use it.
19461 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
19462 VEC_COND_EXPR cannot be lvalues.
19463 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
19465 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19467 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
19468 Restrict the transformation to equal modes.
19470 2013-03-21 Richard Biener <rguenther@suse.de>
19472 PR tree-optimization/39326
19473 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
19474 (MEM_ANALYZABLE): Adjust.
19475 (record_mem_ref_loc): Move bitmap ops ...
19476 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
19477 unanalyzable refs, do not record locations for it.
19478 (analyze_memory_references): Allocate ref zero as shared
19480 (refs_independent_p): Do not test for unanalyzed mems here.
19481 (ref_indep_loop_p_1): Special-case disambiguation against
19482 the unanalyzed ref.
19483 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
19485 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
19487 * config/arm/arm-protos.h (tune_params): Add
19488 prefer_neon_for_64bits field.
19489 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
19490 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
19491 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
19492 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
19493 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
19494 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
19495 (arm_option_override): Handle -mneon-for-64bits new option.
19496 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
19497 (prefer_neon_for_64bits): Declare new variable.
19498 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
19499 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
19500 (arch_enabled): Handle new arch types. Remove support for onlya8
19502 (one_cmpldi2): Use new arch names.
19503 (zero_extend<mode>di2, extend<mode>di2): Ditto.
19504 * config/arm/arm.opt (mneon-for-64bits): Add option.
19505 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
19506 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
19507 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
19509 * doc/invoke.texi (-mneon-for-64bits): Document.
19511 2013-03-21 Richard Biener <rguenther@suse.de>
19513 PR tree-optimization/39326
19514 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
19515 (sort_bbs_in_loop_postorder_cmp): New function.
19516 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
19518 2013-03-21 Richard Biener <rguenther@suse.de>
19520 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
19521 (vect_insert_into_interleaving_chain): Likewise.
19522 (vect_drs_dependent_in_basic_block): Inline ...
19523 (vect_slp_analyze_data_ref_dependence): ... here. New function,
19525 (vect_analyze_data_ref_dependence): ... here. Simplify.
19526 (vect_check_interleaving): Simplify.
19527 (vect_analyze_data_ref_dependences): Likewise. Split out ...
19528 (vect_slp_analyze_data_ref_dependences): ... this new function.
19529 (dr_group_sort_cmp): New function.
19530 (vect_analyze_data_ref_accesses): Compute data-reference groups
19531 here instead of in vect_analyze_data_ref_dependence. Use
19532 a more efficient algorithm.
19533 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
19534 vect_slp_analyze_data_ref_dependences. Call
19535 vect_analyze_data_ref_accesses earlier.
19536 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
19537 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
19538 (vect_slp_analyze_data_ref_dependences): New prototype.
19540 2013-03-21 Richard Biener <rguenther@suse.de>
19542 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
19543 ref is stored in the loop.
19544 (find_refs_for_sm): Walk only over all stores.
19545 (store_motion_loop): Allocate from lim_bitmap_obstack.
19546 (store_motion): Likewise.
19548 2013-03-21 Richard Biener <rguenther@suse.de>
19550 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
19551 Update virtual SSA form.
19553 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19555 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
19556 * configure: Regenerate.
19557 * config.in: Regenerate.
19558 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
19559 if !HAVE_LD_EH_FRAME_CIEV3.
19561 2013-03-21 Richard Biener <rguenther@suse.de>
19563 * tree-cfg.c (verify_expr_no_block): New function.
19564 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
19565 nor DECL_VALUE_EXPR have locations with associated blocks.
19566 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
19567 (clear_unused_block_pointer): Remove code dealing with
19568 blocks in DECL_DEBUG_EXPR locations.
19570 2013-03-21 Richard Biener <rguenther@suse.de>
19572 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
19573 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
19574 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
19575 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
19576 instead of DECL_DEBUG_EXPR_IS_FROM.
19577 * gimplify.c (gimplify_modify_expr): Likewise.
19578 * tree-cfg.c (verify_expr_location_1): Likewise.
19579 * tree-complex.c (create_one_component_var): Likewise.
19580 * tree-sra.c (create_access_replacement): Likewise.
19581 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
19582 (clear_unused_block_pointer): Likewise.
19583 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
19584 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
19585 * var-tracking.c (var_debug_decl): Likewise.
19586 (track_expr_p): Likewise.
19587 * tree-inline.c (add_local_variables): Likewise. Set
19588 DECL_HAS_DEBUG_EXPR_P after copying it.
19589 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
19590 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
19592 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19595 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
19596 * configure: Regenerate.
19597 * config.in: Regenerate.
19598 * config/i386/i386.md (*movdf_internal): Use
19599 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19600 movd instead of movq mnemonic for interunit moves.
19601 (*movdi_internal): Ditto.
19603 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19605 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
19606 (abd<mode>_3): New pattern.
19607 (aba<mode>_3): New pattern.
19608 (fabd<mode>_3): New pattern.
19610 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19612 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
19613 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
19614 occurrence of REGISTER_PREFIX as its empty string.
19616 2013-03-20 Jeff Law <law@redhat.com>
19618 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
19619 addititional equivalences for equality comparisons between an SSA_NAME
19620 and a constant where the SSA_NAME was set from a widening conversion.
19622 2013-03-20 Walter Lee <walt@tilera.com>
19624 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
19626 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19628 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
19629 depending on type attribute.
19630 (*movti_internal): Ditto.
19631 (*movtf_internal): Ditto.
19632 (*movxf_internal): Ditto.
19633 (*movdf_internal): Ditto.
19634 (*movsf_internal): Ditto.
19636 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19638 * config/i386/i386.md (*movti_internal): Set prefix attribute to
19639 maybe_vex for sselog1 and ssemov types.
19640 (*movdi_internal): Reorder operand constraints.
19641 (*movsi_internal): Ditto. Set prefix attribute to
19642 maybe_vex for sselog1 and ssemov types.
19643 (*movtf_internal): Set prefix attribute to maybe_vex
19644 for sselog1 and ssemov types.
19645 (*movdf_internal): Ditto. Set prefix_data16 attribute for
19646 DImode ssemov types. Reorder operand constraints.
19647 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
19648 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
19649 attribute for SImode ssemov types. Reorder operand constraints.
19651 2013-03-20 Martin Jambor <mjambor@suse.cz>
19653 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
19654 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
19656 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
19658 * config/rs6000/predicates.md (indexed_address, update_address_mem
19659 update_indexed_address_mem): New predicates.
19660 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
19661 attribute for load/store instructions.
19662 * config/rs6000/dfp.md (movsd_store): Likewise.
19663 (movsd_load): Likewise.
19664 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
19665 (unnamed HI->DI extend define_insn): Likewise.
19666 (unnamed SI->DI extend define_insn): Likewise.
19667 (unnamed QI->SI extend define_insn): Likewise.
19668 (unnamed QI->HI extend define_insn): Likewise.
19669 (unnamed HI->SI extend define_insn): Likewise.
19670 (unnamed HI->SI extend define_insn): Likewise.
19671 (extendsfdf2_fpr): Likewise.
19672 (movsi_internal1): Likewise.
19673 (movsi_internal1_single): Likewise.
19674 (movhi_internal): Likewise.
19675 (movqi_internal): Likewise.
19676 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
19677 attribute for load/store instructions.
19678 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
19680 (mov<mode>_softfloat): Likewise.
19681 (mov<mode>_hardfloat32): Likewise.
19682 (mov<mode>_hardfloat64): Likewise.
19683 (mov<mode>_softfloat64): Likewise.
19684 (movdi_internal32): Likewise.
19685 (movdi_internal64): Likewise.
19686 (probe_stack_<mode>): Likewise.
19688 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
19690 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
19691 floating point, and decimal floating point to reload iterator.
19693 * config/rs6000/constraints.md (wl constraint): New constraints to
19694 return FLOAT_REGS if certain options are used to reduce the number
19695 of separate patterns that exist in the file.
19696 (wx constraint): Likewise.
19697 (wz constraint): Likewise.
19699 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19700 -mdebug=reg, print wg, wl, wx, and wz constraints.
19701 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19702 Initialize the reload functions for 64-bit binary/decimal floating
19704 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19705 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19706 create the buffer on the stack to overcome not having a 32-bit
19708 (rs6000_emit_move): Likewise.
19709 (rs6000_secondary_memory_needed_rtx): Likewise.
19710 (rs6000_alloc_sdmode_stack_slot): Likewise.
19711 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19712 via xxlxor, just like DFmode 0.0.
19714 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
19715 define as 1 if we are running on a power7 or newer.
19716 (enum r6000_reg_class_enum): Add new constraints.
19718 * config/rs6000/dfp.md (movsd): Delete, combine with binary
19719 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19720 with other moves by using conditional constraits (wg). Use LFIWZX
19721 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
19722 (movsd splitter): Likewise.
19723 (movsd_hardfloat): Likewise.
19724 (movsd_softfloat): Likewise.
19726 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
19727 binary and decimal floating point moves.
19728 (fmove_ok): New attributes to combine binary and decimal floating
19729 point moves, and to combine power6x (mfpgpr) moves along normal
19731 (real_value_to_target): Likewise.
19732 (f32_lr): Likewise.
19733 (f32_lm): Likewise.
19734 (f32_li): Likewise.
19735 (f32_sr): Likewise.
19736 (f32_sm): Likewise.
19737 (f32_si): Likewise.
19738 (movsf): Combine binary and decimal floating point moves. Combine
19739 power6x (mfpgpr) moves with other moves by using conditional
19740 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
19741 (mov<mode> for SFmode/SDmode); Likewise.
19742 (SFmode/SDmode splitters): Likewise.
19743 (movsf_hardfloat): Likewise.
19744 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
19745 (movsf_softfloat): Likewise.
19746 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
19748 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
19749 wx and wz constraints.
19751 * config/rs6000/constraints.md (wg constraint): New constraint to
19752 return FLOAT_REGS if -mmfpgpr (power6x) was used.
19754 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
19757 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19758 -mdebug=reg, print wg, wl, wx, and wz constraints.
19759 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19760 Initialize the reload functions for 64-bit binary/decimal floating
19762 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19763 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19764 create the buffer on the stack to overcome not having a 32-bit
19766 (rs6000_emit_move): Likewise.
19767 (rs6000_secondary_memory_needed_rtx): Likewise.
19768 (rs6000_alloc_sdmode_stack_slot): Likewise.
19769 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19770 via xxlxor, just like DFmode 0.0.
19772 * config/rs6000/dfp.md (movdd): Delete, combine with binary
19773 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19774 with other moves by using conditional constraits (wg). Use LFIWZX
19775 and STFIWX for loading SDmode on power7.
19776 (movdd splitters): Likewise.
19777 (movdd_hardfloat32): Likewise.
19778 (movdd_softfloat32): Likewise.
19779 (movdd_hardfloat64_mfpgpr): Likewise.
19780 (movdd_hardfloat64): Likewise.
19781 (movdd_softfloat64): Likewise.
19783 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
19784 64-bit binary and decimal floating point moves.
19785 (FMOVE64X): Likewise.
19786 (movdf): Combine 64-bit binary and decimal floating point moves.
19787 Combine power6x (mfpgpr) moves with other moves by using
19788 conditional constraits (wg).
19789 (mov<mode> for DFmode/DDmode): Likewise.
19790 (DFmode/DDmode splitters): Likewise.
19791 (movdf_hardfloat32): Likewise.
19792 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
19793 (movdf_softfloat32): Likewise.
19794 (movdf_hardfloat64_mfpgpr): Likewise.
19795 (movdf_hardfloat64): Likewise.
19796 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
19797 (movdf_softfloat64): Likewise.
19798 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
19799 (reload_<mode>_load): Move to later in the file so they aren't in
19800 the middle of the floating point move insns.
19801 (reload_<mode>_store): Likewise.
19803 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
19806 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
19807 constraint if -mdebug=reg.
19808 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
19809 Enable using dd reload support if needed.
19811 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
19812 binary and decimal floating point moves in rs6000.md.
19813 (movtd_internal): Likewise.
19815 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
19816 decimal floating point moves.
19818 (movtf_internal): Likewise.
19819 (mov<mode>_internal, TDmode/TFmode): Likewise.
19820 (movtf_softfloat): Likewise.
19821 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
19823 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
19824 movdi_internal64, using wg constraint for move direct operations.
19825 (movdi_internal64): Likewise.
19827 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
19828 MODES_TIEABLE_P for selected modes. Print the numerical value of
19829 the various virtual registers. Use GPR/FPR first/last values,
19830 instead of hard coding the register numbers. Print which modes
19831 have reload functions registered.
19832 (rs6000_option_override_internal): If -mdebug=reg, trace the options
19833 settings before/after setting cpu, target and subtarget settings.
19834 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
19835 and for secondary reload failures in rs6000_secondary_reload_inner.
19836 (rs6000_secondary_reload_fail): Likewise.
19837 (rs6000_secondary_reload_inner): Likewise.
19839 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
19840 macros for first/last GPR and FPR registers.
19841 (LAST_GPR_REGNO): Likewise.
19842 (FIRST_FPR_REGNO): Likewise.
19843 (LAST_FPR_REGNO): Likewise.
19845 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
19846 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
19847 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
19848 (vcond<mode><mode>): Likewise.
19849 (vcondu<mode><mode>): Likewise.
19850 (vector_gtu<mode>): Likewise.
19851 (vector_gte<mode>): Likewise.
19852 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
19853 to prevent the compiler from converting DImode operations to TImode.
19854 (ior<mode>3): Likewise.
19855 (and<mode>3): Likewise.
19856 (one_cmpl<mode>2): Likewise.
19857 (nor<mode>3): Likewise.
19858 (andc<mode>3): Likewise.
19860 * config/rs6000/constraints.md (wt constraint): New constraint
19861 that returns VSX_REGS if TImode is allowed in VSX registers.
19863 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
19864 constant under VSX.
19866 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
19867 similar to TImode, but it is restricted to being in the GPRs.
19869 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
19870 TImode to occupy a single VSX register.
19872 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
19873 -mvsx-timode for power7/power8.
19874 (power7 cpu): Likewise.
19875 (power8 cpu): Likewise.
19877 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
19878 sure that TFmode/TDmode take up two registers if they are ever
19879 allowed in the upper VSX registers.
19880 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
19882 (rs6000_init_hard_regno_mode_ok): Likewise.
19883 (rs6000_debug_reg_global): Add debugging for PTImode and wt
19884 constraint. Print if LRA is turned on.
19885 (rs6000_option_override_internal): Give an error if -mvsx-timode
19886 and VSX is not enabled.
19887 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
19888 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
19889 to reg+offset addressing. Use PTImode when checking offset
19890 addresses for validity.
19891 (reg_offset_addressing_ok_p): Likewise.
19892 (rs6000_legitimate_offset_address_p): Likewise.
19893 (rs6000_legitimize_address): Likewise.
19894 (rs6000_legitimize_reload_address): Likewise.
19895 (rs6000_legitimate_address_p): Likewise.
19896 (rs6000_eliminate_indexed_memrefs): Likewise.
19897 (rs6000_emit_move): Likewise.
19898 (rs6000_secondary_reload): Likewise.
19899 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
19900 reloads to fpr registers to continue to use reg+offset addressing,
19901 but 64-bit reloads to altivec registers need reg+reg addressing.
19902 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
19903 it. Treat LO_SUM like a PLUS operation.
19904 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
19905 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
19906 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
19907 registers to share a register with a smaller sized type, since VSX
19908 puts scalars in the upper 64-bits.
19909 (print_operand): Add support for PTImode.
19910 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
19911 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
19912 registers, but don't have arithmetic support.
19913 (rs6000_memory_move_cost): Add test for VSX.
19914 (rs6000_opt_masks): Add -mvsx-timode.
19916 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
19919 (VSr): Use wt constraint for TImode.
19920 (VSv): Drop TImode support.
19921 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
19922 (vsx_movti_64bit): Likewise.
19923 (vsx_movti_32bit): Likewise.
19924 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
19925 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
19926 one '?' on the appropriate output constraint. Do not allow TImode
19927 logical operations on 32-bit systems.
19928 (vsx_ior<mode>3): Likewise.
19929 (vsx_xor<mode>3): Likewise.
19930 (vsx_one_cmpl<mode>2): Likewise.
19931 (vsx_nor<mode>3): Likewise.
19932 (vsx_andc<mode>3): Likewise.
19933 (vsx_concat_<mode>): Likewise.
19934 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
19936 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
19937 OPTION_MASK_VSX_TIMODE.
19938 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
19939 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
19941 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
19942 (TI2 iterator): New iterator for TImode, PTImode.
19943 (wd mode attribute): Add values for vector types.
19944 (movti_string): Replace TI move operations with operations for TImode
19945 and PTImode. Add support for TImode being allowed in VSX registers.
19946 (mov<mode>_string, TImode/PTImode): Likewise.
19947 (movti_ppc64): Likewise.
19948 (mov<mode>_ppc64, TImode/PTImode): Likewise.
19949 (TI mode splitters): Likewise.
19951 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
19954 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
19956 PR tree-optimization/56355
19957 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19958 Also handle integers with undefined overflow.
19960 2013-03-20 Catherine Moore <clm@codesourcery.com>
19961 Maciej W. Rozycki <macro@codesourcery.com>
19962 Tom de Vries <tom@codesourcery.com>
19963 Nathan Sidwell <nathan@codesourcery.com>
19964 Iain Sandoe <iain@codesourcery.com>
19965 Nathan Froyd <froydnj@codesourcery.com>
19966 Chao-ying Fu <fu@mips.com>
19968 * doc/extend.texi: (micromips, nomicromips, nocompression):
19969 Document new function attributes.
19970 * doc/invoke.texi (minterlink-compressed, mmicromips,
19971 m14k, m14ke, m14kec): Document new options.
19972 (minterlink-mips16): Update documentation.
19973 * doc/md.texi (ZC, ZD): Document new constraints.
19974 * configure.ac (gcc_cv_as_micromips): Check if linker
19975 supports the .set micromips directive.
19976 * configure: Regenerate.
19977 * config.in: Regenerate.
19978 * config/mips/mips-tables.opt: Regenerate.
19979 * config/mips/micromips.md: New file.
19980 * constraints.md (ZC, ZD): New constraints.
19981 * config/mips/predicates.md (movep_src_register): New predicate.
19982 (movep_src_operand): New predicate.
19983 (non_volatile_mem_operand): New predicate.
19984 * config/mips/mips.md (multimem): New type.
19985 (length): Differentiate between 17-bit and 18-bit branch offsets.
19986 (MOVEP1, MOVEP2): New mode iterator.
19987 (mov_<load>l): Use ZC constraint.
19988 (mov_<load>r): Likewise.
19989 (mov_<store>l): Likewise.
19990 (mov_<store>r): Likewise.
19991 (*branch_equality<mode>_inverted): Add microMIPS support.
19992 (*branch_equality<mode>): Likewise.
19993 (*jump_absolute): Likewise.
19994 (indirect_jump_<mode>): Likewise.
19995 (tablejump_<mode>): Likewise.
19996 (<optab>_internal): Likewise.
19997 (sibcall_internal): Likewise.
19998 (sibcall_value_internal): Likewise.
19999 (prefetch): Use constraint ZD.
20000 * config/mips/mips.opt (minterlink-compressed): New option.
20001 (minterlink-mips16): Now an alias for minterlink-compressed.
20002 (mmicromips): New option.
20003 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20004 (compare_and_swap_12): Likewise.
20005 (sync_add<mode>): Likewise.
20006 (sync_<optab>_12): Likewise.
20007 (sync_old_<optab>_12): Likewise.
20008 (sync_new_<optab>_12): Likewise.
20009 (sync_nand_12): Likewise.
20010 (sync_old_nand_12): Likewise.
20011 (sync_new_nand_12): Likewise.
20012 (sync_sub<mode>): Likewise.
20013 (sync_old_add<mode>): Likewise.
20014 (sync_old_sub<mode>): Likewise.
20015 (sync_new_add<mode>): Likewise.
20016 (sync_new_sub<mode>): Likewise.
20017 (sync_<optab><mode>): Likewise.
20018 (sync_old_<optab><mode>): Likewise.
20019 (sync_new_<optab><mode>): Likewise.
20020 (sync_nand<mode>): Likewise.
20021 (sync_old_nand<mode>): Likewise.
20022 (sync_new_nand<mode>): Likewise.
20023 (sync_lock_test_and_set<mode>): Likewise.
20024 (test_and_set_12): Likewise.
20025 (atomic_compare_and_swap<mode>): Likewise.
20026 (atomic_exchange<mode>_llsc): Likewise.
20027 (atomic_fetch_add<mode>_llsc): Likewise.
20028 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20029 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20030 (umips_save_restore_pattern_p): Likewise.
20031 (umips_load_store_pair_p): Likewise.
20032 (umips_output_load_store_pair): Likewise.
20033 (umips_movep_target_p): Likewise.
20034 (umips_12bit_offset_address_p): Likewise.
20035 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20036 (mips_base_mips16): Rename this...
20037 (mips_base_compression_flags): ...to this. Update all uses.
20038 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20039 (mips_mips16_decl_p): Delete.
20040 (mips_nomips16_decl_p): Delete.
20041 (mips_get_compress_on_flags): New function.
20042 (mips_get_compress_off_flags): New function.
20043 (mips_get_compress_mode): New function.
20044 (mips_get_compress_on_name): New function.
20045 (mips_get_compress_off_name): New function.
20046 (mips_insert_attributes): Support multiple compression types.
20047 (mips_merge_decl_attributes): Likewise.
20048 (umips_12bit_offset_address_p): New function.
20049 (mips_start_function_definition): Emit .set micromips directive.
20050 (mips_call_may_need_jalx_p): New function.
20051 (mips_function_ok_for_sibcall): Add microMIPS support.
20052 (mips_print_operand_punctuation): Support short delay slots and
20054 (umips_swm_mask, umips_swm_encoding): New.
20055 (umips_build_save_restore): New function.
20056 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20057 (was_mips16_p): Remove.
20058 (old_compression_mode): New.
20059 (mips_set_compression_mode): New function.
20060 (mips_set_current_function): Add microMIPS support.
20061 (mips_option_override): Likewise.
20062 (umips_save_restore_pattern_p): New function.
20063 (umips_output_save_restore): New function.
20064 (umips_load_store_pair_p_1): New function.
20065 (umips_load_store_pair_p): New function.
20066 (umips_output_load_store_pair_1): New function.
20067 (umips_output_load_store_pair): New function.
20068 (umips_movep_target_p) New function.
20069 (mips_prepare_pch_save): Add microMIPS support.
20070 * config/mips/mips.h (TARGET_COMPRESSION): New.
20071 (TARGET_CPU_CPP_BUILTINS): Update macro
20072 to use new compression flags and to support microMIPS.
20073 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20074 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20075 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20076 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20077 (ASM_SPEC): Support mmicromips and mno-micromips.
20078 (M16STORE_REG_P): New macro.
20079 (MIPS_CALL): Support TARGET_MICROMIPS.
20080 (MICROMIPS_J): New macro.
20081 (mips_base_mips16): Rename this...
20082 (mips_base_compression_flags): ...to this.
20083 (UMIPS_12BIT_OFFSET_P): New macro.
20084 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20085 (MULTILIB_DIRNAMES): Likewise.
20086 2013-03-20 Richard Biener <rguenther@suse.de>
20088 PR tree-optimization/56661
20089 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
20090 the result does not have to be distinct.
20092 2013-03-20 Richard Biener <rguenther@suse.de>
20094 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
20097 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20098 Steven Bosscher <steven@gcc.gnu.org>
20100 PR rtl-optimization/56605
20101 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
20103 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20106 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
20107 that require movd instead of movq.
20109 2013-03-20 Richard Biener <rguenther@suse.de>
20111 * tree-ssa-structalias.c (struct variable_info): Add pointer
20112 to the first field of an aggregate with sub-vars. Make
20113 this and the pointer to the next subfield its ID.
20114 (vi_next): New function.
20115 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
20116 storedanything_id, integer_id): Increment by one.
20117 (new_var_info, get_call_vi, lookup_call_clobber_vi,
20118 get_call_clobber_vi): Adjust.
20119 (solution_set_expand): Simplify and speedup.
20120 (solution_set_add): Inline into ...
20121 (set_union_with_increment): ... this. Adjust accordingly.
20122 (do_sd_constraint): Likewise.
20123 (do_ds_constraint): Likewise.
20124 (do_complex_constraint): Simplify.
20125 (build_pred_graph): Adjust.
20126 (solve_graph): Likewise. Simplify and speedup.
20127 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
20128 get_constraint_for_component_ref, get_constraint_for_1,
20129 first_vi_for_offset, first_or_preceding_vi_for_offset,
20130 create_function_info_for, create_variable_info_for_1,
20131 create_variable_info_for, intra_create_variable_infos): Adjust.
20132 (init_base_vars): Push NULL for ID zero.
20133 (compute_points_to_sets): Adjust.
20135 2013-03-20 Richard Biener <rguenther@suse.de>
20137 * cfgloop.c (verify_loop_structure): Streamline and avoid
20138 ICEing on corrupt loop tree.
20139 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
20142 2013-03-20 Richard Biener <rguenther@suse.de>
20144 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
20145 check whether an SSA update is needed.
20147 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
20149 * config/mips/constraints.md (T): Rename to...
20153 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
20154 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
20156 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20158 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
20159 (*subsi3_carryin_uxtw): Likewise.
20161 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20163 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
20164 (*rorsi3_insn_uxtw): Likewise.
20166 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20168 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
20169 (*extrsi5_insn_uxtw): Likewise.
20171 2013-03-19 Richard Biener <rguenther@suse.de>
20173 PR tree-optimization/56273
20174 * passes.c (init_optimization_passes): Move second VRP after DOM.
20176 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
20178 * config/i386/i386.md (*movti_internal): Merge from
20179 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
20180 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
20181 nox64 isa attributes.
20183 2013-03-18 Richard Biener <rguenther@suse.de>
20185 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
20187 (merge_node_constraints): Likewise.
20188 (build_succ_graph): Likewise.
20189 (valid_graph_edge): Inline into single caller.
20190 (unify_nodes): Likewise. Use bitmap_set_bit return value
20192 (scc_visit): Fix formatting and variable use.
20193 (do_sd_constraint): Use gcc_checking_assert.
20194 (do_ds_constraint): Likewise.
20195 (do_complex_constraint): Likewise.
20196 (condense_visit): Likewise. Cleanup.
20197 (dump_pred_graph): New function.
20198 (perform_var_substitution): Dump the pred-graph before
20199 variable substitution.
20200 (find_equivalent_node): Use gcc_checking_assert.
20201 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
20203 2013-03-18 Richard Biener <rguenther@suse.de>
20205 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
20206 Remove cond_expr_stmt_list argument and do not gimplify the
20208 (vect_loop_versioning): Adjust.
20209 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
20210 Cleanup to use less temporaries.
20211 (vect_create_data_ref_ptr): Cleanup.
20213 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20215 PR tree-optimization/56635
20216 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
20217 require types_compatible_p types.
20219 2013-03-18 Nick Clifton <nickc@redhat.com>
20221 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
20222 spurious backslash.
20224 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
20225 Add missing line to comment describing function.
20227 2013-03-18 Richard Biener <rguenther@suse.de>
20229 PR tree-optimization/56210
20230 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20231 Handle string / character search functions.
20232 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
20234 2013-03-18 Richard Biener <rguenther@suse.de>
20236 PR middle-end/56483
20237 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
20238 and implement properly.
20239 * gimple.h (gimple_cond_single_var_p): Remove.
20241 2013-03-18 Richard Biener <rguenther@suse.de>
20243 * tree-data-ref.h (find_data_references_in_loop): Declare.
20244 * tree-data-ref.c (get_references_in_stmt): Use a stack
20245 vector pre-allocated in the callers.
20246 (find_data_references_in_stmt): Adjust.
20247 (graphite_find_data_references_in_stmt): Likewise.
20248 (create_rdg_vertices): Likewise.
20249 (find_data_references_in_loop): Export.
20250 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
20251 Compute dependences here...
20252 (vect_analyze_data_refs): ...not here. When we encounter
20253 a non-vectorizable data reference in basic-block vectorization
20254 truncate the data reference vector. Do not bother to
20255 fixup data-dependence information for gather loads.
20256 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
20257 of data references, as reported.
20259 2013-03-18 Richard Biener <rguenther@suse.de>
20261 PR tree-optimization/3713
20262 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
20263 has_constants and expr.
20264 (stmt_has_constants): Properly valueize SSA names when deciding
20265 whether the stmt has constants.
20267 2013-03-18 Richard Biener <rguenther@suse.de>
20269 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
20270 whole function when there is nothing to do.
20271 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
20272 * tree-vectorizer.c (vectorize_loops): Update virtual and
20273 loop-closed SSA once.
20274 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
20276 2013-03-18 Richard Biener <rguenther@suse.de>
20278 PR middle-end/56113
20279 * domwalk.c (bb_postorder): New global static.
20280 (cmp_bb_postorder): New function.
20281 (walk_dominator_tree): Replace scheme imposing an order for
20282 visiting dominator sons by one sorting them at the time they
20283 are pushed on the stack.
20285 2013-03-18 Richard Biener <rguenther@suse.de>
20287 PR tree-optimization/39326
20288 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
20289 (struct mem_ref): Replace mem member with ao_ref typed member.
20290 (MEM_ANALYZABLE): Adjust.
20291 (memref_eq): Likewise.
20292 (mem_ref_alloc): Likewise.
20293 (gather_mem_refs_stmt): Likewise.
20294 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
20295 (execute_sm_if_changed_flag_set): Adjust.
20296 (execute_sm): Likewise.
20297 (ref_always_accessed_p): Likewise.
20298 (refs_independent_p): Likewise.
20299 (can_sm_ref_p): Likewise.
20301 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20304 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
20305 return 1 even for !unsignedp.
20307 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20309 * config/i386/i386.md (isa): Add x64 and nox64.
20310 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
20311 (*pushtf): Enable *roF alternative for x64 isa only.
20312 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
20313 mode attribute of integer alternatives to DImode for TARGET_64BIT.
20314 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
20315 (*movtf_internal): Merge from *movtf_internal_rex64 and
20316 *movtf_internal_sse. Use x64 and nox64 isa attributes.
20317 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
20318 nox64 isa attributes.
20319 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
20320 nox64 isa attributes.
20321 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
20323 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20325 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20327 2013-03-17 Jakub Jelinek <jakub@redhat.com>
20330 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
20331 class names. Remove trailing comma after "ALL_REGS".
20333 2013-03-16 Jan Hubicka <jh@suse.cz>
20335 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
20336 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
20337 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
20338 of cgraph_get_create_node.
20339 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
20341 2013-03-16 Jason Merrill <jason@redhat.com>
20344 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
20345 with DW_AT_default_value.
20347 2013-03-16 Jakub Jelinek <jakub@redhat.com>
20349 * BASE-VER: Set to 4.9.0.
20351 2013-03-14 Andi Kleen <ak@linux.intel.com>
20354 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
20355 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
20356 Document _x* TSX intrinsics.
20358 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
20359 David Holsgrove <david.holsgrove@xilinx.com>
20361 * configure.ac: Add MicroBlaze TLS support detection.
20362 * configure: Regenerate.
20363 * config/microblaze/microblaze-protos.h
20364 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
20365 symbol_mentioned_p, label_mentioned_p): Add prototypes.
20366 * config/microblaze/microblaze.c (microblaze_address_type): Add
20367 ADDRESS_TLS and tls_reloc address types.
20368 (microblaze_address_info): Add tls_reloc.
20369 (TARGET_HAVE_TLS): Define.
20370 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
20371 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
20372 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
20373 load_tls_operand, microblaze_call_tls_get_addr,
20374 microblaze_legitimize_tls_address): New functions.
20375 (microblaze_classify_unspec): Handle UNSPEC_TLS.
20376 (get_base_reg): Use microblaze_tls_symbol_p.
20377 (microblaze_classify_address): Handle TLS.
20378 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
20379 label_mentioned_p and microblaze_tls_referenced_p.
20380 (microblaze_legitimize_address): Handle TLS.
20381 (microblaze_address_insns): Handle ADDRESS_TLS.
20382 (pic_address_needs_scratch): Handle TLS.
20383 (print_operand_address): Handle TLS.
20384 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
20385 (microblaze_expand_move): Handle TLS.
20386 (microblaze_legitimate_constant_p): Check
20387 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
20388 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
20389 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
20390 (PIC_OFFSET_TABLE_REGNUM): Set.
20391 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
20392 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
20393 (addsi3, movsi_internal2, movdf_internal): Update constraints
20394 * config/microblaze/predicates.md (arith_plus_operand): Define
20395 (move_operand): Redefine as move_src_operand,
20396 check microblaze_tls_referenced_p.
20398 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20400 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
20401 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
20403 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20405 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
20408 2013-03-14 Jakub Jelinek <jakub@redhat.com>
20410 PR tree-optimization/53265
20411 * common.opt (Waggressive-loop-optimizations): New option.
20412 * tree-ssa-loop-niter.c: Include tree-pass.h.
20413 (do_warn_aggressive_loop_optimizations): New function.
20414 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
20415 if number_of_latch_executions returned constant.
20416 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
20417 early. If number_of_latch_executions returned constant, set
20418 nb_iterations_upper_bound back to it.
20419 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
20421 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
20422 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
20424 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
20425 (MULTILIB_OSDIRNAMES): Set.
20426 * genmultilib: If defaultosdirname doesn't start with :: , set
20427 defaultosdirname2 instead, clear it and emit two . multilib_raw
20428 entries instead of just one.
20430 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
20432 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20433 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
20434 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20435 (SUBTARGET_OVERRIDE_OPTIONS): New.
20437 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
20440 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
20441 (musermode): Convert to Var(TARGET_USERMODE).
20442 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
20443 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
20444 * config/sh/sh.c (sh_option_override): Use
20445 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
20446 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
20448 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
20450 (udivsi3_i4_single, divsi3_i4_single): Use
20451 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
20453 2013-03-13 Dave Korn <dave.korn.cygwin@....>
20455 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
20458 2013-03-13 Richard Biener <rguenther@suse.de>
20460 PR tree-optimization/56608
20461 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
20462 calls when vectorizing basic-blocks.
20464 2013-03-13 Jakub Jelinek <jakub@redhat.com>
20467 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
20470 2013-03-12 Jakub Jelinek <jakub@redhat.com>
20472 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
20474 2013-03-11 Jan Hubicka <jh@suse.cz>
20477 * lto-streamer-out.c (output_symbol_p): Skip references from
20478 constructors of external variables.
20480 2013-03-11 Jan Hubicka <jh@suse.cz>
20482 PR middle-end/56571
20483 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
20485 * emit-rtl.c (verify_rtx_sharing): Likewise.
20486 (copy_insn_1): Likewise.
20487 * rtl.c (copy_rtx): Likewise.
20489 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
20492 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
20493 output_operand_lossage message.
20495 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
20498 * arm.c (shift_op): Validate RTL pattern on the fly.
20499 (arm_print_operand, case 'S'): Don't use shift_operator to validate
20502 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20505 * config/pa/pa.md (call_value): Check for calls to powf and direct to
20506 new call patterns that clobber %fr12.
20507 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
20508 split and postreload patterns.
20509 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
20510 registers %fr12 and %fr12R as call used.
20512 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
20514 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
20515 (canon_address, record_store, replace_read, check_mem_read_rtx,
20516 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
20517 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
20518 rest_of_handle_dse): Likewise.
20520 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
20522 PR middle-end/56524
20523 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
20525 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
20526 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
20527 (save_optabs_if_changed): Replace with...
20528 (init_tree_optimization_optabs): ...this.
20529 * optabs.c (save_optabs_if_changed): Rename to...
20530 (init_tree_optimization_optabs): ...this. Take the optimization node
20531 as argument. Do nothing if the base optabs are already correct.
20532 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
20533 to recompute optabs.
20534 * function.h (function): Remove optabs field.
20535 * function.c (invoke_set_current_function_hook): Call
20536 init_tree_optimization_optabs. Use the result to initialize
20539 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
20541 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
20542 if GTMA_HAS_NO_INSTRUMENTATION.
20543 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
20544 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
20545 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
20546 * gimple-pretty-print.c (dump_gimple_transaction): Handle
20547 GTMA_HAS_NO_INSTRUMENTATION.
20549 2013-03-08 Jakub Jelinek <jakub@redhat.com>
20551 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
20554 2013-03-08 Marek Polacek <polacek@redhat.com>
20555 Jakub Jelinek <jakub@redhat.com>
20557 PR tree-optimization/56478
20558 * predict.c (is_comparison_with_loop_invariant_p): Change the
20559 type of loop_step to tree.
20560 (predict_loops): Adjust.
20561 (predict_iv_comparison): Perform the computations on double_ints.
20563 2013-03-08 Richard Biener <rguenther@suse.de>
20565 PR tree-optimization/56570
20566 * tree-cfg.c (verify_expr_location_1): Verify locations for
20568 * tree-sra.c (create_access_replacement): Strip locations
20569 from DECL_DEBUG_EXPRs.
20571 2013-03-08 Richard Biener <rguenther@suse.de>
20573 * tree-inline.c (expand_call_inline): Do not associate
20574 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
20575 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
20577 2013-03-08 Richard Biener <rguenther@suse.de>
20579 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
20580 or block changes with -Og. Fix for location / block encoding
20581 changes and PHI arguments with locations.
20583 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
20585 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
20587 (struct output_info): Likewise.
20588 (register_overhead): Remove bad gcc_assert.
20589 (bitmap_find_bit): If there is only a single bitmap element, do not
20590 count a miss as a search.
20591 (print_statistics): Update for counter type changes.
20592 (dump_bitmap_statistics): Likewise. Print headers such that they
20593 are properly lined up with the printed counters.
20595 2013-03-07 Jakub Jelinek <jakub@redhat.com>
20597 PR tree-optimization/56559
20598 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
20599 check that it has only a single use.
20601 2013-03-07 Richard Biener <rguenther@suse.de>
20603 * doc/invoke.texi (fwhole-program): Discourage use in combination
20606 2013-03-06 Jakub Jelinek <jakub@redhat.com>
20608 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
20610 PR tree-optimization/56539
20611 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
20612 instead of GSI_CONTINUE_LINKING as last argument to
20613 force_gimple_operand_gsi. Adjust function comment.
20615 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
20618 PR middle-end/56548
20619 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
20620 promoted mode, convert the result back to the original mode.
20622 2013-03-06 Richard Biener <rguenther@suse.de>
20624 PR middle-end/56294
20625 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
20626 (insert_updated_phi_nodes_compare_uids): New function.
20627 (update_ssa): Sort symbols_to_rename after UID before
20628 traversing it to insert PHI nodes.
20630 2013-03-06 Richard Biener <rguenther@suse.de>
20632 PR middle-end/50494
20633 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
20634 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
20637 2013-02-13 Richard Biener <rguenther@suse.de>
20640 * varasm.c (output_constant_def_1): Get the decl representing
20641 the constant as argument.
20642 (output_constant_def): Wrap output_constant_def_1.
20643 (make_decl_rtl): Use output_constant_def_1 with the decl
20644 representing the constant.
20645 (build_constant_desc): Optionally re-use a decl already
20646 representing the constant.
20647 (tree_output_constant_def): Adjust.
20649 2013-03-06 Joey Ye <joey.ye@arm.com>
20652 * gcc.c (convert_white_space): New function.
20653 (main): Handles white space in function name.
20655 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
20658 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
20659 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
20660 to SH_DIV_CALL_TABLE for TARGET_SH2.
20661 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
20663 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
20664 call-table options.
20666 2013-03-05 Sterling Augustine <saugustine@google.com>
20667 Cary Coutant <ccoutant@google.com>
20670 * dwarf2out.c (resolve_addr): Don't call
20671 remove_loc_list_addr_table_entries a second time for the same
20674 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20677 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
20678 (avoid_complex_debug_insns): New function.
20679 (expand_debug_locations): Call it.
20681 PR rtl-optimization/56484
20682 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
20683 lifetimes of hard registers on small register class machines.
20685 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20687 * config/microblaze/microblaze-protos.h: Rename
20688 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
20689 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
20691 (microblaze_fast_interrupt_function_p): New function.
20692 (microblaze_is_interrupt_handler): Rename to
20693 microblaze_is_interrupt_variant and add fast_interrupt check.
20694 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
20695 (save_restore_insns): Likewise.
20696 (compute_frame_size): Likewise.
20697 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
20698 (microblaze_globalize_label): Likewise.
20699 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
20700 * config/microblaze/microblaze.md: Use wrapper
20701 microblaze_is_interrupt_variant.
20703 2013-03-05 Kai Tietz <ktietz@redhat.com>
20705 * sdbout.c (sdbout_one_type): Switch to current function's section
20706 supporting cold/hot.
20708 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20710 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
20713 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20715 PR middle-end/56461
20716 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
20717 if VALGRIND_GET_VBITS is defined, temporarily make object
20718 memory all defined, and restore previous valgrind addressability
20719 and definability afterwards. Free this_object at the end.
20721 PR middle-end/56461
20722 * lra.c (lra): Call lra_clear_live_ranges if live_p,
20723 right before calling lra_create_live_ranges, also call it
20724 when clearing live_p. Only call lra_clear_live_ranges
20725 at the end if live_p.
20727 PR middle-end/56461
20728 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
20730 2013-03-05 Richard Biener <rguenther@suse.de>
20732 PR tree-optimization/56521
20733 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
20736 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
20739 * except.h (remove_unreachable_eh_regions): New prototype.
20740 * except.c (remove_eh_handler_splicer): New function, split out
20741 of remove_eh_handler.
20742 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
20743 warning about running it on many EH regions one at a time.
20744 (remove_unreachable_eh_regions_worker): New function, walk the
20745 EH tree in depth-first order and remove non-marked regions.
20746 (remove_unreachable_eh_regions): New function.
20747 * tree-eh.c (mark_reachable_handlers): New function, split out
20748 from remove_unreachable_handlers.
20749 (remove_unreachable_handlers): Use mark_reachable_handlers and
20750 remove_unreachable_eh_regions.
20751 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
20752 and remove_unreachable_eh_regions.
20754 2013-03-05 Richard Biener <rguenther@suse.de>
20756 PR middle-end/56525
20757 * loop-init.c (fix_loop_structure): Remove loops in two stages,
20758 not freeing them until the end.
20760 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20762 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
20764 2013-03-05 Richard Biener <rguenther@suse.de>
20766 PR tree-optimization/56270
20767 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
20768 of loads after scheduling an SLP instance.
20770 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20772 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
20774 (check_gcc_parallelize): Run guality.exp as a separate job from
20775 vect.exp with unsorted.exp and $(dg_target_exps) separately from
20776 struct-layout-1.exp with stackalign.exp.
20778 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
20780 PR middle-end/56461
20781 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
20782 load_index sbitmap even if some bit in it isn't set.
20784 PR middle-end/56461
20785 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
20786 (discover_iteration_bound_by_body_walk): Change queues to
20787 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
20788 spelling in comment. Call safe_push on queues[bound_index] directly.
20789 Release queues[queue_index] in every iteration unconditionally.
20790 Release bounds vector.
20792 PR middle-end/56461
20793 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
20794 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
20795 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
20798 2013-03-05 Richard Biener <rguenther@suse.de>
20801 * tree-inline.c (remap_blocks_to_null): New function.
20802 (expand_call_inline): When expanding a call stmt without
20803 an associated block inline remap all callee blocks to NULL.
20805 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20807 PR rtl-optimization/56494
20808 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
20809 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
20810 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
20812 PR middle-end/56461
20813 * sel-sched-ir.c (free_sched_pools): Release
20814 succs_info_pool.stack[succs_info_pool.max_top] vectors too
20815 if succs_info_pool.max_top isn't -1.
20818 * opts.c (opts_obstack, opts_concat): Moved to...
20819 * opts-common.c (opts_obstack, opts_concat): ... here.
20821 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20823 PR middle-end/56461
20824 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
20826 2013-03-04 Martin Jambor <mjambor@suse.cz>
20828 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
20829 all appropriate places.
20831 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
20833 PR tree-optimization/56424
20834 * ipa-split.c (split_function): Do not set the RSO flag if result is
20835 not by reference and its type is a register type.
20837 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
20839 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
20840 (microblaze_legitimate_pic_operand): Likewise
20841 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
20842 new function microblaze_legitimate_pic_operand
20843 * config/microblaze/microblaze-protos.h
20844 (microblaze_legitimate_pic_operand): Declare.
20846 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20848 * config/microblaze/predicates.md (call_insn_simple_operand):
20849 New predicate for supported rtx code types.
20850 * config/microblaze/microblaze.md (call_internal1): Use
20851 call_insn_simple_operand predicate.
20853 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20855 PR middle-end/56461
20856 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
20857 partitions.ordered_remove.
20859 PR middle-end/56461
20860 * tree-vect-stmts.c (vectorizable_conversion): Don't call
20861 vec_oprnds0.create (1) for modifier == NONE.
20863 PR middle-end/56461
20864 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
20865 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
20866 vec_oprnds1 right before pushing anything to it for
20869 PR middle-end/56461
20870 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
20871 set nbbs to 0 instead of having separate code path.
20872 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
20873 instead of false as last argument if returning NULL.
20875 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
20877 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
20878 the attribute is now called "target" instead of "option".
20879 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
20880 * doc/tm.texi.in (Target Attributes): Likewise document the correct
20881 attribute/pragma name for TARGET_OPTION_VALID_P and
20882 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
20883 * doc/tm.texi: Regenerated.
20885 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
20887 * config/microblaze/microblaze.c:
20888 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
20889 * config/microblaze/microblaze.h: Add -mxl-reorder to
20891 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
20892 instructions emitted if TARGET_REORDER.
20893 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
20894 or 0 for -m/-mno case, but initialises as 2 to detect default use case
20897 2013-03-01 Xinliang David Li <davidxl@google.com>
20899 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
20902 2013-03-01 Jakub Jelinek <jakub@redhat.com>
20904 PR middle-end/56461
20905 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
20906 vector even when returning true. Fix up function comment formatting.
20908 PR middle-end/56461
20909 * ira-build.c (ira_loop_nodes_count): New variable.
20910 (create_loop_tree_nodes): Initialize it.
20911 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
20913 PR middle-end/56461
20914 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
20915 method on dr_chain and result_chain.
20916 * tree-vect-stmts.c (vectorizable_store): Only call
20917 result_chain.create if j == 0.
20919 PR middle-end/56461
20920 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
20921 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
20922 before overwriting it.
20924 2013-03-01 Tobias Burnus <burnus@net-b.de>
20926 * doc/extended.texi (C Extensions): Change order in @menu
20928 (Other MIPS Built-in Functions): Move last MIPS entry before
20929 "picoChip Built-in Functions".
20930 (SH Built-in Functions): Move after RX Built-in Functions.
20931 * doc/gcc.texi (Introduction): Change order in @menu
20933 * doc/md.texi (Constraints): Ditto.
20934 * gty.texi (Type Information): Ditto.
20935 (User-provided marking routines for template types): Make
20937 * doc/invoke.texi (AArch64 Options): Move before
20938 "Adapteva Epiphany Options".
20940 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
20941 Jakub Jelinek <jakub@redhat.com>
20944 * asan.c (gate_asan): Lookup no_sanitize_address instead of
20945 no_address_safety_analysis attribute.
20946 * doc/extend.texi (no_address_safety_attribute): Rename to
20947 no_sanitize_address attribute, mention no_address_safety_analysis
20948 attribute as deprecated alias.
20950 2013-02-28 Jakub Jelinek <jakub@redhat.com>
20952 PR middle-end/56461
20953 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
20954 type to vec<vec<tree> > *.
20955 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
20956 to be vec<tree> instead of vec<tree> *, set vec_defs
20957 to vNULL and call vec_defs.create (number_of_vects), adjust other
20959 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20960 vectorizable_condition): Adjust vect_get_slp_defs callers.
20962 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20964 * config/aarch64/aarch64.c
20965 (aarch64_float_const_representable): Remove unused variable.
20967 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20969 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20971 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20973 * config/aarch64/aarch64-builtins.c
20974 (aarch64_init_simd_builtins): Make static.
20976 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20978 * config/aarch64/aarch64.c
20979 (aarch64_simd_make_constant): Make static.
20981 2013-02-28 Martin Jambor <mjambor@suse.cz>
20983 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20984 with no initialization to the RHS of debug statements.
20986 2013-02-28 Martin Jambor <mjambor@suse.cz>
20988 PR tree-optimization/56294
20989 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20991 (get_access_replacement): Do not call create_access_replacement.
20992 Assert a replacement exists.
20993 (get_repl_default_def_ssa_name): Create the replacement declaration
20996 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20998 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20999 final_end_function.
21001 2013-02-28 Marek Polacek <polacek@redhat.com>
21003 PR rtl-optimization/56466
21004 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
21005 if we're changing a loop.
21006 (peel_loops_completely): Likewise.
21008 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
21011 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
21013 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21016 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
21017 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
21018 INTX_FTYPE_FX, FX_FTYPE_INTX.
21019 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
21021 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21023 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
21024 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
21025 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
21026 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
21027 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
21028 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
21029 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
21030 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
21031 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
21032 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
21033 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
21034 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
21035 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
21036 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
21037 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
21038 (avrxmega6): Increase max flash segments from 5 to 6.
21039 * config/avr/t-multilib: Regenerate.
21040 * config/avr/avr-tables.opt: Regenerate.
21041 * doc/avr-mmcu.texi: Regenerate.
21043 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21045 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
21046 (avr_device_to_arch): Rename to avr_device_to_ld.
21047 (avr_device_to_as): New prototype.
21048 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
21049 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
21050 * config/avr/driver-avr.c (avr_device_to_as): New.
21051 (avr_device_to_arch): Rename to avr_device_to_ld.
21053 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21055 PR middle-end/56461
21056 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
21057 method on dr_chain and result_chain.
21059 PR middle-end/56461
21060 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
21061 pointer_set_destroy on not_executed_last_iteration.
21063 PR middle-end/56461
21064 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
21066 PR middle-end/56461
21067 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
21068 FOR_EACH_DEFINED_FUNCTION when freeing state.
21070 PR middle-end/56461
21071 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
21073 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
21076 PR middle-end/56461
21077 * ipa-cp.c (decide_whether_version_node): Call vec_free on
21078 known_aggs[i].items and release known_aggs vector.
21080 PR middle-end/56461
21081 * ipa-reference.c (propagate): Free node_info even for alias nodes.
21083 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21085 * config/microblaze/microblaze.c (microblaze_emit_compare):
21086 Use xor for EQ/NE comparisions.
21087 * config/microblaze/microblaze.md (cstoresf4): Add constraints
21088 (cbranchsf4): Adjust operator to comparison_operator.
21090 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21092 PR middle-end/56461
21093 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
21095 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
21096 vec_safe_push, always update *slot.
21097 (redirect_edge_var_map_clear): Use vec_free.
21098 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
21099 (free_var_map_entry): Use vec_free.
21100 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
21101 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
21103 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
21105 PR middle-end/45472
21106 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
21107 when the may_trap_p bit of the exprs being merged differs.
21108 Reorder tests for speculativeness in the logical and operator.
21110 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21112 * incpath.c (add_standard_paths): Use reconcat instead of concat
21113 where appropriate and avoid leaking memory.
21115 * opts.h: Include obstack.h.
21116 (opts_concat): New prototype.
21117 (opts_obstack): New declaration.
21118 * opts.c (opts_concat): New function.
21119 (opts_obstack): New variable.
21120 (init_options_struct): Call gcc_init_obstack on opts_obstack.
21121 (finish_options): Use opts_concat instead of concat
21122 and XOBNEWVEC instead of XNEWVEC.
21123 * opts-common.c (generate_canonical_option, decode_cmdline_option,
21124 generate_option): Likewise.
21125 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
21126 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
21129 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
21130 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
21132 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21134 PR middle-end/56461
21135 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
21137 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21139 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
21140 (arm_block_move_unaligned_straight): Likewise.
21141 (arm_adjust_block_mem): Likewise.
21143 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21146 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
21147 temp, cond and label.
21148 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
21151 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
21152 * config/c6x/c6x.h (dbx_register_map): Update declaration.
21155 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
21156 of prologue/epilogue functions.
21159 * config/tilegx/tilegx.c (tilegx_expand_prologue):
21160 Remove unused variable cfa_offset.
21161 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
21164 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
21165 type promotion to unsigned.
21168 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
21169 for HOST_WIDE_INT of 32 bit / same size as int.
21170 (arm_block_move_unaligned_straight): Likewise.
21171 (arm_adjust_block_mem): Likewise.
21174 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
21177 2013-02-26 Marek Polacek <polacek@redhat.com>
21179 PR tree-optimization/56426
21180 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
21182 2013-02-26 Richard Biener <rguenther@suse.de>
21185 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
21186 unused variable loops.
21188 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21190 PR tree-optimization/56448
21191 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
21192 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
21193 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
21194 later operands of the references, or even first operand for
21195 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
21197 PR tree-optimization/56443
21198 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
21199 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
21200 to type_for_mode langhook.
21202 2013-02-25 Matt Turner <mattst88@gmail.com>
21204 * doc/invoke.texi: Document r4700.
21206 2013-02-25 Richard Biener <rguenther@suse.de>
21208 PR tree-optimization/56175
21209 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
21211 (simplify_bitwise_binary): ... here. Also guard the conversion
21212 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
21214 2013-02-25 Catherine Moore <clm@codesourcery.com>
21217 2013-02-24 Catherine Moore <clm@codesourcery.com>
21218 Maciej W. Rozycki <macro@codesourcery.com>
21219 Tom de Vries <tom@codesourcery.com>
21220 Nathan Sidwell <nathan@codesourcery.com>
21221 Iain Sandoe <iain@codesourcery.com>
21222 Nathan Froyd <froydnj@codesourcery.com>
21223 Chao-ying Fu <fu@mips.com>
21225 * doc/extend.texi: (micromips, nomicromips, nocompression):
21226 Document new function attributes.
21227 * doc/invoke.texi (minterlink-compressed, mmicromips,
21228 m14k, m14ke, m14kec): Document new options.
21229 (minterlink-mips16): Update documentation.
21230 * doc/md.texi (ZC, ZD): Document new constraints.
21231 * configure.ac (gcc_cv_as_micromips): Check if linker
21232 supports the .set micromips directive.
21233 * configure: Regenerate.
21234 * config.in: Regenerate.
21235 * config/mips/mips-tables.opt: Regenerate.
21236 * config/mips/micromips.md: New file.
21237 * constraints.md (ZC, AD): New constraints.
21238 * config/mips/predicates.md (movep_src_register): New predicate.
21239 (movep_src_operand): New predicate.
21240 (non_volatile_mem_operand): New predicate.
21241 * config/mips/mips.md (multimem): New type.
21242 (length): Differentiate between 17-bit and 18-bit branch offsets.
21243 (MOVEP1, MOVEP2): New mode iterator.
21244 (mov_<load>l): Use ZC constraint.
21245 (mov_<load>r): Likewise.
21246 (mov_<store>l): Likewise.
21247 (mov_<store>r): Likewise.
21248 (*branch_equality<mode>_inverted): Add microMIPS support.
21249 (*branch_equality<mode>): Likewise.
21250 (*jump_absolute): Likewise.
21251 (indirect_jump_<mode>): Likewise.
21252 (tablejump_<mode>): Likewise.
21253 (<optab>_internal): Likewise.
21254 (sibcall_internal): Likewise.
21255 (sibcall_value_internal): Likewise.
21256 (prefetch): Use constraint ZD.
21257 * config/mips/mips.opt (minterlink-compressed): New option.
21258 (minterlink-mips16): Now an alias for minterlink-compressed.
21259 (mmicromips): New option.
21260 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21261 (compare_and_swap_12): Likewise.
21262 (sync_add<mode>): Likewise.
21263 (sync_<optab>_12): Likewise.
21264 (sync_old_<optab>_12): Likewise.
21265 (sync_new_<optab>_12): Likewise.
21266 (sync_nand_12): Likewise.
21267 (sync_old_nand_12): Likewise.
21268 (sync_new_nand_12): Likewise.
21269 (sync_sub<mode>): Likewise.
21270 (sync_old_add<mode>): Likewise.
21271 (sync_old_sub<mode>): Likewise.
21272 (sync_new_add<mode>): Likewise.
21273 (sync_new_sub<mode>): Likewise.
21274 (sync_<optab><mode>): Likewise.
21275 (sync_old_<optab><mode>): Likewise.
21276 (sync_new_<optab><mode>): Likewise.
21277 (sync_nand<mode>): Likewise.
21278 (sync_old_nand<mode>): Likewise.
21279 (sync_new_nand<mode>): Likewise.
21280 (sync_lock_test_and_set<mode>): Likewise.
21281 (test_and_set_12): Likewise.
21282 (atomic_compare_and_swap<mode>): Likewise.
21283 (atomic_exchange<mode>_llsc): Likewise.
21284 (atomic_fetch_add<mode>_llsc): Likewise.
21285 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21286 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21287 (umips_save_restore_pattern_p): Likewise.
21288 (umips_load_store_pair_p): Likewise.
21289 (umips_output_load_store_pair): Likewise.
21290 (umips_movep_target_p): Likewise.
21291 (umips_12bit_offset_address_p): Likewise.
21292 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21293 (mips_base_mips16): Rename this...
21294 (mips_base_compression_flags): ...to this. Update all uses.
21295 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21296 (mips_mips16_decl_p): Delete.
21297 (mips_nomips16_decl_p): Delete.
21298 (mips_get_compress_on_flags): New function.
21299 (mips_get_compress_off_flags): New function.
21300 (mips_get_compress_mode): New function.
21301 (mips_get_compress_on_name): New function.
21302 (mips_get_compress_off_name): New function.
21303 (mips_insert_attributes): Support multiple compression types.
21304 (mips_merge_decl_attributes): Likewise.
21305 (umips_12bit_offset_address_p): New function.
21306 (mips_start_function_definition): Emit .set micromips directive.
21307 (mips_call_may_need_jalx_p): New function.
21308 (mips_function_ok_for_sibcall): Add microMIPS support.
21309 (mips_print_operand_punctuation): Support short delay slots and
21311 (umips_swm_mask, umips_swm_encoding): New.
21312 (umips_build_save_restore): New function.
21313 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21314 (was_mips16_p): Remove.
21315 (old_compression_mode): New.
21316 (mips_set_compression_mode): New function.
21317 (mips_set_current_function): Add microMIPS support.
21318 (mips_option_override): Likewise.
21319 (umips_save_restore_pattern_p): New function.
21320 (umips_output_save_restore): New function.
21321 (umips_load_store_pair_p_1): New function.
21322 (umips_load_store_pair_p): New function.
21323 (umips_output_load_store_pair_1): New function.
21324 (umips_output_load_store_pair): New function.
21325 (umips_movep_target_p) New function.
21326 (mips_prepare_pch_save): Add microMIPS support.
21327 * config/mips/mips.h (TARGET_COMPRESSION): New.
21328 (TARGET_CPU_CPP_BUILTINS): Update macro
21329 to use new compression flags and to support microMIPS.
21330 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21331 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21332 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21333 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21334 (ASM_SPEC): Support mmicromips and mno-micromips.
21335 (M16STORE_REG_P): New macro.
21336 (MIPS_CALL): Support TARGET_MICROMIPS.
21337 (MICROMIPS_J): New macro.
21338 (mips_base_mips16): Rename this...
21339 (mips_base_compression_flags): ...to this.
21340 (UMIPS_12BIT_OFFSET_P): New macro.
21341 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21342 (MULTILIB_DIRNAMES): Likewise.
21344 2013-02-25 Tom de Vries <tom@codesourcery.com>
21346 PR rtl-optimization/56131
21347 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
21348 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
21349 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
21351 2013-02-25 Tobias Burnus <burnus@net-b.de>
21353 * doc/invoke.texi (-fsanitize=): Move from optimization
21354 to debugging options.
21356 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21358 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
21360 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21361 Alexander Monakov <amonakov@ispras.ru>
21363 PR middle-end/56077
21364 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
21365 flush pending lists also on non-jumps. Adjust comment.
21367 2013-02-24 Catherine Moore <clm@codesourcery.com>
21368 Maciej W. Rozycki <macro@codesourcery.com>
21369 Tom de Vries <tom@codesourcery.com>
21370 Nathan Sidwell <nathan@codesourcery.com>
21371 Iain Sandoe <iain@codesourcery.com>
21372 Nathan Froyd <froydnj@codesourcery.com>
21373 Chao-ying Fu <fu@mips.com>
21375 * doc/extend.texi: (micromips, nomicromips, nocompression):
21376 Document new function attributes.
21377 * doc/invoke.texi (minterlink-compressed, mmicromips,
21378 m14k, m14ke, m14kec): Document new options.
21379 (minterlink-mips16): Update documentation.
21380 * doc/md.texi (ZC, ZD): Document new constraints.
21381 * configure.ac (gcc_cv_as_micromips): Check if linker
21382 supports the .set micromips directive.
21383 * configure: Regenerate.
21384 * config.in: Regenerate.
21385 * config/mips/mips-tables.opt: Regenerate.
21386 * config/mips/micromips.md: New file.
21387 * constraints.md (ZC, AD): New constraints.
21388 * config/mips/predicates.md (movep_src_register): New predicate.
21389 (movep_src_operand): New predicate.
21390 (non_volatile_mem_operand): New predicate.
21391 * config/mips/mips.md (multimem): New type.
21392 (length): Differentiate between 17-bit and 18-bit branch offsets.
21393 (MOVEP1, MOVEP2): New mode iterator.
21394 (mov_<load>l): Use ZC constraint.
21395 (mov_<load>r): Likewise.
21396 (mov_<store>l): Likewise.
21397 (mov_<store>r): Likewise.
21398 (*branch_equality<mode>_inverted): Add microMIPS support.
21399 (*branch_equality<mode>): Likewise.
21400 (*jump_absolute): Likewise.
21401 (indirect_jump_<mode>): Likewise.
21402 (tablejump_<mode>): Likewise.
21403 (<optab>_internal): Likewise.
21404 (sibcall_internal): Likewise.
21405 (sibcall_value_internal): Likewise.
21406 (prefetch): Use constraint ZD.
21407 * config/mips/mips.opt (minterlink-compressed): New option.
21408 (minterlink-mips16): Now an alias for minterlink-compressed.
21409 (mmicromips): New option.
21410 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21411 (compare_and_swap_12): Likewise.
21412 (sync_add<mode>): Likewise.
21413 (sync_<optab>_12): Likewise.
21414 (sync_old_<optab>_12): Likewise.
21415 (sync_new_<optab>_12): Likewise.
21416 (sync_nand_12): Likewise.
21417 (sync_old_nand_12): Likewise.
21418 (sync_new_nand_12): Likewise.
21419 (sync_sub<mode>): Likewise.
21420 (sync_old_add<mode>): Likewise.
21421 (sync_old_sub<mode>): Likewise.
21422 (sync_new_add<mode>): Likewise.
21423 (sync_new_sub<mode>): Likewise.
21424 (sync_<optab><mode>): Likewise.
21425 (sync_old_<optab><mode>): Likewise.
21426 (sync_new_<optab><mode>): Likewise.
21427 (sync_nand<mode>): Likewise.
21428 (sync_old_nand<mode>): Likewise.
21429 (sync_new_nand<mode>): Likewise.
21430 (sync_lock_test_and_set<mode>): Likewise.
21431 (test_and_set_12): Likewise.
21432 (atomic_compare_and_swap<mode>): Likewise.
21433 (atomic_exchange<mode>_llsc): Likewise.
21434 (atomic_fetch_add<mode>_llsc): Likewise.
21435 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21436 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21437 (umips_save_restore_pattern_p): Likewise.
21438 (umips_load_store_pair_p): Likewise.
21439 (umips_output_load_store_pair): Likewise.
21440 (umips_movep_target_p): Likewise.
21441 (umips_12bit_offset_address_p): Likewise.
21442 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21443 (mips_base_mips16): Rename this...
21444 (mips_base_compression_flags): ...to this. Update all uses.
21445 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21446 (mips_mips16_decl_p): Delete.
21447 (mips_nomips16_decl_p): Delete.
21448 (mips_get_compress_on_flags): New function.
21449 (mips_get_compress_off_flags): New function.
21450 (mips_get_compress_mode): New function.
21451 (mips_get_compress_on_name): New function.
21452 (mips_get_compress_off_name): New function.
21453 (mips_insert_attributes): Support multiple compression types.
21454 (mips_merge_decl_attributes): Likewise.
21455 (umips_12bit_offset_address_p): New function.
21456 (mips_start_function_definition): Emit .set micromips directive.
21457 (mips_call_may_need_jalx_p): New function.
21458 (mips_function_ok_for_sibcall): Add microMIPS support.
21459 (mips_print_operand_punctuation): Support short delay slots and
21461 (umips_swm_mask, umips_swm_encoding): New.
21462 (umips_build_save_restore): New function.
21463 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21464 (was_mips16_p): Remove.
21465 (old_compression_mode): New.
21466 (mips_set_compression_mode): New function.
21467 (mips_set_current_function): Add microMIPS support.
21468 (mips_option_override): Likewise.
21469 (umips_save_restore_pattern_p): New function.
21470 (umips_output_save_restore): New function.
21471 (umips_load_store_pair_p_1): New function.
21472 (umips_load_store_pair_p): New function.
21473 (umips_output_load_store_pair_1): New function.
21474 (umips_output_load_store_pair): New function.
21475 (umips_movep_target_p) New function.
21476 (mips_prepare_pch_save): Add microMIPS support.
21477 * config/mips/mips.h (TARGET_COMPRESSION): New.
21478 (TARGET_CPU_CPP_BUILTINS): Update macro
21479 to use new compression flags and to support microMIPS.
21480 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21481 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21482 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21483 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21484 (ASM_SPEC): Support mmicromips and mno-micromips.
21485 (M16STORE_REG_P): New macro.
21486 (MIPS_CALL): Support TARGET_MICROMIPS.
21487 (MICROMIPS_J): New macro.
21488 (mips_base_mips16): Rename this...
21489 (mips_base_compression_flags): ...to this.
21490 (UMIPS_12BIT_OFFSET_P): New macro.
21491 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21492 (MULTILIB_DIRNAMES): Likewise.
21494 2013-02-24 Jakub Jelinek <jakub@redhat.com>
21497 * target-globals.c (save_target_globals): For init_reg_sets and
21498 target_reinit remporarily set this_fn_optabs to this_target_optabs.
21500 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
21502 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
21503 * config/aarch64/t-aarch64
21504 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
21506 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
21508 PR inline-asm/56148
21509 * lra-constraints.c (process_alt_operands): Reload operand
21510 conflicting with earlier clobber only if no more other conflicting
21513 2013-02-22 Jakub Jelinek <jakub@redhat.com>
21516 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
21517 if not linking a shared library.
21519 2013-02-22 Seth LaForge <sethml@google.com>
21521 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
21523 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
21525 * config/arm/arm.md (split for extendsidi): Update condition.
21526 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
21527 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
21528 (qhs_zextenddi_cstr): Likewise.
21530 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21532 PR middle-end/56420
21533 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
21534 avoid signed wrapping.
21535 (expand_mult): Handle properly multiplication by
21536 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
21537 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
21538 in the compiler if coeff is HOST_WIDE_INT_MIN.
21539 (expand_divmod): Don't make ext_op1 static, change it's type to
21540 uhwi. Avoid undefined behavior in -INTVAL (op1).
21542 PR rtl-optimization/50339
21543 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
21545 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
21546 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
21547 into splitting_ashiftrt field.
21548 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
21550 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
21553 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
21555 PR middle-end/56108
21556 * trans-mem.c (execute_tm_mark): Do not expand transactions that
21557 are sure to go irrevocable.
21559 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
21561 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
21562 scalars are valid operands.
21564 2013-02-21 Martin Jambor <mjambor@suse.cz>
21566 PR tree-optimization/56310
21567 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
21568 only matching indices and non-negative final offsets.
21569 (intersect_aggregates_with_edge): Pass src_idx to
21570 agg_replacements_to_vector. Pass src_idx insstead of index to
21571 intersect_with_agg_replacements.
21573 2013-02-21 Martin Jambor <mjambor@suse.cz>
21575 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
21576 instead of hard-wired defaults.
21578 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
21580 * doc/invoke.texi (MIPS Options): Update documentation of the
21581 floating-point multiply-accumulate instruction restrictions.
21583 2013-02-21 Kostya Serebryany <kcc@google.com>
21585 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
21586 asan_shadow_offset on x86_64 linux.
21588 2013-02-21 Richard Biener <rguenther@suse.de>
21590 PR tree-optimization/56415
21592 2013-02-11 Richard Biener <rguenther@suse.de>
21594 PR tree-optimization/56273
21595 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21598 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21601 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
21604 PR inline-asm/56405
21605 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
21606 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
21608 2013-02-20 Jan Hubicka <jh@suse.cz>
21610 PR tree-optimization/56265
21611 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
21612 when target is referenced for first time.
21614 2013-02-20 Richard Biener <rguenther@suse.de>
21616 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
21617 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
21618 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
21619 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
21620 not return anything.
21621 (rename_ssa_copies): Do not remove unused locals.
21622 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
21623 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
21624 * passes.c (execute_function_todo): Do not schedule unused locals
21625 removal if cleanup_tree_cfg did something.
21626 * tree-ssa-live.c (remove_unused_locals): Dump statistics
21627 about the number of removed locals.
21629 2013-02-20 Richard Biener <rguenther@suse.de>
21631 PR tree-optimization/56398
21632 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
21634 2013-02-20 Martin Jambor <mjambor@suse.cz>
21636 PR tree-optimization/55334
21637 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
21638 restricted pointers to arrays.
21640 2013-02-20 Richard Biener <rguenther@suse.de>
21641 Jakub Jelinek <jakub@redhat.com>
21643 PR tree-optimization/56396
21644 * tree-ssa-ccp.c (n_const_val): New static variable.
21645 (get_value): Return NULL for SSA names we don't have a lattice
21647 (ccp_initialize): Initialize n_const_val.
21648 * tree-ssa-copy.c (n_copy_of): New static variable.
21649 (init_copy_prop): Initialize n_copy_of.
21650 (get_value): Return NULL_TREE for SSA names we don't have a
21653 2013-02-20 Martin Jambor <mjambor@suse.cz>
21655 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
21657 2013-02-20 Richard Biener <rguenther@suse.de>
21659 * genpreds.c (write_lookup_constraint): Do not compare first
21660 letter of the constraint again.
21662 2013-02-20 Richard Biener <rguenther@suse.de>
21664 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
21666 (get_use_iv_cost): Terminate hashtable walk when coming across
21669 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
21671 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
21672 reassociation for avx2 targets.
21674 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21676 * config/microblaze/microblaze.c: microblaze_has_clz = 0
21677 Add version check for v8.10.a to enable microblaze_has_clz
21678 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
21679 version and TARGET_PATTERN_COMPARE check
21680 * config/microblaze/microblaze.md: New clzsi2 instruction
21682 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21684 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
21685 function before branching.
21687 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21689 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
21691 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
21693 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21695 PR middle-end/55889
21696 * sel-sched.c: Include ira.h.
21697 (implicit_clobber_conflict_p): New function.
21698 (moveup_expr): Use it.
21699 * Makefile.in (sel-sched.o): Depend on ira.h.
21701 2013-02-19 Richard Biener <rguenther@suse.de>
21703 PR tree-optimization/56384
21704 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
21705 (vn_hash_type): Split out from ...
21706 (vn_hash_constant_with_type): ... here.
21707 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
21708 (vn_phi_eq): Compare types from vn_phi_s structure.
21709 (vn_phi_lookup): Populate vn_phi_s type.
21710 (vn_phi_insert): Likewise.
21712 2013-02-19 Jakub Jelinek <jakub@redhat.com>
21714 PR tree-optimization/56350
21715 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
21716 if haven't found reduction or nested cycle operand, rather than
21717 asserting we must find it.
21719 PR tree-optimization/56381
21720 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
21723 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
21724 Jakub Jelinek <jakub@redhat.com>
21727 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
21728 (swap_optab_enable): Same.
21729 (init_all_optabs): Use argument instead of global.
21730 * tree.h (struct tree_optimization_option): New field target_optabs.
21731 * expr.h (init_all_optabs): Add argument to prototype.
21732 (TREE_OPTIMIZATION_OPTABS): New.
21733 (save_optabs_if_changed): Protoize.
21734 * optabs.h: Declare this_fn_optabs.
21735 * optabs.c (save_optabs_if_changed): New.
21736 Declare this_fn_optabs.
21737 (init_optabs): Add argument to init_all_optabs() call.
21738 * function.c (invoke_set_current_function_hook): Handle per
21740 * function.h (struct function): New field optabs.
21741 * config/mips/mips.c (mips_set_mips16_mode): Handle when
21742 optimization_current_node has changed.
21743 * target-globals.h (save_target_globals_default_opts): Protoize.
21744 * target-globals.c (save_target_globals_default_opts): New.
21746 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21749 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
21750 registers %fr12 and %fr12R as call used.
21753 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
21754 and HImode, require all displacements to be an integer multiple of
21756 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
21757 only allow QImode and HImode when reload is in progress and strict is
21758 true. Likewise for symbolic addresses. Use base14_operand to check
21759 displacements in REG+BASE addresses.
21761 2013-02-18 Richard Biener <rguenther@suse.de>
21763 PR tree-optimization/56366
21764 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21765 handle sign-conversion of outer-loop initial induction value.
21767 2013-02-18 Richard Biener <rguenther@suse.de>
21769 PR middle-end/56349
21770 * cfghooks.c (merge_blocks): If we merge a latch into another
21771 block adjust references to it.
21772 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
21773 (verify_loop_structure): Verify that a recorded latch is in fact
21776 2013-02-18 Richard Biener <rguenther@suse.de>
21778 PR tree-optimization/56321
21779 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
21780 order SSA name release and virtual operand unlinking.
21782 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21784 * config/microblaze/microblaze.md (save_stack_block): Define.
21785 (restore_stack_block): Likewise.
21787 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21789 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
21790 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
21791 * config/microblaze/microblaze.c (microblaze_option_override):
21792 Bail out early for PIC modes when target does not support PIC.
21794 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21796 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
21797 Replace with a microblaze version.
21798 (microblaze_trampoline_init): Adapt for microblaze.
21799 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
21802 2013-02-16 Jakub Jelinek <jakub@redhat.com>
21803 Dodji Seketeli <dodji@redhat.com>
21806 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
21807 (instrument_mem_region_access): Do not forget to always put
21808 instrumentation of the of 'base' and 'base + len' in a "if (len !=
21809 0) statement, even for cases where either 'base' or 'base + len'
21810 are not instrumented -- because they have been previously
21811 instrumented. Simplify the logic by putting all the statements
21812 instrument 'base + len' inside a sequence, and then insert that
21813 sequence right before the current insertion point. Then, to
21814 instrument 'base + len', just get an iterator on that statement.
21815 And do not forget to update the pointer to iterator the function
21816 received as argument.
21818 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
21820 PR rtl-optimization/56348
21821 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
21823 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
21825 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
21826 (clean_graph_dump_file): Pass base to start_graph_dump.
21828 2013-02-14 Richard Henderson <rth@redhat.com>
21831 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
21833 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
21835 * collect2-aix.h: Define F_LOADONLY.
21837 2013-02-14 Richard Biener <rguenther@suse.de>
21840 * varasm.c (output_constant_def_1): Get the decl representing
21841 the constant as argument.
21842 (output_constant_def): Wrap output_constant_def_1.
21843 (make_decl_rtl): Use output_constant_def_1 with the decl
21844 representing the constant.
21845 (build_constant_desc): Optionally re-use a decl already
21846 representing the constant.
21847 (tree_output_constant_def): Adjust.
21849 2013-02-14 Dodji Seketeli <dodji@redhat.com>
21852 * asan.c (instrument_builtin_call): Really put the length of the
21853 second source argument into src1_len.
21855 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21857 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
21858 argument. If it is false, don't create edge from then_bb to
21860 (insert_if_then_before_iter): Pass true to it.
21861 (build_check_stmt): Pass false to it.
21862 (transform_statements): Flush hash table only on extended basic
21863 block boundaries, rather than at the beginning of every bb.
21864 Don't flush hash table on nonfreeing_call_p calls.
21865 * tree-flow.h (nonfreeing_call_p): New prototype.
21866 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
21868 2013-02-13 David S. Miller <davem@davemloft.net>
21870 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
21872 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
21875 * ira.c (max_regno_before_ira): Move from ...
21877 (fix_reg_equiv_init): Use max_regno_before_ira instead of
21880 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21882 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
21884 2013-02-13 Richard Biener <rguenther@suse.de>
21887 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
21888 globals in MEM_REFs.
21890 2013-02-13 Richard Biener <rguenther@suse.de>
21892 * loop-init.c (loop_optimizer_init): Clear loop state when
21893 re-initializing preserved loops.
21894 * loop-unswitch.c (unswitch_single_loop): Return whether
21895 we unswitched the loop. Do not verify loop state here.
21896 (unswitch_loops): When we unswitched a loop discover new loops.
21898 2013-02-13 Kostya Serebryany <kcc@google.com>
21900 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
21902 * sanitizer.def: Rename __asan_init to __asan_init_v1.
21904 2013-02-12 Dodji Seketeli <dodji@redhat.com>
21906 Avoid instrumenting duplicated memory access in the same basic block
21907 * Makefile.in (asan.o): Add new dependency on hash-table.h
21908 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
21909 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
21910 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
21911 (free_mem_ref_resources, has_mem_ref_been_instrumented)
21912 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
21913 (get_mem_ref_of_assignment): New functions.
21914 (get_mem_refs_of_builtin_call): Extract from
21915 instrument_builtin_call and tweak a little bit to make it fit with
21917 (instrument_builtin_call): Use the new
21918 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
21919 of is_gimple_builtin_call.
21920 (instrument_derefs, instrument_mem_region_access): Insert the
21921 instrumented memory reference into the hash table.
21922 (maybe_instrument_assignment): Renamed instrument_assignment into
21923 this, and change it to advance the iterator when instrumentation
21924 actually happened and return true in that case. This makes it
21925 homogeneous with maybe_instrument_assignment, and thus give a
21926 chance to callers to be more 'regular'.
21927 (transform_statements): Clear the memory reference hash table
21928 whenever we enter a new BB, when we cross a function call, or when
21929 we are done transforming statements. Use
21930 maybe_instrument_assignment instead of instrumentation. No more
21931 need to special case maybe_instrument_assignment and advance the
21932 iterator after calling it; it's now handled just like
21933 maybe_instrument_call. Update comment.
21935 2013-02-13 Richard Biener <rguenther@suse.de>
21937 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
21938 Fix loop discovery code.
21940 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
21942 PR inline-asm/56148
21943 * lra-constraints.c (process_alt_operands): Match early clobber
21944 operand with itself. Check conflicts with earlyclobber only if
21945 the operand is not reloaded. Prefer to reload conflicting operand
21946 if earlyclobber and matching operands are the same.
21948 2013-02-12 Richard Biener <rguenther@suse.de>
21951 * lto-streamer-out.c (write_symbol): Do not output symbols
21952 for hard register variables.
21954 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
21957 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21958 (umulsidi3_insn, mulsidi3_insn): New insns.
21960 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21962 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21963 (struct tune_params): Add vec_costs field.
21964 * config/arm/arm.c (arm_builtin_vectorization_cost)
21965 (arm_add_stmt_cost): New functions.
21966 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21967 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21968 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21969 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21970 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21971 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21972 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21974 2013-02-12 Richard Biener <rguenther@suse.de>
21977 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21978 decls again if possible.
21980 2013-02-12 Richard Biener <rguenther@suse.de>
21982 PR middle-end/56288
21983 * tree-ssa.c (verify_ssa_name): Fix check, move
21984 SSA_NAME_IN_FREE_LIST check up.
21986 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21987 Steven Bosscher <steven@gcc.gnu.org>
21989 PR rtl-optimization/56151
21990 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21991 equal to op0 or op1, and last_insn pattern is CODE operation
21992 with MEM dest and one of the operands matches that MEM.
21994 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21996 * doc/extend.texi: Document Function Multiversioning and "default"
21997 parameter string to target attribute.
21998 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21999 target attribute parameter is "default".
22000 (ix86_compare_version_priority): Remove checks for target attribute.
22001 (ix86_mangle_function_version_assembler_name): Change error to sorry.
22002 Remove check for target attribute equal to NULL. Add assert.
22003 (ix86_generate_version_dispatcher_body): Change error to sorry.
22005 2013-02-11 Iain Sandoe <iain@codesourcery.com>
22006 Jack Howarth <howarth@bromo.med.uc.edu>
22007 Patrick Marlier <patrick.marlier@gmail.com>
22010 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
22011 define ENDFILE_SPEC as TM_DESTRUCTOR.
22012 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
22014 2013-02-11 Alexander Potapenko <glider@google.com>
22015 Jack Howarth <howarth@bromo.med.uc.edu>
22016 Jakub Jelinek <jakub@redhat.com>
22019 * config/darwin.c (cdtor_record): Rename ctor_record.
22020 (sort_cdtor_records): Rename sort_ctor_records.
22021 (finalize_dtors): New routine to sort destructors by
22022 priority before use in assemble_integer.
22023 (machopic_asm_out_destructor): Use finalize_dtors if needed.
22025 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
22027 PR rtl-optimization/56275
22028 * simplify-rtx.c (avoid_constant_pool_reference): Check that
22029 offset is non-negative and less than cmode size before
22030 calling simplify_subreg.
22032 2013-02-11 Richard Biener <rguenther@suse.de>
22034 PR tree-optimization/56264
22035 * cfgloop.h (fix_loop_structure): Adjust prototype.
22036 * loop-init.c (fix_loop_structure): Return the number of
22037 newly discovered loops.
22038 * tree-cfgcleanup.c (repair_loop_structures): When new loops
22039 are discovered, do a full loop-closed SSA rewrite.
22041 2013-02-11 Richard Biener <rguenther@suse.de>
22043 PR tree-optimization/56273
22044 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22046 (check_array_ref): Fix missing newline in dumps.
22047 (search_for_addr_array): Likewise.
22049 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
22051 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
22053 2013-02-09 Jakub Jelinek <jakub@redhat.com>
22056 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
22058 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
22060 PR rtl-optimization/56246
22061 * lra-constraints.c (simplify_operand_subreg): Try to reuse
22063 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
22064 constraints are satisfied.
22066 2013-02-08 Jeff Law <law@redhat.com>
22069 * emit-rtl.c (reg_is_parm_p): New function.
22070 * regs.h (reg_is_parm_p): New prototype.
22071 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
22072 callee-clobbered registers.
22074 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
22077 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
22078 If there is no implicit builtin declaration, just return NULL.
22080 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22082 * config/i386/sse.md (FMAMODEM): New mode iterator.
22083 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
22084 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
22086 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22088 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
22089 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
22090 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
22092 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22094 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
22095 (microblaze*-*-elf): Likewise.
22096 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
22098 * config/microblaze/microblaze-c.c: Add builtin defines for
22099 _LITTLE_ENDIAN and _BIG_ENDIAN.
22100 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
22101 add to TARGET_DEFAULT flags.
22102 Expand ASM_SPEC and LINK_SPEC.
22103 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
22104 * config/microblaze/microblaze.md: Update extendsidi2 and
22105 movdi_internal instructions to use low-order / high-order reg
22107 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
22108 options and inversemask / mask of LITTLE_ENDIAN.
22109 * config/microblaze/t-microblaze: Expand multilib options to
22110 include mlittle-endian (le) and update exceptions patterns.
22112 2013-02-08 Jakub Jelinek <jakub@redhat.com>
22114 PR rtl-optimization/56195
22115 * lra-constraints.c (get_reload_reg): Don't reuse regs
22116 if they have smaller mode than requested, if they have
22117 wider mode than requested, try to return a SUBREG.
22119 PR tree-optimization/56250
22120 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
22121 if type is unsigned and code isn't MULT_EXPR.
22123 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22125 PR tree-optimization/56064
22126 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
22127 bits according to mode.
22128 * fixed-value.h (fixed_from_double_int)
22129 (const_fixed_from_double_int): Adjust comments.
22131 2013-02-08 Richard Biener <rguenther@suse.de>
22134 * lto-streamer.h (struct data_in): Remove current_file, current_line
22135 and current_col members.
22136 * lto-streamer-out.c (lto_output_location): Stream changed bits
22137 en-block for efficiency.
22138 * lto-streamer-in.c (clear_line_info): Remove.
22139 (lto_input_location): Cache current file, line and column
22140 globally via local statics. Read changed bits en-block.
22141 (input_function): Do not call clear_line_info.
22142 (lto_read_body): Likewise.
22143 (lto_input_toplevel_asms): Likewise.
22145 2013-02-08 Michael Matz <matz@suse.de>
22147 PR tree-optimization/52448
22148 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
22149 (nt_call_phase): New static.
22150 (add_or_mark_expr): Only mark accesses with newer phase than any
22152 (nonfreeing_call_p): New.
22153 (nt_init_block): Update nt_call_phase, mark blocks as visited.
22154 (nt_fini_block): Keep blocks marked as visited.
22155 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
22157 2013-02-08 Richard Biener <rguenther@suse.de>
22159 * ira.c (ira): Free broken dominator information.
22161 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22163 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
22165 2013-02-08 Marek Polacek <polacek@redhat.com>
22167 * cfgloop.c (verify_loop_structure): Add more checking of headers.
22169 2013-02-08 Richard Biener <rguenther@suse.de>
22171 PR middle-end/56181
22172 * cfgloop.h (flow_loops_find): Adjust.
22173 (bb_loop_header_p): Declare.
22174 * cfgloop.c (bb_loop_header_p): New function split out from ...
22175 (flow_loops_find): ... here. Adjust function signature,
22176 support incremental loop structure update.
22177 (verify_loop_structure): Cleanup. Verify a loop is a loop.
22178 * cfgloopmanip.c (fix_loop_structure): Move ...
22179 * loop-init.c (fix_loop_structure): ... here.
22180 (apply_loop_flags): Split out from ...
22181 (loop_optimizer_init): ... here.
22182 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
22183 in incremental mode, only remove dead loops here.
22185 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22188 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
22189 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
22190 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
22191 (*round<mode>3.libgcc): New insns for fixed-modes.
22192 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
22193 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
22194 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
22195 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
22196 implementations. Define to __builtin_avr_absFX,
22197 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
22198 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
22199 __builtin_avr_countlsFX, respectively.
22200 * config/avr/avr-c.c (target.h): Include it.
22201 (enum avr_builtin_id): New enum.
22202 (avr_resolve_overloaded_builtin): New static function.
22203 (avr_register_target_pragmas): Use it to set
22204 targetm.resolve_overloaded_builtin.
22205 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
22206 tree nodes used by DEF_BUILTIN.
22207 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
22208 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
22209 <AVR_BUILTIN_xxBITS>: Same.
22211 2013-02-08 Richard Biener <rguenther@suse.de>
22213 * cfgloop.c (verify_loop_structure): Properly handle
22214 a loop exiting to another loop header.
22215 * ira-int.h (ira_loops): Remove.
22216 * ira.c (ira_loops): Remove.
22217 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
22218 (do_reload): Use loop_optimizer_finalize.
22219 * ira-build.c (create_loop_tree_nodes): Use get_loops and
22220 number_of_loops to access the loop tree.
22221 (more_one_region_p): Likewise.
22222 (finish_loop_tree_nodes): Likewise.
22223 (rebuild_regno_allocno_maps): Likewise.
22224 (mark_loops_for_removal): Likewise.
22225 (mark_all_loops_for_removal): Likewise.
22226 (remove_unnecessary_regions): Likewise.
22227 (ira_build): Likewise.
22228 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22230 2013-02-08 Richard Biener <rguenther@suse.de>
22232 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
22233 * ipa-pure-const.c (analyze_function): Avoid calling
22234 mark_irreducible_loops twice.
22235 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
22237 2013-02-07 David S. Miller <davem@davemloft.net>
22239 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
22241 * var-tracking.c (vt_add_function_parameter): Test the presence of
22242 HAVE_window_save properly and do not remap argument registers when
22243 we have a leaf function.
22245 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
22248 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
22250 * config/i386/i386.c (ix86_print_operand): Ditto.
22252 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
22254 * lra-constraints.c (process_alt_operands): Fix recently added comment.
22256 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
22258 PR rtl-optimization/56225
22259 * lra-constraints.c (process_alt_operands): Check that reload hard
22260 reg can hold value for strict_low_part.
22262 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22265 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
22266 dwarf2out_end_function.
22267 (in_first_function_p, maybe_at_text_label_p,
22268 first_loclabel_num_not_at_text_label): New variables.
22269 (dwarf2out_var_location): In the first function find out
22270 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
22271 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
22274 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
22276 PR rtl-optimization/56178
22277 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
22278 SUBREG of a register. Tidy up related block of code.
22279 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
22280 note if the source is a register or a SUBREG of a register.
22282 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22285 * config/rs6000/rs6000.md (ptrm): New mode attr.
22286 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
22287 call_value_indirect_aix<pttrsize>,
22288 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
22291 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22293 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
22294 if -bnortl. Convert to strcmp and strncmp.
22296 2013-02-07 Alan Modra <amodra@gmail.com>
22299 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
22300 addresses won't wrap when offsetting.
22301 (rs6000_secondary_reload): Provide secondary reloads needed for
22302 wrapping LO_SUM addresses.
22304 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
22306 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
22307 MACH, just __MACH__.
22309 2013-02-06 Richard Biener <rguenther@suse.de>
22311 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
22312 instead of calling fix_loop_structure.
22314 2013-02-06 Jakub Jelinek <jakub@redhat.com>
22316 PR middle-end/56217
22317 * omp-low.c (use_pointer_for_field): Return false if
22318 lower_send_shared_vars doesn't generate any copy-out code.
22320 2013-02-06 Tom de Vries <tom@codesourcery.com>
22322 PR rtl-optimization/56131
22323 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
22324 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
22325 of the label is NULL. Add comment.
22327 2013-02-05 Jakub Jelinek <jakub@redhat.com>
22329 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
22332 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
22333 (STATIC_LIBTSAN_LIBS): Likewise.
22334 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
22335 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
22336 is defined, don't add anything else beyond that.
22337 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
22338 (LINK_COMMAND_SPEC): Use them.
22340 PR tree-optimization/56205
22341 * tree-stdarg.c (check_all_va_list_escapes): Return true if
22342 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
22343 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
22345 2013-02-05 Richard Biener <rguenther@suse.de>
22347 PR tree-optimization/53342
22348 PR tree-optimization/53185
22349 * tree-vectorizer.h (vect_check_strided_load): Remove.
22350 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
22351 not disallow peeling for vectorized strided loads.
22352 (vect_check_strided_load): Make static and simplify.
22353 (vect_analyze_data_refs): Adjust.
22354 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
22355 correctly when vectorizing strided loads.
22357 2013-02-05 Richard Biener <rguenther@suse.de>
22359 * doc/install.texi: Refer to ISL, not PPL.
22361 2013-02-05 Jan Hubicka <jh@suse.cz>
22363 PR tree-optimization/55789
22364 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
22366 2013-02-05 Jan Hubicka <jh@suse.cz>
22368 PR tree-optimization/55789
22369 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
22370 the dead call anyway.
22372 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
22375 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
22377 2013-02-04 Alexander Potapenko <glider@google.com>
22378 Jack Howarth <howarth@bromo.med.uc.edu>
22379 Jakub Jelinek <jakub@redhat.com>
22382 * config/darwin.c (sort_ctor_records): Stabilized qsort
22383 on constructor priority by using original position.
22384 (finalize_ctors): New routine to sort constructors by
22385 priority before use in assemble_integer.
22386 (machopic_asm_out_constructor): Use finalize_ctors if needed.
22388 2013-02-04 Jakub Jelinek <jakub@redhat.com>
22391 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
22392 about visibility on artificial decls.
22393 * config/sol2.c (solaris_assemble_visibility): Likewise.
22395 2013-02-04 Kai Tietz <ktietz@redhat.com>
22398 * config/i386/i386.c (function_value_ms_64): Add additional valtype
22399 argument and improve checking of return-argument types for 16-byte
22401 (ix86_function_value_1): Add additional valtype argument on call
22402 of function_value_64.
22403 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
22404 handling infunction_value_64 function.
22406 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
22408 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
22410 2013-02-04 Richard Biener <rguenther@suse.de>
22412 PR tree-optimization/56188
22413 * tree-ssa-structalias.c (label_visit): Consider case with
22414 initially non-empty points-to set.
22415 (perform_var_substitution): Dump node mapping and clean up.
22417 2013-02-04 Richard Guenther <rguenther@suse.de>
22420 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
22421 node prevail as last resort.
22422 (lto_symtab_merge_decls): Remove guard on LTRANS here.
22423 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
22425 2013-02-04 Richard Biener <rguenther@suse.de>
22427 PR tree-optimization/56113
22428 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
22430 (equiv_class_lookup_or_add): ... this.
22431 (label_visit): Adjust and fix error in previous patch.
22432 (perform_var_substitution): Adjust.
22434 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
22436 * config/sh/divtab.c: Fix formatting and comments throughout the file.
22437 * config/sh/sh4-300.md: Likewise.
22438 * config/sh/sh4a.md: Likewise.
22439 * config/sh/constraints.md: Likewise.
22440 * config/sh/sh.md: Likewise.
22441 * config/sh/netbsd-elf.h: Likewise.
22442 * config/sh/predicates.md: Likewise.
22443 * config/sh/sh-protos.h: Likewise.
22444 * config/sh/ushmedia.h: Likewise.
22445 * config/sh/linux.h: Likewise.
22446 * config/sh/sh.c: Likewise.
22447 * config/sh/superh.h: Likewise.
22448 * config/sh/elf.h: Likewise.
22449 * config/sh/sh4.md: Likewise.
22450 * config/sh/sh.h: Likewise.
22452 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22454 * config/pa/constraints.md: Adjust unused letters. Change "T"
22455 constraint to match_test floating_point_store_memory_operand().
22456 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
22457 (base14_operand): New.
22458 (floating_point_store_memory_operand): New.
22459 (integer_store_memory_operand): Revise to use base14_operand and
22460 reg_plus_base_memory_operand.
22461 (move_dest_operand): Allow symbolic_memory_operands.
22462 (symbolic_memory_operand): Check for LO_SOM.
22463 (symbolic_operand): Change default case to break.
22464 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
22465 CONST_DOUBLE values to be reloaded by putting them into memory when
22466 the destination is a floating point register.
22467 (movdf): Remove code to handle CONST_DOUBLE.
22469 (reload_indf_r1): New.
22470 (reload_insf_r1): New.
22471 Consistently use "Q" and "T" constraints with integer and floating
22472 point move instructions, respectively.
22473 (movdi): Remove FAIL.
22474 Change predicate for source operand unamed DImode move from
22475 general_operand to move_src_operand.
22476 (umulsidi3): Change predicate for destination operand to
22478 Likewise for similar unamed patterns.
22479 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
22480 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
22481 (hppa_legitimize_address): Simplify mask calculation.
22482 (pa_emit_move_sequence): Revised handling of secondary reloads from
22483 REG+D addresses for floating point loads and stores. Directly handle
22484 loading CONST0_RTX (mode) to a floating point register.
22485 (pa_secondary_reload): Handle reloading DF and SFmode constant values
22486 to floating point registers. Don't restrict secondary reloads to
22487 floating point registers to integer modes. Revise some comments and
22489 (TARGET_LEGITIMATE_ADDRESS_P): Define.
22490 (pa_legitimate_address_p): New.
22491 (pa_legitimize_reload_address): New.
22492 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
22493 (STRICT_REG_OK_FOR_BASE_P): New.
22494 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
22495 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
22497 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
22498 Andrew Dixie <andrewd@gentrack.com>
22500 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
22503 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
22505 * expmed.c (extract_bit_field_1): Pass the full width of the
22506 structure to get_best_reg_extraction_insn.
22508 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22511 * configure.ac (use_cxa_atexit): Add AIX.
22512 * configure: Regenerate.
22514 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
22516 2013-02-01 Jakub Jelinek <jakub@redhat.com>
22519 * final.c (need_profile_function): New variable.
22520 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
22521 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
22522 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
22523 notes, targetm.asm_out.function_prologue doesn't emit anything,
22524 HAVE_prologue and profiler should be emitted before prologue,
22525 set need_profile_function instead of emitting it.
22526 (final_scan_insn): If need_profile_function, emit
22527 profile_function on the first NOTE_INSN_BASIC_BLOCK or
22528 NOTE_INSN_FUNCTION_BEG note.
22530 2013-02-01 Richard Henderson <rth@redhat.com>
22532 * config/rs6000/rs6000.md (smulditi3): New.
22535 * config/alpha/alpha.md (umulditi3): New.
22537 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22539 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
22540 (ASM_OUTPUT_ALIGNED_LOCAL): New.
22542 2013-02-01 Richard Biener <rguenther@suse.de>
22544 PR tree-optimization/56113
22545 * tree-ssa-structalias.c (label_visit): Reduce work for
22546 single-predecessor nodes.
22548 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
22550 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
22551 range isn't testing for zero.
22553 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
22555 PR middle-end/56113
22556 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
22558 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
22559 Nick Clifton <nickc@redhat.com>
22561 * config/v850/constraints.md (Q): Define as a memory constraint.
22562 * config/v850/predicates.md (label_ref_operand): New predicate.
22563 (e3v5_shift_operand): New predicate.
22564 (ior_operator): New predicate.
22565 * config/v850/t-v850: Add e3v5 multilib.
22566 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
22567 (v850_gen_movdi): Prototype.
22568 * config/v850/v850.c: Add support for e3v5 architecture.
22569 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
22571 (construct_save_jarl): Add e3v5 long JARL support.
22572 (v850_adjust_insn_length): New function. Adjust length of call
22573 insns when using e3v5 instructions.
22574 (v850_gen_movdi): New function: Generate instructions to move a
22576 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
22577 (CPP_SPEC): Define __v850e3v5__ as appropriate.
22578 (TARGET_USE_FPU): Enable for e3v5.
22579 (CONST_OK_FOR_W): New macro.
22580 (ADJUST_INSN_LENGTH): Define.
22581 * config/v850/v850.md (UNSPEC_LOOP): Define.
22582 (attr cpu): Add v850e3v5.
22583 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
22584 (movdi): New pattern.
22585 (movdi_internal): New pattern.
22586 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
22587 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
22588 (cstoresf4): Likewise.
22589 (cstoredf4): Likewise.
22590 (insv): New pattern.
22591 (rotlso3_a): New pattern.
22592 (rotlsi3_b): New pattern
22593 (rotlsi3_v850e3v5): New pattern.
22594 (doloop_begin): New pattern.
22595 (fix_loop_counter): New pattern.
22596 (doloop_end): New pattern.
22597 (branch_normal): Add e3v5 long branch support.
22598 (branch_invert): Likewise.
22599 (branch_z_normal): Likewise.
22600 (branch_z_invert): Likewise.
22601 (branch_nz_normal): Likewise.
22602 (branch_nz_invert): Likewise.
22603 (call_internal_short): Add e3v5 register-indirect JARL support.
22604 (call_internal_long): Likewise.
22605 (call_value_internal_short): Likewise.
22606 (call_value_internal_long): Likewise.
22607 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
22608 (mloop): New option.
22609 * config.gcc: Add support for configuring v840e3v5 target.
22610 * doc/invoke.texi: Document new v850 specific command line options.
22612 2013-01-31 Paul Koning <ni1d@arrl.net>
22616 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
22617 children if parent is a class.
22618 (prune_unused_types_prune): Don't add DW_AT_declaration.
22620 2013-01-31 Richard Biener <rguenther@suse.de>
22622 PR tree-optimization/56157
22623 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
22624 match up operand with SLP child.
22626 2013-01-31 Jason Merrill <jason@redhat.com>
22629 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
22630 parameters the first time.
22631 (gen_scheduled_generic_parms_dies): Check completeness here.
22633 2013-01-31 Richard Biener <rguenther@suse.de>
22635 PR middle-end/53073
22636 * common.opt (faggressive-loop-optimizations): New flag,
22637 enabled by default.
22638 * doc/invoke.texi (faggressive-loop-optimizations): Document.
22639 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
22640 infer_loop_bounds_from_undefined by it.
22642 2013-01-31 Richard Biener <rguenther@suse.de>
22644 PR tree-optimization/56150
22645 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
22646 visit virtual operands.
22647 (find_uses_to_rename_bb): Likewise.
22649 2013-01-31 Richard Biener <rguenther@suse.de>
22651 PR tree-optimization/56150
22652 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
22653 mixed store non-store stmts.
22655 2013-01-30 Jakub Jelinek <jakub@redhat.com>
22658 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
22659 LIBASAN_EARLY_SPEC is defined.
22660 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
22661 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
22663 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
22666 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
22667 invalid args instead of ICEing on it.
22668 (ix86_valid_target_attribute_tree): Return error_mark_node if
22669 ix86_valid_target_attribute_inner_p failed.
22670 (ix86_valid_target_attribute_p): Return false only if
22671 ix86_valid_target_attribute_tree returned error_mark_node. Allow
22672 target("default") attribute.
22673 (sorted_attr_string): Change argument from const char * to tree,
22674 merge in all target attribute arguments rather than just one.
22675 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
22676 instead of free. Avoid using strcat.
22677 (ix86_mangle_function_version_assembler_name): Mangle
22678 target("default") as if no target attribute is present. Adjust
22679 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
22680 instead of xmalloc and XDELETEVEC instead of free.
22681 (ix86_function_versions): Don't return true if one of the decls
22682 doesn't have target attribute. If they don't and one of the decls
22683 is DECL_FUNCTION_VERSIONED, report an error. Adjust
22684 sorted_attr_string caller. Use XDELETEVEC instead of free.
22685 (ix86_supports_function_versions): Remove.
22686 (make_name): Fix up formatting.
22687 (make_dispatcher_decl): Remove resolver_name and its initialization.
22688 Avoid leaking memory.
22689 (is_function_default_version): Return true if there is
22690 target("default") attribute rather than no target attribute at all.
22691 (make_resolver_func): Avoid leaking memory.
22692 (ix86_generate_version_dispatcher_body): Likewise.
22693 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
22694 * target.def (supports_function_versions): Remove.
22695 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
22696 * doc/tm.texi: Regenerated.
22698 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
22700 PR rtl-optimization/56144
22701 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
22702 for values with side effects.
22704 2013-01-30 Richard Biener <rguenther@suse.de>
22706 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
22707 (sparseset_pop): Likewise.
22708 * cfganal.c (compute_idf): Likewise. Increase work-stack size
22709 to be able to use quick_push in the worker loop.
22711 2013-01-30 Marek Polacek <polacek@redhat.com>
22713 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
22715 2013-01-30 Richard Biener <rguenther@suse.de>
22718 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
22720 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
22722 PR tree-optimization/56064
22723 * fixed-value.c (fixed_from_double_int): New function.
22724 * fixed-value.h (fixed_from_double_int): New prototype.
22725 (const_fixed_from_double_int): New static inline function.
22726 * fold-const.c (native_interpret_fixed): New static function.
22727 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
22728 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
22729 (native_encode_fixed): New static function.
22730 (native_encode_expr) <FIXED_CST>: Use it.
22731 (native_interpret_int): Move double_int worker code to...
22732 * double-int.c (double_int::from_buffer): ...this new static method.
22733 * double-int.h (double_int::from_buffer): Prototype it.
22735 2013-01-30 Richard Biener <rguenther@suse.de>
22737 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
22738 New pointer-map and obstack.
22739 (init_alias_vars): Allocate pointer-map and obstack.
22740 (delete_points_to_sets): Free them.
22741 (find_what_var_points_to): Cache result.
22742 (find_what_p_points_to): Adjust for changed interface of
22743 find_what_var_points_to.
22744 (compute_points_to_sets): Likewise.
22745 (ipa_pta_execute): Likewise.
22747 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22749 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
22750 * configure: Regenerate.
22751 * config.in: Regenerate.
22752 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
22753 #nobits/#progbits if supported.
22755 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
22758 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
22759 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
22760 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
22762 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22764 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
22765 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
22767 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22769 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
22770 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
22772 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22774 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
22776 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
22777 * config/arm/cortex-a7.md: New bypasses using
22778 arm_mac_accumulator_is_result.
22780 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22782 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
22783 (cortex_a7_neon_mla): Likewise.
22784 (cortex_a7_fpfmad): New reservation.
22785 (cortex_a7_fpmacs): Use ffmas and update required units.
22786 (cortex_a7_fpmuld): Update required units and latency.
22787 (cortex_a7_fpmacd): Likewise.
22788 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
22789 (cortex_a7_neon). Likewise.
22790 (bypass) Update participating units.
22792 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22794 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
22795 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
22797 * config/arm/vfp11.md (vfp_farith): Use ffmas.
22798 (vfp_fmul): Use ffmad.
22799 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
22800 (cortex_r4_fmacd): Use ffmad.
22801 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
22802 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
22803 (cortex_a9_fmacd): Use ffmad.
22804 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
22805 (cortex_a8_vfp_macd): Use ffmad.
22806 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
22807 (cortex_a5_fpmacd): Use ffmad.
22808 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
22809 (cortex_a15_vfp_macd): Use ffmad.
22810 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
22812 2013-01-29 Jason Merrill <jason@redhat.com>
22815 * varasm.c (default_assemble_visibility): Don't warn about
22816 visibility on artificial decls.
22818 2013-01-29 Richard Biener <rguenther@suse.de>
22820 PR tree-optimization/56113
22821 * tree-ssa-structalias.c (equiv_class_lookup): Also return
22823 (label_visit): Free duplicate bitmaps and record the leader instead.
22824 (perform_var_substitution): Adjust.
22826 2013-01-29 Richard Biener <rguenther@suse.de>
22828 PR tree-optimization/55270
22829 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
22830 the CFG, schedule loops for fixup.
22832 2013-01-29 Nick Clifton <nickc@redhat.com>
22834 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
22837 2013-01-28 Leif Ekblad <leif@rdos.net>
22839 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
22840 * config/i386/i386.h (TARGET_RDOS): New macro.
22841 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
22842 * config/i386/i386.c (ix86_option_override_internal): For 64bit
22843 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
22844 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
22845 DEFAULT_LARGE_SECTION_THRESHOLD.
22846 * config/i386/i386.md (R14_REG, R15_REG): New constants.
22847 * config/i386/rdos.h: New file.
22848 * config/i386/rdos64.h: New file.
22850 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
22853 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
22856 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22858 PR rtl-optimization/56117
22859 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
22860 call cselib_lookup_from_insn on the MEM before calling
22861 add_insn_mem_dependence.
22863 2013-01-28 Richard Biener <rguenther@suse.de>
22865 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
22866 to a stmt that didn't have one.
22867 (copy_phis_for_bb): Likewise for PHI arguments.
22868 (copy_debug_stmt): Likewise for debug stmts.
22870 2013-01-28 Richard Biener <rguenther@suse.de>
22872 PR tree-optimization/56034
22873 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
22874 (partition_builtin_p): Adjust.
22875 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
22876 it is the last partition.
22877 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
22878 up the vertex for the definition.
22879 (classify_partition): Classify whether a partition is a
22880 PKIND_REDUCTION, thus has uses outside of the loop.
22881 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
22882 Merge all PKIND_REDUCTION partitions into the last partition.
22883 (tree_loop_distribution): Seed partitions from reductions as well.
22885 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22887 PR tree-optimization/56125
22888 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
22889 pow(x,c) into sqrt(x) * powi(x, n/2) or
22890 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
22891 optimizing for size.
22892 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
22893 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
22896 PR tree-optimization/56094
22897 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
22898 to UNKNOWN_LOCATION while gimplifying expr.
22900 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
22903 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
22904 operand 0 in movabs insn template for -masm=intel asm alternative.
22905 (*movabs<mode>_2): Ditto for operand 1.
22907 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
22910 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
22913 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22915 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
22918 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22920 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
22921 Undef to avoid warning.
22923 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22925 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
22926 * configure: Regenerate.
22928 2013-01-25 Jakub Jelinek <jakub@redhat.com>
22930 PR tree-optimization/56098
22931 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
22932 for stmts with volatile ops.
22933 (cond_store_replacement): Don't optimize if assign has volatile ops.
22934 (cond_if_else_store_replacement_1): Don't optimize if either
22935 then_assign or else_assign have volatile ops.
22936 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
22939 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22941 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
22943 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22945 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
22946 missing ':' in asm example.
22948 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
22950 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
22951 entries into lane and laneq entries.
22952 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
22953 Remove AdvSIMD scalar modes.
22954 (aarch64_sq<r>dmulh_laneq<mode>): New.
22955 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
22957 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22958 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22959 * config/aarch64/iterators.md (VCOND): New.
22962 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22965 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22966 Add NULL LIBNAME argument to existing definitions.
22967 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22968 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22969 * config/avr/avr.c (DEF_BUILTIN): Same.
22970 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22971 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22972 implementation is available (DECL_ASSEMBLER_NAME is set).
22973 (avr_fold_absfx): New static function.
22974 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22975 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22976 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22977 AVR_BUILTIN_ABSLLK.
22978 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22979 (abshk, absk, abslk, absllk): Provide as static inline functions.
22981 2013-01-25 Marek Polacek <polacek@redhat.com>
22983 PR tree-optimization/56035
22984 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22986 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22988 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22989 (*movtf_internal_rex64): Add (!o,C) alternative
22990 (*movxf_internal_rex64): Ditto.
22991 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22993 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22995 * doc/invoke.texi: fix typo.
22996 * doc/objc.texi: fix typo.
22998 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
23000 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
23001 for the first two alternatives.
23003 2013-01-24 Diego Novillo <dnovillo@google.com>
23005 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
23006 (ggc-zone.o): Remove.
23007 * configure.ac: Remove option --with-gc.
23008 * configure: Re-generate.
23009 * doc/install.texi: Remove documentation for --with-gc.
23010 * gengtype.c (write_enum_defn): Remove. Update all users.
23011 (write_Types_process_field): Remove generation of gt_e_* argument.
23012 (output_type_enum): Remove. Update all users.
23013 (write_enum_defn): Remove. Update all users.
23014 (enum alloc_zone): Remove. Update all users.
23015 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
23016 * ggc-common.c (ggc_splay_alloc): Remove first argument.
23017 Update all callers.
23018 (struct ptr_data): Remove field TYPE. Update all users.
23019 (gt_pch_note_object): Remove argument TYPE. Update all users.
23020 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
23022 * ggc-none.c (ggc_alloc_typed_stat): Remove.
23023 (struct alloc_zone): Remove.
23024 (ggc_internal_alloc_zone_stat): Remove.
23025 (ggc_internal_cleared_alloc_zone_stat): Remove.
23026 * ggc-page.c (ggc_alloc_typed_stat): Remove.
23027 (ggc_pch_count_object): Remove last argument. Update all users.
23028 (ggc_pch_alloc_object): Remove last argument. Update all users.
23029 (struct alloc_zone): Remove.
23030 * ggc-zone.c: Remove.
23031 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
23032 (struct alloc_zone): Remove.
23033 (ggc_alloc_typed_stat): Remove.
23034 (ggc_alloc_typed): Remove.
23035 (ggc_splay_alloc): Remove first argument.
23036 (rtl_zone): Remove. Update all users.
23037 (tree_zone): Remove. Update all users.
23038 (tree_id_zone): Remove. Update all users.
23039 (ggc_internal_zone_alloc_stat): Remove. Update all users.
23040 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
23041 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
23042 * tree-ssanames.c: Remove references to zone allocator in comments.
23044 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
23046 * config/avr/avr.c (avr_out_fract): Make register numbers that
23047 might be outside of source operand signed.
23049 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
23051 * config/i386/constraints.md (Yf): New constraint.
23052 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
23053 of f constraint to conditionaly disable x87 register preferences.
23054 (*movdf_internal): Ditto.
23055 (*movsf_internal): Ditto.
23057 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
23059 PR inline-asm/55934
23060 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
23061 that have operands with impossible constraints.
23062 Add a FIXME for a speed-up opportunity.
23063 * lra-constraints.c (process_alt_operands): Verify that a class
23064 selected from constraints on asms is valid for the operand mode.
23065 (curr_insn_transform): Remove incorrect comment.
23067 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
23069 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
23070 TOC operand is a valid symbol ref in the constant pool.
23072 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23074 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
23076 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
23079 * config/avr/stdfix.h: New file.
23080 * t-avr (stdfix-gcc.h): New rule to build it.
23081 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
23083 2013-01-23 Kostya Serebryany <kcc@google.com>
23085 * config/darwin.h: remove dependency on
23086 CoreFoundation (asan on Mac OS).
23088 2013-01-23 Jakub Jelinek <jakub@redhat.com>
23091 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
23092 instead of cmpdi_operand for first comparison operand.
23093 Don't assert that comparison operands aren't both constants.
23095 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23097 * doc/install.texi (Downloading the Source): Update references to
23098 downloading separate components.
23100 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23102 * doc/extend.texi (__int128): Improve grammar.
23104 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
23107 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
23108 alternative to (o,r).
23109 (*movdi_internal_rex64): Remove (!o,n) alternative.
23110 (DImode immediate->memory splitter): Remove.
23111 (DImode immediate->memory peephole2): Remove.
23112 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
23113 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
23114 alternative to (!o,*r).
23115 (*movtf_internal_sse): New pattern.
23116 (*movxf_internal_rex64): New pattern.
23117 (*movxf_internal): Disable for TARGET_64BIT.
23118 (*movdf_internal_rex64): Remove (!o,F) alternative.
23120 2013-01-22 Jakub Jelinek <jakub@redhat.com>
23122 PR middle-end/56074
23123 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
23124 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
23125 * tree-vect-loop-manip.c (find_loop_location): Also ignore
23126 stmt locations where LOCATION_LOCUS of the stmt location is
23127 UNKNOWN_LOCATION or BUILTINS_LOCATION.
23130 * config/i386/i386.md (UNSPEC_STOS): New.
23131 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
23132 *strsetqi_1): Add UNSPEC_STOS.
23134 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
23137 * doc/invoke.texi: Remove left over -Wsynth example.
23139 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23141 PR tree-optimization/56051
23142 * fold-const.c (fold_binary_loc): Don't fold
23143 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
23144 a narrowing conversion, or widening conversion from signed
23147 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23149 PR rtl-optimization/56023
23150 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
23151 dependent on debug instruction.
23153 2013-01-21 Martin Jambor <mjambor@suse.cz>
23155 PR middle-end/56022
23156 * function.c (allocate_struct_function): Call
23157 invoke_set_current_function_hook earlier.
23159 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23161 * reload1.c (init_reload): Only initialize reload_obstack
23162 during the first call.
23164 2013-01-21 Marek Polacek <polacek@redhat.com>
23166 * cfgloop.c (verify_loop_structure): Fix up grammar.
23168 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
23170 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
23171 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
23173 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23176 * config/arm/marvell-pj4.md: Update copyright year.
23177 Fix up use of alu to alu_reg and simple_alu_imm.
23179 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23181 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
23183 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
23186 * lra-constraints.c (curr_insn_transform): Don't reuse original
23187 insn for secondary memory move when memory mode should be different.
23189 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23191 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
23192 atomic_storedi_1): New patterns.
23194 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23196 btver2 pipeline descriptions.
23197 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
23199 * config/i386/i386.md (btver2_decode): New type attributes.
23200 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
23202 * config/i386/btver2.md: New file describing btver2 pipelines.
23204 2013-01-19 Andrew Pinski <apinski@cavium.com>
23206 PR tree-optimization/52631
23207 * tree-ssa-sccvn (visit_use): Before looking up the original
23208 statement, try looking up the simplified expression.
23210 2013-01-19 Anthony Green <green@moxielogic.com>
23212 * config/moxie/moxie.c (moxie_expand_prologue): Set
23213 current_function_static_stack_size.
23215 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23217 PR tree-optimization/56029
23218 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
23219 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
23221 2013-01-18 Sharad Singhai <singhai@google.com>
23223 PR tree-optimization/55995
23224 * dumpfile.c (dump_loc): Print location only if available.
23225 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
23227 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
23230 * lra-constraints.c (curr_insn_transform): Reuse original insn for
23231 secondary memory move.
23232 (inherit_reload_reg): Use rclass instead of cl for
23233 check_secondary_memory_needed_p.
23235 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23237 PR middle-end/56015
23238 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
23239 the case where writing real complex part of target modifies op1.
23241 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
23243 * config/aarch64/aarch64-simd.md
23244 (aarch64_vcond_internal<mode>): Handle unordered cases.
23245 * config/aarch64/iterators.md (v_cmp_result): New.
23247 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
23248 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23250 * config/arm/marvell-pj4.md: New file.
23251 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
23252 * config/arm/arm.md (generic_sched): Add marvell_pj4.
23253 (generic_vfp): Likewise.
23254 * config/arm/arm-cores.def: Add marvell-pj4.
23255 * config/arm/arm-tune.md: Regenerate.
23256 * config/arm/arm-tables.opt: Regenerate.
23257 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
23258 * doc/invoke.texi: Document marvell-pj4.
23260 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
23262 * config/aarch64/arm_neon.h: Map scalar types to standard types.
23264 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23269 * var-tracking.c (negative_power_of_two_p): New.
23270 (global_get_addr_cache, local_get_addr_cache): New.
23271 (get_addr_from_global_cache, get_addr_from_local_cache): New.
23272 (vt_canonicalize_addr): Rewrite using the above. Adjust the
23274 (vt_stack_offset_p): Remove.
23275 (vt_canon_true_dep): Always canonicalize loc's address.
23276 (clobber_overlapping_mems): Make sure we have a MEM.
23277 (local_get_addr_clear_given_value): New.
23278 (val_reset): Clear local cached entries.
23279 (compute_bb_dataflow): Create and release the local cache.
23280 Disable duplicate MEMs clobbering.
23281 (emit_notes_in_bb): Clobber MEMs likewise.
23282 (vt_emit_notes): Create and release the local cache.
23283 (vt_initialize, vt_finalize): Create and release the global
23284 cache, respectively.
23285 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
23287 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23289 PR libmudflap/53359
23290 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
23291 not found in the symtab.
23293 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23296 PR rtl-optimization/55547
23297 PR rtl-optimization/53827
23300 * alias.c (offset_overlap_p): New, factored out of...
23301 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
23302 the conservative special case for symbolic constants. Don't
23303 adjust zero sizes on alignment.
23305 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
23307 PR rtl-optimization/52573
23308 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
23309 REG_UNUSED for the same register.
23311 2013-01-17 Richard Biener <rguenther@suse.de>
23312 Marek Polacek <polacek@redhat.com>
23314 PR rtl-optimization/55833
23315 * loop-unswitch.c (unswitch_loops): Move loop verification...
23316 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
23317 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
23318 Set it to true when we're removing a loop from hierarchy tree in
23319 an irreducible region.
23320 (fix_bb_placements): Adjust caller.
23321 (fix_loop_placements): Likewise.
23323 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
23325 * config/avr/builtins.def (DEF_BUILTIN): Factor out
23326 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
23327 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
23328 Remove ID. Adjust comments.
23329 * config/avr/avr-c.c (avr_builtin_name): Remove.
23330 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
23331 * config/avr/avr.c (avr_tolower): New static function.
23332 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
23333 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
23334 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
23337 2013-01-17 Jan Hubicka <jh@suse.cz>
23339 PR tree-optimization/55273
23340 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
23342 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
23345 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
23346 store through atomic_store<mode>_1.
23347 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
23349 2013-01-17 Martin Jambor <mjambor@suse.cz>
23351 PR tree-optimizations/55264
23352 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
23353 for virtual methods.
23354 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
23355 virtual methods before inlining is over.
23356 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
23358 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
23361 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
23363 PR rtl-optimization/56005
23364 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
23365 pending reads for prefetch.
23367 2013-01-16 Ian Bolton <ian.bolton@arm.com>
23369 * config/aarch64/aarch64.md
23370 (*cstoresi_neg_uxtw): New pattern.
23371 (*cmovsi_insn_uxtw): New pattern.
23372 (*<optab>si3_uxtw): New pattern.
23373 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
23374 (*<optab>si3_insn_uxtw): New pattern.
23375 (*bswapsi2_uxtw): New pattern.
23377 2013-01-16 Richard Biener <rguenther@suse.de>
23379 * tree-inline.c (tree_function_versioning): Remove set but
23380 never used variable.
23382 2013-01-16 Richard Biener <rguenther@suse.de>
23384 PR tree-optimization/55964
23385 * tree-flow.h (rename_variables_in_loop): Remove.
23386 (rename_variables_in_bb): Likewise.
23387 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
23388 (copy_loop_before): Adjust and delete update-ssa status.
23389 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
23390 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
23391 (rename_variables_in_loop): Remove.
23392 (slpeel_update_phis_for_duplicate_loop): Likewise.
23393 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
23394 use available cfg machinery instead of duplicating it.
23395 Update PHI nodes and perform poor-mans SSA update here.
23396 (slpeel_tree_peel_loop_to_edge): Adjust.
23398 2013-01-16 Richard Biener <rguenther@suse.de>
23400 PR tree-optimization/54767
23401 PR tree-optimization/53465
23402 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
23403 (vrp_visit_phi_node): For PHI arguments coming via backedges
23404 drop all symbolical range information.
23405 (execute_vrp): Compute backedges.
23407 2013-01-16 Richard Biener <rguenther@suse.de>
23409 * doc/install.texi: Update CLooG and ISL requirements to
23412 2013-01-16 Christian Bruel <christian.bruel@st.com>
23415 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
23416 (broken_move): Handle UNSPECV_SP_SWITCH_B.
23417 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
23419 2013-01-16 DJ Delorie <dj@redhat.com>
23421 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
23422 (UNSPECV_SP_SWITCH_E): New.
23423 (sp_switch_1): Change to an unspec.
23424 (sp_switch_2): Change to an unspec. Don't use post-inc when we
23427 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
23429 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
23430 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
23431 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
23432 (expand_mem_thread_fence): Ditto.
23433 (expand_mem_signal_fence): Ditto.
23434 (expand_atomic_load): Ditto.
23435 (expand_atomic_store): Ditto.
23437 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
23439 PR rtl-optimization/55547
23440 PR rtl-optimization/53827
23443 * alias.c (memrefs_conflict_p): Set sizes to negative after
23446 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23449 * function.c (thread_prologue_and_epilogue_insns): Always
23450 add crtl->drap_reg to set_up_by_prologue.set, even if
23451 stack_realign_drap is false.
23453 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
23455 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
23456 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
23457 *call): Fix indention.
23459 2013-01-15 Tom de Vries <tom@codesourcery.com>
23462 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
23465 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
23467 PR rtl-optimization/55153
23468 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
23470 2013-01-15 Martin Jambor <mjambor@suse.cz>
23472 PR tree-optimization/55920
23473 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
23474 accesses as grp_to_be_debug_replaced.
23476 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23478 PR tree-optimization/55920
23479 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
23480 there is non-useless type conversion needed from debug rhs to lhs,
23481 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
23483 2013-01-15 Joseph Myers <joseph@codesourcery.com>
23484 Mikael Pettersson <mikpe@it.uu.se>
23487 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
23489 (ASM_OUTPUT_CASE_LABEL): Remove.
23490 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
23491 * final.c (shorten_branches): Update alignment of labels before
23492 jump tables if CASE_VECTOR_SHORTEN_MODE.
23494 2013-01-15 Richard Biener <rguenther@suse.de>
23497 * system.h: Do not include gmp.h for building host tools.
23499 2013-01-15 Richard Biener <rguenther@suse.de>
23501 PR middle-end/55882
23502 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
23503 account for bitpos when computing alignment.
23505 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
23507 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
23508 (ix86_target_macros_internal): Likewise.
23510 * config/i386/i386.c (m_CORE2I7): Removed.
23511 (m_CORE_HASWELL): New macro.
23512 (m_CORE_ALL): Likewise.
23513 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
23514 (initial_ix86_arch_features): Likewise.
23515 (processor_target_table): Initializations for Core avx2.
23516 (cpu_names): New names "core-avx2".
23517 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
23518 PROCESSOR_CORE_HASWELL.
23519 (ix86_issue_rate): New case.
23520 (ia32_multipass_dfa_lookahead): Likewise.
23521 (ix86_sched_init_global): Likewise.
23523 * config/i386/i386.h (TARGET_HASWELL): New macro.
23524 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
23525 (processor_type): New PROCESSOR_HASWELL.
23527 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23529 PR tree-optimization/55955
23530 * tree-vect-loop.c (vectorizable_reduction): Give up early on
23531 *SHIFT_EXPR and *ROTATE_EXPR codes.
23533 PR tree-optimization/48766
23534 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
23535 -ftrapv disable -fwrapv.
23537 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23540 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
23541 etc. to 1 and not to __flash.
23542 Use LL suffix for __INT24_MAX__ with -mint8.
23543 Use ULL suffix for __UINT24_MAX__ with -mint8.
23545 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23547 * config/avr/avr-arch.h
23548 (struct base_arch_s): Use typedef avr_arch_t instead.
23549 (struct arch_info_s): Use typedef avr_arch_info_t instead.
23550 (struct mcu_type_s): Use typedef avr_mcu_t instead.
23551 * config/avr/avr.c: Same.
23552 * config/avr/avr-devices.c: Same.
23553 * config/avr/driver-avr.c: Same.
23554 * config/avr/gen-avr-mmcu-texi.c: Same.
23555 * config/avr/avr-mcus.def: Adjust comment.
23557 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
23559 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
23560 * config/aarch64/iterators.md (VALLDI): New.
23562 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
23563 Andi Kleen <ak@linux.intel.com>
23566 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
23567 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
23570 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23572 * config/avr/avr-stdint.h: Remove trailing blanks.
23573 * config/avr/avr-log.h: Same.
23574 * config/avr/avr-arch.h: Same.
23575 * config/avr/avr-devices.c: Same.
23576 * config/avr/avr-dimode.md: Same.
23577 * config/avr/predicates.md: Same.
23578 * config/avr/avr-c.c: Same. And fix typo.
23580 * config/avr/avr-protos.h: Same. And:
23581 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
23582 (init_cumulative_args): Rename to avr_init_cumulative_args.
23583 (expand_prologue): Rename to avr_expand_prologue.
23584 (expand_epilogue): Rename to avr_expand_epilogue.
23585 (adjust_insn_length): Rename to avr_adjust_insn_length.
23586 (notice_update_cc): Rename to avr_notice_update_cc.
23587 (final_prescan_insn): Rename to avr_final_prescan_insn.
23588 * config/avr/avr.c: Same.
23589 * config/avr/avr.h: Same.
23590 * config/avr/avr.md: Remove trailing blanks.
23591 (prologue): Use avr_expand_prologue.
23592 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
23594 2013-01-14 Richard Biener <rguenther@suse.de>
23596 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
23597 verify_location, collect_subblocks): New functions.
23598 (verify_gimple_in_cfg): Verify that locations only reference
23599 BLOCKs in the functions BLOCK tree.
23601 2013-01-14 Richard Biener <rguenther@suse.de>
23603 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
23605 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
23607 (insert_out_of_ssa_copy_on_edge): Likewise.
23608 (rewrite_close_phi_out_of_ssa): Likewise.
23609 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
23611 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
23612 propagated constants.
23613 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
23616 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23618 * config/avr/avr-modes.def: Add GPL copyright notice.
23620 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
23622 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
23623 MEMMODEL_MASK to determine memory model.
23624 (atomic_store<mode>): Ditto from operands[2].
23625 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
23627 2013-01-13 Jakub Jelinek <jakub@redhat.com>
23630 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
23631 (fold_gimple_assign): Don't call unshare_expr here.
23632 (fold_ctor_reference): Call unshare_expr.
23634 2013-01-13 Terry Guo <terry.guo@arm.com>
23636 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
23637 * doc/fragments.texi: Document MULTILIB_REUSE.
23638 * gcc.c (multilib_reuse): New internal spec.
23639 (set_multilib_dir): Also search multilib from multilib_reuse.
23640 * genmultilib (tmpmultilib3): Refactor code.
23641 (tmpmultilib4): Ditto.
23642 (multilib_reuse): New multilib argument.
23644 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
23646 * Makefile.in: Update copyright.
23648 2013-01-12 Tom de Vries <tom@codesourcery.com>
23650 PR middle-end/55890
23651 * calls.c (expand_call): Check if arg_nr is valid.
23653 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
23655 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
23656 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
23657 documentation. Add missing '__' in front of
23658 __builtin_ia32_packssdw256.
23660 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23663 * config/s390/s390.c (s390_preferred_reload_class): Do not return
23664 NO_REGS for larl operands.
23665 (s390_reload_larl_operand): Use s390_load_address instead of
23668 2013-01-11 Richard Biener <rguenther@suse.de>
23670 * tree-cfg.c (verify_node_sharing_1): Split out from ...
23671 (verify_node_sharing): ... here.
23672 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
23674 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
23676 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
23677 Substitute TREECHECKING.
23678 * configure: Regenerate.
23679 * Makefile.in (TREECHECKING): New.
23681 2013-01-11 Richard Guenther <rguenther@suse.de>
23683 PR tree-optimization/44061
23684 * tree-vrp.c (extract_range_basic): Compute zero as
23685 value-range for __builtin_constant_p of function parameters.
23687 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
23689 Update copyright years.
23691 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
23693 PR rtl-optimization/55672
23694 * lra-eliminations.c (mark_not_eliminable): Permit addition with
23695 const to be eliminable.
23697 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
23699 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
23700 * configure: Regenerate.
23702 2013-01-10 Richard Biener <rguenther@suse.de>
23704 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
23706 2013-01-10 Richard Biener <rguenther@suse.de>
23709 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
23710 locations for virtual PHI arguments.
23711 (rewrite_update_phi_arguments): Likewise.
23713 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
23715 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
23718 2013-01-10 Jakub Jelinek <jakub@redhat.com>
23720 PR tree-optimization/55921
23721 * tree-complex.c (expand_complex_asm): New function.
23722 (expand_complex_operations_1): Call it for GIMPLE_ASM.
23724 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23727 * config/s390/s390.c (s390_symref_operand_p)
23728 (s390_loadrelative_operand_p): Merge the two functions.
23729 (s390_check_qrst_address, print_operand_address): Add parameters
23730 to s390_loadrelative_operand_p invokation.
23731 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
23732 (s390_reload_larl_operand, s390_secondary_reload): Use
23733 s390_loadrelative_operand_p instead of s390_symref_operand_p.
23734 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
23736 2013-01-09 Mike Stump <mikestump@comcast.net>
23738 * dse.c (record_store): Remove unnecessary assert.
23740 2013-01-09 Jan Hubicka <jh@suse.cz>
23742 PR tree-optimization/55569
23743 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
23744 * cfgloop.h (scale_loop_profile): Likewise.
23746 2013-01-09 Jan Hubicka <jh@suse.cz>
23749 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
23751 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
23753 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
23755 PR middle-end/55114
23756 * expr.h (maybe_emit_group_store): Declare.
23757 * expr.c (maybe_emit_group_store): New function.
23758 * builtins.c (expand_builtin_int_roundingfn): Call it.
23759 (expand_builtin_int_roundingfn_2): Likewise.
23761 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
23763 PR rtl-optimization/55829
23764 * lra-constraints.c (match_reload): Add code for absent output.
23765 (curr_insn_transform): Add code for reloads of matched inputs
23768 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
23770 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
23771 attribute of movddup insn to DF.
23772 (*vec_interleave_lowv2df): Ditto.
23773 (vec_dupv2df): Ditto.
23775 2013-01-09 Jan Hubicka <jh@suse.cz>
23777 PR tree-optimiation/55875
23778 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
23779 EVERY_ITERATION parameter.
23780 (number_of_iterations_exit): Check if exit is executed every iteration.
23781 (idx_infer_loop_bounds): Similarly here.
23782 (n_of_executions_at_most): Simplify
23783 to only test for cases where statement is dominated by the
23784 particular bound; handle correctly the "postdominance" test.
23785 (scev_probably_wraps_p): Use max loop iterations info
23786 as a global bound first.
23788 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
23789 Nick Clifton <nickc@redhat.com>
23791 * config/v850/v850.md (cbranchsf4): New pattern.
23792 (cstoresf4): New pattern.
23793 (cbranchdf4): New pattern.
23794 (cstoredf4): New pattern.
23795 (movsicc): Disallow floating point comparisons.
23796 (cmpsf_le_insn): Fix order of operators.
23797 (cmpsf_lt_insn): Likewise.
23798 (cmpsf_eq_insn): Likewise.
23799 (cmpdf_le_insn): Likewise.
23800 (cmpdf_lt_insn): Likewise.
23801 (cmpdf_eq_insn): Likewise.
23802 (cmpsf_ge_insn): Use LE comparison.
23803 (cmpdf_ge_insn): Likewise.
23804 (cmpsf_gt_insn): Use LT comparison.
23805 (cmpdf_gt_insn): Likewise.
23806 (cmpsf_ne_insn): Delete pattern.
23807 (cmpdf_ne_insn): Delete pattern.
23808 * config/v850/v850.c (v850_gen_float_compare): Use
23809 gen_cmpdf_eq_insn for NE comparison.
23810 (v850_float_z_comparison_operator)
23811 (v850_float_nz_comparison_operator): Move from here ...
23812 * config/v850/predicates.md: ... to here. Move GT and GE
23813 comparisons into v850_float_z_comparison_operator.
23814 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
23816 (v850_float_nz_comparison_operator): Likewise.
23818 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23820 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
23821 with calls to gen_insvsi/gen_insvdi.
23823 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23825 * config/i386/i386.c (initial_ix86_tune_features): Set up
23826 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
23828 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
23829 Jakub Jelinek <jakub@redhat.com>
23831 PR tree-optimization/48189
23832 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
23833 If nitercst is 0, don't predict the exit edge.
23835 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23837 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
23838 in asm_fprintf with reg_names.
23839 (aarch64_print_operand_address): Likewise.
23840 (aarch64_return_addr): Likewise.
23841 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
23843 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23845 * config/pa/pa.h (VAL_U6_BITS_P): Define.
23846 (INT_U6_BITS): Likewise.
23847 * config/pa/predicates.md (uint6_operand): New predicate.
23848 (shift5_operand, shift6_operand): Likewise.
23849 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
23851 (lshrdi3): Use shift6_operand.
23852 (shrpsi4, shrpdi4): New insn patterns.
23853 (extzv): Delete expander.
23854 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
23855 predicates in unamed zero extract patterns. Tighten common constraint.
23856 (extv): Delete expander.
23857 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
23858 predicates in unamed sign extract patterns. Tighten common constraint.
23859 (insv): Delete expander.
23860 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
23861 predicates in unamed insert patterns. Tighten common constraint.
23862 Change uint32_operand predicate to uint6_operand predicate in unamed
23863 DImode pattern to insert constant values of type 1...1xxxx.
23865 2013-01-04 Jan Hubicka <jh@suse.cz>
23867 PR tree-optimization/55823
23868 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
23871 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23872 Uros Bizjak <ubizjak@gmail.com>
23874 PR rtl-optimization/55845
23875 * df-problems.c (can_move_insns_across): Stop scanning at
23876 volatile_insn_p source instruction or give up if
23877 across_from .. across_to range contains any volatile_insn_p
23880 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23882 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
23883 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
23885 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
23886 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
23888 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23891 * asan.c (asan_clear_shadow): New function.
23892 (asan_emit_stack_protection): Use it.
23894 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23896 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
23897 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
23898 with tab instead of space.
23900 2013-01-08 Nick Clifton <nickc@redhat.com>
23902 * config/rl78/rl78.c (rl78_expand_prologue): Always select
23903 register bank 0 at the start of an interrupt handler.
23904 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
23907 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23909 * config/aarch64/aarch64-simd.md
23910 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
23911 (aarch64_simd_bsl): Likewise.
23912 (aarch64_vcond_internal<mode>): Likewise.
23913 (vcond<mode><mode>): Likewise.
23914 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
23915 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
23917 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23919 * config/aarch64/aarch64-builtins.c
23920 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
23922 2013-01-08 Martin Jambor <mjambor@suse.cz>
23925 * tree-sra.c (analyze_access_subtree): Return true also after
23926 potentially creating a debug-only replacement.
23928 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23930 PR middle-end/55890
23931 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
23933 PR tree-optimization/54120
23934 * tree-vrp.c (range_fits_type_p): Don't allow
23935 src_precision < precision from signed vr to unsigned_p
23936 if vr->min or vr->max is negative.
23937 (simplify_float_conversion_using_ranges): Test can_float_p
23938 against CODE_FOR_nothing.
23940 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23941 Richard Biener <rguenther@suse.de>
23943 PR middle-end/55851
23944 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
23945 types instead of just INTEGER_TYPE types.
23947 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
23949 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
23952 2013-01-07 Steve Ellcey <sellcey@mips.com>
23955 * config/mips/mips.opt: Change mad to mmad to match documentation.
23957 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23960 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23961 .progmemx.data now.
23963 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23966 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23967 (avr_addrspace_t): Add .section_name field.
23968 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23970 (avr_addrspace): Same. Initialize .section_name. Remove last
23971 NULL entry. Put __memx into .progmemx.data.
23972 (progmem_section_prefix): Remove.
23973 (avr_asm_init_sections): No need to initialize progmem_section.
23974 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23975 section name prefix.
23976 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23977 retrieve the progmem section.
23978 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23979 boundary to run over avr_addrspace[].
23980 (avr_register_target_pragmas): Ditto.
23982 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23984 * varasm.c (output_constant_def_contents): For asan_protect_global
23985 protected strings, adjust DECL_ALIGN if needed, before testing for
23987 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23988 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23990 (output_object_block): For asan protected decls, emit asan padding
23991 after their contents.
23992 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23993 (asan_finish_file): Test it here instead.
23995 2013-01-07 Nick Clifton <nickc@redhat.com>
23996 Matthias Klose <doko@debian.org>
23997 Doug Kwan <dougkwan@google.com>
23998 H.J. Lu <hongjiu.lu@intel.com>
24001 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
24003 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
24005 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
24007 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
24009 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
24011 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24014 * doc/install.texi (Cross-Compiler-Specific Options): Document
24017 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
24019 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
24020 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
24021 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
24022 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
24023 vqmovun_high_s64): Fix source operand number and update copyright.
24025 2013-01-07 Richard Biener <rguenther@suse.de>
24027 PR middle-end/55890
24028 * gimple.h (gimple_call_builtin_p): New overload.
24029 * gimple.c (validate_call): New function.
24030 (gimple_call_builtin_p): Likewise.
24031 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
24032 Use gimple_call_builtin_p.
24033 (find_func_clobbers): Likewise.
24034 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
24035 (strlen_optimize_stmt): Likewise.
24037 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24039 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
24040 (vld1q_dup_*): Likewise.
24041 (vld1_*): Likewise.
24042 (vld1q_*): Likewise.
24043 (vld1_lane_*): Likewise.
24044 (vld1q_lane_*): Likewise.
24046 2013-01-07 Richard Biener <rguenther@suse.de>
24048 * lto-streamer.h (LTO_minor_version): Bump to 2.
24050 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24052 * config/aarch64/aarch64-protos.h
24053 (aarch64_const_double_zero_rtx_p): Rename to...
24054 (aarch64_float_const_zero_rtx_p): ...this.
24055 (aarch64_float_const_representable_p): New.
24056 (aarch64_output_simd_mov_immediate): Likewise.
24057 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
24058 move immediate case.
24059 * config/aarch64/aarch64.c
24060 (aarch64_const_double_zero_rtx_p): Rename to...
24061 (aarch64_float_const_zero_rtx_p): ...this.
24062 (aarch64_print_operand): Allow printing of new constants.
24063 (aarch64_valid_floating_const): New.
24064 (aarch64_legitimate_constant_p): Check for valid floating-point
24066 (aarch64_simd_valid_immediate): Likewise.
24067 (aarch64_vect_float_const_representable_p): New.
24068 (aarch64_float_const_representable_p): Likewise.
24069 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
24070 (aarch64_output_simd_mov_immediate): New.
24071 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
24072 (*movdf_aarch64): Likewise.
24073 * config/aarch64/constraints.md (Ufc): New.
24074 (Y): call aarch64_float_const_zero_rtx.
24075 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
24077 2013-01-07 Richard Biener <rguenther@suse.de>
24079 PR tree-optimization/55888
24080 PR tree-optimization/55862
24081 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
24082 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
24083 not if it is contained therein.
24085 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24087 * config/avr/t-avr: Typo.
24089 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24092 * config/avr/t-avr: Don't automatically rebuild
24093 $(srcdir)/config/avr/t-multilib
24094 $(srcdir)/config/avr/avr-tables.opt
24095 $(srcdir)/doc/avr-mmcu.texi
24096 (avr-mcus): New phony target to build them on request.
24097 (s-avr-mlib, s-avr-mmcu-texi): Remove.
24098 * avr/avr-mcus.def: Adjust comments.
24100 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
24102 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
24104 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24106 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
24108 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24110 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
24112 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
24114 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
24115 to generate profiling.
24116 * config/rs6000/aix64.h (LIB_SPEC): Same.
24118 2013-01-04 Andrew Pinski <apinski@cavium.com>
24120 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
24122 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
24124 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
24126 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
24128 (ix86_expand_move): Ditto.
24129 (ix86_zero_extend_to_Pmode): Ditto.
24130 (ix86_expand_call): Ditto.
24131 (ix86_expand_special_args_builtin): Ditto.
24132 (ix86_expand_builtin): Ditto.
24134 2013-01-04 Richard Biener <rguenther@suse.de>
24136 PR tree-optimization/55862
24137 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
24138 translating them through PHI nodes.
24140 2013-01-04 Martin Jambor <mjambor@suse.cz>
24142 PR tree-optimization/55755
24143 * tree-sra.c (sra_modify_assign): Do not check that an access has no
24144 children when trying to avoid producing a VIEW_CONVERT_EXPR.
24146 2013-01-04 Marek Polacek <polacek@redhat.com>
24148 PR middle-end/55859
24149 * opts.c (default_options_optimization): Clarify error message.
24151 2013-01-04 Richard Biener <rguenther@suse.de>
24153 PR middle-end/55863
24154 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
24157 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24160 * config/pa/pa.md (movsi): Revert previous change.
24161 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
24164 2013-01-03 Richard Henderson <rth@redhat.com>
24166 * config/i386/i386.c (ix86_expand_move): Always assign to op1
24167 after eliminating TLS symbols.
24169 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
24172 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
24173 * graphite-poly.c (debug_gmp_value): Likewise.
24175 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
24178 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
24179 selected code model, define __code_mode_small__, __code_model_medium__,
24180 __code_model_large__, __code_model_32__ or __code_model_kernel__.
24181 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
24182 xchg temporary register with %k. Declare temporary register as
24184 [__x86_64__]: For medium and large code models, preserve %rbx register.
24186 2013-01-03 Richard Biener <rguenther@suse.de>
24188 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
24189 (dump_subscript): Adjust.
24190 (finalize_ddr_dependent): Do not dump redundant info.
24191 (analyze_siv_subscript): Adjust.
24192 (subscript_dependence_tester): Likewise.
24193 (compute_affine_dependence): Likewise.
24195 2013-01-03 Richard Biener <rguenther@suse.de>
24198 2013-01-03 Richard Biener <rguenther@suse.de>
24200 PR tree-optimization/55857
24201 * tree-vect-stmts.c (vectorizable_load): Do not setup
24202 re-alignment for invariant loads.
24204 2013-01-02 Richard Biener <rguenther@suse.de>
24206 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24207 invariant load do not generate a vector load from the scalar location.
24209 2013-01-03 Richard Biener <rguenther@suse.de>
24211 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
24212 for not vectorizing.
24213 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
24214 not build INDIRECT_REFs, call get_name once only.
24215 (vect_create_data_ref_ptr): Likewise. Dump base object kind
24216 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
24218 2013-01-03 Richard Biener <rguenther@suse.de>
24220 PR tree-optimization/55857
24221 * tree-vect-stmts.c (vectorizable_load): Do not setup
24222 re-alignment for invariant loads.
24224 2013-01-03 Richard Biener <rguenther@suse.de>
24227 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
24228 prefer a built-in decl.
24230 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24232 * gcc.c (process_command): Update copyright notice dates.
24233 * gcov.c (print_version): Likewise.
24234 * gcov-dump.c (print_version): Likewise.
24236 PR rtl-optimization/55838
24237 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
24238 iv0.step, iv1.step and step.
24240 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24241 Marc Glisse <marc.glisse@inria.fr>
24243 PR tree-optimization/55832
24244 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
24245 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
24246 integer_{one,zero}_node.
24248 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24251 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
24252 * var-tracking.c (reverse_op): Don't add reverse ops to
24253 VALUEs that have already
24254 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
24256 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
24258 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
24260 2013-01-02 Teresa Johnson <tejohnson@google.com>
24262 * dumpfile.c (dump_loc): Print filename with location.
24263 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
24264 new location_t parameter to emit complete unroll message with
24265 new dump framework.
24266 (canonicalize_loop_induction_variables): Compute loops location
24267 and pass to try_unroll_loop_completely.
24268 * loop-unroll.c (report_unroll_peel): New function.
24269 (peel_loops_completely): Use new dump format with location
24270 for main dumpfile message, and invoke report_unroll_peel on success.
24271 (decide_unrolling_and_peeling): Ditto.
24272 (decide_peel_once_rolling): Remove old dumpfile message subsumed
24273 by report_unroll_peel.
24274 (decide_peel_completely): Ditto.
24275 (decide_unroll_constant_iterations): Ditto.
24276 (decide_unroll_runtime_iterations): Ditto.
24277 (decide_peel_simple): Ditto.
24278 (decide_unroll_stupid): Ditto.
24279 * cfgloop.c (get_loop_location): New function.
24280 * cfgloop.h (get_loop_location): Declare.
24282 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24284 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
24287 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24289 PR middle-end/55198
24290 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
24291 BLKmode objects when EXPAND_MEMORY is specified.
24293 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24295 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
24297 (fold_builtin_cpu): Do not share cpu model decls across statements.
24299 2013-01-02 Jason Merrill <jason@redhat.com>
24302 * tree.c (build_array_type_1): Revert earlier change.
24304 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
24306 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
24308 * config/aarch64/aarch64-tune.md: Re-generate.
24310 2013-01-02 Richard Biener <rguenther@suse.de>
24312 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24313 invariant load do not generate a vector load from the scalar location.
24315 2013-01-02 Richard Biener <rguenther@suse.de>
24318 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
24319 * configure: Regenerate.
24321 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24323 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
24324 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
24325 (expand_builtin_int_roundingfn_2): Keep the original target around
24326 for the fallback case.
24328 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24330 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
24331 to be clear for sign changes.
24333 2013-01-01 Jan Hubicka <jh@suse.cz>
24335 * ipa-inline-analysis.c: Fix formatting.
24337 2013-01-01 Jakub Jelinek <jakub@redhat.com>
24339 PR tree-optimization/55831
24340 * tree-vect-loop.c (get_initial_def_for_induction): Use
24341 gsi_after_labels instead of gsi_start_bb.
24343 Copyright (C) 2013 Free Software Foundation, Inc.
24345 Copying and distribution of this file, with or without modification,
24346 are permitted in any medium without royalty provided the copyright
24347 notice and this notice are preserved.