1 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3 * config/aarch64/arm_none.h
4 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6 2013-08-12 Nick Clifton <nickc@redhat.com>
8 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
10 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
13 expand for QI/HImode operand to produce more effictive code for
14 unsigned char(short) --> float(double) conversion.
16 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
18 * doc/invoke.texi: Mention that -ftls-model does not force the final
21 2013-08-12 Marek Polacek <polacek@redhat.com>
22 Marc Glisse <marc.glisse@inria.fr>
24 PR tree-optimization/57980
25 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
26 when creating -1 constant.
28 2013-08-10 Jan Hubicka <jh@suse.cz>
30 Workaround binutils PR14342.
31 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
32 (gimple_init_edge_profiler): Likewise.
33 (gimple_gen_ic_func_profiler): Likewise.
35 2013-08-09 Jan Hubicka <jh@suse.cz>
37 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
39 2013-08-09 Xinliang David Li <davidxl@google.com>
41 * config/i386/stringop.def: New file.
42 * config/i386/stringop.opt: New file.
43 * config/i386/i386-opts.h: Include stringopt.def.
44 * config/i386/i386.opt: Include stringopt.opt.
45 * config/i386/i386.c (ix86_option_override_internal):
46 Override default size based stringop inline strategies
48 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
51 2013-08-09 Jan Hubicka <jh@suse.cz>
53 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
55 2013-08-09 Jan Hubicka <jh@suse.cz>
57 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
59 (dump_cgraph_node): Dump profile-id.
60 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
61 and common_target_probability.
62 * lto-cgraph.c (lto_output_edge): Stream common targets.
63 (lto_output_node): Stream profile ids.
64 (input_node): Stream profile ids.
65 (input_edge): Stream common targets.
66 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
67 * ipa.c: Include value-prof.h
68 (ipa_profile_generate_summary): Turn indirect call statement histograms
70 (ipa_profile): Turn common targets into speculative edges.
72 2013-08-09 Jan Hubicka <jh@suse.cz>
74 * cgraph.h (cgraph_node): Add profile_id.
75 * value-prof.c (cgraph_node_map): Turn into pointer_map.
76 (init_node_map): Rewrite to handle hashes increas of incremental
78 (del_node_map): Update.
79 (find_func_by_funcdef_no): Replace by ...
80 (find_func_by_profile_id): ... this one.
81 (gimple_ic_transform): Do not remove useful histograms when
82 speculation is not done; dump info when indirect call removal
84 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
85 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
86 (__gcov_indirect_call_profiler_v2): .. this one.
87 * profile.h (init_node_map): Update.
88 * coverage.c (coverage_compute_profile_id): New function.
89 * coverage.h (coverage_compute_profile_id): Declare.
90 * tree-profile.c (init_ic_make_global_vars): Make
91 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
92 (gimple_init_edge_profiler): Update prototype of
93 __gcov_indirect_call_profiler.
94 (gimple_gen_ic_func_profiler): Simplify.
95 (tree_profiling): Use init_node_map
97 2013-08-09 Jan Hubicka <jh@suse.cz>
99 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only non-speculative
101 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
102 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
103 (cgraph_set_call_stmt): Likewise.
104 (cgraph_create_edge_1): Fix release checking compilatoin;
106 (cgraph_free_edge): Free indirect info.
107 (cgraph_turn_edge_to_speculative): New function.
108 (cgraph_speculative_call_info): New function.
109 (cgraph_make_edge_direct): Return direct edge; handle speculation.
110 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative
112 (dump_cgraph_node): Dump speculation.
113 (verify_edge_count_and_frequency): Accept speculative edges.
114 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
115 (verify_cgraph_node): Handle speculation.
116 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
117 (cgraph_set_call_stmt): Update prototype.
118 (cgraph_make_edge_direct): Update prototype.
119 (cgraph_speculative_call_info): Declare.
120 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
121 to change; update call of ipa_find_references.
122 * ipa-ref.c (ipa_record_reference): Fix return value; clear
123 lto_stmt_uid and speculative flags.
124 (ipa_dump_references): Dump speculation.
125 (ipa_clone_references): Clone speculative flag.
126 (ipa_clone_referring): Likewise.
127 (ipa_clone_ref): New function.
128 (ipa_find_reference): Look into lto_stmt_uids
129 (ipa_clear_stmts_in_references): Do not clear speculative calls.
130 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
131 (ipa_find_reference): Update declaration.
132 (ipa_clone_ref): Declare.
133 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
134 stream speculative flag.
135 (lto_output_ref): Stream statements uids and speculation.
136 (input_ref): Likewise.
137 (input_edge): Stream speuclation.
138 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
139 (cgraph_set_call_stmt_including_clones): Handle speculation.
140 * ipa-inline.c (heap_edge_removal_hook): New function.
141 (inline_small_functions): Register it.
142 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
143 also initialize refs.
144 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
146 (try_make_edge_direct_simple_call): Likewise.
147 (try_make_edge_direct_simple_call): Likewise.
148 (update_indirect_edges_after_inlining): Likewise.
149 (remove_described_reference): Look proper lto_stmt_uid.
150 (propagate_controlled_uses): Likewise.
151 (propagate_controlled_uses): Liekwise.
152 * tree-inline.c (copy_bb): Copy speculative edges.
153 (redirect_all_calls): New function.
154 (copy_cfg_body): Do redirection after loop info
156 (delete_unreachable_blocks_update_callgraph): Updadte
159 2013-08-09 Jan Hubicka <jh@suse.cz>
161 * lto-streamer-out.c (output_function): Renumber PHIs.
162 * lto-streamer-in.c (input_function): Likewise.
164 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
166 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
167 (get_lane_unsigned): Likewise.
168 (dup_lane_scalar): Likewise.
169 (get_lane): enable for VALL.
170 * config/aarch64/aarch64-simd.md
171 (aarch64_dup_lane_scalar<mode>): Remove.
172 (aarch64_get_lane_signed<mode>): Likewise.
173 (aarch64_get_lane_unsigned<mode>): Likewise.
174 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
175 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
176 (aarch64_get_lane<mode>): Enable for all vector modes.
177 (aarch64_get_lanedi): Remove misleading constraints.
178 * config/aarch64/arm_neon.h
179 (__aarch64_vget_lane_any): Define.
180 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
181 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
182 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
183 * config/aarch64/iterators.md (VDQQH): New.
187 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
189 * configure.ac: Add GAS check for LEON instructions on SPARC.
190 * configure: Regenerate.
191 * config.in: Likewise.
192 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
194 * config/sparc/sparc.opt (LEON, LEON3): New masks.
195 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
197 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
198 (AS_LEON_FLAG): New macro.
199 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
200 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
201 Deal with LEON and LEON3 for the memory model.
202 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
203 (atomic_compare_and_swap<mode>_1): Likewise.
204 (*atomic_compare_and_swap<mode>_1): Likewise.
206 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
208 * config/arm/neon.md (vcond): Fix floating-point vector
209 comparisons against 0.
211 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
213 * lra-constraints.c (emit_spill_move): Remove assert.
214 (process_alt_operands): Add more debugging
215 output. Increase reject for spilling into memory. Decrease
216 reject for reloading scratch.
217 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
219 2013-08-08 Steve Ellcey <sellcey@mips.com>
221 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
222 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
223 micromips incompatible. Add nan2008.
224 (MULTILIB_DIRNAMES): Add nan2008.
225 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
226 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
227 and micromips incompatible. Add nan2008.
228 (MULTILIB_DIRNAMES): Add nan2008.
229 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
231 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
233 PR rtl-optimization/58079
234 * combine.c (combine_simplify_rtx): Avoid using SUBST if
235 simplify_comparison has widened a comparison with an integer.
237 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
239 * config/arm/neon.md (movmisalign<mode>): Disable when we
240 don't allow unaligned accesses.
241 (*movmisalign<mode>_neon_store): Likewise.
242 (*movmisalign<mode>_neon_load): Likewise.
243 (*movmisalign<mode>_neon_store): Likewise.
244 (*movmisalign<mode>_neon_load): Likewise.
246 2013-08-08 Jan Hubicka <jh@suse.cz>
248 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
249 (make_pass_rebuild_cgraph_edges): Also clear references.
250 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
251 * ipa-inline-transform.c (inline_transform): Remove all references
253 * cgraphunit.c (expand_function): Remove all references after expansion.
254 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
255 (ipa_find_reference): Rewrite to iterator.
256 (remove_stmt_references): Likewise.
257 (ipa_clear_stmts_in_references): New function.
258 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
259 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
261 * ipa-split.c (split_function): Remove references in split function.
263 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
266 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
267 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
269 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
272 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
273 suggested mode for the assignment isn't compatible with the
276 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
279 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
281 2013-08-07 Xinliang David Li <davidxl@google.com>
283 * config/i386/i386.opt: New option -mtune-ctrl=.
284 * config/i386/x86-tune.def: New file.
285 * config/i386/i386.h: include x86-tune.def.
286 * config/i386/i386.c (ix86_option_override_internal):
287 Parsing -mtune-ctrl= option and set tune features.
289 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
292 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
293 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
296 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
298 * rtl.h (update_alignments): Declare.
299 * final.c (grow_label_align): New function extracted from...
300 (shorten_branches): ...here. Call it.
301 (update_alignments): New function.
302 * reorg.c (sibling_labels): New variable.
303 (get_label_before): Add SIBLING parameter. If it is non-zero, push
304 the new label along with it onto the sibling_labels vector.
305 (fill_simple_delay_slots): Adjust call to get_label_before.
306 (fill_slots_from_thread): Likewise.
307 (relax_delay_slots): Likewise.
308 (make_return_insns): Likewise.
309 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
311 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
313 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
314 document its semantics.
315 (diagnostic_report_diagnostic): Adjust accordingly.
317 2013-08-07 David Malcolm <dmalcolm@redhat.com>
319 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
320 (sparc_option_override): ...and port to new C++ pass API.
321 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
323 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
325 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
327 2013-08-06 Caroline Tice <cmtice@google.com>
329 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
330 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
331 * tree-pass.h: Add pass_vtable_verify.
332 * varasm.c (assemble_variable): Add code to properly set the comdat
333 section and name for the .vtable_map_vars section.
334 (assemble_vtyv_preinit_initializer): New function.
335 (default_sectin_type_flags): Make sure .vtable_map_vars section has
337 * output.h: Add function decl for assemble_vtv_preinit_initializer.
338 * vtable-verify.c: New file.
339 * vtable-verify.h: New file.
340 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
341 initialiation levels.
342 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
343 * passes.def: Insert pass_vtable_verify.
344 * aclocal.m4: Reorder includes.
345 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
346 -fvtv-counts options.
347 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
348 as appropriate, if -fvtable-verify=... is used.
349 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
350 -fvtable-verify=... is used.
351 * Makefile.in (OBJS): Add vtable-verify.o to list.
352 (vtable-verify.o): Add new build rule.
353 (GTFILES): Add vtable-verify.c to list.
354 * common.opt (fvtable-verify=): New flag.
355 (vtv_priority): Values for fvtable-verify= flag.
356 (fvtv-counts): New flag.
357 (fvtv-debug): New flag.
358 * tree.h (save_vtable_map_decl): New extern function decl.
360 2013-08-07 David Malcolm <dmalcolm@redhat.com>
362 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
363 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
364 (pass_data_rl78_devirt): ...new pass_data instance and...
365 (make_pass_rl78_devirt): ...new function.
366 (rl78_asm_file_start): Port pass registration to new C++ API.
368 2013-08-07 David Malcolm <dmalcolm@redhat.com>
370 * coretypes.h (rtl_opt_pass): Add.
372 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
373 (epiphany_init): Port to new C++ pass API.
374 (epiphany_optimize_mode_switching): Likewise.
375 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
376 (pass_manager::get_pass_mode_switching): New.
377 (pass_manager::get_pass_peephole2): New.
378 * mode-switching.c (pass_mode_switching): Add clone method.
379 * recog.c (pass_peephole2): Add clone method.
380 (pass_split_all_insns): Add clone method.
382 2013-08-06 David Malcolm <dmalcolm@redhat.com>
384 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
385 (mips_option_override): ...here, porting to new C++ API for passes.
387 2013-08-06 Jan Hubicka <jh@suse.cz>
389 * cgraph.c (cgraph_get_body): New function based on lto.c
391 * cgraph.h (cgraph_get_body): Declare.
392 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
394 * cgraphunit.c (expand_function): Get body prior expanding.
395 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
396 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
398 * passes.c (do_per_function_toporder): Get body.
399 * tree-inline.c (expand_call_inline): Get body prior inlining it.
400 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
402 2013-08-06 Martin Jambor <mjambor@suse.cz>
405 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
406 is not re-finalized. Rename second parameter to no_collect.
408 2013-08-06 Martin Jambor <mjambor@suse.cz>
411 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
412 MEM_REF has proper alignment information.
414 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
417 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
419 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
420 rtx (*) (rtx, ...) with insn_gen_fn.
421 * genoutput.c (output_insn_data): Cast gen_? function pointers to
422 insn_gen_fn::stored_funcptr. Add initializer braces.
424 2013-08-05 David Malcolm <dmalcolm@redhat.com>
426 Rewrite how instances of passes are cloned to remove assumptions
427 about their sizes (thus allowing pass subclasses to have
428 additional data fields, albeit non-GC-managed ones at this point).
430 * passes.c (make_pass_instance): Now that passes have clone
431 methods, rewrite this function to eliminate XNEW and memcpy
432 calls that used hardcoded sizes. Since this function no longer
433 creates pass instances, rename it to...
434 (add_pass_instance): ...this. Document the old way that passes were
435 numbered and flagged, and rework this function to continue using it.
436 (next_pass_1): Add an initial_pass argument for use by
438 (position_pass): When adding multiple instances of a pass, use
439 the pass's clone method, rather than relying on the XNEW/memcpy
440 within the former make_pass_instance (now add_pass_instance).
441 (pass_manager::pass_manager): When invoking next_pass_1, also supply
442 the initial instance of the current pass within the pass manager.
444 2013-08-05 David Malcolm <dmalcolm@redhat.com>
446 This is the automated part of the conversion of passes from C
447 structs to C++ classes.
449 Patch autogenerated by refactor_passes.py from
450 https://github.com/davidmalcolm/gcc-refactoring-scripts
451 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
453 * asan.c (pass_asan): Convert from a global struct to a subclass of
454 gimple_opt_pass along with...
455 (pass_data_asan): ...new pass_data instance and...
456 (make_pass_asan): ...new function.
457 (pass_asan_O0): Convert from a global struct to a subclass of
458 gimple_opt_pass along with...
459 (pass_data_asan_O0): ...new pass_data instance and...
460 (make_pass_asan_O0): ...new function.
461 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
462 subclass of rtl_opt_pass along with...
463 (pass_data_inc_dec): ...new pass_data instance and...
464 (make_pass_inc_dec): ...new function.
465 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
466 a subclass of rtl_opt_pass along with...
467 (pass_data_reorder_blocks): ...new pass_data instance and...
468 (make_pass_reorder_blocks): ...new function.
469 (pass_duplicate_computed_gotos): Convert from a global struct to a
470 subclass of rtl_opt_pass along with...
471 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
472 (make_pass_duplicate_computed_gotos): ...new function.
473 (pass_partition_blocks): Convert from a global struct to a subclass of
474 rtl_opt_pass along with...
475 (pass_data_partition_blocks): ...new pass_data instance and...
476 (make_pass_partition_blocks): ...new function.
477 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
478 struct to a subclass of rtl_opt_pass along with...
479 (pass_data_branch_target_load_optimize1): ...new pass_data instance
481 (make_pass_branch_target_load_optimize1): ...new function.
482 (pass_branch_target_load_optimize2): Convert from a global struct to a
483 subclass of rtl_opt_pass along with...
484 (pass_data_branch_target_load_optimize2): ...new pass_data instance
486 (make_pass_branch_target_load_optimize2): ...new function.
487 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
488 of rtl_opt_pass along with...
489 (pass_data_jump): ...new pass_data instance and...
490 (make_pass_jump): ...new function.
491 (pass_jump2): Convert from a global struct to a subclass of
492 rtl_opt_pass along with...
493 (pass_data_jump2): ...new pass_data instance and...
494 (make_pass_jump2): ...new function.
495 * cfgexpand.c (pass_expand): Convert from a global struct to a
496 subclass of rtl_opt_pass along with...
497 (pass_data_expand): ...new pass_data instance and...
498 (make_pass_expand): ...new function.
499 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
500 of rtl_opt_pass along with...
501 (pass_data_free_cfg): ...new pass_data instance and...
502 (make_pass_free_cfg): ...new function.
503 (pass_into_cfg_layout_mode): Convert from a global struct to a
504 subclass of rtl_opt_pass along with...
505 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
506 (make_pass_into_cfg_layout_mode): ...new function.
507 (pass_outof_cfg_layout_mode): Convert from a global struct to a
508 subclass of rtl_opt_pass along with...
509 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
510 (make_pass_outof_cfg_layout_mode): ...new function.
511 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
512 struct to a subclass of gimple_opt_pass along with...
513 (pass_data_build_cgraph_edges): ...new pass_data instance and...
514 (make_pass_build_cgraph_edges): ...new function.
515 (pass_rebuild_cgraph_edges): Convert from a global struct to a
516 subclass of gimple_opt_pass along with...
517 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
518 (make_pass_rebuild_cgraph_edges): ...new function.
519 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
520 subclass of gimple_opt_pass along with...
521 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
523 (make_pass_remove_cgraph_callee_edges): ...new function.
524 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
525 struct to a subclass of rtl_opt_pass along with...
526 (pass_data_stack_adjustments): ...new pass_data instance and...
527 (make_pass_stack_adjustments): ...new function.
528 * combine.c (pass_combine): Convert from a global struct to a subclass
529 of rtl_opt_pass along with...
530 (pass_data_combine): ...new pass_data instance and...
531 (make_pass_combine): ...new function.
532 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
533 global struct to a subclass of rtl_opt_pass along with...
534 (pass_data_compare_elim_after_reload): ...new pass_data instance
536 (make_pass_compare_elim_after_reload): ...new function.
537 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
538 of rtl_opt_pass along with...
539 (pass_data_rtl_cprop): ...new pass_data instance and...
540 (make_pass_rtl_cprop): ...new function.
541 * cse.c (pass_cse): Convert from a global struct to a subclass of
542 rtl_opt_pass along with...
543 (pass_data_cse): ...new pass_data instance and...
544 (make_pass_cse): ...new function.
545 (pass_cse2): Convert from a global struct to a subclass of
546 rtl_opt_pass along with...
547 (pass_data_cse2): ...new pass_data instance and...
548 (make_pass_cse2): ...new function.
549 (pass_cse_after_global_opts): Convert from a global struct to a
550 subclass of rtl_opt_pass along with...
551 (pass_data_cse_after_global_opts): ...new pass_data instance and...
552 (make_pass_cse_after_global_opts): ...new function.
553 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
554 of rtl_opt_pass along with...
555 (pass_data_ud_rtl_dce): ...new pass_data instance and...
556 (make_pass_ud_rtl_dce): ...new function.
557 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
558 rtl_opt_pass along with...
559 (pass_data_fast_rtl_dce): ...new pass_data instance and...
560 (make_pass_fast_rtl_dce): ...new function.
561 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
562 a subclass of rtl_opt_pass along with...
563 (pass_data_df_initialize_opt): ...new pass_data instance and...
564 (make_pass_df_initialize_opt): ...new function.
565 (pass_df_initialize_no_opt): Convert from a global struct to a
566 subclass of rtl_opt_pass along with...
567 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
568 (make_pass_df_initialize_no_opt): ...new function.
569 (pass_df_finish): Convert from a global struct to a subclass of
570 rtl_opt_pass along with...
571 (pass_data_df_finish): ...new pass_data instance and...
572 (make_pass_df_finish): ...new function.
573 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
574 rtl_opt_pass along with...
575 (pass_data_rtl_dse1): ...new pass_data instance and...
576 (make_pass_rtl_dse1): ...new function.
577 (pass_rtl_dse2): Convert from a global struct to a subclass of
578 rtl_opt_pass along with...
579 (pass_data_rtl_dse2): ...new pass_data instance and...
580 (make_pass_rtl_dse2): ...new function.
581 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
582 subclass of rtl_opt_pass along with...
583 (pass_data_dwarf2_frame): ...new pass_data instance and...
584 (make_pass_dwarf2_frame): ...new function.
585 * except.c (pass_set_nothrow_function_flags): Convert from a global
586 struct to a subclass of rtl_opt_pass along with...
587 (pass_data_set_nothrow_function_flags): ...new pass_data instance
589 (make_pass_set_nothrow_function_flags): ...new function.
590 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
591 subclass of rtl_opt_pass along with...
592 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
594 (make_pass_convert_to_eh_region_ranges): ...new function.
595 * final.c (pass_compute_alignments): Convert from a global struct to a
596 subclass of rtl_opt_pass along with...
597 (pass_data_compute_alignments): ...new pass_data instance and...
598 (make_pass_compute_alignments): ...new function.
599 (pass_final): Convert from a global struct to a subclass of
600 rtl_opt_pass along with...
601 (pass_data_final): ...new pass_data instance and...
602 (make_pass_final): ...new function.
603 (pass_shorten_branches): Convert from a global struct to a subclass of
604 rtl_opt_pass along with...
605 (pass_data_shorten_branches): ...new pass_data instance and...
606 (make_pass_shorten_branches): ...new function.
607 (pass_clean_state): Convert from a global struct to a subclass of
608 rtl_opt_pass along with...
609 (pass_data_clean_state): ...new pass_data instance and...
610 (make_pass_clean_state): ...new function.
611 * function.c (pass_instantiate_virtual_regs): Convert from a global
612 struct to a subclass of rtl_opt_pass along with...
613 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
614 (make_pass_instantiate_virtual_regs): ...new function.
615 (pass_leaf_regs): Convert from a global struct to a subclass of
616 rtl_opt_pass along with...
617 (pass_data_leaf_regs): ...new pass_data instance and...
618 (make_pass_leaf_regs): ...new function.
619 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
620 subclass of rtl_opt_pass along with...
621 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
623 (make_pass_thread_prologue_and_epilogue): ...new function.
624 (pass_match_asm_constraints): Convert from a global struct to a
625 subclass of rtl_opt_pass along with...
626 (pass_data_match_asm_constraints): ...new pass_data instance and...
627 (make_pass_match_asm_constraints): ...new function.
628 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
629 subclass of rtl_opt_pass along with...
630 (pass_data_rtl_fwprop): ...new pass_data instance and...
631 (make_pass_rtl_fwprop): ...new function.
632 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
633 rtl_opt_pass along with...
634 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
635 (make_pass_rtl_fwprop_addr): ...new function.
636 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
637 rtl_opt_pass along with...
638 (pass_data_rtl_pre): ...new pass_data instance and...
639 (make_pass_rtl_pre): ...new function.
640 (pass_rtl_hoist): Convert from a global struct to a subclass of
641 rtl_opt_pass along with...
642 (pass_data_rtl_hoist): ...new pass_data instance and...
643 (make_pass_rtl_hoist): ...new function.
644 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
645 subclass of gimple_opt_pass along with...
646 (pass_data_lower_cf): ...new pass_data instance and...
647 (make_pass_lower_cf): ...new function.
648 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
649 from a global struct to a subclass of gimple_opt_pass along with...
650 (pass_data_strength_reduction): ...new pass_data instance and...
651 (make_pass_strength_reduction): ...new function.
652 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
653 of rtl_opt_pass along with...
654 (pass_data_rtl_ifcvt): ...new pass_data instance and...
655 (make_pass_rtl_ifcvt): ...new function.
656 (pass_if_after_combine): Convert from a global struct to a subclass of
657 rtl_opt_pass along with...
658 (pass_data_if_after_combine): ...new pass_data instance and...
659 (make_pass_if_after_combine): ...new function.
660 (pass_if_after_reload): Convert from a global struct to a subclass of
661 rtl_opt_pass along with...
662 (pass_data_if_after_reload): ...new pass_data instance and...
663 (make_pass_if_after_reload): ...new function.
664 * init-regs.c (pass_initialize_regs): Convert from a global struct to
665 a subclass of rtl_opt_pass along with...
666 (pass_data_initialize_regs): ...new pass_data instance and...
667 (make_pass_initialize_regs): ...new function.
668 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
669 of ipa_opt_pass_d along with...
670 (pass_data_ipa_cp): ...new pass_data instance and...
671 (make_pass_ipa_cp): ...new function.
672 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
673 global struct to a subclass of gimple_opt_pass along with...
674 (pass_data_inline_parameters): ...new pass_data instance and...
675 (make_pass_inline_parameters): ...new function.
676 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
677 subclass of gimple_opt_pass along with...
678 (pass_data_early_inline): ...new pass_data instance and...
679 (make_pass_early_inline): ...new function.
680 (pass_ipa_inline): Convert from a global struct to a subclass of
681 ipa_opt_pass_d along with...
682 (pass_data_ipa_inline): ...new pass_data instance and...
683 (make_pass_ipa_inline): ...new function.
684 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
685 struct to a subclass of gimple_opt_pass along with...
686 (pass_data_local_pure_const): ...new pass_data instance and...
687 (make_pass_local_pure_const): ...new function.
688 (pass_ipa_pure_const): Convert from a global struct to a subclass of
689 ipa_opt_pass_d along with...
690 (pass_data_ipa_pure_const): ...new pass_data instance and...
691 (make_pass_ipa_pure_const): ...new function.
692 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
693 to a subclass of ipa_opt_pass_d along with...
694 (pass_data_ipa_reference): ...new pass_data instance and...
695 (make_pass_ipa_reference): ...new function.
696 * ipa-split.c (pass_split_functions): Convert from a global struct to
697 a subclass of gimple_opt_pass along with...
698 (pass_data_split_functions): ...new pass_data instance and...
699 (make_pass_split_functions): ...new function.
700 (pass_feedback_split_functions): Convert from a global struct to a
701 subclass of gimple_opt_pass along with...
702 (pass_data_feedback_split_functions): ...new pass_data instance and...
703 (make_pass_feedback_split_functions): ...new function.
704 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
705 global struct to a subclass of simple_ipa_opt_pass along with...
706 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
708 (make_pass_ipa_function_and_variable_visibility): ...new function.
709 (pass_ipa_free_inline_summary): Convert from a global struct to a
710 subclass of simple_ipa_opt_pass along with...
711 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
712 (make_pass_ipa_free_inline_summary): ...new function.
713 (pass_ipa_whole_program_visibility): Convert from a global struct to a
714 subclass of ipa_opt_pass_d along with...
715 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
717 (make_pass_ipa_whole_program_visibility): ...new function.
718 (pass_ipa_profile): Convert from a global struct to a subclass of
719 ipa_opt_pass_d along with...
720 (pass_data_ipa_profile): ...new pass_data instance and...
721 (make_pass_ipa_profile): ...new function.
722 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
723 ipa_opt_pass_d along with...
724 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
725 (make_pass_ipa_cdtor_merge): ...new function.
726 * ira.c (pass_ira): Convert from a global struct to a subclass of
727 rtl_opt_pass along with...
728 (pass_data_ira): ...new pass_data instance and...
729 (make_pass_ira): ...new function.
730 (pass_reload): Convert from a global struct to a subclass of
731 rtl_opt_pass along with...
732 (pass_data_reload): ...new pass_data instance and...
733 (make_pass_reload): ...new function.
734 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
735 subclass of rtl_opt_pass along with...
736 (pass_data_cleanup_barriers): ...new pass_data instance and...
737 (make_pass_cleanup_barriers): ...new function.
738 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
739 of rtl_opt_pass along with...
740 (pass_data_loop2): ...new pass_data instance and...
741 (make_pass_loop2): ...new function.
742 (pass_rtl_loop_init): Convert from a global struct to a subclass of
743 rtl_opt_pass along with...
744 (pass_data_rtl_loop_init): ...new pass_data instance and...
745 (make_pass_rtl_loop_init): ...new function.
746 (pass_rtl_loop_done): Convert from a global struct to a subclass of
747 rtl_opt_pass along with...
748 (pass_data_rtl_loop_done): ...new pass_data instance and...
749 (make_pass_rtl_loop_done): ...new function.
750 (pass_rtl_move_loop_invariants): Convert from a global struct to a
751 subclass of rtl_opt_pass along with...
752 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
753 (make_pass_rtl_move_loop_invariants): ...new function.
754 (pass_rtl_unswitch): Convert from a global struct to a subclass of
755 rtl_opt_pass along with...
756 (pass_data_rtl_unswitch): ...new pass_data instance and...
757 (make_pass_rtl_unswitch): ...new function.
758 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
759 subclass of rtl_opt_pass along with...
760 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
762 (make_pass_rtl_unroll_and_peel_loops): ...new function.
763 (pass_rtl_doloop): Convert from a global struct to a subclass of
764 rtl_opt_pass along with...
765 (pass_data_rtl_doloop): ...new pass_data instance and...
766 (make_pass_rtl_doloop): ...new function.
767 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
768 a subclass of rtl_opt_pass along with...
769 (pass_data_lower_subreg): ...new pass_data instance and...
770 (make_pass_lower_subreg): ...new function.
771 (pass_lower_subreg2): Convert from a global struct to a subclass of
772 rtl_opt_pass along with...
773 (pass_data_lower_subreg2): ...new pass_data instance and...
774 (make_pass_lower_subreg2): ...new function.
775 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
776 struct to a subclass of ipa_opt_pass_d along with...
777 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
778 (make_pass_ipa_lto_gimple_out): ...new function.
779 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
780 of ipa_opt_pass_d along with...
781 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
782 (make_pass_ipa_lto_finish_out): ...new function.
783 * mode-switching.c (pass_mode_switching): Convert from a global struct
784 to a subclass of rtl_opt_pass along with...
785 (pass_data_mode_switching): ...new pass_data instance and...
786 (make_pass_mode_switching): ...new function.
787 * modulo-sched.c (pass_sms): Convert from a global struct to a
788 subclass of rtl_opt_pass along with...
789 (pass_data_sms): ...new pass_data instance and...
790 (make_pass_sms): ...new function.
791 * omp-low.c (pass_expand_omp): Convert from a global struct to a
792 subclass of gimple_opt_pass along with...
793 (pass_data_expand_omp): ...new pass_data instance and...
794 (make_pass_expand_omp): ...new function.
795 (pass_lower_omp): Convert from a global struct to a subclass of
796 gimple_opt_pass along with...
797 (pass_data_lower_omp): ...new pass_data instance and...
798 (make_pass_lower_omp): ...new function.
799 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
800 of gimple_opt_pass along with...
801 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
802 (make_pass_diagnose_omp_blocks): ...new function.
803 * passes.c (pass_early_local_passes): Convert from a global struct to
804 a subclass of simple_ipa_opt_pass along with...
805 (pass_data_early_local_passes): ...new pass_data instance and...
806 (make_pass_early_local_passes): ...new function.
807 (pass_all_early_optimizations): Convert from a global struct to a
808 subclass of gimple_opt_pass along with...
809 (pass_data_all_early_optimizations): ...new pass_data instance and...
810 (make_pass_all_early_optimizations): ...new function.
811 (pass_all_optimizations): Convert from a global struct to a subclass
812 of gimple_opt_pass along with...
813 (pass_data_all_optimizations): ...new pass_data instance and...
814 (make_pass_all_optimizations): ...new function.
815 (pass_all_optimizations_g): Convert from a global struct to a subclass
816 of gimple_opt_pass along with...
817 (pass_data_all_optimizations_g): ...new pass_data instance and...
818 (make_pass_all_optimizations_g): ...new function.
819 (pass_rest_of_compilation): Convert from a global struct to a subclass
820 of rtl_opt_pass along with...
821 (pass_data_rest_of_compilation): ...new pass_data instance and...
822 (make_pass_rest_of_compilation): ...new function.
823 (pass_postreload): Convert from a global struct to a subclass of
824 rtl_opt_pass along with...
825 (pass_data_postreload): ...new pass_data instance and...
826 (make_pass_postreload): ...new function.
827 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
828 subclass of rtl_opt_pass along with...
829 (pass_data_gcse2): ...new pass_data instance and...
830 (make_pass_gcse2): ...new function.
831 * postreload.c (pass_postreload_cse): Convert from a global struct to
832 a subclass of rtl_opt_pass along with...
833 (pass_data_postreload_cse): ...new pass_data instance and...
834 (make_pass_postreload_cse): ...new function.
835 * predict.c (pass_profile): Convert from a global struct to a subclass
836 of gimple_opt_pass along with...
837 (pass_data_profile): ...new pass_data instance and...
838 (make_pass_profile): ...new function.
839 (pass_strip_predict_hints): Convert from a global struct to a subclass
840 of gimple_opt_pass along with...
841 (pass_data_strip_predict_hints): ...new pass_data instance and...
842 (make_pass_strip_predict_hints): ...new function.
843 * recog.c (pass_peephole2): Convert from a global struct to a subclass
844 of rtl_opt_pass along with...
845 (pass_data_peephole2): ...new pass_data instance and...
846 (make_pass_peephole2): ...new function.
847 (pass_split_all_insns): Convert from a global struct to a subclass of
848 rtl_opt_pass along with...
849 (pass_data_split_all_insns): ...new pass_data instance and...
850 (make_pass_split_all_insns): ...new function.
851 (pass_split_after_reload): Convert from a global struct to a subclass
852 of rtl_opt_pass along with...
853 (pass_data_split_after_reload): ...new pass_data instance and...
854 (make_pass_split_after_reload): ...new function.
855 (pass_split_before_regstack): Convert from a global struct to a
856 subclass of rtl_opt_pass along with...
857 (pass_data_split_before_regstack): ...new pass_data instance and...
858 (make_pass_split_before_regstack): ...new function.
859 (pass_split_before_sched2): Convert from a global struct to a subclass
860 of rtl_opt_pass along with...
861 (pass_data_split_before_sched2): ...new pass_data instance and...
862 (make_pass_split_before_sched2): ...new function.
863 (pass_split_for_shorten_branches): Convert from a global struct to a
864 subclass of rtl_opt_pass along with...
865 (pass_data_split_for_shorten_branches): ...new pass_data instance
867 (make_pass_split_for_shorten_branches): ...new function.
868 * ree.c (pass_ree): Convert from a global struct to a subclass of
869 rtl_opt_pass along with...
870 (pass_data_ree): ...new pass_data instance and...
871 (make_pass_ree): ...new function.
872 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
873 subclass of rtl_opt_pass along with...
874 (pass_data_stack_regs): ...new pass_data instance and...
875 (make_pass_stack_regs): ...new function.
876 (pass_stack_regs_run): Convert from a global struct to a subclass of
877 rtl_opt_pass along with...
878 (pass_data_stack_regs_run): ...new pass_data instance and...
879 (make_pass_stack_regs_run): ...new function.
880 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
881 subclass of rtl_opt_pass along with...
882 (pass_data_cprop_hardreg): ...new pass_data instance and...
883 (make_pass_cprop_hardreg): ...new function.
884 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
885 subclass of rtl_opt_pass along with...
886 (pass_data_reginfo_init): ...new pass_data instance and...
887 (make_pass_reginfo_init): ...new function.
888 * regmove.c (pass_regmove): Convert from a global struct to a subclass
889 of rtl_opt_pass along with...
890 (pass_data_regmove): ...new pass_data instance and...
891 (make_pass_regmove): ...new function.
892 * regrename.c (pass_regrename): Convert from a global struct to a
893 subclass of rtl_opt_pass along with...
894 (pass_data_regrename): ...new pass_data instance and...
895 (make_pass_regrename): ...new function.
896 * reorg.c (pass_delay_slots): Convert from a global struct to a
897 subclass of rtl_opt_pass along with...
898 (pass_data_delay_slots): ...new pass_data instance and...
899 (make_pass_delay_slots): ...new function.
900 (pass_machine_reorg): Convert from a global struct to a subclass of
901 rtl_opt_pass along with...
902 (pass_data_machine_reorg): ...new pass_data instance and...
903 (make_pass_machine_reorg): ...new function.
904 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
905 of rtl_opt_pass along with...
906 (pass_data_sched): ...new pass_data instance and...
907 (make_pass_sched): ...new function.
908 (pass_sched2): Convert from a global struct to a subclass of
909 rtl_opt_pass along with...
910 (pass_data_sched2): ...new pass_data instance and...
911 (make_pass_sched2): ...new function.
912 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
913 to a subclass of rtl_opt_pass along with...
914 (pass_data_stack_ptr_mod): ...new pass_data instance and...
915 (make_pass_stack_ptr_mod): ...new function.
916 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
917 to a subclass of rtl_opt_pass along with...
918 (pass_data_rtl_store_motion): ...new pass_data instance and...
919 (make_pass_rtl_store_motion): ...new function.
920 * tracer.c (pass_tracer): Convert from a global struct to a subclass
921 of gimple_opt_pass along with...
922 (pass_data_tracer): ...new pass_data instance and...
923 (make_pass_tracer): ...new function.
924 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
925 to a subclass of gimple_opt_pass along with...
926 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
927 (make_pass_diagnose_tm_blocks): ...new function.
928 (pass_lower_tm): Convert from a global struct to a subclass of
929 gimple_opt_pass along with...
930 (pass_data_lower_tm): ...new pass_data instance and...
931 (make_pass_lower_tm): ...new function.
932 (pass_tm_init): Convert from a global struct to a subclass of
933 gimple_opt_pass along with...
934 (pass_data_tm_init): ...new pass_data instance and...
935 (make_pass_tm_init): ...new function.
936 (pass_tm_mark): Convert from a global struct to a subclass of
937 gimple_opt_pass along with...
938 (pass_data_tm_mark): ...new pass_data instance and...
939 (make_pass_tm_mark): ...new function.
940 (pass_tm_edges): Convert from a global struct to a subclass of
941 gimple_opt_pass along with...
942 (pass_data_tm_edges): ...new pass_data instance and...
943 (make_pass_tm_edges): ...new function.
944 (pass_tm_memopt): Convert from a global struct to a subclass of
945 gimple_opt_pass along with...
946 (pass_data_tm_memopt): ...new pass_data instance and...
947 (make_pass_tm_memopt): ...new function.
948 (pass_ipa_tm): Convert from a global struct to a subclass of
949 simple_ipa_opt_pass along with...
950 (pass_data_ipa_tm): ...new pass_data instance and...
951 (make_pass_ipa_tm): ...new function.
952 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
953 subclass of gimple_opt_pass along with...
954 (pass_data_call_cdce): ...new pass_data instance and...
955 (make_pass_call_cdce): ...new function.
956 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
957 subclass of gimple_opt_pass along with...
958 (pass_data_build_cfg): ...new pass_data instance and...
959 (make_pass_build_cfg): ...new function.
960 (pass_split_crit_edges): Convert from a global struct to a subclass of
961 gimple_opt_pass along with...
962 (pass_data_split_crit_edges): ...new pass_data instance and...
963 (make_pass_split_crit_edges): ...new function.
964 (pass_warn_function_return): Convert from a global struct to a
965 subclass of gimple_opt_pass along with...
966 (pass_data_warn_function_return): ...new pass_data instance and...
967 (make_pass_warn_function_return): ...new function.
968 (pass_warn_function_noreturn): Convert from a global struct to a
969 subclass of gimple_opt_pass along with...
970 (pass_data_warn_function_noreturn): ...new pass_data instance and...
971 (make_pass_warn_function_noreturn): ...new function.
972 (pass_warn_unused_result): Convert from a global struct to a subclass
973 of gimple_opt_pass along with...
974 (pass_data_warn_unused_result): ...new pass_data instance and...
975 (make_pass_warn_unused_result): ...new function.
976 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
977 a subclass of gimple_opt_pass along with...
978 (pass_data_merge_phi): ...new pass_data instance and...
979 (make_pass_merge_phi): ...new function.
980 * tree-complex.c (pass_lower_complex): Convert from a global struct to
981 a subclass of gimple_opt_pass along with...
982 (pass_data_lower_complex): ...new pass_data instance and...
983 (make_pass_lower_complex): ...new function.
984 (pass_lower_complex_O0): Convert from a global struct to a subclass of
985 gimple_opt_pass along with...
986 (pass_data_lower_complex_O0): ...new pass_data instance and...
987 (make_pass_lower_complex_O0): ...new function.
988 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
989 subclass of gimple_opt_pass along with...
990 (pass_data_lower_eh): ...new pass_data instance and...
991 (make_pass_lower_eh): ...new function.
992 (pass_refactor_eh): Convert from a global struct to a subclass of
993 gimple_opt_pass along with...
994 (pass_data_refactor_eh): ...new pass_data instance and...
995 (make_pass_refactor_eh): ...new function.
996 (pass_lower_resx): Convert from a global struct to a subclass of
997 gimple_opt_pass along with...
998 (pass_data_lower_resx): ...new pass_data instance and...
999 (make_pass_lower_resx): ...new function.
1000 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1001 of gimple_opt_pass along with...
1002 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1003 (make_pass_lower_eh_dispatch): ...new function.
1004 (pass_cleanup_eh): Convert from a global struct to a subclass of
1005 gimple_opt_pass along with...
1006 (pass_data_cleanup_eh): ...new pass_data instance and...
1007 (make_pass_cleanup_eh): ...new function.
1008 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1009 to a subclass of simple_ipa_opt_pass along with...
1010 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1011 (make_pass_ipa_lower_emutls): ...new function.
1012 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1013 a subclass of gimple_opt_pass along with...
1014 (pass_data_if_conversion): ...new pass_data instance and...
1015 (make_pass_if_conversion): ...new function.
1016 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1017 subclass of gimple_opt_pass along with...
1018 (pass_data_build_ssa): ...new pass_data instance and...
1019 (make_pass_build_ssa): ...new function.
1020 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1021 global struct to a subclass of gimple_opt_pass along with...
1022 (pass_data_loop_distribution): ...new pass_data instance and...
1023 (make_pass_loop_distribution): ...new function.
1024 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1025 subclass of gimple_opt_pass along with...
1026 (pass_data_mudflap_1): ...new pass_data instance and...
1027 (make_pass_mudflap_1): ...new function.
1028 (pass_mudflap_2): Convert from a global struct to a subclass of
1029 gimple_opt_pass along with...
1030 (pass_data_mudflap_2): ...new pass_data instance and...
1031 (make_pass_mudflap_2): ...new function.
1032 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1033 subclass of gimple_opt_pass along with...
1034 (pass_data_mudflap_1): ...new pass_data instance and...
1035 (make_pass_mudflap_1): ...new function.
1036 (pass_mudflap_2): Convert from a global struct to a subclass of
1037 gimple_opt_pass along with...
1038 (pass_data_mudflap_2): ...new pass_data instance and...
1039 (make_pass_mudflap_2): ...new function.
1040 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1041 gimple_opt_pass along with...
1042 (pass_data_nrv): ...new pass_data instance and...
1043 (make_pass_nrv): ...new function.
1044 (pass_return_slot): Convert from a global struct to a subclass of
1045 gimple_opt_pass along with...
1046 (pass_data_return_slot): ...new pass_data instance and...
1047 (make_pass_return_slot): ...new function.
1048 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1049 to a subclass of gimple_opt_pass along with...
1050 (pass_data_object_sizes): ...new pass_data instance and...
1051 (make_pass_object_sizes): ...new function.
1052 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1053 global struct to a subclass of gimple_opt_pass along with...
1054 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1056 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1057 (pass_fixup_cfg): Convert from a global struct to a subclass of
1058 gimple_opt_pass along with...
1059 (pass_data_fixup_cfg): ...new pass_data instance and...
1060 (make_pass_fixup_cfg): ...new function.
1061 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1062 (make_pass_mudflap_1): ...new function.
1063 (pass_mudflap_2): Replace declaration with that of...
1064 (make_pass_mudflap_2): ...new function.
1065 (pass_asan): Replace declaration with that of...
1066 (make_pass_asan): ...new function.
1067 (pass_asan_O0): Replace declaration with that of...
1068 (make_pass_asan_O0): ...new function.
1069 (pass_tsan): Replace declaration with that of...
1070 (make_pass_tsan): ...new function.
1071 (pass_tsan_O0): Replace declaration with that of...
1072 (make_pass_tsan_O0): ...new function.
1073 (pass_lower_cf): Replace declaration with that of...
1074 (make_pass_lower_cf): ...new function.
1075 (pass_refactor_eh): Replace declaration with that of...
1076 (make_pass_refactor_eh): ...new function.
1077 (pass_lower_eh): Replace declaration with that of...
1078 (make_pass_lower_eh): ...new function.
1079 (pass_lower_eh_dispatch): Replace declaration with that of...
1080 (make_pass_lower_eh_dispatch): ...new function.
1081 (pass_lower_resx): Replace declaration with that of...
1082 (make_pass_lower_resx): ...new function.
1083 (pass_build_cfg): Replace declaration with that of...
1084 (make_pass_build_cfg): ...new function.
1085 (pass_early_tree_profile): Replace declaration with that of...
1086 (make_pass_early_tree_profile): ...new function.
1087 (pass_cleanup_eh): Replace declaration with that of...
1088 (make_pass_cleanup_eh): ...new function.
1089 (pass_sra): Replace declaration with that of...
1090 (make_pass_sra): ...new function.
1091 (pass_sra_early): Replace declaration with that of...
1092 (make_pass_sra_early): ...new function.
1093 (pass_early_ipa_sra): Replace declaration with that of...
1094 (make_pass_early_ipa_sra): ...new function.
1095 (pass_tail_recursion): Replace declaration with that of...
1096 (make_pass_tail_recursion): ...new function.
1097 (pass_tail_calls): Replace declaration with that of...
1098 (make_pass_tail_calls): ...new function.
1099 (pass_tree_loop): Replace declaration with that of...
1100 (make_pass_tree_loop): ...new function.
1101 (pass_tree_loop_init): Replace declaration with that of...
1102 (make_pass_tree_loop_init): ...new function.
1103 (pass_lim): Replace declaration with that of...
1104 (make_pass_lim): ...new function.
1105 (pass_tree_unswitch): Replace declaration with that of...
1106 (make_pass_tree_unswitch): ...new function.
1107 (pass_predcom): Replace declaration with that of...
1108 (make_pass_predcom): ...new function.
1109 (pass_iv_canon): Replace declaration with that of...
1110 (make_pass_iv_canon): ...new function.
1111 (pass_scev_cprop): Replace declaration with that of...
1112 (make_pass_scev_cprop): ...new function.
1113 (pass_empty_loop): Replace declaration with that of...
1114 (make_pass_empty_loop): ...new function.
1115 (pass_record_bounds): Replace declaration with that of...
1116 (make_pass_record_bounds): ...new function.
1117 (pass_graphite): Replace declaration with that of...
1118 (make_pass_graphite): ...new function.
1119 (pass_graphite_transforms): Replace declaration with that of...
1120 (make_pass_graphite_transforms): ...new function.
1121 (pass_if_conversion): Replace declaration with that of...
1122 (make_pass_if_conversion): ...new function.
1123 (pass_loop_distribution): Replace declaration with that of...
1124 (make_pass_loop_distribution): ...new function.
1125 (pass_vectorize): Replace declaration with that of...
1126 (make_pass_vectorize): ...new function.
1127 (pass_slp_vectorize): Replace declaration with that of...
1128 (make_pass_slp_vectorize): ...new function.
1129 (pass_complete_unroll): Replace declaration with that of...
1130 (make_pass_complete_unroll): ...new function.
1131 (pass_complete_unrolli): Replace declaration with that of...
1132 (make_pass_complete_unrolli): ...new function.
1133 (pass_parallelize_loops): Replace declaration with that of...
1134 (make_pass_parallelize_loops): ...new function.
1135 (pass_loop_prefetch): Replace declaration with that of...
1136 (make_pass_loop_prefetch): ...new function.
1137 (pass_iv_optimize): Replace declaration with that of...
1138 (make_pass_iv_optimize): ...new function.
1139 (pass_tree_loop_done): Replace declaration with that of...
1140 (make_pass_tree_loop_done): ...new function.
1141 (pass_ch): Replace declaration with that of...
1142 (make_pass_ch): ...new function.
1143 (pass_ccp): Replace declaration with that of...
1144 (make_pass_ccp): ...new function.
1145 (pass_phi_only_cprop): Replace declaration with that of...
1146 (make_pass_phi_only_cprop): ...new function.
1147 (pass_build_ssa): Replace declaration with that of...
1148 (make_pass_build_ssa): ...new function.
1149 (pass_build_alias): Replace declaration with that of...
1150 (make_pass_build_alias): ...new function.
1151 (pass_build_ealias): Replace declaration with that of...
1152 (make_pass_build_ealias): ...new function.
1153 (pass_dominator): Replace declaration with that of...
1154 (make_pass_dominator): ...new function.
1155 (pass_dce): Replace declaration with that of...
1156 (make_pass_dce): ...new function.
1157 (pass_dce_loop): Replace declaration with that of...
1158 (make_pass_dce_loop): ...new function.
1159 (pass_cd_dce): Replace declaration with that of...
1160 (make_pass_cd_dce): ...new function.
1161 (pass_call_cdce): Replace declaration with that of...
1162 (make_pass_call_cdce): ...new function.
1163 (pass_merge_phi): Replace declaration with that of...
1164 (make_pass_merge_phi): ...new function.
1165 (pass_split_crit_edges): Replace declaration with that of...
1166 (make_pass_split_crit_edges): ...new function.
1167 (pass_pre): Replace declaration with that of...
1168 (make_pass_pre): ...new function.
1169 (pass_profile): Replace declaration with that of...
1170 (make_pass_profile): ...new function.
1171 (pass_strip_predict_hints): Replace declaration with that of...
1172 (make_pass_strip_predict_hints): ...new function.
1173 (pass_lower_complex_O0): Replace declaration with that of...
1174 (make_pass_lower_complex_O0): ...new function.
1175 (pass_lower_complex): Replace declaration with that of...
1176 (make_pass_lower_complex): ...new function.
1177 (pass_lower_vector): Replace declaration with that of...
1178 (make_pass_lower_vector): ...new function.
1179 (pass_lower_vector_ssa): Replace declaration with that of...
1180 (make_pass_lower_vector_ssa): ...new function.
1181 (pass_lower_omp): Replace declaration with that of...
1182 (make_pass_lower_omp): ...new function.
1183 (pass_diagnose_omp_blocks): Replace declaration with that of...
1184 (make_pass_diagnose_omp_blocks): ...new function.
1185 (pass_expand_omp): Replace declaration with that of...
1186 (make_pass_expand_omp): ...new function.
1187 (pass_expand_omp_ssa): Replace declaration with that of...
1188 (make_pass_expand_omp_ssa): ...new function.
1189 (pass_object_sizes): Replace declaration with that of...
1190 (make_pass_object_sizes): ...new function.
1191 (pass_strlen): Replace declaration with that of...
1192 (make_pass_strlen): ...new function.
1193 (pass_fold_builtins): Replace declaration with that of...
1194 (make_pass_fold_builtins): ...new function.
1195 (pass_stdarg): Replace declaration with that of...
1196 (make_pass_stdarg): ...new function.
1197 (pass_early_warn_uninitialized): Replace declaration with that of...
1198 (make_pass_early_warn_uninitialized): ...new function.
1199 (pass_late_warn_uninitialized): Replace declaration with that of...
1200 (make_pass_late_warn_uninitialized): ...new function.
1201 (pass_cse_reciprocals): Replace declaration with that of...
1202 (make_pass_cse_reciprocals): ...new function.
1203 (pass_cse_sincos): Replace declaration with that of...
1204 (make_pass_cse_sincos): ...new function.
1205 (pass_optimize_bswap): Replace declaration with that of...
1206 (make_pass_optimize_bswap): ...new function.
1207 (pass_optimize_widening_mul): Replace declaration with that of...
1208 (make_pass_optimize_widening_mul): ...new function.
1209 (pass_warn_function_return): Replace declaration with that of...
1210 (make_pass_warn_function_return): ...new function.
1211 (pass_warn_function_noreturn): Replace declaration with that of...
1212 (make_pass_warn_function_noreturn): ...new function.
1213 (pass_cselim): Replace declaration with that of...
1214 (make_pass_cselim): ...new function.
1215 (pass_phiopt): Replace declaration with that of...
1216 (make_pass_phiopt): ...new function.
1217 (pass_forwprop): Replace declaration with that of...
1218 (make_pass_forwprop): ...new function.
1219 (pass_phiprop): Replace declaration with that of...
1220 (make_pass_phiprop): ...new function.
1221 (pass_tree_ifcombine): Replace declaration with that of...
1222 (make_pass_tree_ifcombine): ...new function.
1223 (pass_dse): Replace declaration with that of...
1224 (make_pass_dse): ...new function.
1225 (pass_nrv): Replace declaration with that of...
1226 (make_pass_nrv): ...new function.
1227 (pass_rename_ssa_copies): Replace declaration with that of...
1228 (make_pass_rename_ssa_copies): ...new function.
1229 (pass_sink_code): Replace declaration with that of...
1230 (make_pass_sink_code): ...new function.
1231 (pass_fre): Replace declaration with that of...
1232 (make_pass_fre): ...new function.
1233 (pass_check_data_deps): Replace declaration with that of...
1234 (make_pass_check_data_deps): ...new function.
1235 (pass_copy_prop): Replace declaration with that of...
1236 (make_pass_copy_prop): ...new function.
1237 (pass_vrp): Replace declaration with that of...
1238 (make_pass_vrp): ...new function.
1239 (pass_uncprop): Replace declaration with that of...
1240 (make_pass_uncprop): ...new function.
1241 (pass_return_slot): Replace declaration with that of...
1242 (make_pass_return_slot): ...new function.
1243 (pass_reassoc): Replace declaration with that of...
1244 (make_pass_reassoc): ...new function.
1245 (pass_rebuild_cgraph_edges): Replace declaration with that of...
1246 (make_pass_rebuild_cgraph_edges): ...new function.
1247 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
1248 (make_pass_remove_cgraph_callee_edges): ...new function.
1249 (pass_build_cgraph_edges): Replace declaration with that of...
1250 (make_pass_build_cgraph_edges): ...new function.
1251 (pass_local_pure_const): Replace declaration with that of...
1252 (make_pass_local_pure_const): ...new function.
1253 (pass_tracer): Replace declaration with that of...
1254 (make_pass_tracer): ...new function.
1255 (pass_warn_unused_result): Replace declaration with that of...
1256 (make_pass_warn_unused_result): ...new function.
1257 (pass_diagnose_tm_blocks): Replace declaration with that of...
1258 (make_pass_diagnose_tm_blocks): ...new function.
1259 (pass_lower_tm): Replace declaration with that of...
1260 (make_pass_lower_tm): ...new function.
1261 (pass_tm_init): Replace declaration with that of...
1262 (make_pass_tm_init): ...new function.
1263 (pass_tm_mark): Replace declaration with that of...
1264 (make_pass_tm_mark): ...new function.
1265 (pass_tm_memopt): Replace declaration with that of...
1266 (make_pass_tm_memopt): ...new function.
1267 (pass_tm_edges): Replace declaration with that of...
1268 (make_pass_tm_edges): ...new function.
1269 (pass_split_functions): Replace declaration with that of...
1270 (make_pass_split_functions): ...new function.
1271 (pass_feedback_split_functions): Replace declaration with that of...
1272 (make_pass_feedback_split_functions): ...new function.
1273 (pass_strength_reduction): Replace declaration with that of...
1274 (make_pass_strength_reduction): ...new function.
1275 (pass_ipa_lower_emutls): Replace declaration with that of...
1276 (make_pass_ipa_lower_emutls): ...new function.
1277 (pass_ipa_function_and_variable_visibility): Replace declaration with
1279 (make_pass_ipa_function_and_variable_visibility): ...new function.
1280 (pass_ipa_tree_profile): Replace declaration with that of...
1281 (make_pass_ipa_tree_profile): ...new function.
1282 (pass_early_local_passes): Replace declaration with that of...
1283 (make_pass_early_local_passes): ...new function.
1284 (pass_ipa_whole_program_visibility): Replace declaration with that
1286 (make_pass_ipa_whole_program_visibility): ...new function.
1287 (pass_ipa_lto_gimple_out): Replace declaration with that of...
1288 (make_pass_ipa_lto_gimple_out): ...new function.
1289 (pass_ipa_increase_alignment): Replace declaration with that of...
1290 (make_pass_ipa_increase_alignment): ...new function.
1291 (pass_ipa_inline): Replace declaration with that of...
1292 (make_pass_ipa_inline): ...new function.
1293 (pass_ipa_free_lang_data): Replace declaration with that of...
1294 (make_pass_ipa_free_lang_data): ...new function.
1295 (pass_ipa_free_inline_summary): Replace declaration with that of...
1296 (make_pass_ipa_free_inline_summary): ...new function.
1297 (pass_ipa_cp): Replace declaration with that of...
1298 (make_pass_ipa_cp): ...new function.
1299 (pass_ipa_reference): Replace declaration with that of...
1300 (make_pass_ipa_reference): ...new function.
1301 (pass_ipa_pure_const): Replace declaration with that of...
1302 (make_pass_ipa_pure_const): ...new function.
1303 (pass_ipa_pta): Replace declaration with that of...
1304 (make_pass_ipa_pta): ...new function.
1305 (pass_ipa_lto_finish_out): Replace declaration with that of...
1306 (make_pass_ipa_lto_finish_out): ...new function.
1307 (pass_ipa_tm): Replace declaration with that of...
1308 (make_pass_ipa_tm): ...new function.
1309 (pass_ipa_profile): Replace declaration with that of...
1310 (make_pass_ipa_profile): ...new function.
1311 (pass_ipa_cdtor_merge): Replace declaration with that of...
1312 (make_pass_ipa_cdtor_merge): ...new function.
1313 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
1315 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1316 (pass_init_datastructures): Replace declaration with that of...
1317 (make_pass_init_datastructures): ...new function.
1318 (pass_fixup_cfg): Replace declaration with that of...
1319 (make_pass_fixup_cfg): ...new function.
1320 (pass_expand): Replace declaration with that of...
1321 (make_pass_expand): ...new function.
1322 (pass_instantiate_virtual_regs): Replace declaration with that of...
1323 (make_pass_instantiate_virtual_regs): ...new function.
1324 (pass_rtl_fwprop): Replace declaration with that of...
1325 (make_pass_rtl_fwprop): ...new function.
1326 (pass_rtl_fwprop_addr): Replace declaration with that of...
1327 (make_pass_rtl_fwprop_addr): ...new function.
1328 (pass_jump): Replace declaration with that of...
1329 (make_pass_jump): ...new function.
1330 (pass_jump2): Replace declaration with that of...
1331 (make_pass_jump2): ...new function.
1332 (pass_lower_subreg): Replace declaration with that of...
1333 (make_pass_lower_subreg): ...new function.
1334 (pass_cse): Replace declaration with that of...
1335 (make_pass_cse): ...new function.
1336 (pass_fast_rtl_dce): Replace declaration with that of...
1337 (make_pass_fast_rtl_dce): ...new function.
1338 (pass_ud_rtl_dce): Replace declaration with that of...
1339 (make_pass_ud_rtl_dce): ...new function.
1340 (pass_rtl_dce): Replace declaration with that of...
1341 (make_pass_rtl_dce): ...new function.
1342 (pass_rtl_dse1): Replace declaration with that of...
1343 (make_pass_rtl_dse1): ...new function.
1344 (pass_rtl_dse2): Replace declaration with that of...
1345 (make_pass_rtl_dse2): ...new function.
1346 (pass_rtl_dse3): Replace declaration with that of...
1347 (make_pass_rtl_dse3): ...new function.
1348 (pass_rtl_cprop): Replace declaration with that of...
1349 (make_pass_rtl_cprop): ...new function.
1350 (pass_rtl_pre): Replace declaration with that of...
1351 (make_pass_rtl_pre): ...new function.
1352 (pass_rtl_hoist): Replace declaration with that of...
1353 (make_pass_rtl_hoist): ...new function.
1354 (pass_rtl_store_motion): Replace declaration with that of...
1355 (make_pass_rtl_store_motion): ...new function.
1356 (pass_cse_after_global_opts): Replace declaration with that of...
1357 (make_pass_cse_after_global_opts): ...new function.
1358 (pass_rtl_ifcvt): Replace declaration with that of...
1359 (make_pass_rtl_ifcvt): ...new function.
1360 (pass_into_cfg_layout_mode): Replace declaration with that of...
1361 (make_pass_into_cfg_layout_mode): ...new function.
1362 (pass_outof_cfg_layout_mode): Replace declaration with that of...
1363 (make_pass_outof_cfg_layout_mode): ...new function.
1364 (pass_loop2): Replace declaration with that of...
1365 (make_pass_loop2): ...new function.
1366 (pass_rtl_loop_init): Replace declaration with that of...
1367 (make_pass_rtl_loop_init): ...new function.
1368 (pass_rtl_move_loop_invariants): Replace declaration with that of...
1369 (make_pass_rtl_move_loop_invariants): ...new function.
1370 (pass_rtl_unswitch): Replace declaration with that of...
1371 (make_pass_rtl_unswitch): ...new function.
1372 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
1373 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1374 (pass_rtl_doloop): Replace declaration with that of...
1375 (make_pass_rtl_doloop): ...new function.
1376 (pass_rtl_loop_done): Replace declaration with that of...
1377 (make_pass_rtl_loop_done): ...new function.
1378 (pass_web): Replace declaration with that of...
1379 (make_pass_web): ...new function.
1380 (pass_cse2): Replace declaration with that of...
1381 (make_pass_cse2): ...new function.
1382 (pass_df_initialize_opt): Replace declaration with that of...
1383 (make_pass_df_initialize_opt): ...new function.
1384 (pass_df_initialize_no_opt): Replace declaration with that of...
1385 (make_pass_df_initialize_no_opt): ...new function.
1386 (pass_reginfo_init): Replace declaration with that of...
1387 (make_pass_reginfo_init): ...new function.
1388 (pass_inc_dec): Replace declaration with that of...
1389 (make_pass_inc_dec): ...new function.
1390 (pass_stack_ptr_mod): Replace declaration with that of...
1391 (make_pass_stack_ptr_mod): ...new function.
1392 (pass_initialize_regs): Replace declaration with that of...
1393 (make_pass_initialize_regs): ...new function.
1394 (pass_combine): Replace declaration with that of...
1395 (make_pass_combine): ...new function.
1396 (pass_if_after_combine): Replace declaration with that of...
1397 (make_pass_if_after_combine): ...new function.
1398 (pass_ree): Replace declaration with that of...
1399 (make_pass_ree): ...new function.
1400 (pass_partition_blocks): Replace declaration with that of...
1401 (make_pass_partition_blocks): ...new function.
1402 (pass_match_asm_constraints): Replace declaration with that of...
1403 (make_pass_match_asm_constraints): ...new function.
1404 (pass_regmove): Replace declaration with that of...
1405 (make_pass_regmove): ...new function.
1406 (pass_split_all_insns): Replace declaration with that of...
1407 (make_pass_split_all_insns): ...new function.
1408 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1409 (make_pass_fast_rtl_byte_dce): ...new function.
1410 (pass_lower_subreg2): Replace declaration with that of...
1411 (make_pass_lower_subreg2): ...new function.
1412 (pass_mode_switching): Replace declaration with that of...
1413 (make_pass_mode_switching): ...new function.
1414 (pass_sms): Replace declaration with that of...
1415 (make_pass_sms): ...new function.
1416 (pass_sched): Replace declaration with that of...
1417 (make_pass_sched): ...new function.
1418 (pass_ira): Replace declaration with that of...
1419 (make_pass_ira): ...new function.
1420 (pass_reload): Replace declaration with that of...
1421 (make_pass_reload): ...new function.
1422 (pass_clean_state): Replace declaration with that of...
1423 (make_pass_clean_state): ...new function.
1424 (pass_branch_prob): Replace declaration with that of...
1425 (make_pass_branch_prob): ...new function.
1426 (pass_value_profile_transformations): Replace declaration with that
1428 (make_pass_value_profile_transformations): ...new function.
1429 (pass_postreload_cse): Replace declaration with that of...
1430 (make_pass_postreload_cse): ...new function.
1431 (pass_gcse2): Replace declaration with that of...
1432 (make_pass_gcse2): ...new function.
1433 (pass_split_after_reload): Replace declaration with that of...
1434 (make_pass_split_after_reload): ...new function.
1435 (pass_branch_target_load_optimize1): Replace declaration with that
1437 (make_pass_branch_target_load_optimize1): ...new function.
1438 (pass_thread_prologue_and_epilogue): Replace declaration with that
1440 (make_pass_thread_prologue_and_epilogue): ...new function.
1441 (pass_stack_adjustments): Replace declaration with that of...
1442 (make_pass_stack_adjustments): ...new function.
1443 (pass_peephole2): Replace declaration with that of...
1444 (make_pass_peephole2): ...new function.
1445 (pass_if_after_reload): Replace declaration with that of...
1446 (make_pass_if_after_reload): ...new function.
1447 (pass_regrename): Replace declaration with that of...
1448 (make_pass_regrename): ...new function.
1449 (pass_cprop_hardreg): Replace declaration with that of...
1450 (make_pass_cprop_hardreg): ...new function.
1451 (pass_reorder_blocks): Replace declaration with that of...
1452 (make_pass_reorder_blocks): ...new function.
1453 (pass_branch_target_load_optimize2): Replace declaration with that
1455 (make_pass_branch_target_load_optimize2): ...new function.
1456 (pass_leaf_regs): Replace declaration with that of...
1457 (make_pass_leaf_regs): ...new function.
1458 (pass_split_before_sched2): Replace declaration with that of...
1459 (make_pass_split_before_sched2): ...new function.
1460 (pass_compare_elim_after_reload): Replace declaration with that of...
1461 (make_pass_compare_elim_after_reload): ...new function.
1462 (pass_sched2): Replace declaration with that of...
1463 (make_pass_sched2): ...new function.
1464 (pass_stack_regs): Replace declaration with that of...
1465 (make_pass_stack_regs): ...new function.
1466 (pass_stack_regs_run): Replace declaration with that of...
1467 (make_pass_stack_regs_run): ...new function.
1468 (pass_df_finish): Replace declaration with that of...
1469 (make_pass_df_finish): ...new function.
1470 (pass_compute_alignments): Replace declaration with that of...
1471 (make_pass_compute_alignments): ...new function.
1472 (pass_duplicate_computed_gotos): Replace declaration with that of...
1473 (make_pass_duplicate_computed_gotos): ...new function.
1474 (pass_variable_tracking): Replace declaration with that of...
1475 (make_pass_variable_tracking): ...new function.
1476 (pass_free_cfg): Replace declaration with that of...
1477 (make_pass_free_cfg): ...new function.
1478 (pass_machine_reorg): Replace declaration with that of...
1479 (make_pass_machine_reorg): ...new function.
1480 (pass_cleanup_barriers): Replace declaration with that of...
1481 (make_pass_cleanup_barriers): ...new function.
1482 (pass_delay_slots): Replace declaration with that of...
1483 (make_pass_delay_slots): ...new function.
1484 (pass_split_for_shorten_branches): Replace declaration with that of...
1485 (make_pass_split_for_shorten_branches): ...new function.
1486 (pass_split_before_regstack): Replace declaration with that of...
1487 (make_pass_split_before_regstack): ...new function.
1488 (pass_convert_to_eh_region_ranges): Replace declaration with that
1490 (make_pass_convert_to_eh_region_ranges): ...new function.
1491 (pass_shorten_branches): Replace declaration with that of...
1492 (make_pass_shorten_branches): ...new function.
1493 (pass_set_nothrow_function_flags): Replace declaration with that of...
1494 (make_pass_set_nothrow_function_flags): ...new function.
1495 (pass_dwarf2_frame): Replace declaration with that of...
1496 (make_pass_dwarf2_frame): ...new function.
1497 (pass_final): Replace declaration with that of...
1498 (make_pass_final): ...new function.
1499 (pass_rtl_seqabstr): Replace declaration with that of...
1500 (make_pass_rtl_seqabstr): ...new function.
1501 (pass_release_ssa_names): Replace declaration with that of...
1502 (make_pass_release_ssa_names): ...new function.
1503 (pass_early_inline): Replace declaration with that of...
1504 (make_pass_early_inline): ...new function.
1505 (pass_inline_parameters): Replace declaration with that of...
1506 (make_pass_inline_parameters): ...new function.
1507 (pass_update_address_taken): Replace declaration with that of...
1508 (make_pass_update_address_taken): ...new function.
1509 (pass_convert_switch): Replace declaration with that of...
1510 (make_pass_convert_switch): ...new function.
1511 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1512 to a subclass of simple_ipa_opt_pass along with...
1513 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1514 (make_pass_ipa_tree_profile): ...new function.
1515 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1516 subclass of gimple_opt_pass along with...
1517 (pass_data_sra_early): ...new pass_data instance and...
1518 (make_pass_sra_early): ...new function.
1519 (pass_sra): Convert from a global struct to a subclass of
1520 gimple_opt_pass along with...
1521 (pass_data_sra): ...new pass_data instance and...
1522 (make_pass_sra): ...new function.
1523 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1524 gimple_opt_pass along with...
1525 (pass_data_early_ipa_sra): ...new pass_data instance and...
1526 (make_pass_early_ipa_sra): ...new function.
1527 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1528 subclass of gimple_opt_pass along with...
1529 (pass_data_ccp): ...new pass_data instance and...
1530 (make_pass_ccp): ...new function.
1531 (pass_fold_builtins): Convert from a global struct to a subclass of
1532 gimple_opt_pass along with...
1533 (pass_data_fold_builtins): ...new pass_data instance and...
1534 (make_pass_fold_builtins): ...new function.
1535 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1536 subclass of gimple_opt_pass along with...
1537 (pass_data_copy_prop): ...new pass_data instance and...
1538 (make_pass_copy_prop): ...new function.
1539 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1540 global struct to a subclass of gimple_opt_pass along with...
1541 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1542 (make_pass_rename_ssa_copies): ...new function.
1543 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1544 subclass of gimple_opt_pass along with...
1545 (pass_data_dce): ...new pass_data instance and...
1546 (make_pass_dce): ...new function.
1547 (pass_dce_loop): Convert from a global struct to a subclass of
1548 gimple_opt_pass along with...
1549 (pass_data_dce_loop): ...new pass_data instance and...
1550 (make_pass_dce_loop): ...new function.
1551 (pass_cd_dce): Convert from a global struct to a subclass of
1552 gimple_opt_pass along with...
1553 (pass_data_cd_dce): ...new pass_data instance and...
1554 (make_pass_cd_dce): ...new function.
1555 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1556 subclass of gimple_opt_pass along with...
1557 (pass_data_dominator): ...new pass_data instance and...
1558 (make_pass_dominator): ...new function.
1559 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1560 gimple_opt_pass along with...
1561 (pass_data_phi_only_cprop): ...new pass_data instance and...
1562 (make_pass_phi_only_cprop): ...new function.
1563 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1564 subclass of gimple_opt_pass along with...
1565 (pass_data_dse): ...new pass_data instance and...
1566 (make_pass_dse): ...new function.
1567 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1568 a subclass of gimple_opt_pass along with...
1569 (pass_data_forwprop): ...new pass_data instance and...
1570 (make_pass_forwprop): ...new function.
1571 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1572 struct to a subclass of gimple_opt_pass along with...
1573 (pass_data_tree_ifcombine): ...new pass_data instance and...
1574 (make_pass_tree_ifcombine): ...new function.
1575 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1576 subclass of gimple_opt_pass along with...
1577 (pass_data_ch): ...new pass_data instance and...
1578 (make_pass_ch): ...new function.
1579 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1580 subclass of gimple_opt_pass along with...
1581 (pass_data_tree_loop): ...new pass_data instance and...
1582 (make_pass_tree_loop): ...new function.
1583 (pass_tree_loop_init): Convert from a global struct to a subclass of
1584 gimple_opt_pass along with...
1585 (pass_data_tree_loop_init): ...new pass_data instance and...
1586 (make_pass_tree_loop_init): ...new function.
1587 (pass_lim): Convert from a global struct to a subclass of
1588 gimple_opt_pass along with...
1589 (pass_data_lim): ...new pass_data instance and...
1590 (make_pass_lim): ...new function.
1591 (pass_tree_unswitch): Convert from a global struct to a subclass of
1592 gimple_opt_pass along with...
1593 (pass_data_tree_unswitch): ...new pass_data instance and...
1594 (make_pass_tree_unswitch): ...new function.
1595 (pass_predcom): Convert from a global struct to a subclass of
1596 gimple_opt_pass along with...
1597 (pass_data_predcom): ...new pass_data instance and...
1598 (make_pass_predcom): ...new function.
1599 (pass_vectorize): Convert from a global struct to a subclass of
1600 gimple_opt_pass along with...
1601 (pass_data_vectorize): ...new pass_data instance and...
1602 (make_pass_vectorize): ...new function.
1603 (pass_graphite): Convert from a global struct to a subclass of
1604 gimple_opt_pass along with...
1605 (pass_data_graphite): ...new pass_data instance and...
1606 (make_pass_graphite): ...new function.
1607 (pass_graphite_transforms): Convert from a global struct to a subclass
1608 of gimple_opt_pass along with...
1609 (pass_data_graphite_transforms): ...new pass_data instance and...
1610 (make_pass_graphite_transforms): ...new function.
1611 (pass_check_data_deps): Convert from a global struct to a subclass of
1612 gimple_opt_pass along with...
1613 (pass_data_check_data_deps): ...new pass_data instance and...
1614 (make_pass_check_data_deps): ...new function.
1615 (pass_iv_canon): Convert from a global struct to a subclass of
1616 gimple_opt_pass along with...
1617 (pass_data_iv_canon): ...new pass_data instance and...
1618 (make_pass_iv_canon): ...new function.
1619 (pass_scev_cprop): Convert from a global struct to a subclass of
1620 gimple_opt_pass along with...
1621 (pass_data_scev_cprop): ...new pass_data instance and...
1622 (make_pass_scev_cprop): ...new function.
1623 (pass_record_bounds): Convert from a global struct to a subclass of
1624 gimple_opt_pass along with...
1625 (pass_data_record_bounds): ...new pass_data instance and...
1626 (make_pass_record_bounds): ...new function.
1627 (pass_complete_unroll): Convert from a global struct to a subclass of
1628 gimple_opt_pass along with...
1629 (pass_data_complete_unroll): ...new pass_data instance and...
1630 (make_pass_complete_unroll): ...new function.
1631 (pass_complete_unrolli): Convert from a global struct to a subclass of
1632 gimple_opt_pass along with...
1633 (pass_data_complete_unrolli): ...new pass_data instance and...
1634 (make_pass_complete_unrolli): ...new function.
1635 (pass_parallelize_loops): Convert from a global struct to a subclass
1636 of gimple_opt_pass along with...
1637 (pass_data_parallelize_loops): ...new pass_data instance and...
1638 (make_pass_parallelize_loops): ...new function.
1639 (pass_loop_prefetch): Convert from a global struct to a subclass of
1640 gimple_opt_pass along with...
1641 (pass_data_loop_prefetch): ...new pass_data instance and...
1642 (make_pass_loop_prefetch): ...new function.
1643 (pass_iv_optimize): Convert from a global struct to a subclass of
1644 gimple_opt_pass along with...
1645 (pass_data_iv_optimize): ...new pass_data instance and...
1646 (make_pass_iv_optimize): ...new function.
1647 (pass_tree_loop_done): Convert from a global struct to a subclass of
1648 gimple_opt_pass along with...
1649 (pass_data_tree_loop_done): ...new pass_data instance and...
1650 (make_pass_tree_loop_done): ...new function.
1651 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1652 struct to a subclass of gimple_opt_pass along with...
1653 (pass_data_cse_reciprocals): ...new pass_data instance and...
1654 (make_pass_cse_reciprocals): ...new function.
1655 (pass_cse_sincos): Convert from a global struct to a subclass of
1656 gimple_opt_pass along with...
1657 (pass_data_cse_sincos): ...new pass_data instance and...
1658 (make_pass_cse_sincos): ...new function.
1659 (pass_optimize_bswap): Convert from a global struct to a subclass of
1660 gimple_opt_pass along with...
1661 (pass_data_optimize_bswap): ...new pass_data instance and...
1662 (make_pass_optimize_bswap): ...new function.
1663 (pass_optimize_widening_mul): Convert from a global struct to a
1664 subclass of gimple_opt_pass along with...
1665 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1666 (make_pass_optimize_widening_mul): ...new function.
1667 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1668 subclass of gimple_opt_pass along with...
1669 (pass_data_phiopt): ...new pass_data instance and...
1670 (make_pass_phiopt): ...new function.
1671 (pass_cselim): Convert from a global struct to a subclass of
1672 gimple_opt_pass along with...
1673 (pass_data_cselim): ...new pass_data instance and...
1674 (make_pass_cselim): ...new function.
1675 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1676 subclass of gimple_opt_pass along with...
1677 (pass_data_phiprop): ...new pass_data instance and...
1678 (make_pass_phiprop): ...new function.
1679 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1680 subclass of gimple_opt_pass along with...
1681 (pass_data_pre): ...new pass_data instance and...
1682 (make_pass_pre): ...new function.
1683 (pass_fre): Convert from a global struct to a subclass of
1684 gimple_opt_pass along with...
1685 (pass_data_fre): ...new pass_data instance and...
1686 (make_pass_fre): ...new function.
1687 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1688 subclass of gimple_opt_pass along with...
1689 (pass_data_reassoc): ...new pass_data instance and...
1690 (make_pass_reassoc): ...new function.
1691 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1692 subclass of gimple_opt_pass along with...
1693 (pass_data_sink_code): ...new pass_data instance and...
1694 (make_pass_sink_code): ...new function.
1695 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1696 subclass of gimple_opt_pass along with...
1697 (pass_data_strlen): ...new pass_data instance and...
1698 (make_pass_strlen): ...new function.
1699 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1700 struct to a subclass of gimple_opt_pass along with...
1701 (pass_data_build_alias): ...new pass_data instance and...
1702 (make_pass_build_alias): ...new function.
1703 (pass_build_ealias): Convert from a global struct to a subclass of
1704 gimple_opt_pass along with...
1705 (pass_data_build_ealias): ...new pass_data instance and...
1706 (make_pass_build_ealias): ...new function.
1707 (pass_ipa_pta): Convert from a global struct to a subclass of
1708 simple_ipa_opt_pass along with...
1709 (pass_data_ipa_pta): ...new pass_data instance and...
1710 (make_pass_ipa_pta): ...new function.
1711 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1712 subclass of gimple_opt_pass along with...
1713 (pass_data_uncprop): ...new pass_data instance and...
1714 (make_pass_uncprop): ...new function.
1715 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1716 global struct to a subclass of gimple_opt_pass along with...
1717 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1718 (make_pass_late_warn_uninitialized): ...new function.
1719 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1720 to a subclass of gimple_opt_pass along with...
1721 (pass_data_init_datastructures): ...new pass_data instance and...
1722 (make_pass_init_datastructures): ...new function.
1723 (pass_early_warn_uninitialized): Convert from a global struct to a
1724 subclass of gimple_opt_pass along with...
1725 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1726 (make_pass_early_warn_uninitialized): ...new function.
1727 (pass_update_address_taken): Convert from a global struct to a
1728 subclass of gimple_opt_pass along with...
1729 (pass_data_update_address_taken): ...new pass_data instance and...
1730 (make_pass_update_address_taken): ...new function.
1731 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1732 struct to a subclass of gimple_opt_pass along with...
1733 (pass_data_release_ssa_names): ...new pass_data instance and...
1734 (make_pass_release_ssa_names): ...new function.
1735 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1736 subclass of gimple_opt_pass along with...
1737 (pass_data_stdarg): ...new pass_data instance and...
1738 (make_pass_stdarg): ...new function.
1739 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1740 global struct to a subclass of gimple_opt_pass along with...
1741 (pass_data_convert_switch): ...new pass_data instance and...
1742 (make_pass_convert_switch): ...new function.
1743 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1744 to a subclass of gimple_opt_pass along with...
1745 (pass_data_tail_recursion): ...new pass_data instance and...
1746 (make_pass_tail_recursion): ...new function.
1747 (pass_tail_calls): Convert from a global struct to a subclass of
1748 gimple_opt_pass along with...
1749 (pass_data_tail_calls): ...new pass_data instance and...
1750 (make_pass_tail_calls): ...new function.
1751 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1752 struct to a subclass of gimple_opt_pass along with...
1753 (pass_data_lower_vector): ...new pass_data instance and...
1754 (make_pass_lower_vector): ...new function.
1755 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1756 gimple_opt_pass along with...
1757 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1758 (make_pass_lower_vector_ssa): ...new function.
1759 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1760 to a subclass of gimple_opt_pass along with...
1761 (pass_data_slp_vectorize): ...new pass_data instance and...
1762 (make_pass_slp_vectorize): ...new function.
1763 (pass_ipa_increase_alignment): Convert from a global struct to a
1764 subclass of simple_ipa_opt_pass along with...
1765 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1766 (make_pass_ipa_increase_alignment): ...new function.
1767 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1768 gimple_opt_pass along with...
1769 (pass_data_vrp): ...new pass_data instance and...
1770 (make_pass_vrp): ...new function.
1771 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1772 subclass of simple_ipa_opt_pass along with...
1773 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1774 (make_pass_ipa_free_lang_data): ...new function.
1775 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1776 gimple_opt_pass along with...
1777 (pass_data_tsan): ...new pass_data instance and...
1778 (make_pass_tsan): ...new function.
1779 (pass_tsan_O0): Convert from a global struct to a subclass of
1780 gimple_opt_pass along with...
1781 (pass_data_tsan_O0): ...new pass_data instance and...
1782 (make_pass_tsan_O0): ...new function.
1783 * var-tracking.c (pass_variable_tracking): Convert from a global
1784 struct to a subclass of rtl_opt_pass along with...
1785 (pass_data_variable_tracking): ...new pass_data instance and...
1786 (make_pass_variable_tracking): ...new function.
1787 * web.c (pass_web): Convert from a global struct to a subclass of
1788 rtl_opt_pass along with...
1789 (pass_data_web): ...new pass_data instance and...
1790 (make_pass_web): ...new function.
1791 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1792 declaration with that of...
1793 (make_pass_mode_switch_use): ...new function.
1794 (pass_resolve_sw_modes): Replace declaration with that of...
1795 (make_pass_resolve_sw_modes): ...new function.
1796 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1797 from a global struct to a subclass of rtl_opt_pass along with...
1798 (pass_data_mode_switch_use): ...new pass_data instance and...
1799 (make_pass_mode_switch_use): ...new function.
1800 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1801 from a global struct to a subclass of rtl_opt_pass along with...
1802 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1803 (make_pass_resolve_sw_modes): ...new function.
1804 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1805 struct to a subclass of rtl_opt_pass along with...
1806 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1807 (make_pass_insert_vzeroupper): ...new function.
1808 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1809 global struct to a subclass of rtl_opt_pass along with...
1810 (pass_data_work_around_errata): ...new pass_data instance and...
1811 (make_pass_work_around_errata): ...new function.
1812 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1813 struct to a subclass of rtl_opt_pass along with...
1814 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1815 (make_pass_mips_machine_reorg2): ...new function.
1817 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1819 * passes.c (pass_manager::operator new): New.
1821 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1823 Handwritten part of conversion of passes to C++ classes.
1825 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1826 (toplev.o): Add dep on PASS_MANAGER_H.
1827 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1828 of early local pases to reflect this moving from a global to a
1829 member of gcc::pass_manager.
1830 (cgraph_add_new_function): Likewise.
1831 * lto-cgraph.c (lto_output_node): Update for conversion of
1832 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1833 * passes.c (opt_pass::clone): New.
1834 (opt_pass::gate): New.
1835 (opt_pass::execute): New.
1836 (opt_pass::opt_pass): New.
1837 (pass_manager::execute_early_local_passes): New.
1838 (pass_manager::execute_pass_mode_switching): new.
1839 (finish_optimization_passes): Convert to...
1840 (pass_manager::finish_optimization_passes): ...this.
1841 (finish_optimization_passes): Update for conversion of passes to
1843 (register_dump_files_1): Use has_gate since we cannot portably
1844 check a vtable entry against NULL.
1845 (dump_one_pass): Likewise.
1846 (ipa_write_summaries_2): Likewise.
1847 (ipa_write_optimization_summaries_1): Likewise.
1848 (ipa_read_summaries_1): Likewise.
1849 (ipa_read_optimization_summaries_1): Likewise.
1850 (execute_ipa_stmt_fixups): Likewise.
1851 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1852 pass-creation functions rather than wiring up globals, and
1853 storing the results in fields of pass_manager generated using
1855 (pass_manager::dump_profile_report): Update for conversion of
1856 passes to C++ classes.
1857 (pass_manager::execute_ipa_summary_passes): Likewise.
1858 (execute_one_ipa_transform_pass): Likewise.
1859 (execute_one_pass): Use has_gate and has_execute since we cannot
1860 portably check a vtable entry against NULL.
1861 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1862 (pass_manager): Use pass-instances.def to add fields for the
1863 various pass instances.
1864 * toplev.c (finalize): Update for move of
1865 finish_optimization_passes to a method of gcc::pass_manager.
1866 * toplev.h (finish_optimization_passes): Move to method of class
1868 * tree-pass.h (struct pass_data): New.
1869 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
1870 (opt_pass::gate): Convert to virtual function.
1871 (opt_pass::~opt_pass): New.
1872 (opt_pass::clone): New.
1873 (opt_pass::execute): Convert to virtual function.
1874 (opt_pass::opt_pass): New.
1875 (opt_pass::ctxt_): new.
1876 (gimple_opt_pass): Convert to subclass of opt_pass.
1877 (gimple_opt_pass::gimple_opt_pass): New.
1878 (rtl_opt_pass): Convert to subclass of opt_pass.
1879 (rtl_opt_pass::rtl_opt_pass): New.
1880 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1881 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1882 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1883 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1884 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1885 invocation of pass_mode_switching to reflect this moving from a
1886 global to a member of gcc::pass_manager.
1887 (ix86_option_override): Rework how pass_insert_vzeroupper is
1888 added to the pass_manager to reflect autogenerated changes.
1889 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
1891 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
1893 PR rtl-optimization/57708
1894 * recog.c (peep2_find_free_register): Validate all regs in a
1897 2013-08-05 Jan Hubicka <jh@suse.cz>
1900 * cgraph.c (verify_cgraph_node): Accept local flags from other
1902 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1903 (function_and_variable_visibility): Likewise.
1904 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1906 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1908 * graph.c (init_graph_slim_pretty_print): Remove.
1909 (print_graph_cfg): Do not call it. Use local pretty printer.
1910 (start_graph_dump): Likewise.
1912 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1914 * gimple-pretty-print.c (buffer): Remove.
1915 (initialized): Likewise.
1916 (maybe_init_pretty_print): Likewise.
1917 (print_gimple_stmt): Do not call it. Use non-static local
1918 pretty_printer variable.
1919 (print_gimple_expr): Likewise.
1920 (print_gimple_seq): Likewise.
1921 (gimple_dump_bb): Likewise.
1923 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1925 * asan.c (asan_pp): Remove.
1926 (asan_pp_initialized): Likewise.
1927 (asan_pp_initialize): Likewise.
1928 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1929 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1930 (asan_add_global): Likewise.
1932 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1934 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1935 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1936 * pretty-print.c (pp_formatted_text_data): Likewise.
1937 (pp_indent): Rename from pp_base_indent.
1938 (pp_format): Rename from pp_base_format.
1939 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1940 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1941 (pp_flush): Rename from pp_base_flush.
1942 (pp_set_line_maximum_length): Rename from
1943 pp_base_set_line_maximum_length.
1944 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1945 (pp_set_prefix): Rename from pp_base_set_prefix.
1946 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1947 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1948 (pp_append_text): Rename from pp_base_append_text.
1949 (pp_formatted_text): Rename from pp_base_formatted_text.
1950 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1951 (pp_remaining_character_count_for_line): Rename from
1952 pp_base_remaining_character_count_for_line.
1953 (pp_newline): Rename from pp_base_newline.
1954 (pp_character): Rename from pp_base_character.
1955 (pp_string): Rename from pp_base_string.
1956 (pp_maybe_space): Rename from pp_base_maybe_space.
1957 * asan.c (asan_pp_string): Adjust.
1958 (asan_emit_stack_protection): Likewise.
1959 (asan_add_global): Likewise.
1960 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1961 * tree-mudflap.c (mf_varname_tree): Likewise.
1962 * tree-pretty-print.c (pp_tree_identifier): Rename from
1963 pp_base_tree_identifier.
1964 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1965 Declare as function.
1967 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1969 * pretty-print.h (pp_bar_bar): New.
1970 (pp_ampersand_ampersand): Likewise.
1971 (pp_less_equal): Likewise.
1972 (pp_greater_equal): Likewise.
1973 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1974 printer functions instead of pp_string or operators and punctuators.
1975 (dump_gimple_call): Likewise.
1976 (dump_gimple_omp_for): Likewise.
1977 (dump_gimple_transaction): Likewise.
1978 (dump_gimple_phi): Likewise.
1979 (pp_gimple_stmt_1): Likewise.
1980 * sched-vis.c (print_insn): Likewise.
1981 * tree-mudflap.c (mf_varname_tree): Likewise.
1982 * tree-pretty-print.c (dump_block_node): Likewise.
1983 (dump_generic_node): Likewise.
1985 2013-08-02 Jan Hubicka <jh@suse.cz>
1987 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1989 * lto-streamer-out.c (tree_is_indexable): Results decls and
1990 parm decls are not indexable.
1991 (DFS_write_tree_body): Do not follow args and results.
1992 (hash_tree): Likewise.
1993 (output_functions): Rearrange so struct function is needed
1994 only when real body is output; be able to also ouptut abstract
1995 functions; output DECL_ARGUMENTS and DECL_RESULT.
1996 (lto_output): When not in WPA, ale store abstract functions.
1997 (write_symbol): Do not care about RESULT_DECL.
1998 (output_symbol_p): Handle correctly sbtract decls.
1999 * lto-streamer-in.c (input_function): Rearrange so struct
2000 function can be NULL at entry; allow streaming of
2001 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2002 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2003 sanity check during LTO.
2004 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2005 RESULT_DECl and DECL_ARGUMENTS.
2006 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2009 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2011 * pretty-print.h (pp_underscore): New.
2013 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2014 printer functions instead of pp_character.
2015 (dump_binary_rhs): Likewise.
2016 (dump_ternary_rhs): Likewise.
2017 (dump_gimple_call_args): Likewise.
2018 (pp_points_to_solution): Likewise.
2019 (dump_gimple_call): Likewise.
2020 (dump_gimple_switch): Likewise.
2021 (dump_gimple_cond): Likewise.
2022 (dump_gimple_bind): Likewise.
2023 (dump_gimple_try): Likewise.
2024 (dump_gimple_omp_for): Likewise.
2025 (dump_gimple_omp_continue): Likewise.
2026 (dump_gimple_omp_single): Likewise.
2027 (dump_gimple_omp_sections): Likewise.
2028 (dump_gimple_omp_block): Likewise.
2029 (dump_gimple_omp_critical): Likewise.
2030 (dump_gimple_transaction): Likewise.
2031 (dump_gimple_asm): Likewise.
2032 (dump_gimple_phi): Likewise.
2033 (dump_gimple_omp_parallel): Likewise.
2034 (dump_gimple_omp_task): Likewise.
2035 (dump_gimple_omp_atomic_load): Likewise.
2036 (dump_gimple_omp_atomic_store): Likewise.
2037 (dump_gimple_mem_ops): Likewise.
2038 (pp_gimple_stmt_1): Likewise.
2039 (pp_cfg_jump): Likewise.
2040 (dump_implicit_edges): Likewise.
2041 (gimple_dump_bb_for_graph): Likewise.
2042 * graph.c (draw_cfg_node): Likewise.
2043 * langhooks.c (lhd_print_error_function): Likewise.
2044 * sched-vis.c (print_exp): Likewise.
2045 (print_value): Likewise.
2046 (print_pattern): Likewise.
2047 (print_insn): Likewise.
2048 (rtl_dump_bb_for_graph): Likewise.
2049 * tree-pretty-print.c (dump_function_declaration): Likewise.
2050 (dump_array_domain): Likewise.
2051 (dump_omp_clause): Likewise.
2052 (dump_location): Likewise.
2053 (dump_generic_node): Likewise.
2054 (print_struct_decl): Likewise.
2055 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2057 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2059 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2060 candidate table when replacing a candidate statement.
2061 (replace_rhs_if_not_dup): Likewise.
2062 (replace_one_candidate): Likewise.
2064 2013-08-02 Jan Hubicka <jh@suse.cz>
2065 Martin Liska <marxin.liska@gmail.com>
2067 * cgraphunit.c (add_new_function): Fix logic when adding from
2069 (assemble_thunk): Rename to ...
2070 (expand_thunk); .. this one; export; get it working with
2071 general functions; make produced gimple valid.
2072 * cgraph.h (expand_thunk): Declare.
2074 2013-08-02 Jan Hubicka <jh@suse.cz>
2076 * ipa-cp.c (gather_context_independent_values): Use
2077 ipa_get_param_move_cost.
2078 (get_replacement_map): Remove PARAM; move parameter folding
2080 (create_specialized_node): Update.
2081 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2082 assert that we have gimple body; update move_cost.
2083 (count_formal_params): Assert that we have gimple body.
2084 (ipa_dump_param): New function.
2085 (ipa_alloc_node_params): Break out from ...
2086 (ipa_initialize_node_params): ... here.
2087 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2088 (ipa_write_node_info): Stream move costs.
2089 (ipa_read_node_info): Read move costs.
2090 (ipa_update_after_lto_read): Do not recompute node params.
2091 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2092 (ipa_get_param): Check we are not in WPA.
2093 (ipa_get_param_move_cost): New.
2094 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2095 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2096 parm numbers to be present.
2098 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2100 PR rtl-optimization/58048
2101 * lra-constraints.c (process_alt_operands): Don't check asm
2102 operand on register.
2104 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2106 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2107 the implied StoreLoad barrier for atomic operations if before.
2109 2013-08-02 Jan Hubicka <jh@suse.cz>
2110 Martin Liska <marxin.liska@gmail.com>
2112 * cgraph.c (cgraph_function_body_availability): Do not check
2114 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2115 symtab_node_availability): Declare.
2116 * ipa.c (can_replace_by_local_alias): New.
2117 (function_and_variable_visibility): Use it.
2118 * symtab.c (symtab_for_node_and_aliases,
2119 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2121 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2123 PR rtl-optimization/57963
2124 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2125 (lra_constraints): Use them.
2127 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2129 * config/arm/types.md (define_attr "type"): Add "load_acq"
2131 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2133 (cortex_a53_store1): Likewise.
2135 2013-08-01 Jan Hubicka <jh@suse.cz>
2137 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
2138 partitions are not needed.
2140 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
2142 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
2143 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
2144 MAYBE_NON_Q_CLASS_P where appropriate.
2146 2013-08-01 Jan Hubicka <jh@suse.cz>
2148 * cgraph.h (release_function_body): Declare.
2149 * tree.c (free_lang_data_in_decl): Free, parameters and return values
2150 of unused delcarations.
2152 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2154 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
2155 RTL form when subtracting a constant.
2157 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2159 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
2160 Generate canonical plus rtx with negated immediate instead of minus
2162 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
2164 2013-08-01 Jan Hubicka <jh@suse.cz>
2166 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
2167 (cgraph_release_function_body): Likewise.
2168 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2169 * cgraph.h (cgrpah_node): Rename abstract_and_needed
2170 to used_as_abstract_origin.
2171 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
2172 symbols used as abstract origins.
2173 * cgraphunit.c (analyze_functions): Update.
2174 * ipa.c (symtab_remove_unreachable_nodes): Recompute
2175 used_as_abstract_origin.
2176 * tree-inline.c (tree_function_versioning): Update
2177 used_as_abstract_origin; be ready for DECL_RESULT and
2178 DECL_ARGUMENTS to be NULL.
2180 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
2181 for abstract functions.
2182 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
2185 2013-08-01 Jan Hubicka <jh@suse.cz>
2187 * profile.c (compute_value_histograms): Fix thinko.
2189 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2191 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
2192 aarch-common-protos.h to extra_headers.
2193 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2194 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
2195 * config/aarch64/t-aarch64 (aarch-common.o): Define.
2197 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2199 * config/aarch64/aarch64.md (define_attr "type"): Delete.
2200 Include "../arm/types.md". Define "type" attribute for all patterns.
2201 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
2204 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
2206 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
2207 to support power8 load fusion.
2208 (fusion_gpr_mem_load): Likewise.
2210 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
2212 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
2213 declarations for power8 load fusion.
2214 (emit_fusion_gpr_load): Likewise.
2216 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
2217 tuning for power8, turn on fusion mode by default. Turn on sign
2218 extending fusion mode if normal fusion mode is on, and we are at
2220 (fusion_gpr_load_p): New function, return true if we can fuse an
2221 addis instruction with a dependent load to a GPR.
2222 (emit_fusion_gpr_load): Emit the instructions for power8 load
2225 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
2226 (VSX load fusion peepholes): New peepholes to fuse together an
2227 addi instruction with a VSX load instruction.
2229 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
2230 peepholes to fuse an addis instruction with a load to a GPR base
2231 register. If we are supporting sign extending fusions, convert
2232 sign extending loads to zero extending loads and add an explicit
2235 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2237 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
2238 aarch-common-protos.h to extra_headers.
2239 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2240 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
2241 (arm_early_store_addr_dep): Likewise.
2242 (arm_no_early_alu_shift_dep): Likewise.
2243 (arm_no_early_alu_shift_value_dep): Likewise.
2244 (arm_no_early_mul_dep): Likewise.
2245 (arm_no_early_store_addr_dep): Likewise.
2246 (arm_mac_accumulator_is_mul_result): Likewise.
2247 (arm_mac_accumulator_is_result): Likewise.
2248 * config/arm/aarch-common.c: ... here. New file.
2249 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
2251 (arm_early_store_addr_dep): Likewise.
2252 (arm_no_early_alu_shift_dep): Likewise.
2253 (arm_no_early_alu_shift_value_dep): Likewise.
2254 (arm_no_early_mul_dep): Likewise.
2255 (arm_no_early_store_addr_dep): Likewise.
2256 (arm_mac_accumulator_is_mul_result): Likewise.
2257 (arm_mac_accumulator_is_result): Likewise.
2258 * config/arm/aarch-common-protos.h: ... here. New file.
2259 * config/arm/t-arm (aarch-common.o): Define.
2261 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2263 * config/arm/arm.md: Include new file "types.md".
2264 (define_attr "type"): Move from here to ...
2265 (define_attr "mul32"): Likewise.
2266 (define_attr "mul64"): Likewise.
2267 * config/arm/types.md: ... here. New file.
2269 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
2271 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
2272 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
2274 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2276 * gen-pass-instances.awk: Fix offset of substr().
2278 2013-07-31 David Malcolm <dmalcolm@redhat.com>
2280 * Makefile.in (pass-instances.def): New.
2281 (passes.o): Replace dependency on passes.def with one on
2284 * gen-pass-instances.awk: New.
2286 * passes.c (pass_manager::pass_manager): Use pass-instances.def
2287 rather than passes.def, updating local definition of NEXT_PASS
2288 macro to add an extra NUM parameter (currently unused).
2290 2013-07-30 David Malcolm <dmalcolm@redhat.com>
2292 * Makefile.in (PASS_MANAGER_H): New.
2293 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
2294 (passes.o): Likewise.
2295 (statistics.o): Likewise.
2296 (cgraphunit.o): Likewise.
2297 (context.o): Depend on PASS_MANAGER_H.
2299 * pass_manager.h: New.
2301 * cgraphunit.c (cgraph_add_new_function): Update for moves
2302 of globals to fields of pass_manager.
2303 (analyze_function): Likewise.
2304 (expand_function): Likewise.
2305 (ipa_passes): Likewise.
2306 (compile): Likewise.
2308 * context.c (context::context): New.
2309 * context.h (context::context): New.
2310 (context::get_passes): New.
2311 (context::passes_): New.
2313 * lto-cgraph.c (input_node): Update for moves of globals to
2314 fields of pass_manager.
2316 * passes.c (all_passes): Remove, in favor of a field of the
2317 same name within the new class pass_manager.
2318 (all_small_ipa_passes): Likewise.
2319 (all_lowering_passes): Likewise.
2320 (all_regular_ipa_passes): Likewise.
2321 (all_late_ipa_passes): Likewise.
2322 (all_lto_gen_passes): Likewise.
2323 (passes_by_id): Likewise.
2324 (passes_by_id_size): Likewise.
2325 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2326 the new class pass_manager.
2327 (set_pass_for_id): Convert to...
2328 (pass_manager::set_pass_for_id): ...method.
2329 (get_pass_for_id): Convert to...
2330 (pass_manager::get_pass_for_id): ...method.
2331 (register_one_dump_file): Move body of implementation into...
2332 (pass_manager::register_one_dump_file): ...here.
2333 (register_dump_files_1): Convert to...
2334 (pass_manager::register_dump_files_1): ...method.
2335 (register_dump_files): Convert to...
2336 (pass_manager::register_dump_files): ...method.
2337 (create_pass_tab): Update for moves of globals to fields of
2339 (dump_passes): Move body of implementation into...
2340 (pass_manager::dump_passes): ...here.
2341 (register_pass): Move body of implementation into...
2342 (pass_manager::register_pass): ...here.
2343 (init_optimization_passes): Convert into...
2344 (pass_manager::pass_manager): ...constructor for new
2345 pass_manager class, and initialize the pass_lists array.
2346 (check_profile_consistency): Update for moves of globals to
2347 fields of pass_manager.
2348 (dump_profile_report): Move body of implementation into...
2349 (pass_manager::dump_profile_report): ...here.
2350 (ipa_write_summaries_1): Update for moves of pass lists from
2351 being globals to fields of pass_manager.
2352 (ipa_write_optimization_summaries): Likewise.
2353 (ipa_read_summaries): Likewise.
2354 (ipa_read_optimization_summaries): Likewise.
2355 (execute_all_ipa_stmt_fixups): Likewise.
2357 * statistics.c (statistics_fini): Update for moves of globals to
2358 fields of pass_manager.
2360 * toplev.c (general_init): Replace call to
2361 init_optimization_passes with construction of the pass_manager
2364 * tree-pass.h (all_passes): Remove, in favor of a field of the
2365 same name within the new class pass_manager.
2366 (all_small_ipa_passes): Likewise.
2367 (all_lowering_passes): Likewise.
2368 (all_regular_ipa_passes): Likewise.
2369 (all_lto_gen_passes): Likewise.
2370 (all_late_ipa_passes): Likewise.
2371 (passes_by_id): Likewise.
2372 (passes_by_id_size): Likewise.
2373 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2374 the new class pass_manager.
2375 (get_pass_for_id): Remove.
2377 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2379 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
2382 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2384 * arm.md (mulhi3): New expand pattern.
2386 2013-07-30 Jan Hubicka <jh@suse.cz>
2387 Martin Liska <marxin.liska@gmail.com>
2389 * profile.c (compute_value_histograms): Do not ICE when
2390 there is mismatch only on some counters.
2392 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2394 PR rtl-optimization/57637
2395 * function.c (move_insn_for_shrink_wrap): Also check the
2396 GEN set of the LIVE problem for the liveness analysis
2397 if it exists, otherwise give up.
2399 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2401 PR tree-optimization/57993
2402 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2403 replaced statement in the candidate table.
2404 (phi_add_costs): Return infinite cost when the hidden basis does
2405 not dominate all phis on which the candidate is dependent.
2406 (replace_one_candidate): Record replaced statement in the
2409 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2411 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2412 (ashlv2si3): New expander.
2413 (*ashlv2si3_i): New define_insn_and_split.
2414 * predicates.md (float_operation): Allow patterns with three
2417 PR rtl-optimization/58021
2418 * mode-switching.c (create_pre_exit): Always split off preceding
2419 insns if we are not at the basic block head.
2421 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2423 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2424 (UCLIBC_DYNAMIC_LINKER): New macro.
2425 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2427 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2428 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2430 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2431 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2432 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2433 for SF and DF modes. Use ieee_quad_format for TF mode.
2434 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2435 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2436 (mips_option_override): Handle `-mnan=legacy'.
2437 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2438 `-mabs=2008' and `-mnan=2008'.
2439 (OPTION_DEFAULT_SPECS): Add "nan" default.
2440 (ASM_SPEC): Handle `-mnan='.
2441 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2442 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2443 comment accordingly.
2444 (neg<mode>2): Likewise.
2445 * config/mips/mips.opt (mabs, mnan): New options.
2446 * doc/install.texi (Configuration): Document `--with-nan=' option.
2447 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2449 (MIPS Options): Document them.
2450 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2451 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2452 * configure: Regenerate.
2453 * config.in: Regenerate.
2455 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2457 * config/i386/i386.md (float post-reload splitters): Do not check
2458 for subregs of SSE registers.
2460 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2461 H.J. Lu <hongjiu.lu@intel.com>
2465 * config/i386/i386.md (post-reload splitter
2466 to avoid partial SSE reg dependency stalls): New pattern.
2468 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2470 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2472 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2474 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2475 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2477 2013-07-26 Cary Coutant <ccoutant@google.com>
2479 * dwarf2out.c (die_checksum_ordered): Don't include template
2480 instantiations in signature.
2481 (is_template_parameter): New function.
2482 (is_template_instantiation): New function.
2483 (generate_skeleton_bottom_up): Don't include template instantiations
2485 (generate_skeleton): Likewise.
2486 (break_out_comdat_types): Move recursive call to break out nested
2488 (prune_unused_types_mark_generic_parms_dies): Call
2489 is_template_parameter.
2491 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2493 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2494 uses vector registers.
2495 * config/aarch64/iterators.md: Add attributes rtn and vas.
2497 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2498 Richard Earnshaw <richard.earnshaw@arm.com>
2500 * combine.c (simplify_comparison): Re-canonicalize operands
2502 * config/arm/arm.md (movcond_addsi): New splitter.
2504 2013-07-25 Sterling Augustine <saugustine@google.com>
2506 * dwarf2out.c (size_of_pubnames): Move code to...
2507 (include_pubname_in_output): ...here. New.
2508 (want_pubnames): Rearrange.
2509 (output_pubnames): Call include_pubname_in_output. Move assertion.
2511 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2513 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2515 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2518 * doc/extend.texi: Remove obsolete builtins. Fix
2519 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2521 2013-07-25 Jan Hubicka <jh@suse.cz>
2523 * cgraph.c (release_function_body): Break out from ...
2524 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2526 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2527 old_tree in the map.
2528 (create_specialized_node): Update.
2529 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2531 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2532 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2533 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2534 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2537 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2539 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2540 addsi3_carryin_alt2_<optab>): Correct output template.
2542 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2544 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2545 Adjust for arm_restrict_it.
2546 Remove trailing whitespace.
2548 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2550 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2551 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2553 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2555 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2557 PR rtl-optimization/57960
2558 * lra-constraints.c (process_alt_operands): Use the right mode
2559 when checking strict_low.
2561 2013-07-25 Jan Hubicka <jh@suse.cz>
2563 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2564 * cgraph.c (cgraph_remove_node): Do not release function body
2565 when in cgraph streaming.
2566 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
2567 in other partitions are not considered reachable; fix handling of
2570 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2572 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2574 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2579 * config/arm/arm.md ("*sibcall_insn): Replace use of
2580 Ss with US. Adjust output for v5 and v4t.
2581 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
2583 * config/arm/constraints.md ("Ss"): Rename to US.
2585 2013-07-25 Terry Guo <terry.guo@arm.com>
2587 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2588 shift_add/shift_sub0/shift_sub1 RTXs.
2590 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2591 Anton Blanchard <anton@au1.ibm.com>
2593 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2594 (altivec_vpks<VI_char>ss): Likewise.
2595 (altivec_vpks<VI_char>us): Likewise.
2596 (altivec_vpku<VI_char>us): Likewise.
2597 (altivec_vpku<VI_char>um): Likewise.
2599 2013-07-24 David Malcolm <dmalcolm@redhat.com>
2601 Introduce context class.
2603 * Makefile.in (CONTEXT_H): New.
2604 (OBJS): Add context.o.
2605 (toplev.o): Add CONTEXT_H to dependencies.
2608 * toplev.c (general_init): Create the singleton gcc::context instance.
2614 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2616 PR rtl-optimization/57968
2617 * mode-switching.c (create_pre_exit): Allow instructions that
2618 don't set a return register to need a non-exit mode.
2620 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2621 Anton Blanchard <anton@au1.ibm.com>
2623 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2624 operands to vperm for little endian.
2625 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2626 of lvsl to create the control mask for a vperm for little endian.
2628 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2629 Anton Blanchard <anton@au1.ibm.com>
2631 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2632 two operands for little-endian.
2634 2013-07-23 Steve Ellcey <sellcey@mips.com>
2636 * config/mips/mips.c (mips_case_values_threshold): New.
2637 (TARGET_CASE_VALUES_THRESHOLD): Define.
2639 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2640 Anton Blanchard <anton@au1.ibm.com>
2642 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2643 selection of field for vector splat in little endian mode.
2645 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2647 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2648 expanders to rs6000.md.
2649 (ior<mode>3): Likewise.
2650 (and<mode>3): Likewise.
2651 (one_cmpl<mode>2): Likewise.
2652 (nor<mode>3): Likewise.
2653 (andc<mode>3): Likewise.
2654 (eqv<mode>3): Likewise.
2655 (nand<mode>3): Likewise.
2656 (orc<mode>3): Likewise.
2658 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2661 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2662 to split multi-word logical operations.
2663 (rs6000_split_logical_di): Likewise.
2664 (rs6000_split_logical): Likewise.
2666 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2667 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2668 and allow TImode operations in 32-bit.
2669 (vsx_and<mode>3_64bit): Likewise.
2670 (vsx_ior<mode>3_32bit): Likewise.
2671 (vsx_ior<mode>3_64bit): Likewise.
2672 (vsx_xor<mode>3_32bit): Likewise.
2673 (vsx_xor<mode>3_64bit): Likewise.
2674 (vsx_one_cmpl<mode>2_32bit): Likewise.
2675 (vsx_one_cmpl<mode>2_64bit): Likewise.
2676 (vsx_nor<mode>3_32bit): Likewise.
2677 (vsx_nor<mode>3_64bit): Likewise.
2678 (vsx_andc<mode>3_32bit): Likewise.
2679 (vsx_andc<mode>3_64bit): Likewise.
2680 (vsx_eqv<mode>3_32bit): Likewise.
2681 (vsx_eqv<mode>3_64bit): Likewise.
2682 (vsx_nand<mode>3_32bit): Likewise.
2683 (vsx_nand<mode>3_64bit): Likewise.
2684 (vsx_orc<mode>3_32bit): Likewise.
2685 (vsx_orc<mode>3_64bit): Likewise.
2687 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2688 logical types in GPRs.
2690 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2691 logical insns to rs6000.md, and allow TImode operations in
2693 (altivec_ior<mode>3): Likewise.
2694 (altivec_xor<mode>3): Likewise.
2695 (altivec_one_cmpl<mode>2): Likewise.
2696 (altivec_nor<mode>3): Likewise.
2697 (altivec_andc<mode>3): Likewise.
2699 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2700 attributes for moving the 128-bit logical operations into
2702 (BOOL_REGS_OUTPUT): Likewise.
2703 (BOOL_REGS_OP1): Likewise.
2704 (BOOL_REGS_OP2): Likewise.
2705 (BOOL_REGS_UNARY): Likewise.
2706 (BOOL_REGS_AND_CR0): Likewise.
2707 (one_cmpl<mode>2): Add support for DI logical operations on
2708 32-bit, splitting the operations to 32-bit.
2712 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2713 changes to combine the 32/64-bit code, allow logical operations on
2714 TI mode in 32-bit, and to use similar match_operator patterns like
2715 scalar mode uses. Combine the Altivec and VSX code for logical
2716 operations, and move it here.
2717 (ior<mode>3, 128-bit types): Likewise.
2718 (xor<mode>3, 128-bit types): Likewise.
2719 (one_cmpl<mode>3, 128-bit types): Likewise.
2720 (nor<mode>3, 128-bit types): Likewise.
2721 (andc<mode>3, 128-bit types): Likewise.
2722 (eqv<mode>3, 128-bit types): Likewise.
2723 (nand<mode>3, 128-bit types): Likewise.
2724 (orc<mode>3, 128-bit types): Likewise.
2725 (and<mode>3_internal): Likewise.
2726 (bool<mode>3_internal): Likewise.
2727 (boolc<mode>3_internal1): Likewise.
2728 (boolc<mode>3_internal2): Likewise.
2729 (boolcc<mode>3_internal1): Likewise.
2730 (boolcc<mode>3_internal2): Likewise.
2731 (eqv<mode>3_internal1): Likewise.
2732 (eqv<mode>3_internal2): Likewise.
2733 (one_cmpl1<mode>3_internal): Likewise.
2735 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2737 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2738 Rename flag_stack_usage to flag_stack_usage_info.
2740 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2742 * config/microblaze/sync.md: New file.
2743 * config/microblaze/microblaze.md: Include sync.md
2744 * config/microblaze/microblaze.c: Add print_operand 'y'.
2745 * config/microblaze/constraints.md: Add memory_contraint
2746 'Q' which is a single register.
2748 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2750 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2752 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2754 * reload.c (find_reloads): Exit loop once we find this operand
2755 cannot be reloaded somehow for this alternative.
2757 * reload.c (find_reloads): Exit loop once we find a hard register.
2759 * rtlanal.c (computed_jump_p): Exit loop once we find label
2762 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2764 * cfgloopmanip.c (remove_path): Exit loop after setting
2767 * gensupport.c (subst_dup): Avoid loop if code is not
2768 MATCH_DUP nor MATCH_OP_DUP.
2770 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2772 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2774 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2776 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2777 true for SP_REGNUM if mode == ptr_mode.
2778 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2779 with value R0_REGNUM + 31.
2781 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2783 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2784 pad pointer-typed argument downward.
2786 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2788 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2789 and __ILP32__ when the ILP32 model is in use.
2791 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2793 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2794 (aarch64_load_symref_appropriately): In the case of
2795 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2796 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2797 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2798 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2799 if 'mode' doesn't equal to 'ptr_mode'.
2800 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2801 'vcall_offset'; change to call aarch64_emit_move differently depending
2802 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2803 to calculate the upper bound of 'vcall_offset'.
2804 (aarch64_cannot_force_const_mem): Change to also return true if
2806 (aarch64_legitimize_reload_address): In the case of large
2807 displacements, add new local variable 'xmode' and an assertion
2808 based on it; change to use 'xmode' to generate the new rtx and
2810 (aarch64_asm_trampoline_template): Change to generate the template
2811 differently depending on TARGET_ILP32 or not; change to use
2812 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2813 (aarch64_trampoline_size): Removed.
2814 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2815 and replace immediate literals with it. Change to use 'ptr_mode'
2816 instead of 'DImode' and call convert_memory_address if the mode
2817 of 'fnaddr' doesn't equal to 'ptr_mode'.
2818 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2820 (aarch64_elf_asm_destructor): Likewise.
2821 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2822 on TARGET_ILP32 instead of aarch64_trampoline_size.
2823 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2824 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2825 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2826 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2827 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2828 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2829 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2830 depending on the value of 'mode'.
2831 (add_losym_<mode>): New.
2832 (ldr_got_small_<mode>): New, based on ldr_got_small.
2833 (ldr_got_small): Remove.
2834 (ldr_got_small_sidi): New.
2835 * config/aarch64/iterators.md (P): New.
2836 (PTR): Change to 'ptr_mode' in the condition.
2838 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2840 * config.gcc (aarch64*-*-*): Support --with-abi.
2841 (aarch64*-*-elf): Support --with-multilib-list.
2842 (aarch64*-*-linux*): Likewise.
2843 (supported_defaults): Add abi to aarch64*-*-*.
2844 * configure.ac: Mention AArch64 for --with-multilib-list.
2845 * configure: Re-generated.
2846 * config/aarch64/biarchilp32.h: New file.
2847 * config/aarch64/biarchlp64.h: New file.
2848 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2850 (MULTILIB_DEFAULTS): Ditto.
2851 (DRIVER_SELF_SPECS): Ditto.
2852 (ASM_SPEC): Update to also substitute -mabi.
2853 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2854 file whose name depends on -mabi= and -mbig-endian.
2855 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2857 (POINTER_SIZE): New define.
2858 (POINTERS_EXTEND_UNSIGNED): Ditto.
2859 (enum aarch64_abi_type): New enumeration tag.
2860 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2861 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2862 (TARGET_ILP32): New define.
2863 * config/aarch64/aarch64.opt (mabi): New.
2865 (ilp32, lp64): New values for -mabi.
2866 * config/aarch64/t-aarch64 (comma): New define.
2867 (MULTILIB_OPTIONS): Ditto.
2868 (MULTILIB_DIRNAMES): Ditto.
2869 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2870 * doc/invoke.texi: Document -mabi for AArch64.
2872 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
2874 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2876 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2877 Anton Blanchard <anton@au1.ibm.com>
2879 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2880 endianness when selecting field to splat.
2882 2013-07-22 Eric Christopher <echristo@gmail.com>
2884 * dwarf2out.c (die_odr_checksum): New function to use
2885 CHECKSUM_ macros and ULEB128 for DIE tag.
2886 (generate_type_signature): Use.
2888 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2890 * config.gcc (sparc*-*-*): Accept leon3 processor.
2891 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2892 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2893 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2894 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2895 (mfix-ut699): Adjust comment.
2896 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2897 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2898 (CPP_CPU_SPEC): Likewise.
2899 (ASM_CPU_SPEC): Likewise.
2900 * config/sparc/sparc.c (leon3_cost): New constant.
2901 (sparc_option_override): Add leon3 support.
2902 (mem_ref): New function.
2903 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2904 (sparc_do_work_around_errata): Look into the instruction in the delay
2905 slot and adjust accordingly. Add fix for the data cache nullify issues
2906 of the UT699. Change insertion position for the NOP.
2907 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2908 (leon3_load): New reservation.
2909 (leon_store): Bump latency to 2.
2910 (grfpu): New automaton.
2911 (grfpu_alu): New unit.
2912 (grfpu_ds): Likewise.
2913 (leon_fp_alu): Adjust.
2914 (leon_fp_mult): Delete.
2915 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2916 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2917 * config/sparc/sparc.md (cpu): Add leon3.
2918 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2920 (atomic_test_and_set): Likewise.
2923 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2925 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2926 default for R5900 targets.
2927 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2928 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2929 * config/mips/mips.c (mips_option_override): Report an error for
2930 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2931 for -march=r5900 -mhard-float.
2933 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2935 * df-problems.c (can_move_insns_across): Exit loop once we
2936 find a non-fixed, non-global register.
2938 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2941 * omega.c (omega_eliminate_red): Break after setting red_found.
2942 (omega_problem_has_red_equations): Similarly after setting found.
2943 (omega_query_variable): Similarly after setting coupled.
2945 2013-07-22 Marek Polacek <polacek@redhat.com>
2947 * gimplify.c: Don't include gimple.h twice.
2949 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2951 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2952 instead of TARGET_THUMB1.
2953 (Pz): New constraint.
2954 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2956 (compare_negsi_si): Likewise.
2957 (compare_addsi2_op0): Likewise.
2958 (compare_addsi2_op1): Likewise.
2959 (addsi3_carryin_<optab>): Likewise.
2960 (addsi3_carryin_alt2_<optab>): Likewise.
2961 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2962 for arm_restrict_it.
2963 (subsi3_carryin): Likewise.
2964 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2965 (minmax_arithsi): Disable for arm_restrict_it.
2966 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2967 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2968 (satsi_<SAT:code>_shift): Likewise.
2969 (arm_shiftsi3): Add alternative for 16-bit encoding.
2970 (arm32_movhf): Disable for arm_restrict_it.
2971 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2972 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2974 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2976 * config/arm/arm.md (attribute "insn"): Delete.
2977 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2978 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2979 (not_shiftsi): Update for attribute change.
2980 (not_shiftsi_compare0): Likewise.
2981 (not_shiftsi_compare0_scratch): Likewise.
2982 (arm_one_cmplsi2): Likewise.
2983 (thumb1_one_cmplsi2): Likewise.
2984 (notsi_compare0): Likewise.
2985 (notsi_compare0_scratch): Likewise.
2986 (thumb1_movdi_insn): Likewise.
2987 (arm_movsi_insn): Likewise.
2988 (movhi_insn_arch4): Likewise.
2989 (movhi_bytes): Likewise.
2990 (arm_movqi_insn): Likewise.
2991 (thumb1_movqi_insn): Likewise.
2992 (arm32_movhf): Likewise.
2993 (thumb1_movhf): Likewise.
2994 (arm_movsf_soft_insn): Likewise.
2995 (thumb1_movsf_insn): Likewise.
2996 (thumb_movdf_insn): Likewise.
2997 (movsicc_insn): Likewise.
2998 (movsfcc_soft_insn): Likewise.
2999 (and_scc): Likewise.
3000 (cond_move): Likewise.
3001 (if_move_not): Likewise.
3002 (if_not_move): Likewise.
3003 (if_shift_move): Likewise.
3004 (if_move_shift): Likewise.
3005 (if_shift_shift): Likewise.
3006 (if_not_arith): Likewise.
3007 (if_arith_not): Likewise.
3008 (cond_move_not): Likewise.
3009 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3010 (neon_mov<mode>): Likewise.
3011 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3012 (thumb2_movsi_vfp): Likewise.
3013 (movsf_vfp): Likewise.
3014 (thumb2_movsf_vfp): Likewise.
3015 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3017 (cortexa7_older_only): Likewise.
3018 (cortexa7_younger): Likewise.
3019 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3020 (1020alu_shift_op): Likewise.
3021 (1020alu_shift_reg_op): Likewise.
3022 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3023 (alu_shift_op): Likewise.
3024 (alu_shift_reg_op): Likewise.
3025 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3026 (11_alu_shift_op): Likewise.
3027 (11_alu_shift_reg_op): Likewise.
3028 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3029 (9_alu_shift_reg_op): Likewise.
3030 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3032 (cortex_a15_alu_shift): Likewise.
3033 (cortex_a15_alu_shift_reg): Likewise.
3034 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3035 (cortex_a5_alu_shift): Likewise.
3036 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3038 (cortex_a53_alu_shift): Likewise.
3039 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3041 (cortex_a7_alu_reg): Likewise.
3042 (cortex_a7_alu_shift): Likewise.
3043 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3044 (cortex_a8_alu_shift): Likewise.
3045 (cortex_a8_alu_shift_reg): Likewise.
3046 (cortex_a8_mov): Likewise.
3047 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3048 (cortex_a9_dp_shift): Likewise.
3049 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3050 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3051 (cortex_r4_mov): Likewise.
3052 (cortex_r4_alu_shift): Likewise.
3053 (cortex_r4_alu_shift_reg): Likewise.
3054 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3055 (526_alu_shift_op): Likewise.
3056 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3057 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3058 (626te_alu_shift_op): Likewise.
3059 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3060 (726te_alu_op): Likewise.
3061 (726te_alu_shift_op): Likewise.
3062 (726te_alu_shift_reg_op): Likewise.
3063 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3064 (mp626_alu_shift_op): Likewise.
3065 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3066 (pj4_alu_e1_conds): Likewise.
3067 (pj4_alu): Likewise.
3068 (pj4_alu_conds): Likewise.
3069 (pj4_shift): Likewise.
3070 (pj4_shift_conds): Likewise.
3071 (pj4_alu_shift): Likewise.
3072 (pj4_alu_shift_conds): Likewise.
3074 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3076 * config/arm/predicates.md (shiftable_operator_strict_it):
3078 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3079 Disable cond_exec version for arm_restrict_it.
3080 (thumb2_smaxsi3): Convert to generate cond_exec.
3081 (thumb2_sminsi3): Likewise.
3082 (thumb32_umaxsi3): Likewise.
3083 (thumb2_uminsi3): Likewise.
3084 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3085 (thumb2_neg_abssi2): Likewise.
3086 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3087 (thumb2_movsicc_insn): Adjust alternatives.
3088 (thumb2_mov_negscc): Disable for arm_restrict_it.
3089 (thumb2_mov_negscc_strict_it): New pattern.
3090 (thumb2_mov_notscc_strict_it): New pattern.
3091 (thumb2_mov_notscc): Disable for arm_restrict_it.
3092 (thumb2_ior_scc): Likewise.
3093 (thumb2_ior_scc_strict_it): New pattern.
3094 (thumb2_cond_move): Adjust for arm_restrict_it.
3095 (thumb2_cond_arith): Disable for arm_restrict_it.
3096 (thumb2_cond_arith_strict_it): New pattern.
3097 (thumb2_cond_sub): Adjust for arm_restrict_it.
3098 (thumb2_movcond): Likewise.
3099 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3100 (thumb2_zero_extendhisi2_v6): Likewise.
3101 (thumb2_zero_extendqisi2_v6): Likewise.
3102 (orsi_notsi_si): Likewise.
3103 (orsi_not_shiftsi_si): Likewise.
3105 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3107 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3108 instruction sequence is 1 byte shorter.
3110 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3112 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3113 it is not needed after split.
3115 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3118 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3119 second label for nonlocal goto receivers. Don't output pic base labels
3120 unless we're producing PIC; mark that action unreachable().
3121 (ix86_save_reg): If the function contains a nonlocal label, save the
3123 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3124 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3125 (update_pic_label_number_if_needed): New.
3126 (machopic_output_function_base_name): Adjust for nonlocal receiver
3128 (machopic_should_output_picbase_label): New.
3129 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3130 (nonlocal_goto_receiver): New insn and split.
3132 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3134 * config/aarch64/aarch64-builtins.c
3135 (aarch64_fold_builtin): Fold abs in all modes.
3136 * config/aarch64/aarch64-simd-builtins.def
3137 (abs): Enable for all modes.
3138 * config/aarch64/arm_neon.h
3139 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
3140 (vabs_f64): Add missing intrinsic.
3142 2013-07-19 Ian Bolton <ian.bolton@arm.com>
3144 * config/aarch64/arm_neon.h (vabs_s64): New function
3146 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
3149 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
3150 * config/avr/avr.md (adjust_len): Add `round'.
3151 * config/avr/avr-protos.h (avr_out_round): New prototype.
3152 (avr_out_plus): Add `out_label' argument.
3153 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
3154 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
3155 Handle the case where `insn' is just a pattern.
3156 (avr_out_bitop): Handle the case where `insn' is just a pattern.
3157 (avr_out_round): New function.
3158 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
3160 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
3162 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3163 Add check for flag_stack_usage to handle -fstack-usage support
3165 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
3167 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
3168 interaction for new Power8 flags and VSX.
3170 2013-07-18 Sriraman Tallam <tmsriram@google.com>
3173 * tree-inline.c (expand_call_inline): Emit errors during
3174 early_inlining only if optimization is not turned on.
3176 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3180 * passes.c (init_optimization_passes): Move the construction of
3181 the pass hierarchy into a new passes.def file.
3183 * Makefile.in (passes.o): Add dependency on passes.def.
3185 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3187 * passes.c (init_optimization_passes): Introduce macros for
3188 constructing the tree of passes (INSERT_PASSES_AFTER,
3189 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
3190 TERMINATE_PASS_LIST).
3192 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
3193 Wei Mi <wmi@google.com>
3195 PR rtl-optimization/57878
3196 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
3198 (reload_pseudo_compare_func): Check nregs first for reload
3201 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3203 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
3205 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
3207 * read-rtl.c (validate_const_int): Once an invalid character is
3208 seen, quit the loop.
3210 * gengtype.c (write_roots): Similarly once we find the "deletable"
3211 or "if_marked" option.
3213 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3215 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
3216 "xtab" and "sat". Move value "clz" from here to ...
3217 (attriubte "type"): ... here.
3218 (satsi_<SAT:code>): Delete "insn" attribute.
3219 (satsi_<SAT:code>_shift): Likewise.
3220 (arm_zero_extendqisi2addsi): Likewise.
3221 (arm_extendqisi2addsi): Likewise.
3222 (clzsi2): Update for attribute changes.
3223 (rbitsi2): Likewise.
3224 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
3226 (arm_usatsihi): Likewise.
3227 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3229 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3231 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
3232 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
3233 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
3234 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
3235 in alphabetical order.
3236 (attribute "core_cycles"): Update for attribute changes.
3237 (arm_addsi3): Likewise.
3238 (addsi3_compare0): Likewise.
3239 (addsi3_compare0_scratch): Likewise.
3240 (addsi3_compare_op1): Likewise.
3241 (addsi3_compare_op2): Likewise.
3242 (compare_addsi2_op0): Likewise.
3243 (compare_addsi2_op1): Likewise.
3244 (addsi3_carryin_shift_<optab>): Likewise.
3245 (subsi3_carryin_shift): Likewise.
3246 (rsbsi3_carryin_shift): Likewise.
3247 (arm_subsi3_insn): Likewise.
3248 (subsi3_compare0): Likewise.
3249 (subsi3_compare): Likewise.
3250 (arm_andsi3_insn): Likewise.
3251 (thumb1_andsi3_insn): Likewise.
3252 (andsi3_compare0): Likewise.
3253 (andsi3_compare0_scratch): Likewise.
3254 (zeroextractsi_compare0_scratch
3255 (andsi_not_shiftsi_si): Likewise.
3256 (iorsi3_insn): Likewise.
3257 (iorsi3_compare0): Likewise.
3258 (iorsi3_compare0_scratch): Likewise.
3259 (arm_xorsi3): Likewise.
3260 (thumb1_xorsi3_insn): Likewise.
3261 (xorsi3_compare0): Likewise.
3262 (xorsi3_compare0_scratch): Likewise.
3263 (satsi_<SAT:code>_shift): Likewise.
3265 (arm_shiftsi3): Likewise.
3266 (shiftsi3_compare0): Likewise.
3267 (not_shiftsi): Likewise.
3268 (not_shiftsi_compare0): Likewise.
3269 (not_shiftsi_compare0_scratch): Likewise.
3270 (arm_one_cmplsi2): Likewise.
3271 (thumb_one_complsi2): Likewise.
3272 (notsi_compare0): Likewise.
3273 (notsi_compare0_scratch): Likewise.
3274 (thumb1_zero_extendhisi2): Likewise.
3275 (arm_zero_extendhisi2): Likewise.
3276 (arm_zero_extendhisi2_v6): Likewise.
3277 (arm_zero_extendhisi2addsi): Likewise.
3278 (thumb1_zero_extendqisi2): Likewise.
3279 (thumb1_zero_extendqisi2_v6): Likewise.
3280 (arm_zero_extendqisi2): Likewise.
3281 (arm_zero_extendqisi2_v6): Likewise.
3282 (arm_zero_extendqisi2addsi): Likewise.
3283 (thumb1_extendhisi2): Likewise.
3284 (arm_extendhisi2): Likewise.
3285 (arm_extendhisi2_v6): Likewise.
3286 (arm_extendqisi): Likewise.
3287 (arm_extendqisi_v6): Likewise.
3288 (arm_extendqisi2addsi): Likewise.
3289 (thumb1_extendqisi2): Likewise.
3290 (thumb1_movdi_insn): Likewise.
3291 (arm_movsi_insn): Likewise.
3292 (movsi_compare0): Likewise.
3293 (movhi_insn_arch4): Likewise.
3294 (movhi_bytes): Likewise.
3295 (arm_movqi_insn): Likewise.
3296 (thumb1_movqi_insn): Likewise.
3297 (arm32_movhf): Likewise.
3298 (thumb1_movhf): Likewise.
3299 (arm_movsf_soft_insn): Likewise.
3300 (thumb1_movsf_insn): Likewise.
3301 (movdf_soft_insn): Likewise.
3302 (thumb_movdf_insn): Likewise.
3303 (arm_cmpsi_insn): Likewise.
3304 (cmpsi_shiftsi): Likewise.
3305 (cmpsi_shiftsi_swp): Likewise.
3306 (arm_cmpsi_negshiftsi_si): Likewise.
3307 (movsicc_insn): Likewise.
3308 (movsfcc_soft_insn): Likewise.
3309 (arith_shiftsi): Likewise.
3310 (arith_shiftsi_compare0
3311 (arith_shiftsi_compare0_scratch
3312 (sub_shiftsi): Likewise.
3313 (sub_shiftsi_compare0
3314 (sub_shiftsi_compare0_scratch
3315 (and_scc): Likewise.
3316 (cond_move): Likewise.
3317 (if_plus_move): Likewise.
3318 (if_move_plus): Likewise.
3319 (if_move_not): Likewise.
3320 (if_not_move): Likewise.
3321 (if_shift_move): Likewise.
3322 (if_move_shift): Likewise.
3323 (if_shift_shift): Likewise.
3324 (if_not_arith): Likewise.
3325 (if_arith_not): Likewise.
3326 (cond_move_not): Likewise.
3327 (thumb1_ashlsi3): Set type attribute.
3328 (thumb1_ashrsi3): Likewise.
3329 (thumb1_lshrsi3): Likewise.
3330 (thumb1_rotrsi3): Likewise.
3331 (shiftsi3_compare0_scratch): Likewise.
3332 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
3333 (neon_mov<mode>): Likewise.
3334 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
3336 (thumb2_movsi_insn): Likewise.
3337 (thumb2_cmpsi_neg_shiftsi): Likewise.
3338 (thumb2_extendqisi_v6): Likewise.
3339 (thumb2_zero_extendhisi2_v6): Likewise.
3340 (thumb2_zero_extendqisi2_v6): Likewise.
3341 (thumb2_shiftsi3_short): Likewise.
3342 (thumb2_addsi3_compare0_scratch): Likewise.
3343 (orsi_not_shiftsi_si): Likewise.
3344 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3345 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
3347 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3348 (1020alu_shift_op): Likewise.
3349 (1020alu_shift_reg_op): Likewise.
3350 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3351 (alu_shift_op): Likewise.
3352 (alu_shift_reg_op): Likewise.
3353 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3354 (11_alu_shift_op): Likewise.
3355 (11_alu_shift_reg_op): Likewise.
3356 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3357 (9_alu_shift_reg_op): Likewise.
3358 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3360 (cortex_a15_alu_shift): Likewise.
3361 (cortex_a15_alu_shift_reg): Likewise.
3362 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
3364 (cortex_a5_alu_shift): Likewise.
3365 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
3367 (cortex_a53_alu_shift): Likewise.
3368 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3370 (cortex_a7_alu_reg): Likewise.
3371 (cortex_a7_alu_shift): Likewise.
3372 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
3374 (cortex_a8_alu_shift): Likewise.
3375 (cortex_a8_alu_shift_reg): Likewise.
3376 (cortex_a8_mov): Likewise.
3377 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3378 (cortex_a9_dp_shift): Likewise.
3379 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
3381 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
3383 (cortex_r4_mov): Likewise.
3384 (cortex_r4_alu_shift): Likewise.
3385 (cortex_r4_alu_shift_reg): Likewise.
3386 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3387 (526_alu_shift_op): Likewise.
3388 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3389 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3390 (626te_alu_shift_op): Likewise.
3391 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3392 (726te_alu_op): Likewise.
3393 (726te_alu_shift_op): Likewise.
3394 (726te_alu_shift_reg_op): Likewise.
3395 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3396 (mp626_alu_shift_op): Likewise.
3397 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3398 (pj4_alu_e1_conds): Likewise.
3399 (pj4_alu): Likewise.
3400 (pj4_alu_conds): Likewise.
3401 (pj4_shift): Likewise.
3402 (pj4_shift_conds): Likewise.
3403 (pj4_alu_shift): Likewise.
3404 (pj4_alu_shift_conds): Likewise.
3405 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3407 (cortexa7_older_only): Likewise.
3408 (cortexa7_younger): Likewise.
3410 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3412 * ipa-pure-const.c (generate_summary): Rename to...
3413 (pure_const_generate_summary): ... this.
3415 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3417 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3419 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3422 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3425 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3427 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3428 enabled without -march=zEC12.
3429 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3432 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3434 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3435 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3436 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3437 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3438 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3439 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3440 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3442 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3443 and ISA_HAS_NMADD3_NMSUB3.
3444 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3445 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3446 (nmsub4<mode>, nmsub3<mode>): Likewise.
3447 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3449 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3451 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3452 TARGET_MIPS5400 checking.
3454 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3455 Peter Bergner <bergner@vnet.ibm.com>
3457 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3458 registers in the comment.
3459 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3460 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3461 rather than FIRST_PSEUDO_REGISTERS.
3463 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3465 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3466 enable extra ISA flags with TARGET_HTM.
3468 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3470 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3473 2013-07-15 Cong Hou <congh@google.com>
3475 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3476 in compare function for sorting.
3478 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3480 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3481 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3482 * config/rs6000/rs6000.opt: Add -mhtm option.
3483 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3484 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3485 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3486 __HTM__ if the HTM instructions are available.
3487 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3488 htm_spr_reg_operand): New define_predicates.
3489 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3490 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3492 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3493 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3494 HTM builtin functions.
3495 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3496 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3497 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3498 (rs6000_builtin_mask_calculate): Likewise.
3499 (rs6000_option_override_internal): Likewise.
3500 (bdesc_htm): Add new HTM builtin support.
3501 (htm_spr_num): New function.
3502 (htm_spr_regno): Likewise.
3503 (rs6000_htm_spr_icode): Likewise.
3504 (htm_expand_builtin): Likewise.
3505 (htm_init_builtins): Likewise.
3506 (rs6000_expand_builtin): Add support for HTM builtin functions.
3507 (rs6000_init_builtins): Likewise.
3508 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3510 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3511 (TARGET_HTM, MASK_HTM): Define macros.
3512 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3513 (FIXED_REGISTERS): Likewise.
3514 (CALL_USED_REGISTERS): Likewise.
3515 (CALL_REALLY_USED_REGISTERS): Likewise.
3516 (REG_ALLOC_ORDER): Likewise.
3517 (enum reg_class): Likewise.
3518 (REG_CLASS_NAMES): Likewise.
3519 (REG_CLASS_CONTENTS): Likewise.
3520 (REGISTER_NAMES): Likewise.
3521 (ADDITIONAL_REGISTER_NAMES): Likewise.
3522 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3523 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3524 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3525 * config/rs6000/htm.md: New file.
3526 * config/rs6000/htmintrin.h: New file.
3527 * config/rs6000/htmxlintrin.h: New file.
3529 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3531 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3532 Define SYMBOL_TINY_GOT, update comment.
3533 * config/aarch64/aarch64.c
3534 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3535 (aarch64_expand_mov_immediate): Likewise.
3536 (aarch64_print_operand): Likewise.
3537 (aarch64_classify_symbol): Likewise.
3538 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3539 (ldr_got_tiny): Define.
3541 2013-07-13 Tobias Grosser <tobias@grosser.es>
3543 PR tree-optimization/54094
3544 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3545 scheduling dimension for the parallelism check from the polyhedral
3546 information in the AST.
3547 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3550 2013-07-13 Jason Merrill <jason@redhat.com>
3552 * print-tree.c (debug_vec_tree): Use debug_raw.
3553 (debug_raw (vec<tree, va_gc> &)): New.
3554 (debug_raw (vec<tree, va_gc> *)): New.
3555 * tree.h: Declare them.
3557 2013-07-13 Bin Cheng <bin.cheng@arm.com>
3559 * ifcvt.c (ifcvt_after_combine): New static variable.
3560 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3562 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3563 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3564 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3566 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3568 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3570 2013-07-12 Michael Matz <matz@suse.de>
3573 * convert.c (convert_to_real): Reject non-float inner types.
3575 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3577 * config/aarch64/aarch64-protos.h
3578 (aarch64_simd_immediate_valid_for_move): Remove.
3579 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3580 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3582 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3584 2013-07-11 Steve Ellcey <sellcey@mips.com>
3586 * config/mips/mips.c (mips_conditional_register_usage): Do not
3587 use t[0-7] registers in MIPS16 mode when optimizing for size.
3589 2013-07-11 Sriraman Tallam <tmsriram@google.com>
3591 * config/i386/i386.c (dispatch_function_versions): Fix array
3592 indexing of function_version_info to match actual_versions.
3594 2013-07-11 Teresa Johnson <tejohnson@google.com>
3596 * vec.h (struct va_gc): Move release out-of-line.
3597 (va_gc::release): Call ggc_free on released vec.
3599 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3601 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3602 Require GOT register as additional operand in UNSPEC.
3603 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3604 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3605 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3606 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3607 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3608 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3609 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3611 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
3614 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3615 name seen by assembler/linker rather if available.
3617 2013-07-11 Andreas Schwab <schwab@suse.de>
3619 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3621 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3623 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
3625 2013-07-10 Joseph Myers <joseph@codesourcery.com>
3627 * doc/tm.texi.in: Move hook documentation to ....
3628 * target.def: ... here.
3630 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3632 * doc/tm.texi: Regenerate.
3634 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3637 * doc/invoke.texi: Document -Wconditionally-supported.
3639 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3642 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3645 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3648 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3649 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3650 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3651 Remove duplicate devices.
3652 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3653 * config/avr/t-multilib: Regenerate.
3654 * config/avr/avr-tables.opt: Regenerate.
3655 * doc/avr-mmcu.texi: Regenerate.
3657 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3660 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3662 2013-07-10 Graham Stott <graham.stott@btinternet.com>
3664 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3665 the cost of MULT when optimizing for size.
3667 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3669 * config/cr16/cr16-protos.h: Don't include target.h.
3671 2013-07-09 Joseph Myers <joseph@codesourcery.com>
3673 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3674 adjust register size for TDmode and TFmode for VSX registers.
3676 2013-07-08 Kai Tietz <ktietz@redhat.com>
3679 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3680 hook_bool_const_tree_true.
3682 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3684 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3685 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3686 * config/s390/s390.md: Define FPR*_REGNUM constants.
3687 Fix FPR2_REGNUM constant (18 -> 17).
3688 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3689 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3690 ("trunc<BFP:mode><DFP_ALL:mode>2")
3691 ("trunc<DFP_ALL:mode><BFP:mode>2")
3692 ("*extend<BFP:mode><DFP_ALL:mode>2")
3693 ("*extend<DFP_ALL:mode><BFP:mode>2")
3694 ("extend<BFP:mode><DFP_ALL:mode>2")
3695 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3698 2013-07-08 Graham Stott <graham.stott@btinternet.com>
3700 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
3702 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3704 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3705 and cfun_fpr_bit_p to cfun_fpr_save_p.
3706 (s390_frame_area, s390_register_info, s390_frame_info)
3707 (s390_emit_prologue, s390_emit_epilogue)
3708 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3710 * config/s390/s390.h: Define *_REGNUM macros for floating point
3713 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3715 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3717 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3719 PR rtl-optimization/57786
3720 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3721 and break out of the loop when it is set to false.
3723 2013-07-08 Jakub Jelinek <jakub@redhat.com>
3726 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3727 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3728 (const_int 63)) 0)).
3729 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3730 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3731 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3733 PR rtl-optimization/57829
3734 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3735 mask bits outside of mode are just sign-extension from mode to HWI.
3737 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3739 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3740 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3741 adjust_address instead of change_address to keep info about alignment.
3742 (emit_strmov): Remove.
3743 (emit_memmov): New function.
3744 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3745 (expand_movmem_epilogue): Likewise and return updated rtx for
3747 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3748 destination and source.
3749 (decide_alignment): Refactor, handle vector_loop.
3750 (ix86_expand_movmem): Likewise.
3751 (ix86_expand_setmem): Likewise.
3752 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3754 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
3756 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3757 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3759 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3761 * config/i386/sse.md (sse_movlhps): Change alternative 3
3762 of operand 2 to "m".
3764 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3767 * config/i386/sse.md (iptr): New mode attribute.
3768 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3769 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3770 (<sse>_vmmul<mode>3): Ditto.
3771 (<sse>_vmdiv<mode>3): Ditto.
3772 (sse_vmrcpv4sf2): Ditto.
3773 (<sse>_vmsqrt<mode>2): Ditto.
3774 (sse_vmrsqrtv4sf2): Ditto.
3775 (<sse>_vm<code><mode>3): Ditto.
3776 (avx_vmcmp<mode>3): Ditto.
3777 (<sse>_vmmaskcmp<mode>3): Ditto.
3778 (<sse>_comi): Ditto.
3779 (<sse>_ucomi): Ditto.
3780 (*xop_vmfrcz_<mode>): Ditto.
3781 (*fmai_fmadd_<mode>): Ditto.
3782 (*fmai_fmsub_<mode>): Ditto.
3783 (*fmai_fnmadd_<mode>): Ditto.
3784 (*fmai_fnmsub_<mode>): Ditto.
3785 (*fma4i_vmfmadd_<mode>): Ditto.
3786 (*fma4i_vmfmsub_<mode>): Ditto.
3787 (*fma4i_vmfnmadd_<mode>): Ditto.
3788 (*fma4i_vmfnmsub_<mode>): Ditto.
3789 (*xop_vmfrcz_<mode>): Ditto.
3790 (sse_cvtps2pi): Ditto.
3791 (sse_cvttps2pi): Ditto.
3792 (sse_cvtss2si): Ditto.
3793 (sse_cvtss2si_2): Ditto.
3794 (sse_cvtss2siq_2): Ditto.
3795 (sse_cvttss2si): Ditto.
3796 (sse_cvttss2siq): Ditto.
3797 (sse_cvtsd2si): Ditto.
3798 (sse_cvtsd2si_2): Ditto.
3799 (sse_cvtsd2siq_2): Ditto.
3800 (sse_cvttsd2si): Ditto.
3801 (sse_cvttsd2siq): Ditto.
3802 (sse_cvtsd2ss): Ditto.
3803 (sse_cvtss2sd): Ditto.
3804 (avx2_pbroadcast<mode>): Ditto.
3805 (avx2_pbroadcast<mode>_1): Ditto.
3806 (*avx_vperm_broadcast_v4sf): Ditto.
3808 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3809 (sse_movlhps): Ditto.
3810 (sse_storehps): Ditto.
3811 (sse_loadhps): Ditto.
3812 (sse_storelps): Ditto.
3813 (sse_loadlps): Ditto.
3814 (*vec_concatv4sf): Ditto.
3815 (*vec_interleave_highv2df): Ditto.
3816 (*vec_interleave_lowv2df): Ditto.
3817 (*vec_extractv2df_1_sse): Ditto.
3818 (*vec_extractv2df_0_sse): Ditto.
3819 (sse2_storelpd): Ditto.
3820 (sse2_loadlpd): Ditto.
3821 (sse2_movsd): Ditto.
3822 (*vec_concatv4si): Ditto.
3823 (vec_concatv2di): Ditto.
3825 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3826 for Intel asm dialect.
3827 (mmx_punpcklwd): Ditto.
3828 (mmx_punpckldq): Ditto.
3830 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3831 for intel assembler dialect.
3833 2013-07-06 Jakub Jelinek <jakub@redhat.com>
3836 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3837 for BUILT_IN_C{LZ,LRSB}*.
3838 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3839 * tree-vrp.c (extract_range_basic): Handle
3840 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3841 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3842 fall thru to code calling set_value*.
3843 * builtins.c (expand_builtin): Remove *IMAX cases.
3844 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3845 if width is bigger than 2*HWI.
3847 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3849 PR rtl-optimization/55342
3850 * lra-int.h (lra_subreg_reload_pseudos): New.
3851 * lra.c: Add undoing optional reloads to the block diagram.
3852 (lra_subreg_reload_pseudos): New.
3853 (lra_optional_reload_pseudos): Change comments.
3854 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3855 lra_optional_reload_pseudos after undo transformations.
3856 * lra-assigns.c (pseudo_prefix_title): New.
3857 (lra_setup_reg_renumber): Use it.
3858 (spill_for): Ditto. Check subreg reload pseudos too.
3859 (assign_by_spills): Consider subreg reload pseudos too.
3860 * lra-constraints.c (simplify_operand_subreg): Use
3861 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3862 (curr_insn_transform): Recognize and do optional reloads.
3863 (undo_optional_reloads): New.
3864 (lra_undo_inheritance): Call undo_optional_reloads.
3866 2013-07-05 Thomas Quinot <quinot@adacore.com>
3868 * tree-complex.c (expand_complex_operations_1): Fix typo.
3870 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3872 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3873 (tune_params): New member 'const vec_costs'.
3874 * config/aarch64/aarch64.c (generic_vector_cost): New.
3875 (generic_tunings): New member 'generic_vector_cost'.
3876 (aarch64_builtin_vectorization_cost): New.
3877 (aarch64_add_stmt_cost): New.
3878 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3879 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3881 2013-07-03 Jakub Jelinek <jakub@redhat.com>
3884 * config/i386/predicates.md (vsib_address_operand): Disallow
3885 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3887 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3890 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3891 expand_builtin_setjmp_receiver.
3892 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3893 with NULL for the label parameter.
3894 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3895 the frame-pointer. Adjust comments.
3896 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3897 only if LABEL is non-NULL.
3899 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3901 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3902 (ARM_ABI_AAPCS64): Ditto.
3904 (ARM_DEFAULT_ABI): Ditto.
3906 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3908 * config/aarch64/aarch64-builtins.c
3909 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3910 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3912 * config/aarch64/aarch64-simd.md
3913 (aarch64_ld1<VALL:mode>): New.
3914 (aarch64_st1<VALL:mode>): Likewise.
3915 * config/aarch64/arm_neon.h
3916 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3918 2013-07-02 Sriraman Tallam <tmsriram@google.com>
3920 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3922 (ix86_option_override_internal):Turn on all -mavx target flags by
3923 default as they are dependent on AVX anyway.
3925 2013-07-02 Cary Coutant <ccoutant@google.com>
3927 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3929 (loc_checksum_ordered): Likewise.
3930 (hash_loc_operands): Remove inline keyword.
3932 2013-07-02 Jakub Jelinek <jakub@redhat.com>
3934 PR tree-optimization/57741
3935 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3936 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3937 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3938 Allow REAL_CST step_exprs if flag_associative_math.
3939 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3941 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3943 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
3945 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3947 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3949 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3951 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3953 (iorsi3_insn): Likewise.
3954 (arm_xorsi3): Likewise.
3956 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3958 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3960 (attribute "type"): ... here, and prefix with "wmmx_".
3961 (attribute "core_cycles"): Update for attribute changes.
3962 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3963 (tbcstv4hi): Likewise.
3964 (tbcstv2si): Likewise.
3965 (iwmmxt_iordi3): Likewise.
3966 (iwmmxt_xordi3): Likewise.
3967 (iwmmxt_anddi3): Likewise.
3968 (iwmmxt_nanddi3): Likewise.
3969 (iwmmxt_arm_movdi): Likewise.
3970 (iwmmxt_movsi_insn): Likewise.
3971 (mov<mode>_internal): Likewise.
3972 (and<mode>3_iwmmxt): Likewise.
3973 (ior<mode>3_iwmmxt): Likewise.
3974 (xor<mode>3_iwmmxt): Likewise.
3975 (add<mode>3_iwmmxt): Likewise.
3976 (ssaddv8qi3): Likewise.
3977 (ssaddv4hi3): Likewise.
3978 (ssaddv2si3): Likewise.
3979 (usaddv8qi3): Likewise.
3980 (usaddv4hi3): Likewise.
3981 (usaddv2si3): Likewise.
3982 (sub<mode>3_iwmmxt): Likewise.
3983 (sssubv8qi3): Likewise.
3984 (sssubv4hi3): Likewise.
3985 (sssubv2si3): Likewise.
3986 (ussubv8qi3): Likewise.
3987 (ussubv4hi3): Likewise.
3988 (ussubv2si3): Likewise.
3989 (mulv4hi3_iwmmxt): Likewise.
3990 (smulv4hi3_highpart): Likewise.
3991 (umulv4hi3_highpart): Likewise.
3992 (iwmmxt_wmacs): Likewise.
3993 (iwmmxt_wmacsz): Likewise.
3994 (iwmmxt_wmacu): Likewise.
3995 (iwmmxt_wmacuz): Likewise.
3996 (iwmmxt_clrdi): Likewise.
3997 (iwmmxt_clrv8qi): Likewise.
3998 (iwmmxt_clr4hi): Likewise.
3999 (iwmmxt_clr2si): Likewise.
4000 (iwmmxt_uavgrndv8qi3): Likewise.
4001 (iwmmxt_uavgrndv4hi3): Likewise.
4002 (iwmmxt_uavgv8qi3): Likewise.
4003 (iwmmxt_uavgv4hi3): Likewise.
4004 (iwmmxt_tinsrb): Likewise.
4005 (iwmmxt_tinsrh): Likewise.
4006 (iwmmxt_tinsrw): Likewise.
4007 (iwmmxt_textrmub): Likewise.
4008 (iwmmxt_textrmsb): Likewise.
4009 (iwmmxt_textrmuh): Likewise.
4010 (iwmmxt_textrmsh): Likewise.
4011 (iwmmxt_textrmw): Likewise.
4012 (iwmxxt_wshufh): Likewise.
4013 (eqv8qi3): Likewise.
4014 (eqv4hi3): Likewise.
4015 (eqv2si3): Likewise.
4016 (gtuv8qi3): Likewise.
4017 (gtuv4hi3): Likewise.
4018 (gtuv2si3): Likewise.
4019 (gtv8qi3): Likewise.
4020 (gtv4hi3): Likewise.
4021 (gtv2si3): Likewise.
4022 (smax<mode>3_iwmmxt): Likewise.
4023 (umax<mode>3_iwmmxt): Likewise.
4024 (smin<mode>3_iwmmxt): Likewise.
4025 (umin<mode>3_iwmmxt): Likewise.
4026 (iwmmxt_wpackhss): Likewise.
4027 (iwmmxt_wpackwss): Likewise.
4028 (iwmmxt_wpackdss): Likewise.
4029 (iwmmxt_wpackhus): Likewise.
4030 (iwmmxt_wpackwus): Likewise.
4031 (iwmmxt_wpackdus): Likewise.
4032 (iwmmxt_wunpckihb): Likewise.
4033 (iwmmxt_wunpckihh): Likewise.
4034 (iwmmxt_wunpckihw): Likewise.
4035 (iwmmxt_wunpckilb): Likewise.
4036 (iwmmxt_wunpckilh): Likewise.
4037 (iwmmxt_wunpckilw): Likewise.
4038 (iwmmxt_wunpckehub): Likewise.
4039 (iwmmxt_wunpckehuh): Likewise.
4040 (iwmmxt_wunpckehuw): Likewise.
4041 (iwmmxt_wunpckehsb): Likewise.
4042 (iwmmxt_wunpckehsh): Likewise.
4043 (iwmmxt_wunpckehsw): Likewise.
4044 (iwmmxt_wunpckelub): Likewise.
4045 (iwmmxt_wunpckeluh): Likewise.
4046 (iwmmxt_wunpckeluw): Likewise.
4047 (iwmmxt_wunpckelsb): Likewise.
4048 (iwmmxt_wunpckelsh): Likewise.
4049 (iwmmxt_wunpckelsw): Likewise.
4050 (ror<mode>3): Likewise.
4051 (ashr<mode>3_iwmmxt): Likewise.
4052 (lshr<mode>3_iwmmxt): Likewise.
4053 (ashl<mode>3_iwmmxt): Likewise.
4054 (ror<mode>3_di): Likewise.
4055 (ashr<mode>3_di): Likewise.
4056 (lshr<mode>3_di): Likewise.
4057 (ashl<mode>3_di): Likewise.
4058 (iwmmxt_wmadds): Likewise.
4059 (iwmmxt_wmaddu): Likewise.
4060 (iwmmxt_tmia): Likewise.
4061 (iwmmxt_tmiaph): Likewise.
4062 (iwmmxt_tmiabb): Likewise.
4063 (iwmmxt_tmiatb): Likewise.
4064 (iwmmxt_tmiabt): Likewise.
4065 (iwmmxt_tmiatt): Likewise.
4066 (iwmmxt_tmovmskb): Likewise.
4067 (iwmmxt_tmovmskh): Likewise.
4068 (iwmmxt_tmovmskw): Likewise.
4069 (iwmmxt_waccb): Likewise.
4070 (iwmmxt_wacch): Likewise.
4071 (iwmmxt_waccw): Likewise.
4072 (iwmmxt_waligni): Likewise.
4073 (iwmmxt_walignr): Likewise.
4074 (iwmmxt_walignr0): Likewise.
4075 (iwmmxt_walignr1): Likewise.
4076 (iwmmxt_walignr2): Likewise.
4077 (iwmmxt_walignr3): Likewise.
4078 (iwmmxt_wsadb): Likewise.
4079 (iwmmxt_wsadh): Likewise.
4080 (iwmmxt_wsadbz): Likewise.
4081 (iwmmxt_wsadhz): Likewise.
4082 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4083 (iwmmxt_wabsdiffb): Likewise.
4084 (iwmmxt_wabsdiffh): Likewise.
4085 (iwmmxt_wabsdiffw): Likewise.
4086 (iwmmxt_waddsubhx): Likewise
4087 (iwmmxt_wsubaddhx): Likewise.
4088 (addc<mode>3): Likewise.
4089 (iwmmxt_avg4): Likewise.
4090 (iwmmxt_avg4r): Likewise.
4091 (iwmmxt_wmaddsx): Likewise.
4092 (iwmmxt_wmaddux): Likewise.
4093 (iwmmxt_wmaddsn): Likewise.
4094 (iwmmxt_wmaddun): Likewise.
4095 (iwmmxt_wmulwsm): Likewise.
4096 (iwmmxt_wmulwum): Likewise.
4097 (iwmmxt_wmulsmr): Likewise.
4098 (iwmmxt_wmulumr): Likewise.
4099 (iwmmxt_wmulwsmr): Likewise.
4100 (iwmmxt_wmulwumr): Likewise.
4101 (iwmmxt_wmulwl): Likewise.
4102 (iwmmxt_wqmulm): Likewise.
4103 (iwmmxt_wqmulwm): Likewise.
4104 (iwmmxt_wqmulmr): Likewise.
4105 (iwmmxt_wqmulwmr): Likewise.
4106 (iwmmxt_waddbhusm): Likewise.
4107 (iwmmxt_waddbhusl): Likewise.
4108 (iwmmxt_wqmiabb): Likewise.
4109 (iwmmxt_wqmiabt): Likewise.
4110 (iwmmxt_wqmiatb): Likewise.
4111 (iwmmxt_wqmiatt): Likewise.
4112 (iwmmxt_wqmiabbn): Likewise.
4113 (iwmmxt_wqmiabtn): Likewise.
4114 (iwmmxt_wqmiatbn): Likewise.
4115 (iwmmxt_wqmiattn): Likewise.
4116 (iwmmxt_wmiabb): Likewise.
4117 (iwmmxt_wmiabt): Likewise.
4118 (iwmmxt_wmiatb): Likewise.
4119 (iwmmxt_wmiatt): Likewise.
4120 (iwmmxt_wmiabbn): Likewise.
4121 (iwmmxt_wmiabtn): Likewise.
4122 (iwmmxt_wmiatbn): Likewise.
4123 (iwmmxt_wmiattn): Likewise.
4124 (iwmmxt_wmiawbb): Likewise.
4125 (iwmmxt_wmiawbt): Likewise.
4126 (iwmmxt_wmiawtb): Likewise.
4127 (iwmmxt_wmiawtt): Likewise.
4128 (iwmmxt_wmiawbbn): Likewise.
4129 (iwmmxt_wmiawbtn): Likewise.
4130 (iwmmxt_wmiawtbn): Likewise.
4131 (iwmmxt_wmiawttn): Likewise.
4132 (iwmmxt_wmerge): Likewise.
4133 (iwmmxt_tandc<mode>3): Likewise.
4134 (iwmmxt_torc<mode>3): Likewise.
4135 (iwmmxt_torvsc<mode>3): Likewise.
4136 (iwmmxt_textrc<mode>3): Likewise.
4137 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
4138 (wmmxt_pack): Likewise.
4139 (wmmxt_mult_c1): Likewise.
4140 (wmmxt_mult_c2): Likewise.
4141 (wmmxt_alu_c1): Likewise.
4142 (wmmxt_alu_c2): Likewise.
4143 (wmmxt_alu_c3): Likewise.
4144 (wmmxt_transfer_c1): Likewise.
4145 (wmmxt_transfer_c2): Likewise.
4146 (wmmxt_transfer_c3): Likewise.
4147 (marvell_f_iwmmxt_wstr): Likewise.
4148 (marvell_f_iwmmxt_wldr): Likewise.
4150 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
4152 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
4154 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4157 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4158 * lra-constraints.c (need_for_split_p): Check call used hard regs
4159 living through calls.
4161 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
4162 call used regs for call insn.
4164 2013-06-28 Jakub Jelinek <jakub@redhat.com>
4167 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
4168 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
4170 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
4172 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
4174 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4176 * lra-constraints.c (need_for_split_p): Check call used hard regs
4177 living through calls.
4179 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
4182 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
4183 to tie with any other modes. Eliminate Altivec vector mode tests,
4184 since these are a subset of ALTIVEC or VSX vector modes. Simplify
4185 code, to return 0 if testing MODE2 for a condition, if we've
4186 already tested MODE1 for the same condition.
4188 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4190 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
4193 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4195 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4196 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
4198 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4200 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
4202 (aarch64_symbolic_constant_p): Remove.
4203 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
4204 static. Fix line length and white space.
4205 (aarch64_symbolic_constant_p): Remove.
4206 * config/aarch64/predicates.md (aarch64_valid_symref):
4207 Use aarch64_classify_symbol_expression.
4209 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4211 * config/arm/constraints.md (Ts): New constraint.
4212 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
4214 (compare_scc): Use "Ts" constraint for operand 0.
4215 (ior_scc_scc): Likewise.
4216 (and_scc_scc): Likewise.
4217 (and_scc_scc_nodom): Likewise.
4218 (ior_scc_scc_cmp): Likewise for operand 7.
4219 (and_scc_scc_cmp): Likewise.
4220 * config/arm/thumb2.md (thumb2_movsi_insn):
4221 Add alternatives for 16-bit encodings.
4222 (thumb2_movhi_insn): Likewise.
4223 (thumb2_movsicc_insn): Likewise.
4224 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
4225 (thumb2_negscc): Use "Ts" constraint.
4226 Move mvn instruction outside cond_exec block.
4227 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
4228 for 16-bit encodings.
4230 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4232 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
4234 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
4235 (mulsi3subsi): Likewise.
4236 (mulsidi3adddi): Likewise.
4237 (mulsidi3_v6): Likewise.
4238 (umulsidi3_v6): Likewise.
4239 (umulsidi3adddi_v6): Likewise.
4240 (smulsi3_highpart_v6): Likewise.
4241 (umulsi3_highpart_v6): Likewise.
4242 (mulhisi3tb): Likewise.
4243 (mulhisi3bt): Likewise.
4244 (mulhisi3tt): Likewise.
4245 (maddhisi4): Likewise.
4246 (maddhisi4tb): Likewise.
4247 (maddhisi4tt): Likewise.
4248 (maddhidi4): Likewise.
4249 (maddhidi4tb): Likewise.
4250 (maddhidi4tt): Likewise.
4251 (zeroextractsi_compare0_scratch): Likewise.
4252 (insv_zero): Likewise.
4253 (insv_t2): Likewise.
4254 (anddi_notzesidi_di): Likewise.
4255 (anddi_notsesidi_di): Likewise.
4256 (andsi_notsi_si): Likewise.
4257 (iordi_zesidi_di): Likewise.
4258 (xordi_zesidi_di): Likewise.
4259 (andsi_iorsi3_notsi): Likewise.
4261 (smax_m1): Likewise.
4263 (not_shiftsi): Likewise.
4264 (unaligned_loadsi): Likewise.
4265 (unaligned_loadhis): Likewise.
4266 (unaligned_loadhiu): Likewise.
4267 (unaligned_storesi): Likewise.
4268 (unaligned_storehi): Likewise.
4269 (extv_reg): Likewise.
4270 (extzv_t2): Likewise.
4272 (udivsi3): Likewise.
4273 (arm_zero_extendhisi2addsi): Likewise.
4274 (arm_zero_extendqisi2addsi): Likewise.
4275 (compareqi_eq0): Likewise.
4276 (arm_extendhisi2_v6): Likewise.
4277 (arm_extendqisi2addsi): Likewise.
4278 (arm_movt): Likewise.
4279 (thumb2_ldrd): Likewise.
4280 (thumb2_ldrd_base): Likewise.
4281 (thumb2_ldrd_base_neg): Likewise.
4282 (thumb2_strd): Likewise.
4283 (thumb2_strd_base): Likewise.
4284 (thumb2_strd_base_neg): Likewise.
4285 (arm_negsi2): Add alternative for 16-bit encoding.
4286 (arm_one_cmplsi2): Likewise.
4288 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4290 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
4291 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
4292 (movdfcc): Likewise.
4293 * config/arm/vfp.md (*thumb2_movsf_vfp):
4294 Disable predication for arm_restrict_it.
4295 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
4296 (*thumb2_movdfcc_vfp): Likewise.
4297 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
4298 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
4299 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
4300 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
4301 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
4302 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
4303 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
4304 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
4305 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
4306 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
4307 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
4308 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
4309 Disable predication for arm_restrict_it.
4311 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
4313 * config/i386/bmiintrin.h (_bextr_u32): New.
4314 (_bextr_u64): Ditto.
4316 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
4318 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
4319 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
4320 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
4321 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
4322 * config/mips/n32-elf.h: ...this new file.
4324 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
4327 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
4328 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
4330 2013-06-27 Catherine Moore <clm@codesourcery.com>
4332 * config/mips/mips-tables.opt: Regenerate.
4333 * config/mips/mips-cpus.def: Add m14ke and m14kec.
4334 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
4335 * doc/invoke.texi: Add -m14kc.
4337 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4340 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
4341 constraints of operand 1 and 2.
4344 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
4345 to match RTL canonicalization. Swap predicates and
4346 constraints of operand 1 and 2.
4348 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
4350 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
4351 Process OP_INOUT regs for splitting too.
4353 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4355 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
4356 decl before the loop, initialize to NULL.
4357 (vectorizable_load): Initialize ptr_incr to NULL.
4359 2013-06-27 Martin Jambor <mjambor@suse.cz>
4362 * ipa-ref.h (ipa_maybe_record_reference): Declare.
4363 * ipa-ref.c (ipa_maybe_record_reference): New function.
4364 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
4365 * ipa-cp.c (create_specialized_node): Record potential references from
4367 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
4369 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4371 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
4372 parameter 'mode' of type 'enum machine_mode mode'; change to pass
4373 'mode' to force_reg.
4374 (aarch64_add_offset): Update calls to aarch64_force_temporary.
4375 (aarch64_expand_mov_immediate): Likewise.
4377 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4379 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
4380 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
4382 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4384 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
4385 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4386 (struct machine_function): Add tbegin_p.
4387 (s390_canonicalize_comparison): Fold CC mode compares to
4388 conditional jump if possible.
4389 (s390_emit_jump): Return the emitted jump.
4390 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4391 Handle CCRAWmode compares.
4392 (s390_option_override): Default to -mhtm if available.
4393 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4394 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4395 FPRs instead of df_regs_ever_live_p.
4396 (s390_optimize_nonescaping_tx): New function.
4397 (s390_init_frame_layout): Extend clobbered_regs array to cover
4399 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4400 (s390_expand_tbegin): New function.
4401 (enum s390_builtin): New enum definition.
4402 (code_for_builtin): New array definition.
4403 (s390_init_builtins): New function.
4404 (s390_expand_builtin): New function.
4405 (TARGET_INIT_BUILTINS): Define.
4406 (TARGET_EXPAND_BUILTIN): Define.
4407 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4408 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4409 (s390_alc_comparison): Likewise.
4410 * config/s390/s390-modes.def: Add CCRAWmode.
4411 * config/s390/s390.h (processor_flags): Add PF_TX.
4412 (TARGET_CPU_HTM): Define macro.
4413 (TARGET_HTM): Define macro.
4414 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4415 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4416 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4417 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4418 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4419 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4420 ("*cc_to_int"): Move up.
4421 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4422 constants other than 0.
4423 ("*ccraw_to_int"): New insn and splitter definition.
4424 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4425 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4426 ("tx_assist"): New expander.
4427 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4428 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4429 * config/s390/s390.opt: Add -mhtm option.
4430 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4431 * config/s390/htmxlintrin.h: New file.
4432 * config/s390/htmintrin.h: New file.
4433 * config/s390/s390intrin.h: New file.
4434 * doc/extend.texi: Document htm builtins.
4435 * config.gcc: Add the new header files to extra_headers.
4437 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4439 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4440 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4442 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4443 Pat Haugen <pthaugen@us.ibm.com>
4444 Peter Bergner <bergner@vnet.ibm.com>
4446 * config/rs6000/power8.md: New.
4447 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4448 setting for power8 entry.
4449 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4450 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4451 test for Power4/Power5 only.
4452 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4454 (force_new_group): Adjust comment.
4455 * config/rs6000/rs6000.md: Include power8.md.
4457 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4459 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4460 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4462 (tune_params): Update comment.
4463 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4464 (arm_max_conditional_execute): New function.
4465 (thumb2_final_prescan_insn): Use max_insn_skipped and
4466 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4468 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4470 PR tree-optimization/57705
4471 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4472 SSA_NAME step, provided that it is not defined inside the loop.
4473 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4474 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4476 2013-06-25 Martin Jambor <mjambor@suse.cz>
4479 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4480 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4482 (ipa_note_param_call): Initialize member_ptr flag.
4483 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4484 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4485 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4486 (ipa_read_indirect_edge_info): Likewise.
4488 2013-06-25 Richard Biener <rguenther@suse.de>
4491 * passes.c (init_optimization_passes): Move pass_fold_builtins
4492 and pass_dce earlier with -Og.
4494 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4496 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4497 <BIT_FIELD_REF>: Remove trailing TAB.
4498 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4501 2013-06-24 Martin Jambor <mjambor@suse.cz>
4503 PR tree-optimization/57358
4504 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4505 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4506 (ipa_analyze_params_uses): Generate pessimistic info when true.
4508 2013-06-24 Martin Jambor <mjambor@suse.cz>
4510 PR tree-optimization/57539
4511 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4512 global.inlined_to of the new node to it. All callers changed.
4513 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4514 inlining_into, pass it to cgraph_clone_node.
4515 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4516 ipa_free_edge_args_substructures.
4517 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4518 rdesc linked list. Do not assert rdesc edges have inlined caller.
4519 Assert we have found an rdesc in the rdesc list.
4521 2013-06-24 Richard Biener <rguenther@suse.de>
4523 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4524 (pointer_set_lookup): Declare.
4525 (class pointer_map): New template class implementing a
4526 generic pointer to T map.
4527 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4528 pointer_map<T>::contains, pointer_map<T>::insert,
4529 pointer_map<T>::traverse): New functions.
4530 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4531 (pointer_set_lookup): New function.
4532 (pointer_set_contains): Use pointer_set_lookup.
4533 (pointer_set_insert): Likewise.
4534 (insert_aux): Remove.
4535 (struct pointer_map_t): Embed a pointer_set_t.
4536 (pointer_map_create): Adjust.
4537 (pointer_map_destroy): Likewise.
4538 (pointer_map_contains): Likewise.
4539 (pointer_map_insert): Likewise.
4540 (pointer_map_traverse): Likewise.
4541 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4542 pointer_map<unsigned> instead of a pointer_map_t.
4543 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4544 (streamer_tree_cache_lookup): Likewise.
4545 (streamer_tree_cache_create): Likewise.
4546 (streamer_tree_cache_delete): Likewise.
4547 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4548 pointer_map<unsigned> instead of a pointer_map_t.
4549 (lto_init_tree_ref_encoder): Adjust.
4550 (lto_destroy_tree_ref_encoder): Likewise.
4551 * lto-section-out.c (lto_output_decl_index): Likewise.
4552 (lto_record_function_out_decl_state): Likewise.
4553 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4555 2013-06-24 Richard Biener <rguenther@suse.de>
4557 PR tree-optimization/57488
4558 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4560 2013-06-24 Alan Modra <amodra@gmail.com>
4562 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4563 (gen_easy_altivec_constant): Likewise.
4564 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4565 easy_vector_constant_msb): Likewise.
4567 2013-06-23 Jakub Jelinek <jakub@redhat.com>
4570 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4571 add missing return true.
4573 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4576 * config/sh/predicates.md (general_extend_operand): Invoke
4577 general_movsrc_operand for memory operands.
4578 (general_movsrc_operand): Allow reg+reg addressing, do not use
4579 general_operand for memory operands.
4581 2013-06-23 Sriraman Tallam <tmsriram@google.com>
4583 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4584 when current target options does not apply.
4585 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4586 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
4587 * config/i386/bmiintrin.h: Pass appropriate target
4588 attributes to header.
4589 * config/i386/mmintrin.h: Ditto.
4590 * config/i386/nmmintrin.h: Ditto.
4591 * config/i386/avx2intrin.h: Ditto.
4592 * config/i386/fxsrintrin.h: Ditto.
4593 * config/i386/tbmintrin.h: Ditto.
4594 * config/i386/xsaveintrin.h: Ditto.
4595 * config/i386/f16cintrin.h: Ditto.
4596 * config/i386/xtestintrin.h: Ditto.
4597 * config/i386/xsaveoptintrin.h: Ditto.
4598 * config/i386/bmi2intrin.h: Ditto.
4599 * config/i386/lzcntintrin.h: Ditto.
4600 * config/i386/smmintrin.h: Ditto.
4601 * config/i386/wmmintrin.h: Ditto.
4602 * config/i386/x86intrin.h: Remove all header include guards.
4603 * config/i386/prfchwintrin.h: Ditto.
4604 * config/i386/pmmintrin.h: Ditto.
4605 * config/i386/tmmintrin.h: Ditto.
4606 * config/i386/xmmintrin.h: Ditto.
4607 * config/i386/popcntintrin.h: Ditto.
4608 * config/i386/rdseedintrin.h: Ditto.
4609 * config/i386/ammintrin.h: Ditto.
4610 * config/i386/emmintrin.h: Ditto.
4611 * config/i386/immintrin.h: Remove all header include guards.
4612 * config/i386/fma4intrin.h: Ditto.
4613 * config/i386/lwpintrin.h: Ditto.
4614 * config/i386/xopintrin.h: Ditto.
4615 * config/i386/ia32intrin.h: Ditto.
4616 * config/i386/avxintrin.h: Ditto.
4617 * config/i386/rtmintrin.h: Ditto.
4618 * config/i386/fmaintrin.h: Ditto.
4619 * config/i386/mm3dnow.h: Ditto.
4621 2013-06-22 Sriraman Tallam <tmsriram@google.com>
4623 * common/config/i386/i386-common.c: Handle LZCNT.
4625 2013-06-22 Andi Kleen <ak@linux.intel.com>
4627 * doc/extend.texi: Use __atomic_store_n instead of
4628 __atomic_store in HLE example.
4630 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4632 * config/sh/sh.c: Remove <cstdlib> workaround.
4634 2013-06-21 Andi Kleen <ak@linux.intel.com>
4636 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
4638 2013-06-21 Andi Kleen <ak@linux.intel.com>
4640 * doc/extend.texi: Document that __atomic_clear and
4641 __atomic_test_and_set should only be used with bool.
4643 2013-06-20 Jan Hubicka <jh@suse.cz>
4645 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4647 * tree.c (decls_same_for_odr): New function.
4648 (same_for_edr): New function.
4649 (types_same_for_odr): New function.
4650 (get_binfo_at_offset): Use it.
4651 * tree.h (types_same_for_odr): Declare.
4653 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
4654 Jason Merrill <jason@redhat.com>
4656 * system.h: Include <cstdlib> as well as <stdlib.h>.
4658 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
4661 * config/i386/i386.c (construct_container): Report error if
4662 long double is used with disabled x87 float returns.
4664 2013-06-20 Jan Hubicka <jh@suse.cz>
4666 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4668 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4670 PR rtl-optimization/57425
4671 PR rtl-optimization/57569
4672 * alias.c (write_dependence_p): Remove parameters mem_mode and
4673 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4674 Changed all callers.
4675 (canon_anti_dependence): Get comments and semantics in sync.
4676 Add parameter mem_canonicalized. Changed all callers.
4677 * rtl.h (canon_anti_dependence): Update prototype.
4679 2013-06-20 Richard Biener <rguenther@suse.de>
4681 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4682 byte case, inline streamer_read_uchar and defer section
4685 2013-06-20 Richard Biener <rguenther@suse.de>
4687 PR tree-optimization/57584
4688 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4689 SSA names into the expanded expression that take part in
4690 abnormal coalescing.
4692 2013-06-19 Sharad Singhai <singhai@google.com>
4694 * gcov.c (print_usage): Handle new option.
4695 (process_args): Ditto.
4696 (get_gcov_intermediate_filename): New function.
4697 (output_intermediate_file): New function.
4698 (output_gcov_file): New function
4699 (generate_results): Handle new option.
4700 (release_function): Relase demangled name.
4701 (read_graph_file): Handle demangled name.
4702 (output_lines): Ditto.
4703 * doc/gcov.texi: Document gcov intermediate format.
4705 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4708 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4709 (lra_emit_add): Use the functions. Add comment about Y as an
4712 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
4715 * collect2.c (collect_atexit): New.
4716 (collect_exit): Delete.
4717 (main): Register collect_atexit with atexit.
4718 (collect_wait): Change collect_exit to exit.
4720 * collect2.h (collect_exit): Delete.
4721 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4723 2013-06-19 Wei Mi <wmi@google.com>
4725 PR rtl-optimization/57518
4726 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4727 if regno is used in paradoxical subreg.
4728 (update_equiv_regs): Check pdx_subregs[regno] before
4729 set a reg to be equivalent with a mem.
4731 2013-06-19 Matthias Klose <doko@ubuntu.com>
4734 * file-find.h (find_a_file): Add a mode parameter.
4735 * file-find.c (find_a_file): Likewise.
4736 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4737 with X_OK for the executables.
4738 * collect2.c (main): Call find_a_file with X_OK.
4740 2013-06-19 Steve Ellcey <sellcey@mips.com>
4743 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4744 Use NEXT_INSN instead of next_real_insn.
4746 2013-06-19 Jan Hubicka <jh@suse.cz>
4748 * cgraph.h (const_value_known_p): Replace by ...
4749 (ctor_for_folding): .. this one.
4750 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4751 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4752 * expr.c (expand_expr_real_1): Likewise.
4753 (string_constant): Likewise.
4754 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4755 * ipa.c (process_references): Likewise.
4756 (symtab_remove_unreachable_nodes): Likewise.
4757 * ipa-inline-analysis.c (param_change_prob): Likewise.
4758 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4759 (get_base_constructor): Likwise.
4760 * varpool.c (varpool_remove_node): Likewise.
4761 (varpool_remove_initializer): LIkewise.
4762 (dump_varpool_node): LIkwise.
4763 (const_value_known_p): Rewrite to ...
4764 (ctor_for_folding): ... this one.
4766 2013-06-19 Jakub Jelinek <jakub@redhat.com>
4769 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4770 PERSONALITY in $PATH derived prefixes.
4772 2013-06-19 Jeff Law <law@redhat.com>
4774 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4777 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4778 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4781 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4783 * config/arm/vfp.md: Move VFP instruction classification documentation
4785 * config/arm/arm.md: ... here. Update instruction classification
4788 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
4790 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4791 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4792 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4794 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4797 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4798 with NEXT_INSN. Use tablejump_p to check for jump table data
4801 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4804 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4805 that now in C++ the value is correct per the C++ standards.
4807 2013-06-19 Richard Biener <rguenther@suse.de>
4809 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4812 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4815 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4818 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4819 with next_active_insn.
4821 2013-06-18 Sriraman Tallam <tmsriram@google.com>
4823 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4824 functions are inlined during failures to flag an error.
4825 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4826 in early inline pass.
4828 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4830 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4833 2013-06-18 Julian Brown <julian@codesourcery.com>
4835 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4836 Permit virtual register pre-reload if !strict.
4837 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4839 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4841 * config/arm/neon.md (movmisalign<mode>): Use
4842 neon_perm_struct_or_reg_operand instead of
4843 neon_struct_or_register_operand.
4844 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4845 neon_permissive_struct_operand instead of neon_struct_operand.
4846 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4847 neon_vector_mem_operand.
4848 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4849 neon_vector_mem_operand.
4850 (neon_permissive_struct_operand): New.
4851 (neon_struct_or_register_operand): Rename to...
4852 (neon_perm_struct_or_reg_operand): This. Adjust call to
4853 neon_vector_mem_operand.
4855 2013-06-18 Richard Biener <rguenther@suse.de>
4857 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4858 * lto-streamer.h: Include pointer-set.h.
4859 (struct lto_decl_slot): Remove.
4860 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4861 Remove next_index entry.
4862 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4863 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4864 (lto_init_tree_ref_encoder): Adjust.
4865 (lto_destroy_tree_ref_encoder): Likewise.
4866 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4867 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4868 (lto_output_decl_index): Adjust.
4869 (lto_new_out_decl_state): Likewise.
4870 (lto_record_function_out_decl_state): Likewise.
4871 * lto-streamer-out.c (copy_function): Likewise.
4873 2013-06-18 Richard Biener <rguenther@suse.de>
4875 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4876 * cgraphunit.c: Include cfgloop.h.
4877 (init_lowered_empty_function): Initialize the loop tree.
4878 (assemble_thunk): Insert new BBs into loops.
4880 2013-06-18 Richard Biener <rguenther@suse.de>
4882 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4883 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4884 the map from cache entry to cache index optional.
4885 (streamer_tree_cache_replace_tree): Adjust accordingly.
4886 (streamer_tree_cache_append): Likewise.
4887 (streamer_tree_cache_delete): Likewise.
4888 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4889 streamer cache map from cache entry to cache index.
4890 * lto-streamer-out.c (create_output_block): Adjust.
4892 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4894 * config/arm/arm.md (attribute "insn"): Move multiplication and
4895 division attributes to...
4896 (attribute "type"): ... here. Remove mult.
4897 (attribute "mul32"): New attribute.
4898 (attribute "mul64"): Add umaal.
4899 (*arm_mulsi3): Update attributes.
4900 (*arm_mulsi3_v6): Likewise.
4901 (*thumb_mulsi3): Likewise.
4902 (*thumb_mulsi3_v6): Likewise.
4903 (*mulsi3_compare0): Likewise.
4904 (*mulsi3_compare0_v6): Likewise.
4905 (*mulsi_compare0_scratch): Likewise.
4906 (*mulsi_compare0_scratch_v6): Likewise.
4907 (*mulsi3addsi): Likewise.
4908 (*mulsi3addsi_v6): Likewise.
4909 (*mulsi3addsi_compare0): Likewise.
4910 (*mulsi3addsi_compare0_v6): Likewise.
4911 (*mulsi3addsi_compare0_scratch): Likewise.
4912 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4913 (*mulsi3subsi): Likewise.
4914 (*mulsidi3adddi): Likewise.
4915 (*mulsi3addsi_v6): Likewise.
4916 (*mulsidi3adddi_v6): Likewise.
4917 (*mulsidi3_nov6): Likewise.
4918 (*mulsidi3_v6): Likewise.
4919 (*umulsidi3_nov6): Likewise.
4920 (*umulsidi3_v6): Likewise.
4921 (*umulsidi3adddi): Likewise.
4922 (*umulsidi3adddi_v6): Likewise.
4923 (*smulsi3_highpart_nov6): Likewise.
4924 (*smulsi3_highpart_v6): Likewise.
4925 (*umulsi3_highpart_nov6): Likewise.
4926 (*umulsi3_highpart_v6): Likewise.
4927 (mulhisi3): Likewise.
4928 (*mulhisi3tb): Likewise.
4929 (*mulhisi3bt): Likewise.
4930 (*mulhisi3tt): Likewise.
4931 (maddhisi4): Likewise.
4932 (*maddhisi4tb): Likewise.
4933 (*maddhisi4tt): Likewise.
4934 (maddhidi4): Likewise.
4935 (*maddhidi4tb): Likewise.
4936 (*maddhidi4tt): Likewise.
4938 (udivsi3): Likewise.
4939 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4940 (thumb2_mulsi_short_compare0): Likewise.
4941 (thumb2_mulsi_short_compare0_scratch): Likewise.
4942 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4943 (1020mult2): Likewise.
4944 (1020mult3): Likewise.
4945 (1020mult4): Likewise.
4946 (1020mult5): Likewise.
4947 (1020mult6): Likewise.
4948 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4950 (cortex_a15_mult64): Likewise.
4951 (cortex_a15_sdiv): Likewise.
4952 (cortex_a15_udiv): Likewise.
4953 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4959 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4960 (pj4_ir_div): Likewise.
4961 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4962 (11_mult2): Likewise.
4963 (11_mult3): Likewise.
4964 (11_mult4): Likewise.
4965 (11_mult5): Likewise.
4966 (11_mult6): Likewise.
4967 (11_mult7): Likewise.
4968 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4969 (cortex_a8_mla): Likewise.
4970 (cortex_a8_mull): Likewise.
4971 (cortex_a8_smulwy): Likewise.
4972 (cortex_a8_smlald): Likewise.
4973 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4974 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4975 (cortex_r4_mul_3): Likewise.
4976 (cortex_r4_mla_4): Likewise.
4977 (cortex_r4_mla_3): Likewise.
4978 (cortex_r4_smlald): Likewise.
4979 (cortex_r4_mull): Likewise.
4980 (cortex_r4_sdiv): Likewise.
4981 (cortex_r4_udiv): Likewise.
4982 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4983 (cortex_a7_idiv): Likewise.
4984 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4985 (9_mult2): Likewise.
4986 (9_mult3): Likewise.
4987 (9_mult4): Likewise.
4988 (9_mult5): Likewise.
4989 (9_mult6): Likewise.
4990 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4991 (cortex_a53_sdiv): Likewise.
4992 (cortex_a53_udiv): Likewise.
4993 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
4994 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
4995 (mp626_mult2): Likewise.
4996 (mp626_mult3): Likewise.
4997 (mp626_mult4): Likewise.
4998 * config/arm/fa526.md (526_mult1): Update attribute change.
4999 (526_mult2): Likewise.
5000 * config/arm/arm-generic.md (mult): Update attribute change.
5001 (mult_ldsched_strongarm): Likewise.
5002 (mult_ldsched): Likewise.
5003 (multi_cycle): Likewise.
5004 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5005 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5006 (606te_mult2): Likewise.
5007 (606te_mult3): Likewise.
5008 (606te_mult4): Likewise.
5009 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5010 (cortex_a9_mac16): Likewise.
5011 (cortex_a9_multiply): Likewise.
5012 (cortex_a9_mac): Likewise.
5013 (cortex_a9_multiply_long): Likewise.
5014 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5015 (626te_mult2): Likewise.
5016 (626te_mult3): Likewise.
5017 (626te_mult4): Likewise.
5019 2013-06-18 Richard Biener <rguenther@suse.de>
5022 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5024 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5027 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5028 with next_active_insn.
5030 2013-06-18 Alan Modra <amodra@gmail.com>
5032 * config/rs6000/rs6000.h (enum data_align): New.
5033 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5034 (DATA_ABI_ALIGNMENT): Define.
5035 (CONSTANT_ALIGNMENT): Correct comment.
5036 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5037 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5039 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5041 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5042 ATTRIBUTE_UNUSED marking.
5044 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5046 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5047 alternative and update.
5048 (aarch64_dup_lanedi): Delete.
5049 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5050 * config/aarch64/aarch64-simd-builtins.def: Update.
5052 2013-06-17 Richard Biener <rguenther@suse.de>
5054 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5055 (lto_input_scc): Declare.
5056 (lto_input_tree_1): Likewise.
5057 (struct lto_stats_d): Add num_tree_bodies_output and
5058 num_pickle_refs_output.
5059 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5060 (lto_read_tree_1): Split out from ...
5061 (lto_read_tree): ... this.
5062 (lto_input_scc): New function.
5063 (lto_input_tree_1): Split out from ...
5064 (lto_input_tree): ... this. Handle LTO_tree_scc.
5065 (lto_data_in_create): Create the streamer cache without hashes.
5066 * lto-streamer-out.c (create_output_block): Create the streamer
5067 cache with hashes when not doing WPA.
5068 (lto_write_tree_1): Split out from ...
5069 (lto_write_tree): ... this.
5070 (get_symbol_initial_value): New function.
5071 (lto_output_tree_1): Split out from ...
5072 (lto_output_tree): ... this. Write trees as series of SCCs
5073 using a DFS walk via DFS_write_tree.
5074 (struct sccs, struct scc_entry): New types.
5075 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5076 (DFS_write_tree_body): New function.
5077 (DFS_write_tree): Likewise.
5078 (hash_tree): Likewise.
5079 (scc_entry_compare): Likewise.
5080 (hash_scc): Likewise.
5081 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5082 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5083 TREE_CHAIN as regular reference.
5084 (streamer_read_integer_cst): Remove.
5085 (streamer_get_pickled_tree): Adjust.
5086 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5087 of DECL_EXTERNALs in BLOCK_VARS for now.
5088 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5090 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5091 Add hash value argument and record that if hashes are recorded
5093 (streamer_tree_cache_insert_1): Adjust.
5094 (streamer_tree_cache_insert): Likewise.
5095 (streamer_tree_cache_insert_at): Rename to ...
5096 (streamer_tree_cache_replace_tree): ... this and adjust.
5097 (streamer_tree_cache_append): Adjust.
5098 (record_common_node): Likewise.
5099 (streamer_tree_cache_create): Add argument whether to
5100 record hash values together with trees.
5101 (streamer_tree_cache_delete): Adjust.
5102 * tree-streamer.h (struct streamer_tree_cache_d): Add
5104 (streamer_read_integer_cst): Remove.
5105 (streamer_tree_cache_insert): Adjust.
5106 (streamer_tree_cache_append): Likewise.
5107 (streamer_tree_cache_insert_at): Rename to ...
5108 (streamer_tree_cache_replace_tree): ... this and adjust.
5109 (streamer_tree_cache_create): Add argument whether to record hashes.
5110 (streamer_tree_cache_get): Rename to ...
5111 (streamer_tree_cache_get_tree): ... this.
5112 (streamer_tree_cache_get_hash): New function.
5113 * tree.c (cache_integer_cst): New function.
5114 * tree.h (cache_integer_cst): Declare.
5115 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5116 * lto-symtab.c (lto_varpool_replace_node): Only release
5117 DECL_INITIAL of non-prevailing decls.
5118 * varpool.c (varpool_remove_initializer): Do not release
5119 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5121 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5123 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5124 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5125 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5126 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5127 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5128 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5129 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5130 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5131 instead of TARGET_64BIT.
5132 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5133 Require ISA_HAS_<D>DIV.
5135 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
5137 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
5138 (mips*-*-linux*): Move default with_llsc setting to where other
5140 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
5142 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
5143 Likewise. Remove default with_tune setting. Move default float
5144 setting to its own block. Handle with_llsc in the same block as above.
5146 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
5148 PR rtl-optimization/57425
5149 PR rtl-optimization/57569
5150 * alias.c (write_dependence_p): Add new parameters mem_mode,
5151 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
5152 Changed all callers.
5153 (canon_anti_dependence): New function.
5154 * cse.c (check_dependence): Use canon_anti_dependence.
5155 * cselib.c (cselib_invalidate_mem): Likewise.
5156 * rtl.h (canon_anti_dependence): Declare.
5158 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5160 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
5161 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
5162 ".set mips3" for 64-bit targets.
5164 2013-06-15 Dehao Chen <dehao@google.com>
5166 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
5167 * gimple-low.c (gimple_check_call_matching_types): Likewise.
5168 (gimple_check_call_args): Likewise.
5169 * value-prof.c (check_ic_target): Likewise.
5170 * ipa-inline.c (early_inliner): Likewise.
5171 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
5172 * cgraph.c (cgraph_create_edge_1): Likewise.
5173 (cgraph_make_edge_direct): Likewise.
5175 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
5178 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
5179 rs6000_output_move_128bit to handle emitting quad memory
5180 operations. Set attribute length to 8 bytes.
5182 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
5184 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
5186 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
5187 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
5188 (aarch64_<su>mlsl<mode>): Likewise.
5190 2013-06-14 Mike Stump <mikestump@comcast.net>
5192 * Makefile.in (TARGET_H): Add insn-codes.h.
5194 2013-06-14 Alan Modra <amodra@gmail.com>
5198 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
5199 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
5200 bitfield expansion when EXPAND_MEMORY.
5201 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
5203 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
5205 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
5206 test for clearing quad memory on 32-bit later.
5208 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5210 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
5211 (fold_negate_expr): Likewise.
5212 (fold_real_zero_addition_p): Handle vectors.
5213 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
5215 2013-06-14 Alan Modra <amodra@gmail.com>
5217 * varasm.c (force_const_mem): Revert 2013-06-07 change.
5219 2013-06-13 Jan Hubicka <jh@suse.cz>
5221 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
5222 Local comdats are not externally visible.
5223 * symtab.c (dump_symtab_base): Dump externally visible.
5224 (verify_symtab_base): Verify back links in the symtab hash.
5226 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5228 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
5229 CONVERT_EXPR as equal nodes.
5231 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5233 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
5235 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5237 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
5238 Generalize to complex and vector.
5239 * tree.c (build_all_ones_cst): New function.
5240 * tree.h (build_all_ones_cst): Declare it.
5242 2013-06-13 Alan Modra <amodra@gmail.com>
5244 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
5245 * config/rs6000/rs6000.md (signbittf2): New insn.
5246 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
5247 (abstf2_internal, cmptf_internal2): Likewise.
5248 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
5250 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5251 Pat Haugen <pthaugen@us.ibm.com>
5252 Peter Bergner <bergner@vnet.ibm.com>
5254 * config/rs6000/rs6000.c (emit_load_locked): Add support for
5255 power8 byte, half-word, and quad-word atomic instructions.
5256 (emit_store_conditional): Likewise.
5257 (rs6000_expand_atomic_compare_and_swap): Likewise.
5258 (rs6000_expand_atomic_op): Likewise.
5260 * config/rs6000/sync.md (larx): Add new modes for power8.
5262 (AINT): New mode iterator to include TImode as well as normal
5263 integer modes on power8.
5264 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
5265 that VSX registers are not considered. Use AINT mode iterator
5266 instead of INT1 to allow inclusion of quad word atomic operations
5268 (load_locked<mode>): Likewise.
5269 (store_conditional<mode>): Likewise.
5270 (atomic_compare_and_swap<mode>): Likewise.
5271 (atomic_exchange<mode>): Likewise.
5272 (atomic_nand<mode>): Likewise.
5273 (atomic_fetch_<fetchop_name><mode>): Likewise.
5274 (atomic_nand_fetch<mode>): Likewise.
5275 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
5277 (ATOMIC): On power8, add QImode, HImode modes.
5278 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
5279 modes that promote to SImode.
5280 (load_lockedti): Convert TImode arguments to PTImode, so that we
5281 get a guaranteed even/odd register pair.
5282 (load_lockedpti): Likewise.
5283 (store_conditionalti): Likewise.
5284 (store_conditionalpti): Likewise.
5286 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
5287 atomic load/store instructions.
5290 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
5292 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
5294 (insn_count): New attribute, with most cases extracted from...
5295 (length): ...here. Redefine most cases in terms of insn_count.
5296 (single_insn): Delete.
5297 (can_delay): Use insn_count to check for single instructions.
5298 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
5299 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
5300 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
5301 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
5302 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
5303 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
5304 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
5305 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
5306 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
5307 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
5308 rather than "length".
5309 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
5310 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
5311 Use "insn_count" rather than "length".
5312 * config/mips/mips-dsp.md
5313 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
5314 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
5317 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
5319 PR tree-optimization/57361
5320 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
5322 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
5324 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
5326 (aarch64_simd_combine<mode>): New instruction expansion.
5327 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
5329 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
5330 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
5332 2013-06-12 Jan Hubicka <jh@suse.cz>
5334 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
5335 decl has when in streaming stage.
5336 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
5337 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
5339 2013-06-12 Roland Stigge <stigge@antcom.de>
5342 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
5344 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5346 PR tree-optimization/57537
5347 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
5348 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
5350 2013-06-12 Richard Biener <rguenther@suse.de>
5352 * data-streamer.h (streamer_write_char_stream): CSE
5353 obs->current_pointer.
5354 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
5355 streamer_write_char_stream manually and optimize the resulting loop.
5356 (streamer_write_hwi_stream): Likewise.
5358 2013-06-12 Jan Hubicka <jh@suse.cz>
5360 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
5361 * cgraph.h (varpool_create_empty_node): Declare.
5362 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
5364 * symtab.c (symtab_unregister_node): Be lax about missin entries
5366 (symtab_get_node): Update comment.
5367 * varpool.c (varpool_create_empty_node): Break out from ...
5368 (varpool_node_for_decl): ... here.
5369 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
5371 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
5373 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
5374 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
5375 part. Use straight-line flow at the end.
5376 <COMPONENT_REF>: Remove superfluous else.
5377 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
5379 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5382 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
5383 target hook even for !TREE_PUBLIC decls. If no resolution info
5384 is available, return false for common and external decls.
5386 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5388 * config/rl78/constraints.md (U): New constraint.
5389 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5392 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5395 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5396 to allow returning address to AT_PLATFORM name.
5398 2013-06-11 Jan Hubicka <jh@suse.cz>
5400 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5401 * cgraph.h (symtab_node_base): Add weakref flag.
5402 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5403 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5404 (output_weakrefs): Use weakref flag.
5405 * fold-const.c (simple_operand_p): Handle WEAK.
5406 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5407 * ipa.c (varpool_externally_visible_p): Drop weakref.
5408 (function_and_variable_visibility): Update comment; fix weakref
5409 sanity checks; do not clear DECL_WEAK on them.
5410 * lto-cgraph.c (lto_output_node): update.
5411 (lto_output_varpool_node): Update.
5412 (input_overwrite_node): Update.
5413 (input_node): Update.
5414 (input_varpool_node): Update.
5415 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5416 (lto_symtab_merge_symbols): Add sanity check.
5417 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5418 * passes.c (rest_of_decl_compilation): Set static flag, too.
5419 * symtab.c (dump_symtab_base): Dump weakref.
5420 (verify_symtab_base): Sanity check weakrefs.
5421 (symtab_make_decl_local): Remove duplicated code.
5422 (symtab_alias_ultimate_target): Simplify.
5423 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5425 2013-06-11 DJ Delorie <dj@redhat.com>
5427 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5428 (rl78_unwind_word_mode): New.
5430 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5432 * final.c (debug_prefix_maps): Make static.
5434 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5436 * function.c (initial_trampoline): Remove stray copy.
5438 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5440 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5442 2013-06-11 Martin Jambor <mjambor@suse.cz>
5444 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5445 within bounds at the beginning of the function.
5447 2013-06-11 Alan Modra <amodra@gmail.com>
5449 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5451 (get_named_section): Don't NULL !DECL_P decl.
5453 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5455 * doc/invoke.texi (core-avx2): Document.
5457 (atom): Updated with MOVBE.
5459 2013-06-11 Richard Biener <rguenther@suse.de>
5461 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5463 2013-06-11 Anton Blanchard <anton@samba.org>
5465 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5466 correct shift value in little-endian mode.
5468 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5471 * varasm.c (get_variable_align): Move #endif to the right place.
5473 2013-06-10 Cary Coutant <ccoutant@google.com>
5475 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5476 for hash so that hash table traversal order is deterministic.
5478 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5479 Pat Haugen <pthaugen@us.ibm.com>
5480 Peter Bergner <bergner@vnet.ibm.com>
5482 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5483 of vectors in GPRS if they are direct moves or quad word load or
5486 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5488 (direct_move_p): Likewise.
5489 (quad_load_store_p): Likewise.
5491 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5492 classes into bins based on the physical register type.
5493 (reg_class_to_reg_type): Likewise.
5494 (IS_STD_REG_TYPE): Likewise.
5495 (IS_FP_VECT_REG_TYPE): Likewise.
5496 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5497 use direct move instructions.
5498 (reload_gpr_vsx): Likewise.
5499 (reload_vsx_gpr): Likewise.
5500 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5501 information that is a simplification of register classes. Also
5502 precalculate direct move reload helpers.
5503 (direct_move_p): New function to return true if the operation can
5504 be done as a direct move instruciton.
5505 (quad_load_store_p): New function to return true if the operation
5506 is a quad memory operation.
5507 (rs6000_legitimize_address): If quad memory, only allow register
5508 indirect for TImode addresses.
5509 (rs6000_legitimate_address_p): Likewise.
5510 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5511 (rs6000_reload_register_type): Likewise.
5512 (register_to_reg_type): Return register type.
5513 (rs6000_secondary_reload_simple_move): New helper function for
5514 secondary reload and secondary memory needed to identify anything
5515 that is a simple move, and does not need reloading.
5516 (rs6000_secondary_reload_direct_move): New helper function for
5517 secondary reload to identify cases that can be done with several
5518 instructions via the direct move instructions.
5519 (rs6000_secondary_reload_move): New helper function for secondary
5520 reload to identify moves between register types that can be done.
5521 (rs6000_secondary_reload): Add support for quad memory operations
5522 and for direct move.
5523 (rs6000_secondary_memory_needed): Likewise.
5524 (rs6000_debug_secondary_memory_needed): Change argument names.
5525 (rs6000_output_move_128bit): New function to return the move to
5526 use for 128-bit moves, including knowing about the various
5527 limitations of quad memory operations.
5529 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5530 memory operations. call rs6000_output_move_128bit for the actual
5531 instruciton(s) to generate.
5532 (vsx_movti_64bit): Likewise.
5534 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5535 (UNSPEC_P8V_MTVSRWZ): Likewise.
5536 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5537 (UNSPEC_P8V_MTVSRD): Likewise.
5538 (UNSPEC_P8V_XXPERMDI): Likewise.
5539 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5540 (UNSPEC_FUSION_GPR): Likewise.
5541 (FMOVE128_GPR): New iterator for direct move.
5542 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
5545 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5546 loads and direct move instructions.
5547 (zero_extendsidi2_lfiwzx): Likewise.
5548 (extendsidi2_lfiwax): Likewise.
5549 (extendsidi2_nocell): Likewise.
5550 (floatsi<mode>2_lfiwax): Likewise.
5552 (floatunssi<mode>2_lfiwzx): Likewise.
5554 (fix_trunc<mode>_stfiwx): Likewise.
5555 (fixuns_trunc<mode>_stfiwx): Likewise.
5556 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5557 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5558 (parity<mode>2_cmpb): Set length/type attr.
5559 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5560 for 'mr.' to fast_compare.
5561 (bpermd_<mode>): Change type attr to popcnt.
5562 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5563 (p8_mtvsrwz_1): Likewise.
5564 (p8_mtvsrwz_2): Likewise.
5565 (reload_fpr_from_gpr<mode>): Likewise.
5566 (p8_mtvsrd_1): Likewise.
5567 (p8_mtvsrd_2): Likewise.
5568 (p8_xxpermdi_<mode>): Likewise.
5569 (reload_vsx_from_gpr<mode>): Likewise.
5570 (reload_vsx_from_gprsf): Likewise.
5571 (p8_mfvsrd_3_<mode>): LIkewise.
5572 (reload_gpr_from_vsx<mode>): Likewise.
5573 (reload_gpr_from_vsxsf): Likewise.
5574 (p8_mfvsrd_4_disf): Likewise.
5575 (multi-word GPR splits): Do not split direct moves or quad memory
5578 2013-06-10 David Malcolm <dmalcolm@redhat.com>
5580 * tree-into-ssa.c (interesting_blocks): Make static.
5582 2013-06-10 Jakub Jelinek <jakub@redhat.com>
5585 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5586 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5587 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5588 (get_variable_align): New function.
5589 (get_variable_section, emit_bss, emit_common,
5590 assemble_variable_contents, place_block_symbol): Use
5591 get_variable_align instead of DECL_ALIGN.
5592 (assemble_noswitch_variable): Add align argument, use it
5593 instead of DECL_ALIGN.
5594 (assemble_variable): Adjust caller. Use get_variable_align
5595 instead of DECL_ALIGN.
5596 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5598 (DATA_ABI_ALIGNMENT): Define.
5599 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5600 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5601 opt is false, only return the psABI mandated alignment increase.
5602 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5603 (DATA_ABI_ALIGNMENT): ... this.
5604 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5605 (DATA_ABI_ALIGNMENT): ... this.
5606 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5607 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5608 (DATA_ABI_ALIGNMENT): ... this.
5609 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5610 * doc/tm.texi: Regenerated.
5612 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
5614 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5615 cmp_code to construct REG_EQUAL note.
5617 2013-06-09 Jakub Jelinek <jakub@redhat.com>
5620 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5621 that operands[2] doesn't overlap with operands[0].
5623 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
5624 Jan Hubicka <jh@suse.cz>
5626 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5627 hack to mark symbols as used.
5629 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5631 PR rtl-optimization/57559
5632 * lra-constraints.c (process_alt_operands): Don't discourage
5633 memory with known offset for offsetable memory constraint.
5634 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5636 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5638 * varasm.c (struct oc_local_state): Reorder fields.
5639 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5640 and adjust accordingly.
5641 (output_constructor): Reorder initialization code and adjust call to
5642 output_constructor_bitfield.
5644 2013-06-07 Jan Hubicka <jh@suse.cz>
5646 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5648 2013-06-07 David Malcolm <dmalcolm@redhat.com>
5650 * tree-object-size.c (unknown): Make const.
5652 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5654 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5655 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5656 for last alternative in the cpu_facility attribute.
5658 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5661 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5662 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5663 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5664 * config/arm/constraints.md (Dg): New constraint.
5665 * config/arm/neon.md (xordi3_neon): Remove.
5666 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5667 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5669 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5671 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5672 Clean up alternatives.
5674 2013-06-07 Alan Modra <amodra@gmail.com>
5676 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5679 2013-06-07 Alan Modra <amodra@gmail.com>
5681 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5683 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5685 * config/arm/constraints.md (Df): New constraint.
5686 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5687 Correct length attribute for last two alternatives.
5689 2013-06-07 Alan Modra <amodra@gmail.com>
5691 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5692 override user -mfp-in-toc.
5693 (offsettable_ok_by_alignment): Consider just the current access
5694 rather than the whole object, unless BLKmode. Handle
5695 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5696 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5697 for -mcmodel=medium.
5698 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5699 override user -mfp-in-toc or -msum-in-toc. Default to
5700 -mno-fp-in-toc for -mcmodel=medium.
5702 2013-06-06 DJ Delorie <dj@redhat.com>
5704 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5705 TARGET_VALID_POINTER_MODE.
5707 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5708 Pat Haugen <pthaugen@us.ibm.com>
5709 Peter Bergner <bergner@vnet.ibm.com>
5711 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5712 Document new power8 builtins.
5714 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5715 condition code register, to allow 128-bit logical operations to be
5716 done in the VSX or GPR registers.
5717 (nor<mode>3): Use the canonical form for nor.
5718 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5719 vclz*, and vpopcnt* vector instructions.
5720 (nand<mode>3): Likewise.
5721 (orc<mode>3): Likewise.
5722 (clz<mode>2): LIkewise.
5723 (popcount<mode>2): Likewise.
5725 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5726 that only the GPRs are recognized.
5728 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5729 support for new power8 builtins.
5731 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5733 (xscvdpspn): Likewise.
5739 (vpopcnt): Likewise.
5740 (vpopcntb): Likewise.
5741 (vpopcnth): Likewise.
5742 (vpopcntw): Likewise.
5743 (vpopcntd): Likewise.
5748 (eqv_v16qi3): Likewise.
5749 (eqv_v8hi3): Likewise.
5750 (eqv_v4si3): Likewise.
5751 (eqv_v2di3): Likewise.
5752 (eqv_v4sf3): Likewise.
5753 (eqv_v2df3): Likewise.
5755 (nand_v16qi3): Likewise.
5756 (nand_v8hi3): Likewise.
5757 (nand_v4si3): Likewise.
5758 (nand_v2di3): Likewise.
5759 (nand_v4sf3): Likewise.
5760 (nand_v2df3): Likewise.
5762 (orc_v16qi3): Likewise.
5763 (orc_v8hi3): Likewise.
5764 (orc_v4si3): Likewise.
5765 (orc_v2di3): Likewise.
5766 (orc_v4sf3): Likewise.
5767 (orc_v2df3): Likewise.
5769 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5770 allow power8 quad mode in 64-bit.
5771 (rs6000_builtin_vectorized_function): Add support to vectorize
5772 ISA 2.07 count leading zeros, population count builtins.
5773 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5774 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5775 (builtin_function_type): Add vgbbd builtin function which takes an
5777 (altivec_expand_vec_perm_const): Add support for new power8 merge
5780 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5781 that does not include TImdoe for use with 32-bit.
5782 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5784 (UNSPEC_VSX_CVDPSPN): Likewise.
5785 (vsx_xscvdpspn): Likewise.
5786 (vsx_xscvspdpn): Likewise.
5787 (vsx_xscvdpspn_scalar): Likewise.
5788 (vsx_xscvspdpn_directmove): Likewise.
5789 (vsx_and<mode>3): Split logical operations into 32-bit and
5790 64-bit. Add support to do logical operations on TImode as well as
5791 VSX vector types. Allow logical operations to be done in either
5792 VSX registers or in general purpose registers in 64-bit mode. Add
5793 splitters if GPRs were used. For AND, add clobber of CCmode to
5794 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5796 (vsx_and<mode>3_32bit): Likewise.
5797 (vsx_and<mode>3_64bit): Likewise.
5798 (vsx_ior<mode>3): Likewise.
5799 (vsx_ior<mode>3_32bit): Likewise.
5800 (vsx_ior<mode>3_64bit): Likewise.
5801 (vsx_xor<mode>3): Likewise.
5802 (vsx_xor<mode>3_32bit): Likewise.
5803 (vsx_xor<mode>3_64bit): Likewise.
5804 (vsx_one_cmpl<mode>2): Likewise.
5805 (vsx_one_cmpl<mode>2_32bit): Likewise.
5806 (vsx_one_cmpl<mode>2_64bit): Likewise.
5807 (vsx_nor<mode>3): Likewise.
5808 (vsx_nor<mode>3_32bit): Likewise.
5809 (vsx_nor<mode>3_64bit): Likewise.
5810 (vsx_andc<mode>3): Likewise.
5811 (vsx_andc<mode>3_32bit): Likewise.
5812 (vsx_andc<mode>3_64bit): Likewise.
5813 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5814 and xxlorc instructions.
5815 (vsx_eqv<mode>3_64bit): Likewise.
5816 (vsx_nand<mode>3_32bit): Likewise.
5817 (vsx_nand<mode>3_64bit): Likewise.
5818 (vsx_orc<mode>3_32bit): Likewise.
5819 (vsx_orc<mode>3_64bit): Likewise.
5821 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5823 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5825 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5826 (p8_vmrgow): Likewise.
5827 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5828 GPRs to be split under VSX.
5829 (p8v_clz<mode>2): Add power8 count leading zero support.
5830 (p8v_popcount<mode>2): Add power8 population count support.
5831 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5834 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5837 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5839 (vec_nand): Likewise.
5840 (vec_vclz): Likewise.
5841 (vec_vclzb): Likewise.
5842 (vec_vclzd): Likewise.
5843 (vec_vclzh): Likewise.
5844 (vec_vclzw): Likewise.
5845 (vec_vgbbd): Likewise.
5846 (vec_vmrgew): Likewise.
5847 (vec_vmrgow): Likewise.
5848 (vec_vpopcnt): Likewise.
5849 (vec_vpopcntb): Likewise.
5850 (vec_vpopcntd): Likewise.
5851 (vec_vpopcnth): Likewise.
5852 (vec_vpopcntw): Likewise.
5854 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5856 PR rtl-optimization/57468
5857 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5860 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5862 PR rtl-optimization/57459
5863 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5864 type when setting live regs.
5866 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5868 * config/s390/s390.opt (mlra): New option.
5869 * config/s390/s390.c (s390_decompose_address): Check displacement
5870 for all registers for LRA.
5871 (s390_secondary_reload): Don't used secondary reloads for LRA.
5872 (s390_lra_p): New function.
5873 (TARGET_LRA_P): Define.
5874 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5875 of attribute cpu_facility to zarch for the last alternative.
5876 (*cmpmem_short): Ditto.
5878 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5880 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5881 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5882 (arm_expand_prologue): Likewise.
5884 2013-06-06 Teresa Johnson <tejohnson@google.com>
5887 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5888 as this is now done by redirect_edge_and_branch_force.
5889 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5890 barriers, and fix interaction with splitting.
5891 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5892 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5893 reflect changes made in the routine.
5894 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5895 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5896 since this is called in cfglayout mode, and replace partition fixup
5897 with assert as that is now done by force_nonfallthru_and_redirect.
5898 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5899 already be marked with region crossing note.
5900 (insert_section_boundary_note): Make non-static, gate on flag
5901 has_bb_partition, rewrite to also check for multiple partitions.
5902 (rest_of_handle_reorder_blocks): Remove call to
5903 insert_section_boundary_note, now done later during free_cfg.
5904 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5905 * bb-reorder.h (insert_section_boundary_note): Declare.
5906 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5907 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5908 invoke insert_section_boundary_note.
5909 (try_redirect_by_replacing_jump): Remove unnecessary
5910 check for region crossing note.
5911 (fixup_partition_crossing): New function.
5912 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5913 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5914 in non-cfglayout mode.
5915 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5916 remove old code that tried to do this. Emit barrier correctly
5917 when we are in cfglayout mode.
5918 (last_bb_in_partition): New function.
5919 (rtl_split_edge): Correctly fixup partition boundaries.
5920 (commit_one_edge_insertion): Remove old code that tried to
5921 fixup region crossing edge since this is now handled in
5922 split_block, and set up insertion point correctly since
5923 block may now end in a jump.
5924 (verify_hot_cold_block_grouping): Guard against checking when not in
5925 linearized RTL mode.
5926 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5928 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5929 rtl_verify_flow_info, so not called in cfglayout mode.
5930 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5931 (fixup_reorder_chain): Remove old code that attempted to fixup region
5932 crossing note as this is now handled in force_nonfallthru_and_redirect.
5933 (duplicate_insn_chain): Don't duplicate switch section notes.
5934 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5936 * basic-block.h (emit_barrier_after_bb): Declare.
5938 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5940 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5941 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5942 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5944 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5946 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5948 * config/arm/ldmstm.md: Regenerate.
5950 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5952 * config/arm/sync.md (atomic_loaddi_1):
5953 Disable predication for arm_restrict_it.
5954 (arm_load_exclusive<mode>): Likewise.
5955 (arm_load_exclusivesi): Likewise.
5956 (arm_load_exclusivedi): Likewise.
5957 (arm_load_acquire_exclusive<mode>): Likewise.
5958 (arm_load_acquire_exclusivesi): Likewise.
5959 (arm_load_acquire_exclusivedi): Likewise.
5960 (arm_store_exclusive<mode>): Likewise.
5961 (arm_store_exclusive<mode>): Likewise.
5962 (arm_store_release_exclusivedi): Likewise.
5963 (arm_store_release_exclusive<mode>): Likewise.
5965 2013-06-06 Richard Biener <rguenther@suse.de>
5967 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5969 (lto_tag_is_tree_code_p): Adjust.
5970 (lto_tag_is_gimple_code_p): Likewise.
5971 (lto_gimple_code_to_tag): Likewise.
5972 (lto_tag_to_gimple_code): Likewise.
5973 (lto_tree_code_to_tag): Likewise.
5974 (lto_tag_to_tree_code): Likewise.
5975 * data-streamer.h (streamer_write_hwi_in_range): Use
5976 uhwi streaming to stream the normalized range.
5977 (streamer_read_hwi_in_range): Likewise.
5979 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5981 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5982 (predicable_short_it): Likewise.
5983 (predicated): Likewise.
5984 (enabled): Handle above.
5985 (define_cond_exec): Set predicated attribute to yes.
5987 2013-06-05 Mike Stump <mikestump@comcast.net>
5989 * gdbinit.in (__FUNCTION__): Add.
5991 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
5993 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
5994 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
5996 2013-06-05 Jan Hubicka <jh@suse.cz>
5998 * varasm.c (mark_decl_referenced): Revert the removal until targets
6001 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6003 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6004 instead of mark_decl_referenced.
6006 2013-06-05 Jan Hubicka <jh@suse.cz>
6008 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6009 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6010 and symtab_used_from_object_file_p.
6011 (cgraph_make_node_local_1): Clear forced_by_abi.
6012 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6013 * cgraph.h (symtab_node_base): Add forced_by_abi.
6014 (decide_is_variable_needed): Remove.
6015 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6016 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6017 (decide_is_symbol_needed): ... this one; handle symbols in general;
6018 always analyze virtuals; honnor forced_by_abi.
6019 (cgraph_finalize_function): Update.
6020 (varpool_finalize_decl): Update.
6021 (symbol_defined_and_needed): Remove.
6022 (analyze_functions): Update.
6023 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6024 output_refs, input_overwrite_node): Handle forced_by_abi.
6025 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6026 (address_taken_from_non_vtable_p): ... this one.
6027 (comdat_can_be_unshared_p_1): New function.
6028 (cgraph_comdat_can_be_unshared_p): Rename to ...
6029 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6030 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6031 (function_and_variable_visibility): Clear forced_by_abi as needed.
6032 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6033 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6034 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6035 * varpool.c (decide_is_variable_needed): Remove.
6037 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6039 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6040 (arm_option_override): Override arm_restrict_it where appropriate.
6041 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6042 * config/arm/arm.opt (mrestrict-it): New command-line option.
6043 * doc/invoke.texi: Document -mrestrict-it.
6045 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6047 * tsan.c (tsan_atomic_table): Make const.
6049 2013-06-05 Richard Biener <rguenther@suse.de>
6051 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6052 index associated with the tree we are supposed to replace.
6053 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6054 TREE_ASM_WRITTEN as zero for everything but SSA names.
6056 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6058 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6060 2013-06-04 Jan Hubicka <jh@suse.cz>
6062 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6063 (input_node, input_varpool_node): Handle correctly external same
6065 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6066 nodes at ltrans stage.
6068 2013-06-04 Jan Hubicka <jh@suse.cz>
6070 * ipa-inline.c (update_caller_keys): Fix availability test.
6071 (update_callee_keys): Likewise.
6072 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6073 to follow ELF standard.
6075 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6077 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6078 (mips64r5900el-*-elf*): New configurations.
6079 * config/mips/mips-cpus.def (r5900): New processor.
6080 * config/mips/mips-tables.opt: Regenerate.
6081 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6082 (mips_issue_rate): Handle PROCESSOR_R5900.
6083 (mips_reorg_process_insns): Force reorder mode for the R5900.
6084 * config/mips/mips.h (TARGET_MIPS5900): Define.
6085 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6087 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6089 * config/mips/mips.md (processor): Add r5900.
6090 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6092 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6094 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6095 into function to generate MOVI instruction.
6096 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6097 (aarch64_preferred_simd_mode): Turn into wrapper.
6098 (aarch64_output_scalar_simd_mov_immediate): New function.
6099 * config/aarch64/aarch64-protos.h: Add prototype for above.
6101 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6103 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6104 element_char member.
6105 (sizetochar): Return signed char.
6106 (aarch64_simd_valid_immediate): Remove elchar and other
6107 unnecessary variables.
6108 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6109 Calculate element_char as required.
6110 * config/aarch64/aarch64-protos.h: Update and move prototype
6111 for aarch64_output_simd_mov_immediate.
6112 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6115 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6117 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6118 information completed by aarch64_simd_valid_immediate.
6119 (aarch64_legitimate_constant_p): Update arguments.
6120 (aarch64_simd_valid_immediate): Work with struct rather than many
6122 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6123 (aarch64_simd_make_constant): Update arguments.
6124 (aarch64_output_simd_mov_immediate): Work with struct rather than
6125 many pointers. Output immediate directly rather than as operand.
6126 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6128 * config/aarch64/constraints.md (Dn): Update arguments.
6130 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6132 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6134 (aarch64_simd_immediate_valid_for_move): Remove.
6135 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
6136 (aarch64_simd_make_constant): Update call.
6137 (aarch64_output_simd_mov_immediate): Update call.
6138 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6140 * config/aarch64/constraints.md (Dn): Update call.
6142 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6144 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
6145 return type to bool for prototype.
6146 (aarch64_legitimate_constant_p): Check for true instead of not -1.
6147 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
6148 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
6150 2013-06-04 Catherine Moore <clm@codesourcery.com>
6152 * config/mips/mips.opt (meva): New.
6153 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
6154 (ASM_SPEC): Handle -meva.
6155 * doc/invoke.texi (meva): Document.
6157 2013-06-04 Alan Modra <amodra@gmail.com>
6159 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
6162 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6164 * rtl.def: Add extra fourth optional field to define_cond_exec.
6165 * gensupport.c (process_one_cond_exec): Process attributes from
6167 * doc/md.texi: Document fourth field in define_cond_exec.
6169 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
6171 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
6172 out the processing order as in store_bit_field_1.
6174 2013-06-04 Jan Hubicka <jh@suse.cz>
6177 * cgraphunit.c (cgraph_process_same_body_aliases): Create
6178 non-VAR_DECL node if it does not exist yet.
6180 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
6182 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
6183 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
6184 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
6185 target_cpu_default setting.
6187 2013-06-03 Teresa Johnson <tejohnson@google.com>
6189 * dumpfile.c (opt_info_switch_p): Change -fopt-info
6190 default to -fopt-info=optimized instead of all.
6191 * doc/invoke.texi: Ditto.
6192 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
6193 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
6194 (execute_vect_slp): Emit BB vectorization success under
6195 MSG_OPTIMIZED_LOCATIONS.
6196 * tree-vect-slp.c (vect_slp_transform_bb): Change
6197 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
6198 * tree-vect-loop.c (vect_transform_loop): Ditto.
6200 2013-06-03 Jason Merrill <jason@redhat.com>
6203 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
6204 Use TARGET_EXPR for C++.
6206 2013-06-03 Jakub Jelinek <jakub@redhat.com>
6208 PR rtl-optimization/57268
6209 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
6210 if DEBUG_INSN_P (insn).
6213 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6215 PR rtl-optimization/57268
6216 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6217 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6219 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6221 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
6222 (ix86_avoid_lea_for_addr): Likewise.
6223 (exact_dependency_1): Likewise.
6224 (ix86_adjust_cost): Likewise.
6225 (swap_top_of_ready_list): Fix formatting and !reload_completed check
6227 (do_reorder_for_imul): Fix typo, formatting and
6228 !reload_completed check removed.
6229 (ix86_sched_reorder): Fix typo and formatting.
6230 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
6233 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
6235 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
6237 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
6239 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
6240 <STRING_CST>: Likewise.
6241 <VECTOR_CST>: Likewise.
6243 2013-06-01 Janus Weil <janus@gcc.gnu.org>
6244 Mikael Morin <mikael@gcc.gnu.org>
6246 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
6247 * config.in: Regenerated.
6248 * configure: Regenerated.
6250 2013-06-01 Jan Hubicka <jh@suse.cz>
6253 * cgraphunit.c (compile): When weakref is not supported,
6254 set up transparent aliases before final output pass.
6255 * varasm.c (assemble_alias): Do not try to do it here.
6257 2013-06-01 Jan Hubicka <jh@suse.cz>
6260 * passes.c (for_per_function): Skip unanalyzed functions.
6262 2013-06-01 Jan Hubicka <jh@suse.cz>
6264 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
6265 (lto_symtab_merge_symbols_1): ... this one.
6266 (lto_symtab_merge_cgraph_nodes): Rename to ...
6267 (lto_symtab_merge_symbols): ... this one; simplify.
6268 * cgraph.c (same_body_aliases_done): Rename to ...
6269 (cpp_implicit_aliases_done): ... this one.
6270 (cgraph_create_function_alias): Update.
6271 (cgraph_same_body_alias): Update.
6272 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
6273 (verify_edge_corresponds_to_fndecl): Simplify.
6274 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
6275 (cgraph_node): Remove same_body_alias.
6276 (varpool_node): Remove alias_of and extra_name_alias.
6277 (same_body_aliases_done): Rename to ..
6278 (cpp_implicit_aliases_done): ... this one.
6279 (symtab_alias_ultimate_target): Add default parameter.
6280 (symtab_resolve_alias): New function.
6281 (fixup_same_cpp_alias_visibility): Declare.
6282 (cgraph_function_node): Add default parameter.
6283 (cgraph_node_asm_name): Likewise.
6284 (cgraph_function_or_thunk_node): Add default parameter; do
6285 not ICE when it is NULL.
6286 (varpool_variable_node): Likewise.
6287 * tree-emutls.c (create_emultls_var): Update.
6288 (ipa_lower_emutls): Update.
6289 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
6290 (cgraph_reset_node): Reset alias info.
6291 (cgraph_finalize_function): Update.
6292 (fixup_same_cpp_alias_visibility): Move to symtab.c.
6293 (analyze_function): Simplify.
6294 (cgraph_process_same_body_aliases): Simplify.
6295 (analyze_functions): Fixup same body aliases.
6296 (handle_alias_pairs): Simplify.
6297 (assemble_thunk): Update.
6298 (assemble_thunks_and_aliases): Update.
6299 (output_weakrefs): Rewrite.
6300 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
6301 (lto_output_varpool_node): Likewise.
6302 (compute_ltrans_boundary): Remve assert.
6303 (get_alias_symbol): New functoin.
6304 (input_node): Rewrite alias handling.
6305 (input_varpool_node): Likewise.
6306 * ipa-pure-const.c (propagate_pure_const): Fix formating.
6307 * ipa.c (process_references): Handle weakrefs correctly.
6308 (symtab_remove_unreachable_nodes): Likewise.
6309 * trans-mem.c (get_cg_data): Update.
6310 (ipa_tm_create_version_alias): Update.
6311 (ipa_tm_execute): Update.
6312 * symtab.c (dump_symtab_base): Dump aliases.
6313 (verify_symtab_base): Verify aliases.
6314 (symtab_node_availability): New function.
6315 (symtab_alias_ultimate_target): Simplify.
6316 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
6317 handle all the fixup cases.
6318 (symtab_resolve_alias): New function.
6319 * passes.c (ipa_write_summaries): Handle weakrefs.
6320 * varpool.c (varpool_analyze_node): Simplify.
6321 (assemble_aliases): Update.
6322 (varpool_create_variable_alias): Simplify.
6323 (varpool_extra_name_alias): Simplify.
6324 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
6325 (lto_symtab_merge_symbols): ... this one.
6327 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
6330 PR rtl-optimization/57268
6331 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6332 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6334 2013-06-01 Tobias Burnus <burnus@net-b.de>
6337 2013-05-31 Tobias Burnus <burnus@net-b.de>
6340 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
6343 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6345 PR rtl-optimization/57268
6346 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6347 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6349 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6351 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
6352 unordered comparison operators when -fno-trapping-math is in effect
6354 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
6355 and implement unordered comparison operators properly on the e500.
6357 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6359 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
6360 for constant scalar integers.
6361 (simplify_relational_operation_1): Likewise.
6363 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
6365 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
6366 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
6369 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6370 Igor Zamyatin <igor.zamyatin@intel.com>
6372 Silvermont (SLM) architecture performance tuning.
6373 * config/i386/i386.h (enum ix86_tune_indices): Add
6374 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
6375 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
6377 * config/i386/i386.c (initial_ix86_tune_features)
6378 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
6379 (ix86_lea_outperforms): Handle Silvermont tuning.
6380 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
6382 (ix86_use_lea_for_mov): Likewise.
6383 (ix86_avoid_lea_for_addr): Likewise.
6384 (ix86_lea_for_add_ok): Likewise.
6385 (exact_dependency_1): New function.
6386 (exact_store_load_dependency): Likewise.
6387 (ix86_adjust_cost): Handle Silvermont tuning.
6388 (do_reoder_for_imul): Likewise.
6389 (swap_top_of_ready_list): New function.
6390 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6392 * config/i386/i386.md (peepholes that split memory operand in fp
6395 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6397 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6398 Remove un-necessary braces.
6400 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6402 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6403 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6405 2013-05-31 Tobias Burnus <burnus@net-b.de>
6408 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6410 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6413 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6414 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6415 * config/arm/neon.md (iordi3_neon): Remove.
6416 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6417 * config/arm/predicates.md (imm_for_neon_logic_operand):
6418 Move to earlier in the file.
6419 (neon_logic_op2): Likewise.
6420 (arm_iordi_operand_neon): New predicate.
6422 2013-05-31 Richard Biener <rguenther@suse.de>
6424 PR tree-optimization/57478
6425 PR tree-optimization/57453
6426 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6429 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6431 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6432 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6434 2013-05-30 Tobias Burnus <burnus@net-b.de>
6435 Thomas Koenig <tkoenig@gcc.gnu.org>
6438 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6439 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6441 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6443 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6445 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6447 * target.def (register_usage_leveling_p): New hook.
6448 * targhooks.c (default_register_usage_leveling_p): New.
6449 * targhooks.h (default_register_usage_leveling_p): New prototype.
6450 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6451 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6452 * doc/tm.texi: Update.
6453 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6455 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6457 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6458 (*insv_reg<mode>): New define_insn.
6460 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6462 PR rtl-optimization/57439
6463 * postreload.c (move2add_valid_value_p): Check that we have
6464 a zero subreg_regno_offset when accessing the register in
6467 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6468 Igor Zamyatin <igor.zamyatin@intel.com>
6470 Silvermont (SLM) architecture pipeline model, tuning and
6472 * config.gcc: Add slm config options and target.
6474 * config/i386/slm.md: New.
6476 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6478 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6480 (ix86_target_macros_internal): Likewise.
6482 * gcc/config/i386/i386.c (slm_cost): New cost.
6483 (m_SLM): New macro flag.
6484 (initial_ix86_tune_features): Set m_SLM.
6485 (x86_accumulate_outgoing_args): Likewise.
6486 (x86_arch_always_fancy_math_387): Likewise.
6487 (processor_target_table): Add slm cost.
6488 (cpu_names): Add slm cpu name.
6489 (x86_option_override_internal): Set SLM ISA.
6490 (ix86_issue_rate): New case PROCESSOR_SLM.
6491 (ia32_multipass_dfa_lookahead): Likewise.
6492 (fold_builtin_cpu): Add slm.
6494 * config/i386/i386.h (TARGET_SLM): New target macro.
6495 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6496 (processor_type): Add PROCESSOR_SLM.
6498 * config/i386/i386.md (cpu): Add new value "slm".
6499 (slm.md): Include slm.md.
6501 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6502 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6504 * config/arm/arm-protos.h: Add and update function protos.
6505 * config/arm/arm.c (use_simple_return_p): New added.
6506 (thumb2_expand_return): Check simple_return flag.
6507 * config/arm/arm.md: Add simple_return and conditional simple_return.
6508 * config/arm/iterators.md: Add iterator for return and simple_return.
6510 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6512 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6513 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6514 (arm_emit_vfp_multi_reg_pop): Likewise.
6515 (thumb2_emit_ldrd_pop): Likewise.
6516 (arm_expand_epilogue): Add misc REG_CFA notes.
6517 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6519 2013-05-29 Lawrence Crowl <crowl@google.com>
6521 * config/arm/t-arm: Update for below.
6523 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6524 Change type to hash_table. Update dependent calls and types.
6526 * config/i386/t-cygming: Update for below.
6528 * config/i386/t-interix: Update for below.
6530 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6531 Change type to hash_table. Update dependent calls and types.
6532 (i386_find_on_wrapper_list::wrappers): Likewise.
6534 * config/ia64/t-ia64: Update for below.
6536 * config/ia64/ia64.c (bundle_state_table):
6537 Change type to hash_table. Update dependent calls and types.
6539 * config/mips/mips.c (mips_reorg_process_insns::htab):
6540 Change type to hash_table. Update dependent calls and types.
6542 * config/sol2.c (solaris_comdat_htab):
6543 Change type to hash_table. Update dependent calls and types.
6545 * config/t-sol2: Update for above.
6547 2013-05-29 Teresa Johnson <tejohnson@google.com>
6549 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6550 functions are not yet marked as defined.
6552 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6553 Pat Haugen <pthaugen@us.ibm.com>
6554 Peter Bergner <bergner@vnet.ibm.com>
6556 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6560 (vrotl<mode>3): Likewise.
6561 (vashl<mode>3): Likewise.
6562 (vlshr<mode>3): Likewise.
6563 (vashr<mode>3): Likewise.
6565 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6566 support for power8 V2DI builtins.
6568 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6569 power8 V2DI builtins.
6570 (vupkhsw): Likewise.
6571 (vupklsw): Likewise.
6572 (vaddudm): Likewise.
6577 (vpkudum): Likewise.
6578 (vpksdss): Likewise.
6579 (vpkudus): Likewise.
6580 (vpksdus): Likewise.
6585 (vsubudm): Likewise.
6586 (vcmpequd): Likewise.
6587 (vcmpgtsd): Likewise.
6588 (vcmpgtud): Likewise.
6589 (vcmpequd_p): Likewise.
6590 (vcmpgtsd_p): Likewise.
6591 (vcmpgtud_p): Likewise.
6592 (vupkhsw): Likewise.
6593 (vupklsw): Likewise.
6594 (vaddudm): Likewise.
6599 (vpksdss): Likewise.
6600 (vpksdus): Likewise.
6601 (vpkudum): Likewise.
6602 (vpkudus): Likewise.
6607 (vsubudm): Likewise.
6609 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6610 support for power8 V2DI instructions.
6612 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6613 power8 V2DI instructions. Combine pack and unpack insns to use an
6614 iterator for each mode. Check whether a particular mode supports
6615 Altivec instructions instead of just checking TARGET_ALTIVEC.
6616 (UNSPEC_VPKUWUM): Likewise.
6617 (UNSPEC_VPKSHSS): Likewise.
6618 (UNSPEC_VPKSWSS): Likewise.
6619 (UNSPEC_VPKUHUS): Likewise.
6620 (UNSPEC_VPKSHUS): Likewise.
6621 (UNSPEC_VPKUWUS): Likewise.
6622 (UNSPEC_VPKSWUS): Likewise.
6623 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6624 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6625 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6626 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6627 (UNSPEC_VUPKHSB): Likewise.
6628 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6629 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6630 (UNSPEC_VUPKHSH): Likewise.
6631 (UNSPEC_VUPKLSB): Likewise.
6632 (UNSPEC_VUPKLSH): Likewise.
6634 (VI_char): Likewise.
6635 (VI_scalar): Likewise.
6636 (VI_unit): Likewise.
6638 (VP_small): Likewise.
6639 (VP_small_lc): Likewise.
6640 (VU_char): Likewise.
6641 (add<mode>3): Likewise.
6642 (altivec_vaddcuw): Likewise.
6643 (altivec_vaddu<VI_char>s): Likewise.
6644 (altivec_vadds<VI_char>s): Likewise.
6645 (sub<mode>3): Likewise.
6646 (altivec_vsubcuw): Likewise.
6647 (altivec_vsubu<VI_char>s): Likewise.
6648 (altivec_vsubs<VI_char>s): Likewise.
6649 (altivec_vavgs<VI_char>): Likewise.
6650 (altivec_vcmpbfp): Likewise.
6651 (altivec_eq<mode>): Likewise.
6652 (altivec_gt<mode>): Likewise.
6653 (altivec_gtu<mode>): Likewise.
6654 (umax<mode>3): Likewise.
6655 (smax<mode>3): Likewise.
6656 (umin<mode>3): Likewise.
6657 (smin<mode>3): Likewise.
6658 (altivec_vpkuhum): Likewise.
6659 (altivec_vpkuwum): Likewise.
6660 (altivec_vpkshss): Likewise.
6661 (altivec_vpkswss): Likewise.
6662 (altivec_vpkuhus): Likewise.
6663 (altivec_vpkshus): Likewise.
6664 (altivec_vpkuwus): Likewise.
6665 (altivec_vpkswus): Likewise.
6666 (altivec_vpks<VI_char>ss): Likewise.
6667 (altivec_vpks<VI_char>us): Likewise.
6668 (altivec_vpku<VI_char>us): Likewise.
6669 (altivec_vpku<VI_char>um): Likewise.
6670 (altivec_vrl<VI_char>): Likewise.
6671 (altivec_vsl<VI_char>): Likewise.
6672 (altivec_vsr<VI_char>): Likewise.
6673 (altivec_vsra<VI_char>): Likewise.
6674 (altivec_vsldoi_<mode>): Likewise.
6675 (altivec_vupkhsb): Likewise.
6676 (altivec_vupkhs<VU_char>): Likewise.
6677 (altivec_vupkls<VU_char>): Likewise.
6678 (altivec_vupkhsh): Likewise.
6679 (altivec_vupklsb): Likewise.
6680 (altivec_vupklsh): Likewise.
6681 (altivec_vcmpequ<VI_char>_p): Likewise.
6682 (altivec_vcmpgts<VI_char>_p): Likewise.
6683 (altivec_vcmpgtu<VI_char>_p): Likewise.
6684 (abs<mode>2): Likewise.
6685 (vec_unpacks_hi_v16qi): Likewise.
6686 (vec_unpacks_hi_v8hi): Likewise.
6687 (vec_unpacks_lo_v16qi): Likewise.
6688 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6689 (vec_unpacks_lo_v8hi): Likewise.
6690 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6691 (vec_pack_trunc_v8h): Likewise.
6692 (vec_pack_trunc_v4si): Likewise.
6693 (vec_pack_trunc_<mode>): Likewise.
6695 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6697 (vec_vmaxsd): Likewise.
6698 (vec_vmaxud): Likewise.
6699 (vec_vminsd): Likewise.
6700 (vec_vminud): Likewise.
6701 (vec_vpksdss): Likewise.
6702 (vec_vpksdus): Likewise.
6703 (vec_vpkudum): Likewise.
6704 (vec_vpkudus): Likewise.
6705 (vec_vrld): Likewise.
6706 (vec_vsld): Likewise.
6707 (vec_vsrad): Likewise.
6708 (vec_vsrd): Likewise.
6709 (vec_vsubudm): Likewise.
6710 (vec_vupkhsw): Likewise.
6711 (vec_vupklsw): Likewise.
6713 2013-05-29 Jan Hubicka <jh@suse.cz>
6715 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6716 flags; reorder rest of fields in more consistent way.
6717 (varpool_node): Remove analyzed, finalized and alias.
6718 (cgraph_ndoe): Likewise.
6719 (symtab_alias_ultimate_target): New function.
6720 (cgraph_function_node): Move offline.
6721 (cgraph_reset_node): Declare.
6722 (cgraph_comdat_can_be_unshared_p): Remove.
6723 (varpool_remove_initializer): Declare.
6724 (varpool_first_defined_variable, varpool_next_defined_variable
6725 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6726 (cgraph_function_with_gimple_body_p): Update.
6727 (varpool_all_refs_explicit_p): Update.
6728 (symtab_alias_target): New function.
6729 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6730 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
6731 (cgraph_function_or_thunk_node): Simplify using
6732 symtab_alias_ultimate_target.
6733 (varpool_variable_node): Likewise.
6734 * cgraph.c (cgraph_create_function_alias): Update.
6735 (cgraph_add_thunk): Update.
6736 (cgraph_remove_node): Update.
6737 (dump_cgraph_node): Do not dump removed flags.
6738 (cgraph_function_body_availability): Update.
6739 (cgraph_propagate_frequency): Update.
6740 (verify_cgraph_node): Check sanity of local flag.
6741 (cgraph_function_node): Move here from cgraph.h; revamp for
6742 cgraph_function_or_thunk_node.
6743 * lto-symtab.c (lto_varpool_replace_node): Update.
6744 (lto_symtab_resolve_can_prevail_p): Update.
6745 (lto_symtab_merge_cgraph_nodes): Update.
6746 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6747 propagate_constants_accross_call, devirtualization_time_bonus,
6748 ipcp_propagate_stage): Update.
6749 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
6750 * ipa-inline-transform.c (clone_inlined_nodes,
6751 preserve_function_body_p): Update.
6752 * ipa-reference.c (propagate): Update.
6753 (write_node_summary_p): Update.
6754 * toplev.c (wrapup_global_declaration_2): Update.
6755 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6756 (analyze_function) ... this one.
6757 (cgraph_process_new_functions): Update.
6758 (cgraph_reset_node): Export.
6759 (cgraph_finalize_function): Update.
6760 (cgraph_add_new_function): Update.
6761 (process_function_and_variable_attributes): Update.
6762 (varpool_finalize_decl): Update.
6763 (symbol_finalized): Remove.
6764 (symbol_finalized_and_needed): Rename to ...
6765 (symbol_defined_and_needed): ... update.
6766 (cgraph_analyze_functions): Update.
6767 (handle_alias_pairs): Update.
6768 (mark_functions_to_output): Update.
6769 (assemble_thunk): Update.
6770 (output_in_order): Update.
6771 (output_weakrefs): Update.
6772 (finalize_compilation_unit): Update.
6773 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6774 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6775 input_node, input_varpool_node): Update.
6776 * dbxout.c (dbxout_expand_expr): Update.
6777 * cgraphclones.c (cgraph_clone_node): Update.
6778 (cgraph_copy_node_for_versioning): Update.
6779 (cgraph_materialize_clone): Update.
6780 (cgraph_materialize_all_clones): Update.
6781 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6782 propagate_pure_const, propagate_nothrow): Update.
6783 * lto-streamer-out.c (lto_output, write_symbol): Update.
6784 * ipa-utils.c (ipa_reverse_postorder): Update.
6785 * ipa-inline.c (can_inline_edge_p): Update.
6786 (update_caller_keys, ipa_inline): Update.
6787 * dwarf2out.c (reference_to_unused,
6788 premark_types_used_by_global_vars_helper): Update.
6789 * tree-eh.c (tree_could_trap_p): Update.
6790 * ipa-split.c (consider_split, execute_split_functions): Update.
6791 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
6792 has_addr_references_p): Update; move ahead in file for better
6794 (process_references): Simplify.
6795 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6797 (cgraph_comdat_can_be_unshared_p): Make static.
6798 (cgraph_externally_visible_p): Update.
6799 (varpool_externally_visible_p): Update.
6800 (function_and_variable_visibility): Update.
6801 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6802 ipa_tm_mark_force_output_node): Update.
6803 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6804 estimate_edge_devirt_benefit, inline_generate_summary,
6805 inline_write_summary): Update.
6806 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6807 * ipa-prop.c (ipa_compute_jump_functions): Update.
6808 (ipa_print_node_params, ipa_prop_read_section,
6809 ipa_update_after_lto_read, read_replacements_section): Update.
6810 * varasm.c (mark_decl_referenced): Update.
6811 (assemble_alias, dump_tm_clone_pairs): Update.
6812 * tree-inline.c (copy_bb): Update.
6813 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6815 * symtab.c (dump_symtab_base): Print new flags.
6816 (verify_symtab_base): Verify new flags.
6817 (symtab_alias_ultimate_target): New function.
6818 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6819 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6821 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6823 * i386.c (ix86_get_function_versions_dispatcher,
6824 ix86_generate_version_dispatcher_body): Update.
6825 (fold_builtin_cpu): Use varpool_add_new_variable.
6826 * varpool.c (varpool_remove_initializer): Break out from ...
6827 (varpool_remove_node): ... this one.
6828 (dump_varpool_node, varpool_node_for_asm,
6829 cgraph_variable_initializer_availability, varpool_analyze_node,
6830 varpool_assemble_decl, varpool_remove_unreferenced_decls,
6831 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6834 2013-05-29 Jan Hubicka <jh@suse.cz>
6836 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6838 2013-05-29 Easwaran Raman <eraman@google.com>
6840 PR tree-optimization/57442
6841 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6842 when control exits the main loop.
6844 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6846 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6848 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
6849 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6850 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6853 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6855 PR tree-optimization/57441
6856 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6857 Don't limit size of incr_vec to number of candidates.
6859 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
6861 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6862 and mips16 directories.
6863 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
6864 (MULTILIB_DIRNAMES): Ditto.
6865 (MULTILIB_EXCEPTIONS): Add new exceptions.
6866 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6867 (MULTILIB_DIRNAMES): Ditto.
6868 (MULTILIB_EXCEPTIONS): Add new exceptions.
6870 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6871 Marcus Shawcroft <marcus.shawcroft@arm.com>
6873 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6874 SYMBOL_TINY_ABSOLUTE.
6875 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6876 SYMBOL_TINY_ABSOLUTE.
6877 (aarch64_expand_mov_immediate): Likewise.
6878 (aarch64_classify_symbol): Likewise.
6879 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6880 Permit SYMBOL_TINY_ABSOLUTE.
6881 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6883 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6884 Marcus Shawcroft <marcus.shawcroft@arm.com>
6886 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6887 Refactor if/switch. Replace gcc_assert with if.
6889 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6891 * config/i386/i386.c (initial_ix86_tune_features): Enable
6892 FP Reassociation for AMD bdver1 and bdver2.
6894 2013-05-29 Martin Jambor <mjambor@suse.cz>
6896 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6897 and IMAGPART_EXPR do not occur within other handled_components.
6899 2013-05-29 Richard Biener <rguenther@suse.de>
6901 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6902 access on whether the use is in the BB we currently try to
6904 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6905 vectorize to vect_bb_slp_scalar_cost.
6907 2013-05-29 Richard Biener <rguenther@suse.de>
6909 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6910 computing scalar cost offsetted by stmts that are kept live
6912 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6913 for computation of scalar cost.
6915 2013-05-28 Steve Ellcey <sellcey@mips.com>
6917 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6919 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6921 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6922 array notation built-in reduction functions.
6923 * doc/passes.texi (Passes): Added documentation about changes done
6925 * doc/invoke.texi (C Dialect Options): Added documentation about
6926 the -fcilkplus flag.
6927 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6928 (BUILTINS_DEF): Depend on cilkplus.def.
6929 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6930 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6931 * cilkplus.def: New file.
6933 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6935 PR rtl-optimization/57439
6936 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6938 2013-05-28 Easwaran Raman <eraman@google.com>
6940 PR tree-optimization/57337
6941 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6942 (find_insert_point): Correctly identify the insertion point
6943 when two statements with the same UID is compared.
6945 2013-05-28 Richard Biener <rguenther@suse.de>
6947 PR tree-optimization/56787
6948 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6949 from the list of data references.
6950 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6952 (vect_analyze_loop_operations): Likewise.
6953 (vect_transform_loop): Remove clobbers.
6955 2013-05-28 Martin Jambor <mjambor@suse.cz>
6957 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6958 and REALPART_EXPRs have scalar type.
6960 2013-05-28 Richard Biener <rguenther@suse.de>
6962 PR tree-optimization/57411
6963 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6965 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6966 virtual operand propagation.
6968 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6970 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6971 destination register for bmasksi_vis.
6972 (vector_init_bshuffle): Likewise.
6973 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6975 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6977 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6978 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6979 mode if the instruction isn't available in the original mode.
6980 * config/sparc/sparc.opt (mfix-ut699): New option.
6981 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6982 (divdf3): Turn into expander.
6983 (divdf3_nofix): New insn.
6984 (divdf3_fix): Likewise.
6985 (divsf3): Disable if -mfix-ut699.
6986 (sqrtdf2): Turn into expander.
6987 (sqrtdf2_nofix): New insn.
6988 (sqrtdf2_fix): Likewise.
6989 (sqrtsf2): Disable if -mfix-ut699.
6991 2013-05-27 Richard Biener <rguenther@suse.de>
6994 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
6995 block for the new loop.
6997 2013-05-27 Richard Biener <rguenther@suse.de>
6999 PR tree-optimization/57343
7000 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7001 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7002 (number_of_iterations_cond): Do not build the folded tree.
7004 2013-05-27 Richard Biener <rguenther@suse.de>
7008 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7009 OEP_CONSTANT_ADDRESS_OF retained.
7011 PR tree-optimization/57417
7012 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7014 (set_ssa_val_to): Compare addresses using
7015 get_addr_base_and_unit_offset.
7017 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7019 PR rtl-optimization/56833
7020 * postreload.c (move2add_record_mode): New function.
7021 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7022 (move2add_use_add2_insn): Use move2add_record_sym_value.
7023 (move2add_use_add3_insn): Likewise.
7024 (reload_cse_move2add): Use move2add_valid_value_p and
7025 move2add_record_mode. Invalidate call-clobbered and REG_INC
7026 affected regs by setting reg_mode to VOIDmode.
7027 (move2add_note_store): Don't pretend the inside of a SUBREG is
7028 the actual destination. Invalidate single/leading registers by
7029 setting reg_mode to VOIDmode.
7030 Use move2add_record_sym_value, move2add_valid_value_p and
7031 move2add_record_mode.
7033 2013-05-27 Richard Biener <rguenther@suse.de>
7035 PR tree-optimization/57396
7036 * tree-affine.c (double_int_constant_multiple_p): Properly
7037 return false for val == 0 and div != 0.
7039 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7041 * config/mips/mips.h: Use #elif in preprocessor conditions.
7043 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7046 * config/mips/constraints.md (kl): New constraint.
7047 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7048 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7049 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7051 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7053 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7056 * config/mips/mips.c (mips_can_inline_p): New function.
7057 (TARGET_CAN_INLINE_P): Define.
7059 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7061 * sched-int.h (ds_t, dw_t): Make unsigned int.
7062 Fix documentation that describes how all the ds_t bits are used.
7063 Reserve the last bit for delayed-branch scheduling.
7064 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7065 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7066 (gen_dep_weak_1): Remove prototype.
7067 * sched-deps.c (get_dep_weak_1): Make static.
7068 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7069 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7070 * doc/tm.texi: Regenerate.
7071 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7073 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7076 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7078 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7079 Sandra Loosemore <sandra@codesourcery.com>
7081 * config.gcc (powerpc-*): Allow native for with-cpu.
7083 2013-05-24 Jeff Law <law@redhat.com>
7085 PR tree-optimization/57124
7086 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7087 conversion feeding a condition if the range has an overflow
7088 if -fstrict-overflow. Add warnings for when we do make the
7091 2013-05-24 Dehao Chen <dehao@google.com>
7093 * tree-cfg.c (locus_discrim_map): Fix the typo.
7094 (locus_discrim_hasher): Likewise.
7095 (locus_discrim_hasher::hash): Likewise.
7096 (locus_discrim_hasher::equal): Likewise.
7098 2013-05-24 Martin Jambor <mjambor@suse.cz>
7100 PR tree-optimization/57294
7101 * cgraph.h (ipa_record_stmt_references): Declare.
7102 * cgraphbuild.c (ipa_record_stmt_references): New function.
7103 (build_cgraph_edges): Use ipa_record_stmt_references.
7104 (rebuild_cgraph_edges): Likewise.
7105 (cgraph_rebuild_references): Likewise.
7106 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7107 associated with the old statement and build references from the
7108 newly built statements.
7109 * ipa-ref.c (ipa_remove_stmt_references): New function.
7110 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7112 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7114 * lra-constraints.c (emit_spill_move): Use smaller mode for
7116 (check_and_process_move): Consider mem-reg moves for secondary
7118 (curr_insn_transform): Don't lose insns emitted before for
7119 secondary memory moves.
7120 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7121 reg set up in the current insn.
7123 2013-05-24 Dehao Chen <dehao@google.com>
7125 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7127 (locus_descrim_hasher::equal): Likewise.
7128 (build_gimple_cfg): New discriminator assignment algorithm.
7129 (make_edges): Likewise.
7130 (next_discriminator_for_locus): Likewise.
7131 (same_line_p): Likewise.
7132 (assign_discriminators): Likewise.
7133 (make_cond_expr_edges): Likewise.
7134 (make_gimple_switch_edges): Likewise.
7135 (make_goto_expr_edges): Likewise.
7136 (make_gimple_asm_edges): Likewise.
7138 2013-05-24 Ian Bolton <ian.bolton@arm.com>
7140 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
7141 X format specifier to only display bottom 16 bits.
7142 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
7143 immediate to match for operand 2, since it will be masked.
7145 2013-05-24 Richard Biener <rguenther@suse.de>
7147 PR tree-optimization/57287
7148 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
7149 all SSA names that occur in abnormal PHIs.
7151 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
7154 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
7155 that index is not negative.
7157 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
7159 PR rtl-optimization/55177
7160 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
7161 (simplify_byte_swapping_operation): New.
7162 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
7163 (simplify_relational_operation_1): Deal with BSWAP.
7165 2013-05-23 Richard Henderson <rth@redhat.com>
7168 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
7169 (ix86_reorg): Call it.
7171 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
7174 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
7175 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
7176 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
7178 2013-05-23 Christian Bruel <christian.bruel@st.com>
7181 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
7183 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7184 Marcus Shawcroft <marcus.shawcroft@arm.com>
7186 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
7187 * config/aarch64/constraints.md (Usa): Remove.
7188 * doc/md.texi (AArch64 Usa): Remove.
7190 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7191 Marcus Shawcroft <marcus.shawcroft@arm.com>
7193 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
7194 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
7195 * config/aarch64/predicates.md (aarch64_const_address): Remove.
7196 (aarch64_mov_operand): Use aarch64_mov_operand_p.
7198 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
7200 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
7201 instruction (AdvSIMD).
7202 * config/aarch64/aarch64-builtins.c
7203 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
7204 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
7206 2013-05-23 Martin Jambor <mjambor@suse.cz>
7209 * tree.h (contains_bitfld_component_ref_p): Declare.
7210 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
7211 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
7213 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
7214 not access a bit-field. Assert all final offsets are byte-aligned.
7216 2013-05-23 Richard Biener <rguenther@suse.de>
7218 PR tree-optimization/57380
7219 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
7220 least one invariant or re-used load.
7221 * passes.c (init_optimization_passes): Move pass_phiprop before
7224 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
7226 * config/aarch64/aarch64-simd.md
7227 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
7229 2013-05-23 Richard Biener <rguenther@suse.de>
7232 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7233 OEP_CONSTANT_ADDRESS_OF retained.
7235 2013-05-23 Jakub Jelinek <jakub@redhat.com>
7238 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
7239 don't lower unit. Handle unit not being always BITS_PER_WORD.
7241 2013-05-23 Richard Biener <rguenther@suse.de>
7243 PR rtl-optimization/57341
7244 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
7245 instead of true_dependence.
7247 2013-05-22 David Malcolm <dmalcolm@redhat.com>
7249 * bb-reorder.c (branch_threshold): Make const.
7250 (exec_threshold): Ditto.
7252 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7253 Pat Haugen <pthaugen@us.ibm.com>
7254 Peter Bergner <bergner@vnet.ibm.com>
7256 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
7257 documentation for the power8 crypto builtins.
7259 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
7261 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
7262 macros for defining power8 builtin functions.
7263 (BU_P8V_AV_2): Likewise.
7264 (BU_P8V_AV_P): Likewise.
7265 (BU_P8V_VSX_1): Likewise.
7266 (BU_P8V_OVERLOAD_1): Likewise.
7267 (BU_P8V_OVERLOAD_2): Likewise.
7268 (BU_CRYPTO_1): Likewise.
7269 (BU_CRYPTO_2): Likewise.
7270 (BU_CRYPTO_3): Likewise.
7271 (BU_CRYPTO_OVERLOAD_1): Likewise.
7272 (BU_CRYPTO_OVERLOAD_2): Likewise.
7273 (XSCVSPDP): Fix typo, point to the correct instruction.
7274 (VCIPHER): Add power8 crypto builtins.
7275 (VCIPHERLAST): Likewise.
7276 (VNCIPHER): Likewise.
7277 (VNCIPHERLAST): Likewise.
7278 (VPMSUMB): Likewise.
7279 (VPMSUMH): Likewise.
7280 (VPMSUMW): Likewise.
7281 (VPERMXOR_V2DI): Likewise.
7282 (VPERMXOR_V4SI: Likewise.
7283 (VPERMXOR_V8HI: Likewise.
7284 (VPERMXOR_V16QI: Likewise.
7285 (VSHASIGMAW): Likewise.
7286 (VSHASIGMAD): Likewise.
7288 (VPERMXOR): Likewise.
7289 (VSHASIGMA): Likewise.
7291 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7292 __CRYPTO__ if the crypto instructions are available.
7293 (altivec_overloaded_builtins): Add support for overloaded power8
7296 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
7297 support for power8 crypto builtins.
7298 (builtin_function_type): Likewise.
7299 (altivec_init_builtins): Add support for builtins that take vector
7300 long long (V2DI) arguments.
7302 * config/rs6000/crypto.md: New file, define power8 crypto
7305 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7306 Pat Haugen <pthaugen@us.ibm.com>
7307 Peter Bergner <bergner@vnet.ibm.com>
7309 * doc/invoke.texi (Option Summary): Add power8 options.
7310 (RS/6000 and PowerPC Options): Likewise.
7312 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
7313 constraints.md instead of rs6000.h. Reorder w* constraints. Add
7314 wm, wn, wr documentation.
7316 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
7317 registers if direct move instructions are enabled.
7318 (wn): New constraint for no registers.
7319 (wq): New constraint for quad word even GPR registers.
7320 (wr): New constraint if 64-bit instructions are enabled.
7321 (wv): New constraint if power8 vector instructions are enabled.
7322 (wQ): New constraint for quad word memory locations.
7324 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
7325 constraint for 0..15 for crypto instructions.
7326 (gpc_reg_operand): If VSX allow registers in VSX registers as well
7327 as GPR and floating point registers.
7328 (int_reg_operand): New predicate to match only GPR registers.
7329 (base_reg_operand): New predicate to match base registers.
7330 (quad_int_reg_operand): New predicate to match even GPR registers
7331 for quad memory operations.
7332 (vsx_reg_or_cint_operand): New predicate to allow vector logical
7333 operations in both GPR and VSX registers.
7334 (quad_memory_operand): New predicate for quad memory operations.
7335 (reg_or_indexed_operand): New predicate for direct move support.
7337 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
7338 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
7339 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
7340 (POWERPC_MASKS): Add power8 options.
7341 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
7344 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7345 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
7347 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
7348 (-mpower8-fusion): New power8 options.
7349 (-mpower8-fusion-sign): Likewise.
7350 (-mpower8-vector): Likewise.
7351 (-mcrypto): Likewise.
7352 (-mdirect-move): Likewise.
7353 (-mquad-memory): Likewise.
7355 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
7357 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
7359 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
7360 (rs6000_debug_vector_unit): Add p8_vector.
7361 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
7362 definitions. Also print fusion state.
7363 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
7364 (rs6000_builtin_mask_calculate): Add power8 builtin support.
7365 (rs6000_option_override_internal): Add support for power8.
7366 (rs6000_common_init_builtins): Add debugging for skipped builtins
7368 (rs6000_adjust_cost): Add power8 support.
7369 (rs6000_issue_rate): Likewise.
7370 (insn_must_be_first_in_group): Likewise.
7371 (insn_must_be_last_in_group): Likewise.
7372 (force_new_group): Likewise.
7373 (rs6000_register_move_cost): Likewise.
7374 (rs6000_opt_masks): Likewise.
7376 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
7377 power8 capable assembler, default to power7 options.
7378 (TARGET_DIRECT_MOVE): Likewise.
7379 (TARGET_CRYPTO): Likewise.
7380 (TARGET_P8_VECTOR): Likewise.
7381 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
7382 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
7383 (VECTOR_MEM_P8_VECTOR_P): Likewise.
7384 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
7385 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
7386 (TARGET_XSCVDPSPN): Likewise.
7387 (TARGET_XSCVSPDPN): Likewsie.
7388 (TARGET_SYNC_HI_QI): Likewise.
7389 (TARGET_SYNC_TI): Likewise.
7390 (MASK_CRYPTO): Likewise.
7391 (MASK_DIRECT_MOVE): Likewise.
7392 (MASK_P8_FUSION): Likewise.
7393 (MASK_P8_VECTOR): Likewise.
7394 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7395 temporary used by some of the direct move instructions to get two FP
7396 temporary registers does not force creation of a stack frame.
7397 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7398 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7399 that any VSX registers are tieable, even if they are also an
7400 Altivec vector mode.
7401 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7402 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7403 (RS6000_BTM_CRYPTO): Likewise.
7404 (RS6000_BTM_COMMON): Likewise.
7406 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7407 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7408 (enum rs6000_vector): Add power8 vector support.
7410 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7414 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7415 (any_sibcall_could_use_r3): this and handle indirect calls.
7416 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7418 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7420 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7422 2013-05-22 Richard Biener <rguenther@suse.de>
7425 * profile.c (branch_prob): Do not split blocks that are
7426 abnormally receiving from ECF_RETURNS_TWICE functions.
7428 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7430 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7431 address mode. Move pointer mode initialization to the same place.
7433 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7435 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7436 while it has any effect.
7438 2013-05-21 Easwaran Raman <eraman@google.com>
7440 PR tree-optimization/57322
7441 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7442 UID of the statement added to the BB to be 1.
7444 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7446 PR tree-optimization/57331
7447 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7448 of conversion from pointer type to integral type with integer.
7450 2013-05-21 Martin Jambor <mjambor@suse.cz>
7453 * ipa-prop.c (ipa_read_node_info): Process param_used and
7454 controlled_uses in the same order as when writing.
7456 2013-05-21 Magnus Granberg <baldrick@free.fr>
7459 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7461 2013-05-21 Richard Biener <rguenther@suse.de>
7463 PR tree-optimization/57318
7464 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7465 estimate stmts with side-effects as likely eliminated.
7467 2013-05-21 Richard Biener <rguenther@suse.de>
7469 PR tree-optimization/57330
7470 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7471 preserve the call stmts fntype.
7473 2013-05-21 Richard Biener <rguenther@suse.de>
7475 PR tree-optimization/57303
7476 * tree-ssa-sink.c (statement_sink_location): Improve killing
7477 stmt detection and properly handle self-assignments.
7479 2013-05-21 Christian Bruel <christian.bruel@st.com>
7481 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7482 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7483 registers. Set register size out of the PARALLEL loop.
7485 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7488 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7489 (*fmasf4, *fmasf4_media): New insns.
7491 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7493 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7494 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7495 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7496 (mips_idiv_insns): Update the comments to say that the returned
7497 instruction counts are in units of BASE_INSN_LENGTH.
7498 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7499 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7500 using 2 rather than 4 as the length of indirect MIPS16 and
7501 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7502 length of a NOP. Don't divide MIPS16 lengths by 2.
7503 (mips16_split_long_branches): Assume a branch is long if the
7504 length is greater than 4 rather than 8.
7505 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7506 rather than multiplying them by 2. Multiply instruction counts
7507 by BASE_INSN_LENGTH rather than 4.
7508 (*jump_mips16, tls_get_tp_mips16_<mode>)
7509 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7511 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7513 * config/mips/mips.md (extended_mips16): Remove branch case.
7514 (length): Remove duplicated extended_mips16 test.
7516 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7518 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7520 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7522 * recog.h (Recog_data): Rename to...
7523 (recog_data_d): ...this.
7524 (recog_data): Update accordingly.
7525 * recog.c (recog_data): Likewise.
7526 * reload.c (save_recog_data): Likewise.
7527 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7528 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7530 2013-05-17 Julian Brown <julian@codesourcery.com>
7532 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7533 found in a REG_EQUAL note, invalidate it.
7535 2013-05-17 Easwaran Raman <eraman@google.com>
7537 * tree-ssa-reassoc.c (find_insert_point): New function.
7538 (insert_stmt_after): Likewise.
7539 (get_def_stmt): Likewise.
7540 (ensure_ops_are_available): Likewise.
7541 (not_dominated_by): Likewise.
7542 (rewrite_expr_tree): Do not move statements beyond what is
7543 necessary. Remove call to swap_ops_for_binary_stmt...
7544 (reassociate_bb): ... and move it here.
7545 (build_and_add_sum): Assign UIDs for new statements.
7546 (linearize_expr): Likewise.
7547 (do_reassoc): Renumber gimple statement UIDs.
7549 2013-05-17 Jan Hubicka <jh@suse.cz>
7551 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7553 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
7554 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7556 (output_weakrefs): Update.
7558 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7559 Martin Jambor <mjambor@suse.cz>
7562 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7563 value that corresponds to the given aggval is found in values vector.
7565 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
7567 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7568 sse, sse2, sse3, ssse3 and sse4a flags to options.
7570 2013-05-17 David Malcolm <dmalcolm@redhat.com>
7572 * gengtype-state.c: (s_expr_writer): New class, to handle
7573 prettifying of output layout of s-expressions.
7574 (state_writer): New class, to write out gtype.state.
7575 (state_written_type_count): Move this variable into member data of
7577 (s_expr_writer::s_expr_writer): New code: constructor for new class
7578 (state_writer::state_writer(): ditto
7579 (s_expr_writer::write_new_line): New function
7580 (s_expr_writer::write_any_indent): ditto
7581 (s_expr_writer::begin_s_expr): ditto
7582 (s_expr_writer::end_s_expr): ditto
7583 (write_state_fileloc): convert to method of state_writer...
7584 (state_writer:: write_state_fileloc): ...and use methods of
7585 s_expr_writer to write indentation into the gtype.state output file
7586 to visually represent the hierarchical structure of the list
7588 (write_state_fields): ditto, renaming to...
7589 (state_writer::write_state_fields)
7590 (write_state_a_string): ditto, renaming to...
7591 (state_writer::write_state_a_string)
7592 (write_state_string_option): ditto, renaming to...
7593 (state_writer::write_state_string_option)
7594 (write_state_type_option): ditto, renaming to...
7595 (state_writer::write_state_type_option)
7596 (write_state_nested_option): ditto, renaming to...
7597 (state_writer::write_state_nested_option)
7598 (write_state_option): ditto, renaming to...
7599 (state_writer::write_state_option)
7600 (write_state_options): ditto, renaming to...
7601 (state_writer::write_state_options)
7602 (write_state_lang_bitmap): ditto, renaming to...
7603 (state_writer::write_state_lang_bitmap)
7604 (write_state_version): ditto, renaming to...
7605 (state_writer::write_state_version)
7606 (write_state_scalar_type): ditto, renaming to...
7607 (state_writer::write_state_scalar_type)
7608 (write_state_string_type): ditto, renaming to...
7609 (state_writer::write_state_string_type)
7610 (write_state_undefined_type): ditto, renaming to...
7611 (state_writer::write_state_undefined_type)
7612 (write_state_struct_union_type): ditto, renaming to...
7613 (state_writer::write_state_struct_union_type)
7614 (write_state_struct_type): ditto, renaming to...
7615 (state_writer::write_state_struct_type)
7616 (write_state_user_struct_type): ditto, renaming to...
7617 (state_writer::write_state_user_struct_type)
7618 (write_state_lang_struct_type): ditto, renaming to...
7619 (state_writer::write_state_lang_struct_type)
7620 (write_state_param_struct_type): ditto, renaming to...
7621 (state_writer::write_state_param_struct_type)
7622 (write_state_pointer_type): ditto, renaming to...
7623 (state_writer::write_state_pointer_type)
7624 (write_state_array_type): ditto, renaming to...
7625 (state_writer::write_state_array_type)
7626 (write_state_gc_used): ditto, renaming to...
7627 (state_writer::write_state_gc_used)
7628 (write_state_common_type_content): ditto, renaming to...
7629 (state_writer::write_state_common_type_content)
7630 (write_state_type): ditto, renaming to...
7631 (state_writer::write_state_type)
7632 (write_state_pair_list): ditto, renaming to...
7633 (state_writer::write_state_pair_list)
7634 (write_state_pair): ditto, renaming to...
7635 (state_writer::write_state_pair)
7636 (write_state_typedefs): ditto, renaming to...
7637 (state_writer::write_state_typedefs)
7638 (write_state_structures): ditto, renaming to...
7639 (state_writer::write_state_structures)
7640 (write_state_param_structs): ditto, renaming to...
7641 (state_writer::write_state_param_structs)
7642 (write_state_variables): ditto, renaming to...
7643 (state_writer::write_state_variables)
7644 (write_state_srcdir): ditto, renaming to...
7645 (state_writer::write_state_srcdir)
7646 (write_state_files_list): ditto, renaming to...
7647 (state_writer::write_state_files_list)
7648 (write_state_languages): ditto, renaming to...
7649 (state_writer::write_state_languages)
7650 (write_state): create a state_writer instance and use it when
7651 writing out the state file
7653 2013-05-17 Mike Stump <mikestump@comcast.net>
7655 PR rtl-optimization/57304
7656 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7657 accessing DF_REF_REAL_LOC.
7659 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7661 PR rtl-optimization/57281
7662 PR rtl-optimization/57300
7663 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7664 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7665 what the other splitter did if the registers are dead.
7667 2013-05-17 Richard Biener <rguenther@suse.de>
7669 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7672 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7674 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7677 2013-05-17 Marek Polacek <polacek@redhat.com>
7679 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7680 length when doing non-zero store of storing '\0' to '\0'.
7682 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7684 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7685 vect_external_def oprnd1 with loop_vinfo, try to emit
7686 optional cast, negation and and stmts on the loop preheader
7687 edge instead of into the pattern def seq.
7689 PR tree-optimization/57051
7690 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7691 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7693 2013-05-16 Nick Clifton <nickc@redhat.com>
7695 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7696 (rl78_is_naked_func): New function.
7697 (rl78_expand_prologue): Skip prologue generation for naked functions.
7698 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
7699 * doc/extend.texi (naked): Add RL78 to the list of processors
7700 that supports this attribute.
7702 2013-05-16 Jeff Law <law@redhat.com>
7704 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7706 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7708 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7709 cache parameters using detect_caches_amd also for CYRIX,
7710 NSC and TM2 signatures.
7712 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7713 Dzianis Kahanovich <mahatma@eu.by>
7717 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7718 VIA/Centaur processors and determine their cache parameters
7719 using detect_caches_amd.
7721 2013-05-16 Teresa Johnson <tejohnson@google.com>
7723 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7724 (rtl_verify_edges): New function.
7725 (rtl_verify_bb_insns): Ditto.
7726 (rtl_verify_bb_pointers): Ditto.
7727 (rtl_verify_bb_insn_chain): Ditto.
7728 (rtl_verify_fallthru): Ditto.
7729 (rtl_verify_bb_layout): Ditto.
7730 (rtl_verify_flow_info_1): Outline checks into new functions.
7731 (rtl_verify_flow_info): Ditto.
7733 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
7735 * cfghooks.c (copy_bbs): Add update_dominance argument.
7736 * cfghooks.h (copy_bbs): Update prototype.
7737 * tree-cfg.c (gimple_duplicate_sese_region):
7738 Add update_dominance argument.
7739 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7740 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7741 gimple_duplicate_sese_region call.
7742 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7743 Update copy_bbs call.
7744 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7745 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7747 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7749 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7750 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7751 vect_recog_rotate_pattern.
7752 (vect_recog_rotate_pattern): New function.
7754 2013-05-16 Jason Merrill <jason@redhat.com>
7756 * Makefile.in (LLINKER): New variable.
7757 (mostlyclean): Remove link mutex.
7758 * configure.ac: Handle --enable-link-mutex.
7759 * lock-and-run.sh: New script.
7761 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7764 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7767 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7769 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7771 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7773 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7774 * config/arm/arm.c (next_consecutive_mem): New function.
7775 (gen_movmem_ldrd_strd): Likewise.
7776 * config/arm/arm.md (movmemqi): Update condition and code.
7777 (unaligned_loaddi, unaligned_storedi): New patterns.
7779 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7781 * config.gcc: Obsolete *-*-solaris2.9*.
7782 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7784 2013-05-16 Richard Biener <rguenther@suse.de>
7786 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7787 earlier, after GRAPHITE transforms and IV canonicalization.
7789 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7791 * omp-low.c (extract_omp_for_data): For collapsed loops,
7792 if at least one of the loops is known at compile time to
7793 iterate zero times, set count to 0.
7794 (expand_omp_regimplify_p): New function.
7795 (expand_omp_for_generic): For collapsed loops, if at least
7796 one of the loops isn't known to iterate at least once,
7797 add runtime check with setting count to 0.
7798 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7799 For unsigned types if it isn't known at compile time that
7800 the loop will iterate at least once, add runtime check to bypass
7801 the whole loop if initial condition isn't true.
7803 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7805 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7807 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
7810 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7811 transformations to avoid an infinite loop.
7813 2013-05-16 Marek Polacek <polacek@redhat.com>
7815 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7817 2013-05-15 Leif Ekblad <leif@rdos.net>
7819 * config/i386/i386.c (ix86_decompose_address): Use
7820 DEFAULT_TLS_SEG_REG to access TLS segment register.
7821 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7822 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7823 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7825 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7828 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7829 sibling calls to functions that would normally be lazily bound,
7830 unless $gp is call-clobbered.
7832 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
7834 * config/i386/i386.c (ix86_option_override_internal): Update
7835 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
7836 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7837 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
7838 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7840 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7842 2013-05-15 Andreas Schwab <schwab@suse.de>
7844 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7845 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7848 2013-05-15 Teresa Johnson <tejohnson@google.com>
7850 * loop-unroll.c (report_unroll_peel): Check decision before
7851 emitting unroll/peel message.
7853 2013-05-15 Teresa Johnson <tejohnson@google.com>
7855 * function.h (has_bb_partition): New rtl_data flag.
7856 (bb_reorder_complete): Ditto.
7857 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7858 instead of flag_reorder_blocks_and_partition.
7859 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7860 with some enhancements.
7861 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7862 * bb-reorder.c (connect_traces): Check for has_bb_partition
7863 instead of flag_reorder_blocks_and_partition.
7864 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7865 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7866 verify_hot_cold_block_grouping.
7867 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7869 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7872 * config/arm/predicates.md (call_insn_operand): New predicate.
7873 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7874 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7875 if insn is not a tail call.
7876 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7878 * config/arm/arm.h (enum reg_class): New caller save register class.
7879 (REG_CLASS_NAMES): Likewise.
7880 (REG_CLASS_CONTENTS): Likewise.
7881 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7884 2013-05-15 Richard Biener <rguenther@suse.de>
7886 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7887 of MSG_OPTIMIZED_LOCATIONS.
7888 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7889 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7891 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7892 of MSG_OPTIMIZED_LOCATIONS.
7893 (execute_vect_slp): Likewise.
7894 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7895 (vect_create_cond_for_alias_checks): Likewise.
7896 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7897 (vect_recog_widen_mult_pattern): Likewise.
7898 (vect_recog_widen_sum_pattern): Likewise.
7899 (vect_recog_over_widening_pattern): Likewise.
7900 (vect_recog_widen_shift_pattern): Likewise.
7901 (vect_recog_vector_vector_shift_pattern): Likewise.
7902 (vect_recog_divmod_pattern): Likewise.
7903 (vect_recog_mixed_size_cond_pattern): Likewise.
7904 (vect_recog_bool_pattern): Likewise.
7905 (vect_pattern_recog_1): Likewise.
7907 2013-05-15 Martin Jambor <mjambor@suse.cz>
7909 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7910 non-functions to builtin_unreachable.
7911 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7912 correct when new direct edges were discovered.
7914 2013-05-15 Martin Jambor <mjambor@suse.cz>
7916 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7917 header, print symbol order instead of node uid, print more information
7918 about indirect edge targets.
7919 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7921 (ipa_make_edge_direct_to_target): Likewise.
7922 (remove_described_reference): Likewise.
7923 (propagate_controlled_uses): Likewise.
7924 (ipa_print_node_params): Also print symbol order.
7925 (ipcp_transform_function): Print symbol order instead of node uids.
7926 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7927 (cgraph_get_create_real_symbol_node): Likewise.
7928 * ipa-cp.c (print_lattice): Likewise.
7929 (print_all_lattices): Likewise.
7930 (determine_versionability): Likewise.
7931 (initialize_node_lattices): Likewise.
7932 (estimate_local_effects): Likewise.
7933 (update_profiling_info): Likewise.
7934 (create_specialized_node): Likewise.
7935 (perhaps_add_new_callers): Likewise.
7936 (decide_about_value): Likewise.
7937 (decide_whether_version_node): Likewise.
7938 (identify_dead_nodes): Likewise.
7939 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7940 (dump_inline_summary): Likewise.
7941 (estimate_node_size_and_time): Likewise.
7942 (inline_analyze_function): Likewise.
7943 * ipa-inline.c (report_inline_failed_reason): Likewise.
7944 (want_early_inline_function_p): Likewise.
7945 (edge_badness): Likewise.
7946 (update_edge_key): Likewise.
7947 (inline_small_functions): Likewise. Add dumping of order to two other
7949 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7950 instead of node uids.
7951 (propagate_pure_const): Likewise.
7952 (propagate_pure_const): Likewise.
7953 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7954 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7955 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7957 * tree-pretty-print.c (dump_function_header): Likewise.
7958 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7959 Print symbol order instead of node uids.
7961 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7963 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7964 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7966 2013-05-15 Richard Biener <rguenther@suse.de>
7968 PR tree-optimization/57275
7969 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7970 return value for fail to do runtime alias checks for gather loads.
7972 2013-05-15 Jan Hubicka <jh@suse.cz>
7976 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7977 weakrefs are not external.
7978 (lto_symtab_merge_decls): Fix thinko when dealing with
7979 non-lto_symtab decls.
7980 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7981 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7982 * varpool.c (dump_varpool_node): Dump more flags.
7984 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7986 * config/i386/i386.c (processor_alias_table): Add instruction
7987 FSGSBASE for AMD bdver3 architecture.
7989 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7991 * tree.c (warn_deprecated_use): Print file:line using locus color.
7992 * diagnostic.c (diagnostic_report_current_module): Print file:line
7993 and file:line:column using locus color.
7995 2013-05-14 Mike Stump <mikestump@comcast.net>
7997 * gdbinit.in: Add __null.
7999 2013-05-14 Mike Stump <mikestump@comcast.net>
8001 * recog.h: Rename struct recog_data to Recog_data.
8002 * recog.c: Likewise.
8003 * reload.c (can_reload_into): Likewise.
8004 * config/picochip/picochip.c: Likewise.
8006 2013-05-14 Mike Stump <mikestump@comcast.net>
8008 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8010 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8012 * resource.h (struct resources): Remove unch_memory member.
8013 (CLEAR_RESOURCE): Don't clear unch_memory.
8014 * resource.c (mark_referenced_resources): Don't set it.
8015 (mark_set_resources): Likewise.
8016 (mark_target_live_regs): Don't clear it.
8017 (init_resource_info): Likewise.
8018 * reorg.c (resource_conflicts_p): Don't compare it.
8019 (redundant_insn): Don't set it.
8021 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8023 * emit-rtl.c (next_label): Remove unused function.
8024 (skip_consecutive_labels, link_cc0_insns): Move to ...
8025 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8026 only place where these functions are used, and make them static.
8028 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8030 * fold-const.c (fold_negate_expr): Handle vectors.
8031 (fold_truth_not_expr): Make it static.
8032 (fold_invert_truthvalue): New static function.
8033 (invert_truthvalue_loc): Handle vectors. Do not call
8034 fold_truth_not_expr directly.
8035 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8036 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8037 (fold_comparison): Handle vector constants.
8038 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8039 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8040 * tree.h (fold_truth_not_expr): Remove declaration.
8042 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8044 * config/aarch64/aarch64-simd.md
8045 (aarch64_vcond_internal<mode>): Rename to...
8046 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8047 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8048 float modes. Clarify all iterator modes.
8049 (vcond<mode><mode>): Use new name for vcond expanders.
8050 (vcond<v_cmp_result><mode>): Likewise.
8051 (vcondu<mode><mode>: Likewise.
8052 * config/aarch64/iterators.md (VDQF_COND): New.
8054 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8057 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8058 variable for the shift amount. Check that we shift by non-negative
8061 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8064 * config/arm/arm.h (EPILOGUE_USES): Only return true
8065 for LR_REGNUM after epilogue_completed.
8067 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8069 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8072 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8075 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8076 and Solaris 11+/x86 with gld.
8077 * configure: Regenerate.
8079 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8081 * expmed.c (expand_shift_1): Canonicalize rotates by
8082 constant bitsize / 2 to bitsize - 1.
8083 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8084 case ROTATERT>: Likewise.
8087 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8089 * config/i386/i386.md (rotateinv): New code attr.
8090 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8091 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8092 roll $31, %eax, etc.
8094 2013-05-14 Richard Biener <rguenther@suse.de>
8097 * tree-eh.c (sink_clobbers): Give up for successors with
8098 multiple predecessors and no virtual uses.
8100 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8102 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8103 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8105 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8108 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8109 the case when both op0 and op1 have VOIDmode.
8111 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8113 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8114 in multiply-accumulate mode.
8116 2013-05-13 Guozhi Wei <carrot@google.com>
8118 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8120 2013-05-13 Kai Tietz <ktietz@redhat.com>
8123 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8124 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8125 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8126 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8127 (ix86_expand_prologue): Likewise.
8128 (ix86_expand_split_stack_prologue): Likewise.
8129 (legitimate_pic_address_disp_p): Likewise.
8130 (legitimize_pic_address): Likewise.
8131 (legitimize_tls_address): Likewise.
8132 (legitimize_pe_coff_symbol): Likewise.
8133 (output_pic_addr_const): Likewise.
8134 (construct_plt_address): Likewise.
8135 (ix86_expand_call): Likewise.
8136 (x86_output_mi_thunk): Likewise.
8137 (x86_function_profiler): Likewise.
8139 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
8141 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
8142 similar switch cases.
8143 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
8144 (aarch64_simd_mov_to_<mode>low): Delete.
8145 (aarch64_simd_mov_to_<mode>high): Delete.
8146 (move_lo_quad_<mode>): Add w<-r alternative.
8147 (aarch64_simd_move_hi_quad_<mode>): Likewise.
8148 (aarch64_simd_mov_from_*): Update type attribute.
8149 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
8152 2013-05-13 Jan Hubicka <jh@suse.cz>
8154 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
8155 * config/i386/i386.c (ix86_compute_frame_layout,
8156 ix86_expand_epilogue, emit_i387_cw_initialization,
8157 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
8158 ix86_local_alignment): Fix use of size/speed predicates.
8160 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8162 PR tree-optimization/45216
8163 PR tree-optimization/57157
8164 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
8165 the (-Y) & (B - 1) variant if OP is |.
8166 * expmed.c (expand_shift_1): For rotations by const0_rtx just
8167 return shifted. Use (-op1) & (prec - 1) as other_amount
8168 instead of prec - op1.
8170 2013-05-13 Martin Jambor <mjambor@suse.cz>
8173 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
8174 (ipa_constant_data): New type.
8175 (ipa_jump_func): Use ipa_constant_data to hold information about
8176 constant jump functions.
8177 (ipa_get_jf_constant): Adjust to jump function type changes.
8178 (ipa_get_jf_constant_rdesc): New function.
8179 (ipa_param_descriptor): New field controlled_uses.
8180 (ipa_get_controlled_uses): New function.
8181 (ipa_set_controlled_uses): Likewise.
8182 * ipa-ref.h (ipa_find_reference): Declare.
8183 * ipa-prop.c (ipa_cst_ref_desc): New type.
8184 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
8186 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
8187 New parameter cs. Adjust all callers.
8188 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
8189 (remove_described_reference): New function.
8190 (jfunc_rdesc_usable): Likewise.
8191 (try_make_edge_direct_simple_call): Decrement controlled use count,
8192 attempt to remove reference if it hits zero.
8193 (combine_controlled_uses_counters): New function.
8194 (propagate_controlled_uses): Likewise.
8195 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
8196 (ipa_edge_duplication_hook): Duplicate reference descriptions.
8197 (ipa_print_node_params): Print described use counter.
8198 (ipa_write_jump_function): Adjust to jump function type changes.
8199 (ipa_read_jump_function): New parameter CS, pass it to
8200 ipa_set_jf_constant. Adjust caller.
8201 (ipa_write_node_info): Stream controlled use count
8202 (ipa_read_node_info): Likewise.
8203 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
8205 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
8206 count. Remove cloning-added reference if it reaches zero.
8207 * ipa-ref.c (ipa_find_reference): New function.
8209 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8211 * config/i386/i386.c (processor_target_table): Modified default
8212 alignment values for AMD BD and BT architectures.
8214 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
8216 * tree-vect-generic.c (uniform_vector_p): Move ...
8217 * tree.c (uniform_vector_p): ... here.
8218 * tree.h (uniform_vector_p): Declare it.
8219 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
8222 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8224 PR tree-optimization/57230
8225 * tree-ssa-strlen.c (handle_char_store): Record length for
8226 array store from STRING_CST.
8228 PR tree-optimization/57230
8229 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
8232 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
8234 * config/epiphany/epiphany.c (epiphany_init): Check size of
8235 NUM_MODES_FOR_MODE_SWITCHING.
8236 (epiphany_expand_prologue):
8237 Remove CONFIG_REGNUM initial value handling code.
8238 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
8239 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
8240 (emit_set_fp_mode, epiphany_mode_after): Likewise.
8241 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
8242 Don't return 1 for FP_MODE_NONE.
8243 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
8244 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
8245 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
8246 * config/epiphany/epiphany.md (save_config): New pattern.
8248 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
8250 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
8252 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
8254 * config/i386/i386.md (memory): Handle sseishft1.
8255 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
8256 (*vec_extractv2di_1): Ditto.
8258 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
8260 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
8263 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
8265 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
8266 Add mthumb/march=armv7-a multilib.
8267 Add mthumb/march=armv7-r multilib.
8268 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
8270 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
8272 * config/v850/t-rtems: Add more multilibs.
8274 2013-05-10 Richard Biener <rguenther@suse.de>
8276 PR tree-optimization/57214
8277 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
8278 not propagate from SSA names that occur in abnormal PHI nodes.
8280 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
8282 * stor-layout.c (element_precision): New function.
8283 * machmode.h (element_precision): Declare it.
8284 * tree.c (build_minus_one_cst): New function.
8285 (element_precision): Likewise.
8286 * tree.h (build_minus_one_cst): Declare new function.
8287 (element_precision): Likewise.
8288 * fold-const.c (operand_equal_p): Use element_precision.
8289 (fold_binary_loc): Handle vector types.
8290 * convert.c (convert_to_integer): Use element_precision.
8291 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
8294 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
8296 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
8297 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
8298 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
8299 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8300 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
8301 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
8302 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
8303 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8304 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
8305 (Uuw8): New constraints.
8306 (Usb4): Move into alphabetical order.
8307 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
8308 (sd8_operand, ub8_operand, uw8_operand): New predicates.
8309 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
8310 previously unnamed patterns.
8311 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
8312 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
8313 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
8314 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
8315 of set_attr_alternative/if_then_else. Use extended_mips16 instead
8316 of specific lengths.
8318 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8320 * config/i386/i386.md (rotateinv): New code attr.
8321 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8322 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8323 roll $31, %eax, etc.
8325 PR tree-optimization/45216
8326 PR tree-optimization/57157
8327 * tree-ssa-forwprop.c (simplify_rotate): New function.
8328 (ssa_forward_propagate_and_combine): Call it.
8330 2013-05-10 Richard Biener <rguenther@suse.de>
8332 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
8333 disable peeling when we version for aliasing.
8334 (vector_alignment_reachable_p): Honor explicit user alignment.
8335 (vect_supportable_dr_alignment): Likewise.
8336 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
8337 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
8338 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
8339 then peeling to arrange for the cost-model check to come first.
8341 2013-05-10 Alan Modra <amodra@gmail.com>
8343 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
8344 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
8345 * configure: Regenerate.
8347 2013-05-10 Alan Modra <amodra@gmail.com>
8350 * varasm.c (default_elf_select_section): Move !DECL_P check..
8351 (get_named_section): ..to here before calling get_section_name.
8353 (default_section_type_flags): Add DECL_P check.
8354 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
8355 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
8357 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
8359 * config/epiphany/epiphany.c (epiphany_expand_prologue):
8360 When using gen_stack_adjust_str with a register offset, add a
8361 REG_FRAME_RELATED_EXPR note.
8363 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
8365 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
8366 (*vec_extractv4si_zext_mem): Ditto.
8367 (*vec_extractv2di): Add 0->x and x->x alternatives.
8368 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
8369 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
8371 2013-05-09 Jason Merrill <jason@redhat.com>
8373 N3639 C++1y VLA support
8374 * gimplify.c (gimplify_vla_decl): Don't touch an existing
8377 * tree.c (build_constructor_va): New.
8378 * tree.h: Declare it.
8380 2013-05-09 Martin Jambor <mjambor@suse.cz>
8383 * gimple-fold.c (canonicalize_constructor_val): Call
8384 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
8386 2013-05-09 Jan Hubicka <jh@suse.cz>
8387 Richard Biener <rguenther@suse.de>
8390 * symtab.c (symtab_make_decl_local): Do not add private names.
8392 2013-05-09 Jan Hubicka <jh@suse.cz>
8395 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8396 (unlink_from_assembler_name_hash): Likewise.
8397 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8398 symtab_unregister_node, symtab_initialize_asm_name_hash,
8399 change_decl_assembler_name): Update.
8401 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8403 * config/aarch64/aarch64.md: New movtf split.
8404 (*movtf_aarch64): Update.
8405 (aarch64_movdi_tilow): Handle TF modes and rename to
8406 aarch64_movdi_<mode>low.
8407 (aarch64_movdi_tihigh): Handle TF modes and rename to
8408 aarch64_movdi_<mode>high
8409 (aarch64_movtihigh_di): Handle TF modes and rename to
8410 aarch64_mov<mode>high_di
8411 (aarch64_movtilow_di): Handle TF modes and rename to
8412 aarch64_mov<mode>low_di
8413 (aarch64_movtilow_tilow): Remove spurious whitespace.
8414 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8416 (aarch64_print_operand): Update.
8418 2013-05-09 Alan Modra <amodra@gmail.com>
8420 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8422 * configure: Regenerate.
8424 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8426 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8427 splitter preparation statements.
8428 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8429 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8432 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8434 * gimple-ssa-strength-reduction.c (count_candidates): Change
8435 return value to int.
8436 (analyze_candidates_and_replace): Change type of length to int.
8438 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8440 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8441 (*vec_extract<mode>): Use VI12_128 mode iterator.
8442 (*vec_extract<mode>_mem): Ditto.
8443 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8446 2013-05-08 Diego Novillo <dnovillo@google.com>
8451 2012-08-17 Diego Novillo <dnovillo@google.com>
8454 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8455 * config.in: Regenerate.
8456 * configure: Regenerate.
8457 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8459 2013-05-08 Jan Hubicka <jh@suse.cz>
8462 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8463 * cgraph.h (symtab_node_base): Add unique_name.
8464 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8465 input_overwrite_node, input_varpool_node): Stream unique_name.
8466 * cgraphclones.c (cgraph_create_virtual_clone,
8467 cgraph_function_versioning): Set unique_name.
8468 * ipa.c (function_and_variable_visibility): Set unique_name.
8470 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8472 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8473 (alloc_cand_and_find_basis): Restrict conditional candidate
8474 processing to CAND_MULTs.
8476 2013-05-08 Jan Hubicka <jh@suse.cz>
8479 lto-symtab.c (lto_symtab_symbol_p): New function.
8480 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8481 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8482 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8483 Skip static symbols.
8485 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8487 PR tree-optimization/57200
8488 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8489 Only call inform if the preceding warning_at returns true.
8491 2013-05-07 Han Shen <shenhan@google.com>
8493 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8494 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8495 * common.opt (fstack-protector-strong): New option.
8496 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8497 * doc/invoke.texi (Optimization Options): Document
8498 "-fstack-protector-strong".
8499 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8501 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8503 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8505 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8507 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8508 it is not lhs of the insn.
8509 (setup_reg_equiv): Remove insn having equiv memory which it is not
8511 * lra-constraints.c (process_address): Try to improve generation
8512 code for address base + disp.
8513 (lra_constraints): Make correct the code for checking insn setting
8514 up backward equivalence. Remove insn only if it is in the init
8516 * lra-eliminations.c (update_reg_eliminate): Change return value.
8517 (lra_eliminate): Use the result.
8519 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8521 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8522 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8523 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8524 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8525 register target operands.
8526 (*vec_extractv8hi_sse2): New pattern.
8527 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8528 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8529 (*vec_extract<mode>_mem): New insn and split pattern.
8531 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8533 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8534 (TARGET_ASAN_SHADOW_OFFSET): Define.
8535 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8536 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8538 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8540 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8541 (incr_vec_index): Return -1 if increment not found.
8542 (create_add_on_incoming_edge): Assert if increment not found.
8543 (record_increment): Limit number of increments recorded.
8544 (all_phi_incrs_profitable): Return false if an increment not found.
8545 (replace_profitable_candidates): Don't process increments that were
8547 (analyze_candidates_and_replace): Limit size of incr_vec.
8549 2013-05-07 Richard Biener <rguenther@suse.de>
8551 * calls.c (special_function_p): setjmp-like functions are leaf.
8552 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8553 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8555 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8557 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8558 (aarch64_simd_mov<mode>): New expander.
8559 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8560 (aarch64_simd_mov_to_<mode>high): Likewise.
8561 (aarch64_simd_mov_from_<mode>low): Likewise.
8562 (aarch64_simd_mov_from_<mode>high): Likewise.
8563 (aarch64_dup_lane<mode>): Update.
8564 (aarch64_dup_lanedi): New instruction pattern.
8565 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8566 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8568 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8570 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8571 (replace_mult_candidate): Remove unnecessary argument; remove
8572 unnecessary parameter from call to introduce_cast_before_cand.
8573 (replace_unconditional_candidate): Remove unnecessary parameter
8574 from call to replace_mult_candidate.
8575 (replace_conditional_candidate): Likewise.
8576 (insert_initializers): Use make_temp_ssa_name.
8577 (introduce_cast_before_cand): Remove unnecessary argument; use
8579 (replace_one_candidate): Remove unnecessary argument; remove
8580 unnecessary parameter from calls to introduce_cast_before_cand.
8581 (replace_profitable_candidates): Remove unnecessary parameters
8582 from calls to replace_one_candidate.
8584 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8586 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8587 phi def as possibly hiding a basis for a CAND_ADD whose operands
8588 have been commuted in the analysis.
8589 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8591 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8593 * config/aarch64/aarch64.md
8594 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8595 shift value between 0-4.
8597 2013-05-07 Richard Biener <rguenther@suse.de>
8599 * double-int.h (rshift): New overload.
8600 * double-int.c (rshift): New function.
8601 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8602 (create_reference_ops_from_ref): Remove.
8603 (vn_reference_insert): Use shared ops for constructing the
8604 reference and copy it.
8606 2013-05-07 Richard Biener <rguenther@suse.de>
8609 * tree-eh.c (sink_clobbers): Properly propagate
8610 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8612 2013-05-07 Jakub Jelinek <jakub@redhat.com>
8614 PR tree-optimization/57149
8615 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8616 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8617 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8618 uninit_undefined_value_p instead of ssa_undefined_value_p.
8621 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8622 for modifier == EXPAND_INITIALIZER.
8624 2013-05-07 Anton Blanchard <anton@samba.org>
8626 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8627 for powerpc64 little endian.
8628 * configure: Regenerate.
8630 2013-05-06 Graham Stott <grahams@btinternet.com>
8632 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8633 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8634 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8635 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8637 2013-05-06 Graham Stott <grahams@btinternet.com>
8639 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8640 codes which allow non-lvalues.
8642 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
8644 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8645 components are all 1s.
8646 (integer_minus_onep): New function.
8647 * tree.h (integer_minus_onep): Declare it.
8648 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8649 integer_minus_onep instead of integer_all_onesp.
8651 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8654 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8655 variations of these patterns.
8657 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8659 * config/i386/i386.md (isa): Add x64_sse4 member.
8660 (enabled): Handle x64_sse4.
8661 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8662 instruction for 64bit SSE4_1 targets. Update insn attributes.
8663 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8664 instruction for SSE4_1 targets. Update insn attributes.
8665 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8666 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8667 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8669 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8670 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8671 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8673 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8676 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8678 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8680 * final.c (do_assembler_dialects): Don't handle curly braces and
8681 vertical bar escaped by % as dialect delimiters.
8682 (output_asm_insn): Print curly braces and vertical bar if escaped
8683 by % and ASSEMBLER_DIALECT defined.
8684 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8685 * doc/tm.texi: Regenerated.
8687 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8689 * config/mips/mips.c: Include tree-pass.h.
8690 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8691 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8692 (pass_mips_machine_reorg2): New machine specific pass.
8693 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8694 (mips_option_override): Register the new pass.
8695 * rtl.h (cleanup_barriers): Remove prototype.
8696 (dbr_schedule): Likewise.
8697 * jump.c (cleanup_barriers): Make static.
8698 * reorg.c (dbr_schedule): Likewise.
8700 2013-05-06 Richard Biener <rguenther@suse.de>
8702 PR tree-optimization/57185
8703 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8706 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8709 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8710 between "first_arg" and "insn", not "last" and "insn".
8712 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8714 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8715 (find_candidates_in_block): Re-enable slsr_process_phi.
8716 (create_phi_basis): Fix double counting of candidate adjustment.
8718 2013-05-06 Richard Biener <rguenther@suse.de>
8721 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8722 the edge is also fallthru, preserve it and just clear the
8724 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8725 also complex, preserve that and just clear the fallthru flag.
8726 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8727 update virtual operands.
8729 2013-05-06 Alan Modra <amodra@gmail.com>
8731 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8732 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8733 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8734 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8735 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8736 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8738 2013-05-06 Alan Modra <amodra@gmail.com>
8740 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8741 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8742 (DEFAULT_ASM_ENDIAN): Define.
8743 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8744 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8745 Update -K PIC clause from sysv4.h.
8746 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8747 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8749 2013-05-06 Alan Modra <amodra@gmail.com>
8751 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8752 twice for little-endian.
8753 (ashrdi3_no_power, ashrdi3): Support little-endian.
8755 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8758 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8759 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8761 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8762 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8763 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8766 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8767 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8769 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8770 * config/pa/pa.opt: Make mbig-switch a no-op.
8771 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8772 (CASE_VECTOR_MODE): Always return SImode.
8773 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8774 for the !TARGET_BIG_SWITCH case.
8775 * config/pa/pa-linux.h: Likewise.
8776 * config/pa/pa-openbsd.h: Likewise.
8777 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8778 * config/pa/pa.md (short_jump): Remove define_insn.
8779 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8780 (casesi0): Remove define_insn.
8781 (type): Remove btable_branch.
8782 (pa_combine_type): Likewise.
8783 (in_nullified_branch_delay): Likewise.
8784 (in_call_delay): Likewise.
8785 (define_delay): Likewise.
8786 (define_insn_reservation "Z3"): Likewise.
8787 (define_insn_reservation "Z4"): Likewise.
8788 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8789 (pa_adjust_insn_length): Remove adjustment for btable branches.
8790 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8793 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
8795 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8796 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8797 Add m->r,x alternatives.
8798 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8799 splitters using SWI48x mode iterator.
8800 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8801 TARGET_64BIT. Add m->x alternative.
8802 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8803 Add o->x alternative. Enable for TARGET_SSE.
8804 (sse_storeq): Remove expander.
8805 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8806 with memory input operand.
8807 (*vec_extractv2di_1 splitter): New.
8808 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8809 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8811 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8813 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8814 (INT_LOWPART): Delete.
8815 (extract_MB): Adjust.
8816 (extract_ME): Adjust.
8817 (print_operand): Adjust.
8819 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8821 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8822 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8823 (reg_or_logical_cint_operand, easy_fp_constant,
8824 logical_const_operand): Delete "CONST_DOUBLE" case.
8825 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8826 "HOST_BITS_PER_WIDE_INT == 64" test.
8827 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
8828 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8829 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8830 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8832 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8833 CONST_DOUBLE DImode/VOIDmode case.
8834 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8835 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8836 CONST_DOUBLE VOIDmode case.
8837 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8838 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8839 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8840 Delete CONST_DOUBLE case.
8841 (splitters for mov FMOVE64 const_double): Delete
8842 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8843 "HOST_BITS_PER_WIDE_INT >= 64" test.
8844 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8846 (mov DI const_double): Delete.
8848 2013-05-04 Jakub Jelinek <jakub@redhat.com>
8850 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8851 on op shows all bits zero in mode of a lowpart subreg, return zero.
8853 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8856 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8857 to save TFmode registers and DImode to save TImode registers for
8858 caller save operations.
8859 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8860 mark being partially clobbered since they only use the first
8863 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8864 and TDmode only use the upper 64-bits of each VSX register.
8866 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8868 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8869 (find_candidates_in_block): Disable slsr_process_phi.
8871 2013-05-03 Guozhi Wei <carrot@google.com>
8873 * coverage.c (coverage_obj_init): Move the construction of gcov
8875 (build_init_ctor): ... here.
8877 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8879 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8880 (slsr_cand_d): Redefine def_phi.
8881 (stride_status, phi_adjust_status, count_phis_status): New enums.
8882 (find_phi_def): New.
8883 (find_basis_for_base_expr): New.
8884 (find_basis_for_candidate): Handle hidden bases.
8885 (alloc_cand_and_find_basis): Handle phi candidates.
8886 (slsr_process_phi): New.
8887 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8888 (create_mul_imm_cand): Likewise.
8889 (create_add_ssa_cand): Exclude phi base candidates.
8890 (create_add_imm_cand): Likewise.
8891 (slsr_process_cast): Likewise.
8892 (slsr_process_copy): Likewise.
8893 (find_candidates_in_block): Handle phi candidates.
8894 (dump_candidate): Likewise.
8895 (unconditional_cands): Delete.
8896 (unconditional_cands_with_known_stride_p): Delete.
8897 (phi_dependent_cand_p): New.
8898 (cand_increment): Handle phi-dependent candidates.
8899 (replace_dependent): Delete.
8900 (replace_mult_candidate): New.
8901 (replace_unconditional_candidate): New.
8902 (incr_vec_index): Move to avoid forward reference.
8903 (create_add_on_incoming_edge): New.
8904 (create_phi_basis): New.
8905 (replace_dependents): Delete.
8906 (replace_conditional_candidate): New.
8907 (phi_add_costs): New.
8908 (replace_uncond_cands_and_profitable_phis): New.
8909 (record_increment): Handle phi adjustments.
8910 (record_phi_increments): New.
8911 (record_increments): Handle phi adjustments.
8912 (phi_incr_cost): New.
8913 (lowest_cost_path): Handle phis.
8914 (total_savings): Likewise.
8915 (analyze_increments): Likewise.
8916 (ncd_with_phi): New.
8917 (ncd_of_cand_and_phis): New.
8918 (nearest_common_dominator_for_cands): Handle phi increments.
8919 (all_phi_incrs_profitable): New.
8920 (replace_profitable_candidates): Handle phi-dependent candidates.
8921 (analyze_candidates_and_replace): Likewise.
8923 2013-05-03 Teresa Johnson <tejohnson@google.com>
8926 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8927 do not exceed REG_BR_PROB_BASE.
8929 2013-05-03 Jeff Law <law@redhat.com>
8931 PR tree-optimization/57144
8932 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8933 operand of the condition will bit into the new type when eliminating
8934 a cast feeding a condition.
8936 2013-05-03 Jakub Jelinek <jakub@redhat.com>
8938 PR rtl-optimization/57130
8939 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8940 of COMPARE as in_code to the recursive call if needed.
8942 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
8944 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8945 (enabled): Handle new members.
8946 * config/i386/sse.md (*vec_concatv2si): Merge from
8947 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8948 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8950 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8952 PR tree-optimization/57027
8953 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8954 for fnms opportunity, check we got the prerequisite kind
8955 of tree / gimple before using accessor functions.
8957 2013-05-03 Richard Biener <rguenther@suse.de>
8959 * double-int.h (lshift): New overload without precision
8961 (operator *=, operator +=, operator -=): Move ...
8962 * double-int.c (operator *=, operator +=, operator -=): ... here
8963 and implement more efficiently.
8964 (mul_double_with_sign): Remove.
8965 (lshift_double): Adjust to take unsinged shift argument, push
8966 dispatching code to callers.
8967 (mul_double_wide_with_sign): Add early out for callers that
8968 are not interested in high parts or overflow.
8969 (lshift): New function.
8970 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8972 (lrotate, rrotate): Use logical shifts.
8973 * expr.c (get_inner_reference): Use lshift.
8974 * fixed-value.c (do_fixed_divide): Likewise.
8975 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8976 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8977 (indirect_refs_may_alias_p): Likewise.
8978 (stmt_kills_ref_p_1): Likewise.
8980 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8982 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8984 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8986 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8987 scalar form of FABD instruction.
8989 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8991 * lra-constraints.c (process_alt_operands): Add checking alt
8992 number to choose the best alternative.
8994 2013-05-02 Richard Biener <rguenther@suse.de>
8996 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
8997 bitmap and its handling.
8998 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9000 2013-05-02 Richard Biener <rguenther@suse.de>
9003 * tree-inline.c (copy_loops): Properly handle removed loops.
9004 (copy_cfg_body): Mark destination loops for fixup if source
9007 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9010 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9011 generating simple_return for naked functions.
9013 2013-05-02 Martin Jambor <mjambor@suse.cz>
9016 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9017 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9019 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9020 ipa_agg_replacement_value structures.
9021 (known_aggs_to_agg_replacement_list): Likewise.
9022 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9023 (read_agg_replacement_chain): Likewise.
9024 (ipcp_transform_function): Also check that by_ref flags match.
9026 2013-05-02 Richard Biener <rguenther@suse.de>
9028 * graphds.h (struct graph): Add obstack member.
9029 * graphds.c (new_graph): Initialize obstack and allocate
9031 (add_edge): Allocate edge from the obstack.
9032 (free_graph): Free the obstack instead of all edges and vertices.
9034 2013-05-02 Teresa Johnson <tejohnson@google.com>
9036 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9038 * cfg.c (update_bb_profile_for_threading): Ditto.
9039 * tree-inline.c (copy_bb): Ditto.
9040 (copy_edges_for_bb): Ditto.
9041 (initialize_cfun): Ditto.
9042 (copy_cfg_body): Ditto.
9043 (expand_call_inline): Ditto.
9044 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9045 (estimate_node_size_and_time): Ditto.
9046 (inline_merge_summary): Ditto.
9047 * cgraphclones.c (cgraph_clone_edge): Ditto.
9048 (cgraph_clone_node): Ditto.
9049 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9050 (compute_trg_info): Ditto.
9052 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9054 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9055 S reg when fp attribute set.
9056 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9058 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9060 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9062 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9063 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9064 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9066 2013-05-02 Richard Biener <rguenther@suse.de>
9068 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9069 (struct instantiate_cache_entry): New type.
9070 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9071 (struct instantiate_cache_type): New type.
9072 (set_instantiated_value, get_instantiated_value): Remove.
9073 (get_instantiated_value_entry): New function.
9074 (instantiate_scev_name): Use the new cache and adjust.
9075 (instantiate_scev_poly): Adjust.
9076 (instantiate_scev_binary): Likewise.
9077 (instantiate_array_ref): Likewise.
9078 (instantiate_scev_convert): Likewise.
9079 (instantiate_scev_not): Likewise.
9080 (instantiate_scev_3): Likewise.
9081 (instantiate_scev_2): Likewise.
9082 (instantiate_scev_r): Likewise.
9083 (instantiate_scev): Likewise.
9084 (resolve_mixers): Likewise.
9086 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9089 * lra-constraints.c (best_small_class_operands_num): Remove.
9090 (process_alt_operands): Remove small_class_operands_num. Take
9091 small classes operands into losers and only if the operand is not
9092 matched. Modify debugging output.
9093 (curr_insn_transform): Remove best_small_class_operands_num.
9096 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9098 * config/aarch64/aarch64-builtins.c
9099 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9100 * config/aarch64/aarch64-simd-builtins.def
9101 (reduc_splus_): Add new modes.
9102 (reduc_uplus_): New.
9103 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9104 (reduc_uplus_v4sf): Likewise.
9105 (reduc_splus_v4sf): Likewise.
9106 (aarch64_addv<mode>): Likewise.
9107 (reduc_uplus_<mode>): Likewise.
9108 (reduc_splus_<mode>): Likewise.
9109 (aarch64_addvv2di): Likewise.
9110 (reduc_uplus_v2di): Likewise.
9111 (reduc_splus_v2di): Likewise.
9112 (aarch64_addvv2si): Likewise.
9113 (reduc_uplus_v2si): Likewise.
9114 (reduc_splus_v2si): Likewise.
9115 (reduc_<sur>plus_<mode>): New.
9116 (reduc_<sur>plus_v2di): Likewise.
9117 (reduc_<sur>plus_v2si): Likewise.
9118 (reduc_<sur>plus_v4sf): Likewise.
9119 (aarch64_addpv4sf): Likewise.
9120 * config/aarch64/arm_neon.h
9121 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9122 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9123 add UNSPEC_SADDV, UNSPEC_UADDV.
9125 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9127 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9129 * config/aarch64/arm_neon.h
9130 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9132 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9134 * config/aarch64/aarch64-builtins
9135 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
9137 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9139 * config/aarch64/aarch64-simd-builtins.def
9141 (reduc_smin_): Likewise.
9142 (reduc_umax_): Likewise.
9143 (reduc_umin_): Likewise.
9144 (reduc_smax_nan_): Likewise.
9145 (reduc_smin_nan_): Likewise.
9148 (smax): Update for V2SF, V4SF and V2DF modes.
9151 (smin_nan): Likewise.
9152 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
9153 (<su><maxmin><mode>3): ...This, refactor.
9154 (s<maxmin><mode>3): New.
9155 (<maxmin_uns><mode>3): Likewise.
9156 (reduc_<maxmin_uns>_<mode>): Refactor.
9157 (reduc_<maxmin_uns>_v4sf): Likewise.
9158 (reduc_<maxmin_uns>_v2si): Likewise.
9159 (aarch64_<fmaxmin><mode>: Remove.
9160 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
9162 (vmin<q>_f<32,64>): Likewise.
9163 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9165 (su): Add mappings for smax, smin, umax, umin.
9167 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9168 (FMAXMIN): Rename as...
9169 (FMAXMIN_UNS): ...This.
9171 (fmaxminv): Likewise.
9172 (fmaxmin): Likewise.
9174 (maxmin_uns_op): Likewise.
9176 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9178 * config/aarch64/arm_neon.h
9179 (vac<ge, gt><sd>_f<32, 64>): Rename to...
9180 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
9181 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
9183 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9185 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
9186 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
9188 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9190 * config/aarch64/aarch64-simd.md
9191 (vcond<mode>_internal): Handle special cases for constant masks.
9192 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
9193 (vcondu<mode><mode>): Likewise.
9194 (vcond<v_cmp_result><mode>): New.
9196 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9198 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
9199 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
9200 * config/aarch64/aarch64-simd-builtins.def
9201 (cmeq): Update to BUILTIN_VALLDI.
9206 * config/aarch64/arm_neon.h
9207 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
9208 to builtins or C as appropriate.
9210 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9212 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
9214 (cmhi): Rename to...
9216 * config/aarch64/aarch64-simd.md
9217 (simd_mode): Add SF.
9218 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
9219 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
9220 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
9221 (cstore<mode>_neg): ...This.
9222 * config/aarch64/iterators.md
9224 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
9226 (UCOMPARISONS): Likewise.
9227 (optab): Add missing comparisons.
9235 (V_cmp_result): Add DF, SF modes.
9236 (v_cmp_result): Likewise.
9239 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
9241 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
9243 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
9244 define_insn to define_insn_and_split.
9245 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
9246 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
9247 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
9248 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
9249 (thumb2_negscc): Likewise.
9251 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9253 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
9255 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9257 * config/arm/thumb2.md: Remove trailing whitespaces.
9259 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9261 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
9262 Use gen_int_mode rather than GEN_INT.
9264 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
9266 * value-prof.c (stream_in_histogram_value): Remove the strayed
9269 2013-04-30 Richard Biener <rguenther@suse.de>
9272 * cfghooks.c (split_edge): Properly check for the loop latch edge.
9274 2013-04-30 Richard Biener <rguenther@suse.de>
9277 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
9279 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
9281 PR rtl-optimization/56957
9282 PR rtl-optimization/57105
9283 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
9284 variable. Use just INSN_UID for determining whether an insn
9285 should be only disconnected from the insn stream.
9286 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
9288 2013-04-30 Jakub Jelinek <jakub@redhat.com>
9290 PR tree-optimization/57104
9291 * tsan.c (instrument_expr): Don't instrument accesses to
9292 DECL_HARD_REGISTER VAR_DECLs.
9294 2013-04-30 Richard Biener <rguenther@suse.de>
9296 * function.h (loops_for_fn): New inline function.
9297 (set_loops_for_fn): Likewise.
9298 * cfgloop.h (place_new_loop): Add struct function parameter.
9299 (get_loop): Likewise.
9300 (get_loops): Likewise.
9301 (number_of_loops): Likewise.
9303 (fel_init): Likewise.
9304 * cfg.c (get_loop_copy): Adjust.
9305 * cfgloop.c (flow_loops_dump): Likewise.
9306 (record_loop_exits): Likewise.
9307 (verify_loop_structure): Likewise.
9308 * cfgloopanal.c (mark_irreducible_loops): Likewise.
9309 (estimate_reg_pressure_cost): Likewise.
9310 (mark_loop_exit_edges): Likewise.
9311 * cfgloopmanip.c (place_new_loop): Likewise.
9312 (add_loop): Likewise.
9313 (duplicate_loop): Likewise.
9314 * graph.c (draw_cfg_nodes): Likewise.
9315 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
9316 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
9317 (extract_affine_chrec): Likewise.
9318 (build_scop_iteration_domain): Likewise.
9319 * graphite.c (graphite_initialize): Likewise.
9320 * ira-build.c (create_loop_tree_nodes): Likewise.
9321 (more_one_region_p): Likewise.
9322 (rebuild_regno_allocno_maps): Likewise.
9323 (mark_loops_for_removal): Likewise.
9324 (mark_all_loops_for_removal): Likewise.
9325 (remove_unnecessary_regions): Likewise.
9326 (ira_build): Likewise.
9327 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
9328 * loop-init.c (fix_loop_structure): Likewise.
9329 (gate_rtl_move_loop_invariants): Likewise.
9330 (gate_rtl_unswitch): Likewise.
9331 (gate_rtl_unroll_and_peel_loops): Likewise.
9332 (rtl_doloop): Likewise.
9333 * lto-streamer-in.c (input_cfg): Likewise.
9334 * lto-streamer-out.c (output_cfg): Likewise.
9335 * modulo-sched.c (sms_schedule): Likewise.
9336 * predict.c (tree_estimate_probability): Likewise.
9337 (tree_estimate_probability_driver): Likewise.
9338 (estimate_loops): Likewise.
9339 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
9340 (move_sese_region_to_fn): Likewise.
9341 (debug_loop_num): Likewise.
9342 * tree-chrec.c (chrec_evaluate): Likewise.
9343 (hide_evolution_in_other_loops_than_loop): Likewise.
9344 (chrec_component_in_loop_num): Likewise.
9345 (reset_evolution_in_loop): Likewise.
9346 (evolution_function_is_invariant_rec_p): Likewise.
9347 * tree-if-conv.c (main_tree_if_conversion): Likewise.
9348 * tree-inline.c (copy_loops): Likewise.
9349 (copy_cfg_body): Likewise.
9350 (tree_function_versioning): Likewise.
9351 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
9352 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
9354 (add_to_evolution_1): Likewise.
9355 (scev_const_prop): Likewise.
9356 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
9357 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
9358 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
9359 (tree_ssa_lim_initialize): Likewise.
9360 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
9361 (verify_loop_closed_ssa): Likewise.
9362 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
9363 (tree_ssa_loop_im): Likewise.
9364 (tree_ssa_loop_unswitch): Likewise.
9365 (tree_vectorize): Likewise.
9366 (check_data_deps): Likewise.
9367 (tree_ssa_loop_ivcanon): Likewise.
9368 (tree_ssa_loop_bounds): Likewise.
9369 (tree_complete_unroll): Likewise.
9370 (tree_complete_unroll_inner): Likewise.
9371 (tree_parallelize_loops): Likewise.
9372 (tree_ssa_loop_prefetch): Likewise.
9373 (tree_ssa_loop_ivopts): Likewise.
9374 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
9375 * tree-vectorizer.c (vectorize_loops): Likewise.
9377 2013-04-29 Mike Frysinger <vapier@gentoo.org>
9379 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
9380 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
9381 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
9382 with EABI_LINK_SPEC.
9384 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9387 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9390 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9393 * lra-constraints.c (process_alt_operands): Discourage a bit more
9394 using memory for pseudos. Print cost dump for alternatives.
9395 Modify cost values for conflicts with early clobbers.
9396 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9398 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9401 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9403 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9405 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9407 (movdi_aarch64): Support LDR/STR from/to D register.
9409 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9411 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9412 or higher by default.
9414 2013-04-29 Richard Biener <rguenther@suse.de>
9417 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9418 even if not adding abnormal edges for calls that can make
9421 2013-04-29 Richard Biener <rguenther@suse.de>
9424 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9426 (move_stmt_r): Remove redundant checking.
9428 2013-04-29 Teresa Johnson <tejohnson@google.com>
9431 * basic-block.h (apply_scale): New function.
9432 (apply_probability): Use apply_scale.
9433 * gimple-streamer-in.c (input_bb): Ditto.
9434 * lto-streamer-in.c (input_cfg): Ditto.
9435 * lto-cgraph.c (merge_profile_summaries): Ditto.
9436 * tree-optimize.c (execute_fixup_cfg): Ditto.
9437 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9438 (copy_edges_for_bb): Ditto.
9439 (copy_cfg_body): Ditto.
9441 2013-04-29 Tom de Vries <tom@codesourcery.com>
9443 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9444 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9445 (tail_merge_optimize): Handle current_loops == NULL.
9447 2013-04-26 Jeff Law <law@redhat.com>
9449 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9450 (simplify_cond_using_ranges): Generalize code to simplify
9451 COND_EXPRs where one argument is a constant and the other
9452 is an SSA_NAME created by an integral type conversion.
9454 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9456 * config/arm/arm.md (store_minmaxsi): Use only when
9457 optimize_insn_for_size_p.
9459 2013-04-29 Christian Bruel <christian.bruel@st.com>
9462 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9464 2013-04-29 Richard Biener <rguenther@suse.de>
9467 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9468 loop tree make sure to schedule a fixup for the child as well.
9469 (expand_omp_for_generic): Properly add loops.
9470 (expand_omp_for_static_nochunk): Likewise.
9471 (expand_omp_for_static_chunk): Likewise.
9472 (expand_omp_for): For the degenerate case fixup loops.
9473 (expand_omp_sections): Fix default bb placement in loops.
9474 (expand_omp_atomic_pipeline): Properly add loops.
9476 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9478 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9480 2013-04-29 Tom de Vries <tom@codesourcery.com>
9482 * tree-ssa-tail-merge.c: Update header comment.
9484 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9486 * config/aarch64/arm_neon.h
9487 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9488 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9489 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9490 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9491 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9492 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9493 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9494 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9496 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9498 * config/aarch64/aarch64-simd.md
9499 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9500 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9501 fix_trunc, fixuns_trunc.
9502 (ftrunc<VDQF:mode>2): New.
9503 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9504 (fix_trunc_optab): New.
9506 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9508 * config/aarch64/aarch64-builtins.c
9509 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9510 iceilf, lround, iroundf.
9512 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9515 * config/i386/i386.h (enum ix86_tune_indices)
9516 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9517 New, split from X86_TUNE_INTER_UNIT_MOVES.
9518 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9519 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9520 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9521 (TARGET_INTER_UNIT_MOVES): Remove.
9522 * config/i386/i386.c (initial_ix86_tune_features): Update.
9523 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9524 (ix86_expand_convert_uns_didf_sse): Use
9525 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9526 (ix86_expand_vector_init_one_nonzero): Ditto.
9527 (ix86_expand_vector_init_interleave): Ditto.
9528 (inline_secondary_memory_needed): Return true for moves from SSE class
9529 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9530 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9531 * config/i386/constraints.md (Yi, Ym): Depend on
9532 TARGET_INTER_UNIT_MOVES_TO_VEC.
9533 (Yj, Yn): New constraints.
9534 * config/i386/i386.md (*movdi_internal): Change constraints of
9535 operand 1 from Yi to Yj and from Ym to Yn.
9536 (*movsi_internal): Ditto.
9537 (*movdf_internal): Ditto.
9538 (*movsf_internal): Ditto.
9539 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9540 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9541 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9542 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9543 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9544 * config/i386/sse.md (movdi_to_sse): Ditto.
9545 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9546 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9547 TARGET_INTER_UNIT_MOVES.
9548 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9549 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9550 instead of TARGET_INTER_UNIT_MOVES.
9551 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9552 operand 1 from Yi to Yj and from Ym to Yn.
9554 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9556 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9557 (float_truncate_hi_): Likewise.
9558 (float_extend_lo_): Likewise.
9559 (float_truncate_lo_): Likewise.
9560 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9561 (aarch64_float_extend_lo_v2df): Likewise.
9562 (vec_unpacks_hi_v4sf): Likewise.
9563 (aarch64_float_truncate_lo_v2sf): Likewise.
9564 (aarch64_float_truncate_hi_v4sf): Likewise.
9565 (vec_pack_trunc_v2df): Likewise.
9566 (vec_pack_trunc_df): Likewise.
9568 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9570 * config/aarch64/aarch64-builtins.c
9571 (aarch64_fold_builtin): Fold float conversions.
9572 * config/aarch64/aarch64-simd-builtins.def
9573 (floatv2si, floatv4si, floatv2di): New.
9574 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9575 * config/aarch64/aarch64-simd.md
9576 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9577 * config/aarch64/iterators.md (FLOATUORS): New.
9578 (optab): Add float, floatuns.
9579 (su_optab): Likewise.
9581 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9583 * config/aarch64/aarch64-builtins.c
9584 (aarch64_builtin_vectorized_function): Use new names for
9586 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9587 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9588 (fcvtzu): Split as...
9589 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9590 (fcvtas): Split as...
9591 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9592 (fcvtau): Split as...
9593 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9594 (fcvtps): Split as...
9595 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9596 (fcvtpu): Split as...
9597 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9598 (fcvtms): Split as...
9599 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9600 (fcvtmu): Split as...
9601 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9602 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9603 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9604 (lfrintnusf, lfrintnudf): Likewise.
9605 * config/aarch64/aarch64-simd.md
9606 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9608 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9609 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9610 (fcvt_pattern): Likewise.
9612 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9614 * config/aarch64/aarch64-simd.md
9615 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9616 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9618 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9620 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9621 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9622 (vrnd<a,m,n,p>_f32): Implement using builtins.
9623 (vrnd<i,x><q>_f<32, 64>): New.
9625 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9627 * config/aarch64/aarch64-builtins.c
9628 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9629 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9630 (frintz): Rename to...
9632 (frintp): Rename to...
9634 (frintm): Rename to...
9636 (frinti): Rename to...
9637 (nearbyint): ...this.
9638 (frintx): Rename to...
9640 (frinta): Rename to...
9642 * config/aarch64/aarch64-simd.md
9643 (aarch64_frint<frint_suffix><mode>): Delete.
9644 (<frint_pattern><mode>2): Convert to insn.
9645 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9646 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9647 (frint_pattern): Likewise.
9648 (frint_suffix): Likewise.
9650 2013-04-29 Richard Biener <rguenther@suse.de>
9652 PR tree-optimization/57081
9653 * loop-init.c: Include tree-flow.h.
9654 (loop_optimizer_finalize): Free number of iteration estimates.
9655 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9657 2013-04-29 Jakub Jelinek <jakub@redhat.com>
9659 PR tree-optimization/57083
9660 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9661 non-singleton shift count range, zero extend low_bound for uns case.
9663 * config/i386/predicates.md (general_vector_operand): New predicate.
9664 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9665 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9666 if they aren't nonimmediate operands. If their original values
9667 satisfy const_vector_equal_evenodd_p, don't shift them.
9668 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9669 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9670 if not nonimmediate_operand.
9671 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9672 instead of register_operand.
9673 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9675 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
9677 * stor-layout.c (finalize_size_functions): Allocate a structure and
9678 reset cfun before dumping the functions.
9680 2013-04-27 Jakub Jelinek <jakub@redhat.com>
9682 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9685 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9686 use xop_pmacsdqh if uns_p.
9687 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9688 the immediate rotate count.
9690 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9692 * rtl.h (struct rtx_def): Add comment for field jump.
9693 (LRA_SUBREG_P): New macro.
9694 * recog.c (register_operand): Check LRA_SUBREG_P.
9695 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9697 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9698 (lra_final_code_change): Skip subreg change for operators.
9699 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9700 if there are no operand changes.
9701 * lra-constraints.c (curr_insn_set): New.
9702 (match_reload): Set LRA_SUBREG_P.
9703 (emit_spill_move): Ditto.
9704 (check_and_process_move): Use curr_insn_set. Process only single
9705 set insns. Don't initialize sec_mem_p and change_p.
9706 (simplify_operand_subreg): Use LRA_SUBREG_P.
9707 (reg_in_class_p): New function.
9708 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9709 of #ifdef. Add code to remove cycling.
9710 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9711 non-null disp. Reload inner instead of disp when base and index
9712 are null. Try to put lo_sum into register.
9713 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9714 (check_and_process_move): Move code for move cost check to
9715 simple_move_p. Remove equiv_substitution.
9716 (simple_move_p): New function.
9717 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9718 curr_insn_set. Call check_and_process_move only for single set
9719 insns. Use the new function. Move call of check_and_process_move
9720 after operand equiv substitution and address process.
9722 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9725 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9726 with nonlocal goto receivers or returns twice calls, ignore
9727 unininitialized values from abnormal edges to nl goto receiver
9728 or returns twice call.
9730 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9732 PR tree-optimization/57051
9733 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9734 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9737 2013-04-26 Richard Biener <rguenther@suse.de>
9739 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9740 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9741 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9742 (fixup_loop_arrays_after_move): New function.
9743 (move_sese_region_to_fn): Properly outline the loop tree parts
9746 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
9748 * config/i386/i386.md (type, unit): Fix long lines.
9750 2013-04-26 Richard Biener <rguenther@suse.de>
9752 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9753 (lto-streamer-out.o): Likewise.
9754 * cfgloop.c (init_loops_structure): Export, add struct function
9755 argument and adjust.
9756 (flow_loops_find): Adjust.
9757 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9758 (init_loops_structure): Declare.
9759 * lto-streamer-in.c: Include cfgloop.h.
9760 (input_cfg): Input the loop tree.
9761 * lto-streamer-out.c: Include cfgloop.h.
9762 (output_cfg): Output the loop tree.
9763 (output_struct_function_base): Do not drop PROP_loops.
9765 2013-03-26 Richard Biener <rguenther@suse.de>
9767 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9768 (pass_build_cfg): Provide PROP_loops.
9769 (move_sese_region_to_fn): Remove loops that are outlined into fn
9771 * tree-inline.c: Include cfgloop.h.
9772 (initialize_cfun): Do not drop PROP_loops.
9773 (copy_loops): New function.
9774 (copy_cfg_body): Copy loop structure.
9775 (tree_function_versioning): Initialize destination loop tree.
9776 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9777 (pass_parallelize_loops): Do IL verification.
9778 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9779 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9780 the CFG make sure we fixup loops as well.
9781 * tree-ssa-tail-merge.c: Include cfgloop.h.
9782 (replace_block_by): When merging loop latches mark loops for fixup.
9783 * lto-streamer-out.c (output_struct_function_base): Drop
9785 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9786 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9787 * ipa-split.c: Include cfgloop.h.
9788 (split_function): Add the new return block to the loop tree root.
9789 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9790 whether we have removed the forwarder block.
9791 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9792 * cfgloop.h (place_new_loop): Declare.
9793 * cfgloopmanip.c (place_new_loop): Export.
9794 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9795 (tree-switch-conversion.o): Likewise.
9796 (tree-complex.o): Likewise.
9797 (tree-inline.o): Likewise.
9798 (tree-ssa-tailmerge.o): Likewise.
9799 (ipa-split.o): Likewise.
9800 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9801 (tree-ssa-copy.o): Likewise.
9802 * tree-switch-conversion.c: Include cfgloop.h
9803 (process_switch): If we emit a bit-test cascade, schedule loops
9805 * tree-complex.c: Include cfgloop.h.
9806 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9807 * asan.c: Include cfgloop.h.
9808 (create_cond_insert_point): Properly add new basic-blocks to
9809 loops, schedule loop fixup.
9810 * cfgloop.c (verify_loop_structure): Check that looks are not
9812 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9814 (expand_omp_for_generic): Likewise.
9815 (expand_omp_sections): Likewise.
9816 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9817 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9818 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9819 is initialized, not when loops are present.
9820 * tree-parloops.c (parallelize_loops): Remove checking here.
9821 * passes.c (init_optimization_passes): Schedule a copy-propagation
9822 pass before complete unrolling of inner loops.
9824 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9826 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9827 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9828 (colorize_init): Add argument to _WIN32 version.
9829 * toplev.c: Include diagnostic-color.h.
9830 (process_options): Default to -fdiagnostics-color=auto if
9831 GCC_COLORS env var is in the environment.
9832 * common.opt (fdiagnostics-color=): Add Var and Init.
9833 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9834 env var is in the environment, the default is auto rather than never.
9836 * diagnostic.h (file_name_as_prefix): Add context argument.
9837 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9838 the string as locus.
9839 * langhooks.c (lhd_print_error_function): Adjust caller.
9841 2013-04-25 Lawrence Crowl <crowl@google.com>
9843 * var-tracking.c (shared_hash_def::htab):
9844 Change type to hash_table. Update dependent calls and types.
9846 2013-04-25 Lawrence Crowl <crowl@google.com>
9848 * Makefile.in: Update as needed below.
9850 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9851 Move declaration to after the type's method definitons.
9853 * attribs.c (htab_t scoped_attributes::attribute_hash):
9854 Change type to hash_table. Update dependent calls and types.
9856 * bitmap.c (htab_t bitmap_desc_hash):
9857 Change type to hash_table. Update dependent calls and types.
9859 * cselib.c (htab_t cselib_hash_table):
9860 Change type to hash_table. Update dependent calls and types.
9862 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9863 (hash_string_slot_node): Move implementation into lto-streamer.h
9864 struct string_slot_hasher.
9865 (eq_string_slot_node): Likewise.
9867 * data-streamer-out.c: Update output_block::string_hash_table
9868 dependent calls and types.
9870 * dwarf2cfi.c (htab_t trace_index):
9871 Change type to hash_table. Update dependent calls and types.
9873 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9874 Change type to hash_table. Update dependent calls and types.
9875 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9876 (htab_t optimize_external_refs::map): Likewise.
9877 (htab_t output_comp_unit::extern_map): Likewise.
9878 (htab_t output_comdat_type_unit::extern_map): Likewise.
9879 (htab_t output_macinfo::macinfo_htab): Likewise.
9880 (htab_t optimize_location_lists::htab): Likewise.
9881 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9883 * except.c (htab_t ehspec_hash_type):
9884 Change type to hash_table. Update dependent calls and types.
9885 (assign_filter_values::ttypes): Likewise.
9886 (assign_filter_values::ehspec): Likewise.
9887 (sjlj_assign_call_site_values::ar_hash): Likewise.
9888 (convert_to_eh_region_ranges::ar_hash): Likewise.
9890 * gcse.c (htab_t pre_ldst_table):
9891 Change type to hash_table. Update dependent calls and types.
9893 * ggc-common.c (htab_t saving_htab):
9894 Change type to hash_table. Update dependent calls and types.
9895 (htab_t loc_hash): Likewise.
9896 (htab_t ptr_hash): Likewise.
9897 (call_count): Rename ggc_call_count.
9898 (call_alloc): Rename ggc_call_alloc.
9899 (loc_descriptor): Rename make_loc_descriptor.
9900 (add_statistics): Rename ggc_add_statistics.
9902 * ggc-common.c (saving_htab):
9903 Change type to hash_table. Update dependent calls and types.
9905 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9906 (push_gimplify_context): Likewise.
9907 (pop_gimplify_context): Likewise.
9908 (struct gimple_temp_hash_elt): Added.
9909 (struct gimplify_hasher): Likewise.
9910 (struct gimplify_ctx.temp_htab):
9911 Change type to hash_table. Update dependent calls and types.
9913 * gimple-fold.c: Include gimplify-ctx.h.
9915 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9916 Change type to hash_table. Update dependent calls and types.
9917 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9918 avoid potential global name collision.
9920 * gimplify.c: Include gimplify-ctx.h.
9921 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9922 (htab_t gimplify_ctx::temp_htab):
9923 Update dependent calls and types for new type hash_table.
9924 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9925 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9927 * gimplify-ctx.h: New.
9928 (struct gimple_temp_hash_elt): Move from gimplify.c.
9929 (class gimplify_hasher): New.
9930 (struct gimplify_ctx): Move from gimple.h.
9931 (htab_t gimplify_ctx::temp_htab):
9932 Change type to hash_table. Update dependent calls and types.
9934 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9935 (htab_t ivs_params::newivs_index):
9936 Change type to hash_table. Update dependent calls and types.
9937 (htab_t ivs_params::params_index): Likewise.
9938 (htab_t print_generated_program::params_index): Likewise.
9939 (htab_t gloog::newivs_index): Likewise.
9940 (htab_t gloog::params_index): Likewise.
9942 * graphite.c: Include graphite-htab.h.
9943 4htab_t graphite_transform_loops::bb_pbb_mapping):
9944 Change type to hash_table. Update dependent calls and types.
9946 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9947 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9948 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9950 * graphite-dependences.c: Include graphite-htab.h.
9951 (loop_is_parallel_p): Change hash table type of parameter.
9953 * graphite-htab.h: New.
9954 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9955 (extern find_pbb_via_hash): Move from graphite-poly.h.
9956 (extern loop_is_parallel_p): Move from graphite-poly.h.
9957 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9959 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9960 (extern loop_is_parallel_p): Move to graphite-htab.h.
9961 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9963 * haifa-sched.c (htab_t delay_htab):
9964 Change type to hash_table. Update dependent calls and types.
9965 (htab_t delay_htab_i2): Likewise.
9967 * ira-color.c (htab_t allocno_hard_regs_htab):
9968 Change type to hash_table. Update dependent calls and types.
9970 * ira-costs.c (htab_t cost_classes_htab):
9971 Change type to hash_table. Update dependent calls and types.
9973 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9974 Change type to hash_table. Update dependent calls and types.
9976 * loop-iv.c (htab_t bivs):
9977 Change type to hash_table. Update dependent calls and types.
9979 * loop-unroll.c (htab_t opt_info::insns_to_split):
9980 Change type to hash_table. Update dependent calls and types.
9981 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9983 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9984 (struct string_slot_hasher): New.
9985 (htab_t output_block::string_hash_table):
9986 Change type to hash_table. Update dependent calls and types.
9988 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9989 (htab_t file_name_hash_table):
9990 Change type to hash_table. Update dependent calls and types.
9992 * lto-streamer-out.c: Update output_block::string_hash_table dependent
9995 * lto-streamer.c (htab_t tree_htab):
9996 Change type to hash_table. Update dependent calls and types.
9998 * omp-low.c: Include gimplify-ctx.h.
10000 * passes.c (htab_t name_to_pass_map):
10001 Change type to hash_table. Update dependent calls and types.
10002 (pass_traverse): Rename to passes_pass_traverse.
10004 * plugin.c (htab_t event_tab):
10005 Change type to hash_table. Update dependent calls and types.
10007 * postreload-gcse.c (htab_t expr_table):
10008 Change type to hash_table. Update dependent calls and types.
10009 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10011 * sese.c (debug_rename_map_1): Make extern.
10012 (htab_t copy_bb_and_scalar_dependences::rename_map):
10013 Change type to hash_table. Update dependent calls and types.
10015 * sese.h (extern debug_rename_map): Move to .c file.
10017 * store-motion.c (htab_t store_motion_mems_table):
10018 Change type to hash_table. Update dependent calls and types.
10020 * trans-mem.c (htab_t tm_new_mem_hash):
10021 Change type to hash_table. Update dependent calls and types.
10023 * tree-browser.c (htab_t TB_up_ht):
10024 Change type to hash_table. Update dependent calls and types.
10026 * tree-cfg.c (htab_t discriminator_per_locus):
10027 Change type to hash_table. Update dependent calls and types.
10029 * tree-complex.c: Include tree-hasher.h
10030 (htab_t complex_variable_components):
10031 Change type to hash_table. Update dependent calls and types.
10033 * tree-eh.c (htab_t finally_tree):
10034 Change type to hash_table. Update dependent calls and types.
10036 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10037 struct int_tree_hasher.
10038 (extern int_tree_map_eq): Likewise.
10039 (uid_decl_map_hash): Removed.
10040 (extern decl_tree_map_eq): Likewise.
10042 * tree-hasher.h: New.
10043 (struct int_tree_hasher): New.
10044 (typedef int_tree_htab_type): New.
10046 * tree-inline.c: Include gimplify-ctx.h.
10048 * tree-mudflap.c: Include gimplify-ctx.h.
10050 * tree-parloops.c: Include tree-hasher.h.
10051 (htab_t eliminate_local_variables_stmt::decl_address):
10052 Change type to hash_table. Update dependent calls and types.
10053 (htab_t separate_decls_in_region::decl_copies): Likewise.
10055 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10056 Change type to hash_table. Update dependent calls and types.
10058 * tree-sra.c (candidates):
10059 Change type to hash_table. Update dependent calls and types.
10061 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10063 (int_tree_map_hash): Likewise.
10065 * tree-ssa-dom.c (htab_t avail_exprs):
10066 Change type to hash_table. Update dependent calls and types.
10068 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10069 Change type to hash_table. Update dependent calls and types.
10071 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10072 Change type to hash_table. Update dependent calls and types.
10074 * tree-ssa-phiopt.c (seen_ssa_names):
10075 Change type to hash_table. Update dependent calls and types.
10077 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10078 Change type to hash_table. Update dependent calls and types.
10080 * tree-ssa-uncprop.c (equiv):
10081 Change type to hash_table. Update dependent calls and types.
10083 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10085 PR rtl-optimization/57003
10086 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10087 call note_stores with kill_clobbered_value callback again after
10088 killing regs_invalidated_by_call.
10090 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10092 * config/aarch64/aarch64-simd.md
10093 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10094 (aarch64_simd_bsl<mode>): Likewise.
10095 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10097 2013-04-25 Marek Polacek <polacek@redhat.com>
10099 PR tree-optimization/57066
10100 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10102 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10104 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10106 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10108 * config/aarch64/aarch64-builtins.c
10109 (aarch64_fold_builtin): New.
10110 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10111 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10112 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10113 * config/aarch64/arm_neon.h
10114 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10116 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10117 Tejas Belagod <tejas.belagod@arm.com>
10119 * config/aarch64/aarch64-builtins.c
10120 (aarch64_gimple_fold_builtin): New.
10121 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10122 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10123 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10124 (addvv4sf): Update.
10125 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10127 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10129 * config/aarch64/aarch64.md
10130 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10132 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10134 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10135 (*ngcsi_uxtw): New pattern.
10137 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10138 Julian Brown <julian@codesourcery.com>
10140 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
10141 (TB_DREG): Add T_V4HF.
10142 (v4hf_UP): New macro.
10143 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10144 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10145 Handle initialisation of V4HF. Adjust initialisation of reinterpret
10147 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10148 (arm_vector_mode_supported_p): Handle V4HF.
10149 (arm_mangle_map): Handle V4HFmode.
10150 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
10151 * config/arm/arm_neon_builtins.def: Add entries for
10152 vcvtv4hfv4sf, vcvtv4sfv4hf.
10153 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
10154 (neon_vcvtv4hfv4sf): Likewise.
10155 * config/arm/neon-gen.ml: Handle half-precision floating point
10157 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
10158 * config/arm/arm_neon.h: Regenerate.
10159 * config/arm/neon.ml (type elts): Add F16.
10160 (type vectype): Add T_float16x4, T_floatHF.
10161 (type vecmode): Add V4HF.
10162 (type features): Add Requires_FP_bit feature.
10163 (elt_width): Handle F16.
10164 (elt_class): Likewise.
10165 (elt_of_class_width): Likewise.
10166 (mode_of_elt): Refactor.
10167 (type_for_elt): Handle F16, fix error messages.
10168 (vectype_size): Handle T_float16x4.
10169 (vcvt_sh): New function.
10170 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
10171 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
10172 (string_of_mode): Handle V4HF.
10173 * doc/arm-neon-intrinsics.texi: Regenerate.
10175 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10177 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
10178 format specifier in 'X' case.
10180 2013-04-25 Alan Modra <amodra@gmail.com>
10183 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
10184 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
10185 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
10186 Repeat for many other rotate/shift and mask patterns using subregs.
10187 Name lshiftrt insns.
10188 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
10189 on WORDS_BIG_ENDIAN.
10191 2013-04-25 Alan Modra <amodra@gmail.com>
10193 * config.gcc: Support little-endian powerpc-linux targets.
10194 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
10195 (LINK_OS_LINUX_SPEC): Define.
10196 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
10197 Preserve MASK_LITTLE_ENDIAN.
10198 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
10199 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
10200 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
10201 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
10202 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
10203 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
10204 Correct fp word order for little-endian. Don't shift toc entries
10205 smaller than a word for little-endian.
10206 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
10207 (bswapdi2 splits): Correct low-part subreg for little-endian.
10208 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
10209 low/high where such is correct only for be.
10210 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
10211 little-endian for -mcall-aixdesc.
10213 2013-04-25 Alan Modra <amodra@gmail.com>
10215 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10216 replace_equiv_address_nv.
10218 2013-04-25 Alan Modra <amodra@gmail.com>
10220 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
10222 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10225 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10226 * rtl.h (struct rtx_def): ...
10228 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10230 PR rtl-optimizations/57046
10231 * lra-constraints (split_reg): Set up lra_risky_transformations_p
10232 for multi-reg splits.
10234 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
10236 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
10238 2013-04-24 Sterling Augustine <saugustine@google.com>
10240 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
10241 (comp_dir_string, debug_str_dwo_section): New.
10242 (DEBUG_STR_DWO_SECTION): Rename to ...
10243 (DEBUG_DWO_STR_SECTION): ... this.
10244 (DEBUG_NORM_STR_SECTION): Delete.
10245 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
10246 (DEBUG_STR_DWO_SECTION_FLAGS): New.
10247 (find_AT_string): Move most logic to ...
10248 (find_AT_string_in_table): ... here. New.
10249 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
10250 add_skeleton_AT_string. Delete logic.
10251 (output_skeleton_debug_sections): Remove call to
10252 add_top_level_skeleton_die_attrs.
10253 (add_comp_dir_attribute): Move logic to comp_dir_string.
10254 (dwarf2out_init): Initialize debug_str_dwo_section.
10255 (output_indirect_string): Call find_string_form.
10256 (output_indirect_strings): Rewrite.
10257 (prune_unused_types): Empty skeleton_debug_str_hash.
10258 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
10259 (dwarf2out_finish): Call output_indirect_strings.
10261 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10263 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
10265 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10267 * rtl.h (struct rtx_def): Add comment for field jump.
10268 (LRA_SUBREG_P): New macro.
10269 * recog.c (register_operand): Check LRA_SUBREG_P.
10270 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10272 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10273 (lra_final_code_change): Skip subreg change for operators.
10274 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10275 if there are no operand changes.
10276 * lra-constraints.c (curr_insn_set): New.
10277 (match_reload): Set LRA_SUBREG_P.
10278 (emit_spill_move): Ditto.
10279 (check_and_process_move): Use curr_insn_set. Process only single
10280 set insns. Don't initialize sec_mem_p and change_p.
10281 (simplify_operand_subreg): Use LRA_SUBREG_P.
10282 (reg_in_class_p): New function.
10283 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10284 of #ifdef. Add code to remove cycling.
10285 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10286 non-null disp. Reload inner instead of disp when base and index
10287 are null. Try to put lo_sum into register.
10288 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10289 (check_and_process_move): Move code for move cost check to
10290 simple_move_p. Remove equiv_substitution.
10291 (simple_move_p): New function.
10292 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10293 curr_insn_set. Call check_and_process_move only for single set
10294 insns. Use the new function. Move call of check_and_process_move
10295 after operand equiv substitution and address process.
10297 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
10299 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
10300 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
10301 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
10303 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10305 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
10307 2013-04-24 Marek Polacek <polacek@redhat.com>
10309 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
10310 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
10311 (select_loops_exit_conditions): Likewise.
10312 (number_of_iterations_for_all_loops): Likewise.
10313 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
10314 (scev_analysis): Likewise.
10316 2013-04-02 Catherine Moore <clm@codesourcery.com>
10317 Chao-ying Fu <fu@mips.com>
10319 * config/mips/micromips.md (jraddiusp): New pattern.
10320 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
10321 instruction if possible.
10323 2013-04-24 Alan Modra <amodra@gmail.com>
10325 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
10327 2013-04-24 Julian Brown <julian@codesourcery.com>
10328 Chung-Lin Tang <cltang@codesourcery.com>
10330 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
10331 dependency behavior in enumeration type DIE generation. Add TODO note
10332 to comments about future DW_FORM_sdata/udata re-work of related code.
10334 2013-04-23 Lawrence Crowl <crowl@google.com>
10336 * Makefile.in: Update as needed below.
10338 * hash-table.h (class hash_table):
10339 Correct many methods with parameter types compare_type to the correct
10340 value_type. (Correct code was unlikely to notice the change.)
10341 (hash_table::elements_with_deleted) New.
10342 (class hashtable::iterator): New.
10343 (hashtable::begin()): New.
10344 (hashtable::end()): New.
10345 (FOR_EACH_HASH_TABLE_ELEMENT): New.
10347 * statistics.c (statistics_hashes):
10348 Change type to hash_table. Update dependent calls and types.
10350 * tree-into-ssa.c (var_infos):
10351 Change type to hash_table. Update dependent calls and types.
10353 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
10354 Change type to hash_table. Update dependent calls and types.
10356 * tree-ssa-loop-im.c (struct mem_ref.refs):
10357 Change type to hash_table. Update dependent calls and types.
10359 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
10360 Change type to hash_table. Update dependent calls and types.
10362 * tree-ssa-sccvn.c (vn_tables_s::nary):
10363 Change type to hash_table. Update dependent calls and types.
10364 (vn_tables_s::phis): Likewise.
10365 (vn_tables_s::references): Likewise.
10367 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
10368 (vn_reference_eq): Update parameter and return types.
10370 * tree-ssa-structalias.c (pointer_equiv_class_table):
10371 Change type to hash_table. Update dependent calls and types.
10372 (location_equiv_class_table): Likewise.
10374 * tree-vect-data-refs.c: Consequential changes for making
10375 peeling a hash_table.
10377 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
10378 (destroy_loop_vec_info): Dependent hash_table update.
10380 * tree-vectorizer.h (peeling_htab):
10381 Change type to hash_table. Update dependent calls and types.
10383 2013-04-23 Shiva Chen <shiva0217@gmail.com>
10385 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
10386 to check the register content is equal or not.
10387 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10388 to assign register content record.
10389 * lra-eliminations.c (update_reg_eliminate): Use
10390 lra_update_reg_val_offset to update register content offset.
10391 * lra-int.h (struct lra_reg): Add offset member.
10392 (lra_reg_val_equal_p): New static inline function.
10393 (lra_update_reg_val_offset): New static inline function.
10394 (lra_assign_reg_val): New static inline function.
10395 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10396 to assign register content record.
10397 (initialize_lra_reg_info_element): Initial offset to zero.
10399 2013-04-23 Catherine Moore <clm@codesourcery.com>
10401 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10402 operands. Record compression.
10404 2013-04-23 Xinliang David Li <davidxl@google.com>
10406 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10408 2013-04-23 Richard Biener <rguenther@suse.de>
10410 PR middle-end/57036
10411 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10412 parameter, only add abnormal goto edges from the copied body
10413 if the call could perform abnormal gotos.
10414 (copy_cfg_body): Adjust.
10416 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10418 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10420 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10422 * coretypes.h (gimple_stmt_iterator): Add struct to make
10425 2013-04-23 Richard Biener <rguenther@suse.de>
10427 PR tree-optimization/57026
10428 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10429 from SSA names occuring in abnormal PHI nodes.
10431 2013-04-22 Andi Kleen <ak@linux.intel.com>
10433 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10435 2013-04-22 Andi Kleen <ak@linux.intel.com>
10437 * lto/lto.c (print_lto_report_1): Declare early.
10438 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10440 2013-04-22 Andi Kleen <ak@linux.intel.com>
10442 * common.opt (-flto-report-wpa): Add.
10443 * doc/invoke.texi (-flto-report-wpa): Add.
10444 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10447 2013-04-22 Xinliang David Li <davidxl@google.com>
10449 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10450 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10451 * Makefile.in: New dependency
10453 David Daney <ddaney.cavm@gmail.com>
10455 * configure.ac (gcc_cv_as_micromips_support): Use the
10456 --fatal-warnings option.
10457 * configure: Regenerate.
10459 2013-04-22 Marek Polacek <polacek@redhat.com>
10462 * tsan.c (instrument_expr): Don't instrument expression
10463 in case its size is zero.
10465 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10469 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10471 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10473 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10475 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10476 (gimple_stmt_iterator): New typedef.
10477 * gimple.h (gimple_stmt_iterator): Rename to...
10478 (gimple_stmt_iterator_d): ... This.
10479 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10480 trees be valid for GIMPLE and GENERIC.
10481 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10482 * gimple-fold.c (gimple_fold_call): Call target hook
10483 gimple_fold_builtin.
10484 * hooks.c (hook_bool_gsiptr_false): New.
10485 * hooks.h (hook_bool_gsiptr_false): New.
10486 * target.def (fold_stmt): New.
10487 * doc/tm.texi: Regenerate.
10489 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10492 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10493 a set sp if no stack realignment.
10495 2013-04-22 Nick Clifton <nickc@redhat.com>
10497 * config.gcc (tilegx-linux): Extend extra_objs rather than
10499 (tilepro-linux): Likewise.
10501 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10503 * config/aarch64/aarch64-builtins.c
10505 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10506 (VAR<1-12>): Add MAP parameter.
10507 (BUILTIN_*): Likewise.
10508 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10509 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10510 (aarch64_ushl_n<mode>): Likewise.
10511 (aarch64_sshr_n<mode>): Likewise.
10512 (aarch64_ushr_n<mode>): Likewise.
10513 (aarch64_<maxmin><mode>): Likewise.
10514 (aarch64_sqrt<mode>): Likewise.
10515 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10516 (vshr<q>_n_*): Likewise.
10518 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10520 * config/aarch64/aarch64-builtins.c
10521 (aarch64_simd_builtin_type_mode): Handle SF types.
10523 (BUILTIN_GPF): Define.
10524 (aarch64_init_simd_builtins): Handle SF types.
10525 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10526 (frecps): Likewise.
10527 (frecpx): Likewise.
10528 * config/aarch64/aarch64-simd.md
10529 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10530 (aarch64_frecpe<mode>): New.
10531 (aarch64_frecps<mode>): Likewise.
10532 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10533 (v8type): Add frecp<esx>.
10534 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10535 (aarch64_frecps<mode>): Likewise.
10536 * config/aarch64/iterators.md (FRECP): New.
10537 (frecp_suffix): Likewise.
10538 * config/aarch64/arm_neon.h
10539 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10541 2013-04-22 Christian Bruel <christian.bruel@st.com>
10544 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10545 (REG_CLASS_NAMES): Idem.
10546 (REG_CLASS_CONTENTS): Idem.
10547 (REGCLASS_HAS_FP_REG): Idem.
10548 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10549 (sh_conditional_register_usage): Idem.
10551 2013-04-21 Jeff Law <law@redhat.com>
10553 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10554 (ssa_forward_propagate_and_combine): Use it.
10556 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10558 * lra.c: Update the flow chart diagram.
10560 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10562 PR rtl-optimization/56847
10563 * lra-constraints.c (process_alt_operands): Discourage alternative
10564 with non-matche doffsettable memory constraint fro memory with
10567 2013-04-19 Richard Biener <rguenther@suse.de>
10569 PR tree-optimization/56982
10570 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10572 * gimplify.c (gimplify_call_expr): Notice special calls.
10573 (gimplify_modify_expr): Likewise.
10574 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10575 abnormal control flow receivers.
10576 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10577 in the same way as cfun->has_nonlocal_labels.
10578 (gimple_purge_dead_abnormal_call_edges): Likewise.
10579 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10580 receivers start a basic-block.
10582 2013-04-19 Richard Biener <rguenther@suse.de>
10584 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10586 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10587 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10588 (SLP_TREE_LOAD_PERMUTATION): Add.
10589 (vect_transform_slp_perm_load): Adjust prototype.
10590 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10591 (vect_free_slp_instance): Likewise.
10592 (vect_create_new_slp_node): Likewise.
10593 (vect_supported_slp_permutation_p): Remove.
10594 (vect_slp_rearrange_stmts): Adjust.
10595 (vect_supported_load_permutation_p): Likewise. Inline
10596 vect_supported_slp_permutation_p here.
10597 (vect_analyze_slp_instance): Compute load permutations per
10598 slp node instead of per instance.
10599 (vect_get_slp_defs): Adjust.
10600 (vect_transform_slp_perm_load): Likewise.
10601 (vect_schedule_slp_instance): Remove redundant code.
10602 (vect_schedule_slp): Remove hack for PR56270, add it ...
10603 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10604 CSE loads for SLP. Adjust.
10606 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10608 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10609 spelling in two comments.
10611 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10614 * config/arm/arm.c (load_multiple_sequence): Require SP
10615 as base register for loads if SP is in the register list.
10617 2013-04-19 Martin Jambor <mjambor@suse.cz>
10619 PR tree-optimization/56718
10620 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10621 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10622 and made public. Adjusted all callers.
10623 (ipa_intraprocedural_devirtualization): New function.
10624 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10625 (ipa_intraprocedural_devirtualization): Likewise.
10626 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10628 2013-04-19 Richard Biener <rguenther@suse.de>
10630 PR tree-optimization/57000
10631 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10633 2013-04-19 Terry Guo <terry.guo@arm.com>
10635 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10637 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10638 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10639 (cortex_m4_fmacs): Use new reservations.
10640 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10642 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10644 PR rtl-optimization/56999
10645 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10647 (lra_coalesce): Remove split_origin_bitmap and related code.
10648 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10649 ranges if necessary.
10651 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
10653 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10655 (ix86_expand_call): Remove clobbered_registers array and use
10656 x86_64_ms_sysv_extra_clobbered_registers instead.
10657 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10659 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10661 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10662 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10663 (*call_value_rex64_ms_sysv): Ditto.
10665 2013-04-18 Cary Coutant <ccoutant@google.com>
10667 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10668 parent instead of die_mark.
10670 2013-04-18 Diego Novillo <dnovillo@google.com>
10672 * gimple.c (create_gimple_tmp): New.
10673 (get_expr_type): New.
10674 (build_assign): New.
10675 (build_type_cast): New.
10676 * gimple.h (enum ssa_mode): Define.
10677 (gimple_seq_set_location): New.
10678 * asan.c (build_check_stmt): Change some gimple_build_* calls
10679 to use build_assign and build_type_cast.
10681 2013-04-18 Richard Biener <rguenther@suse.de>
10683 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10684 handle negative step. Remove redundant checks.
10685 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10686 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10687 for negative step and grouped loads fail to vectorize.
10689 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10691 * emit-rtl.c (reset_insn_used_flags): New function.
10692 (reset_all_used_flags): Use it.
10693 (verify_insn_sharing): New function.
10694 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10696 2013-04-18 Jakub Jelinek <jakub@redhat.com>
10698 PR tree-optimization/56984
10699 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10700 and (x >> M) >= N don't register any assertion if N << M is the
10703 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10705 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10706 do not delete_insn insns that have not yet been emitted, only
10707 unlink them with remove_insn.
10708 * df-scan.c (df_insn_delete): Revert r197492.
10710 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10712 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10713 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10715 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10717 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10718 define_insn_and_split.
10719 (and_scc,ior_scc,negscc): Likewise.
10720 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10722 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10724 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10725 can benefit from using a sequence of LDRD instructions in epilogue
10726 instead of a single LDM instruction.
10728 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10731 * doc/extend.texi: Fix typo.
10733 2013-04-17 Richard Biener <rguenther@suse.de>
10735 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10736 (vect_build_slp_tree): ... here.
10737 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10738 match. Remove special-casing of mismatched complex loads.
10739 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10740 re-try the match with swapped commutative operands.
10741 (vect_supported_load_permutation_p): Remove special-casing of
10742 mismatched complex loads.
10743 (vect_analyze_slp_instance): Adjust.
10745 2013-04-17 Richard Biener <rguenther@suse.de>
10747 PR rtl-optimization/56921
10748 * cfgloop.h (struct loop): Add simple_loop_desc member.
10749 (struct niter_desc): Mark with GTY(()).
10750 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10751 * loop-iv.c (get_simple_loop_desc): Likewise.
10752 (free_simple_loop_desc): Likewise.
10755 2013-04-16 Richard Biener <rguenther@suse.de>
10757 PR rtl-optimization/56921
10758 * loop-init.c (pass_rtl_move_loop_invariants): Add
10759 TODO_do_not_ggc_collect to todo_flags_finish.
10760 (pass_rtl_unswitch): Same.
10761 (pass_rtl_unroll_and_peel_loops): Same.
10762 (pass_rtl_doloop): Same.
10764 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10766 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10767 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10768 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10770 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10771 * tree-streamer.c (record_common_node): Adjust reference in comment.
10773 2013-04-17 Terry Guo <terry.guo@arm.com>
10775 * config/arm/cortex-m4.md: Add a new bypass.
10777 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10779 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10781 (*subs_<optab><mode>_multp2): New pattern.
10782 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10783 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10785 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10787 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10788 (*subs_mul_imm_<mode>): New pattern.
10790 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
10793 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10794 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10795 (vsx_movti_32bit): Change j->wa to O->wa.
10797 2013-04-16 Richard Biener <rguenther@suse.de>
10799 PR rtl-optimization/56921
10800 * loop-init.c (pass_rtl_move_loop_invariants): Add
10801 TODO_do_not_ggc_collect to todo_flags_finish.
10802 (pass_rtl_unswitch): Same.
10803 (pass_rtl_unroll_and_peel_loops): Same.
10804 (pass_rtl_doloop): Same.
10806 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
10808 * config/arm/arm.c (emit_multi_reg_push): New declaration
10809 for an existing function.
10810 (arm_emit_strd_push): New function.
10811 (arm_expand_prologue): Used here.
10812 (arm_emit_ldrd_pop): New function.
10813 (arm_expand_epilogue): Used here.
10814 (arm_get_frame_offsets): Update condition.
10815 (arm_emit_multi_reg_pop): Add a special case for load of a single
10816 register with writeback.
10818 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
10820 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10823 2013-04-16 Richard Biener <rguenther@suse.de>
10825 PR tree-optimization/56756
10826 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10827 (first_mem_ref_loc): New.
10828 (execute_sm): Place the load temporarily before a previous
10829 access instead of in the latch edge to ensure its SSA dependencies
10830 are defined at points dominating the load.
10832 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10834 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10835 correct fix by moving header and footer insn to the footer of
10836 the merged basic block. Clear BB_END of the merged-away block.
10838 PR middle-end/43631
10839 * emit-rtl.c (make_note_raw): New function.
10840 (link_insn_into_chain): New static inline function.
10841 (add_insn): Use it.
10842 (add_insn_before, add_insn_after): Factor insn chain linking code...
10843 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10844 using link_insn_into_chain.
10845 (note_outside_basic_block_p): New helper function for emit_note_after
10846 and emit_note_before.
10847 (emit_note_after): Use nobb variant of add_insn_after if the note
10848 should not be contained in a basic block.
10849 (emit_note_before): Use nobb variant of add_insn_before if the note
10850 should not be contained in a basic block.
10851 (emit_note_copy): Use make_note_raw.
10852 (emit_note): Likewise.
10853 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10854 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10855 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10856 the moved barrier the tail of the basic block it follows.
10857 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10859 2013-04-15 Jakub Jelinek <jakub@redhat.com>
10861 PR tree-optimization/56962
10862 * gimple-ssa-strength-reduction.c (record_increment): Only set
10863 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10864 either rhs1 or rhs2 is equal to c->base_expr.
10866 2013-04-15 Richard Biener <rguenther@suse.de>
10868 PR tree-optimization/56933
10869 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10871 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10872 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10873 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10874 dependence check ...
10875 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10877 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10878 GROUP_READ_WRITE_DEPENDENCE.
10880 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10882 * emit-rtl.c (reset_all_used_flags): New function.
10883 (verify_rtl_sharing): Call reset_all_used_flags before and after
10884 performing the checks.
10886 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10888 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10889 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10890 * config/arm/constraints.md (De): New constraint.
10891 * config/arm/neon.md (anddi3_neon): Delete.
10892 (neon_vand<mode>): Expand to standard anddi3 pattern.
10893 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10894 Move earlier in the file.
10895 (neon_inv_logic_op2): Likewise.
10896 (arm_anddi_operand_neon): New predicate.
10898 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10900 * configure.ac (gcc_cv_ld_as_needed): Set
10901 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10902 Use -z ignore, -z record on *-*-solaris2*.
10903 (HAVE_LD_AS_NEEDED): Update comment.
10904 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10905 * configure: Regenerate.
10906 * config.in: Regenerate.
10907 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10908 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10909 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10910 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10911 equivalents. Fix markup.
10912 * doc/tm.texi: Regenerate.
10914 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10916 * config/i386/i386.opt: New option mstack-protector-guard=.
10917 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10918 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10919 TARGET_SSP_TLS_GUARD.
10920 * config/i386/i386.c (ix86_option_override_internal): Set
10921 ix86_stack_protector_guard.
10922 * config/i386/i386.md (stack_protect_set): Enable for
10923 TARGET_SSP_TLS_GUARD only.
10924 (stack_protect_set_<mode>): Ditto.
10925 (stack_protect_test): Ditto.
10926 (stack_protect_test_<mode>): Ditto.
10927 * doc/invoke.texi (i386 Option): Document.
10929 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10932 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10933 (S_MODES): Set H_MODE bit.
10934 (SF_MODES): Set only S_MODE and SF_MODE bits.
10935 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10936 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10937 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10938 <MODE_FLOAT>: Likewise.
10940 2013-04-15 Joey Ye <joey.ye@arm.com>
10942 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10944 2013-04-15 Joey Ye <joey.ye@arm.com>
10946 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10948 (thumb_far_jump_used_p): Count instruction size and set
10951 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10953 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10954 * resource.c (mark_target_live_regs): Reformat conditional block.
10956 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10958 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10959 notes, they are emitted only just before final.
10960 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10962 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10964 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10965 * cfgrtl.c (delete_insn): Call it here instead.
10966 * lra-spills.c (lra_final_code_change): Use delete_insn.
10967 * haifa-sched.c (sched_remove_insn): Likewise.
10968 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10969 returning to the nop pool.
10970 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10971 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10973 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10975 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10976 * doc/tm.texi: Regenerated.
10978 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
10980 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10983 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10985 * df-core.c (df_find_def): Compare register numbers.
10986 (df_find_use): Likewise.
10988 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10991 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
10992 lra_in_progress for return.
10994 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10996 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
10997 define_insn into define_insn_and_split and emit movsicc patterns.
10999 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11001 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11003 2013-04-12 Richard Biener <rguenther@suse.de>
11005 * tree-pass.h (TODO_do_not_ggc_collect): New.
11006 * passes.c (execute_one_ipa_transform_pass): Honor
11007 TODO_do_not_ggc_collect.
11008 (execute_one_pass): Likewise.
11011 2013-04-10 Richard Biener <rguenther@suse.de>
11013 * passes.c (init_optimization_passes): Remove reload pass.
11014 * ira.c (do_reload): Merge into ...
11016 (rest_of_handle_reload): Remove.
11017 (pass_reload): Likewise.
11018 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11019 of reload for vzeroupper pass placement.
11021 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11023 PR tree-optimization/56918
11024 PR tree-optimization/56920
11025 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11026 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11027 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11028 use wide_mul_with_sign method.
11030 2013-04-12 Richard Biener <rguenther@suse.de>
11032 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11033 not be considered a gimple constant.
11035 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11037 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11038 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11041 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11042 Jakub Jelinek <jakub@redhat.com>
11044 * opts.c: Include diagnostic-color.h.
11045 (common_handle_option): Handle OPT_fdiagnostics_color_.
11046 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11047 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11048 (diagnostic-color.o): New.
11049 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11050 (diagnostic_color_rule): New enum.
11051 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11052 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11053 the location string.
11054 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11055 either NULL, or color kind.
11056 * diagnostic-color.c: New file.
11057 * diagnostic-color.h: New file.
11058 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11060 * doc/invoke.texi (-fdiagnostics-color): Document.
11061 * pretty-print.h (pp_show_color): Define.
11062 (struct pretty_print_info): Add show_color field.
11063 * diagnostic.c: Include diagnostic-color.h.
11064 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11065 macros. Colorize error:, warning: etc. strings and also the location
11067 (diagnostic_show_locus): Colorize the caret line.
11068 * pretty-print.c: Include diagnostic-color.h.
11069 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11070 inside of %< %> quotes or quoted through q format modifier.
11072 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11074 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11076 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11078 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11079 code in CC_NZ mode.
11080 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11083 2013-04-11 Marek Polacek <polacek@redhat.com>
11085 PR tree-optimization/48184
11086 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11088 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11090 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11091 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11092 (skip_simple_arithmetic): Tidy up.
11093 * tree.h (skip_simple_constant_arithmetic): Declare.
11095 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11097 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11099 2013-04-11 Richard Biener <rguenther@suse.de>
11101 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11102 generate vector constants.
11104 2013-04-11 Richard Biener <rguenther@suse.de>
11106 PR tree-optimization/56878
11107 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11108 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11110 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11111 Prefer to align the DR with the most invariant base address.
11113 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11115 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11118 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11120 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11121 floating-point vector comparisons against 0.
11123 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11125 PR tree-optimization/56899
11126 * fold-const.c (extract_muldiv_1): Apply distributive law
11127 only if TYPE_OVERFLOW_WRAPS (ctype).
11129 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11132 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11133 loading has side effect.
11135 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
11137 * config/sparc/sparc.c: Include tree-pass.h.
11138 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
11139 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
11140 head of file. Change return type. Split off gate function.
11141 (sparc_gate_work_around_errata): New function.
11142 (pass_work_around_errata): New pass definition.
11143 (insert_pass_work_around_errata) New pass insert definition to
11144 insert pass_work_around_errata just after delayed-branch scheduling.
11145 (sparc_option_override): Insert the pass.
11146 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
11148 2013-04-10 David S. Miller <davem@davemloft.net>
11150 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
11151 or -mcpu=hypersparc.
11153 * target.def (cstore_mode): New hook.
11154 * target.h: Include insn-codes.h
11155 * targhooks.c: Likewise.
11156 (default_cstore_mode): New function.
11157 * targhooks.h: Declare it.
11158 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
11159 * doc/tm.texi: Rebuild.
11160 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
11161 target hook, rather than inspecting the insn_data.
11162 * config/sparc/sparc.c (sparc_cstore_mode): New function.
11163 (TARGET_CSTORE_MODE): Redefine.
11164 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
11166 * config/sparc/predicates.md (cstore_result_operand): New special
11168 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
11169 Use it for operand 0.
11170 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
11171 (*snesi_special): Likewise.
11172 (*snesi_zero): Likewise.
11173 (*seqsi_zero): Likewise.
11174 (*sltu_insn): Likewise.
11175 (*sgeu_insn): Likewise.
11176 (*seqdi_special): Make operand 0 and comparison operation be of
11178 (*snedi_special): Likewise.
11179 (*snedi_special_vis3): Likewise.
11180 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
11181 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
11182 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
11183 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
11184 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
11185 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
11186 (*sltu_extend_sp64): Likewise.
11187 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
11188 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
11189 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
11190 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
11191 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
11193 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
11195 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
11196 (aarch64_start_file): Use the new function.
11198 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11199 Jason Merrill <jason@redhat.com>
11201 * common.opt: Add -gdwarf.
11202 * opts.c (common_handle_option): Handle it.
11203 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
11205 2013-04-10 Richard Biener <rguenther@suse.de>
11207 * passes.c (execute_todo): Do not call ggc_collect conditional here.
11208 (execute_one_ipa_transform_pass): But unconditionally here.
11209 (execute_one_pass): And here.
11210 (init_optimization_passes): Remove reload pass.
11211 * tree-pass.h (TODO_ggc_collect): Remove.
11212 (pass_reload): Likewise.
11213 * ira.c (do_reload): Merge into ...
11215 (rest_of_handle_reload): Remove.
11216 (pass_reload): Likewise.
11217 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11218 of reload for vzeroupper pass placement.
11219 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
11220 and todo_flags_finish of all passes.
11222 2013-04-10 Richard Biener <rguenther@suse.de>
11224 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
11225 first_const_oprnd field, rename first_def_type to first_op_type.
11226 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
11227 (vect_get_and_check_slp_defs): Always use the type of the
11228 operand. Allow mixed vect_external_def, vect_constant_def types.
11229 (vect_get_constant_vectors): Handle mixed vect_external_def,
11230 vect_constant_def types.
11232 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
11234 PR tree-optimization/55524
11235 * tree-ssa-math-opts.c
11236 (convert_mult_to_fma): Don't use an fms construct
11237 when we don't have an fms operation, but fnma, and it looks
11238 likely that we'll be able to use the latter.
11240 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11242 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
11244 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
11245 inline fail caused by overwritable functions.
11247 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
11249 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
11250 unnecessary bits in the constant power of two case.
11252 2013-04-10 Richard Biener <rguenther@suse.de>
11254 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
11255 broken code swapping operands.
11256 (vect_build_slp_tree): Do not compute load permutations here.
11257 (vect_analyze_slp_instance): Compute load permutations here,
11258 after building the SLP tree.
11260 2013-04-09 Christian Bruel <christian.bruel@st.com>
11262 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
11263 of next/prev_real_insn.
11265 2013-04-09 Jan Hubicka <jh@suse.cz>
11267 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11268 Drop aliased parameter.
11269 (function_and_variable_visibility): Do not handle alias pairs.
11270 * cgraph.c (varpool_externally_visible_p): Update prototype.
11271 * varpool.c (varpool_add_new_variable): Update.
11273 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11275 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
11277 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
11279 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
11281 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
11283 2013-04-09 Marek Polacek <polacek@redhat.com>
11285 PR tree-optimization/48762
11286 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
11288 2013-04-09 Richard Biener <rguenther@suse.de>
11290 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
11292 (vect_build_slp_tree): Likewise.
11293 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
11294 calculating the cost of a SLP instance.
11295 (vect_analyze_slp_instance): Use it from here, after building
11298 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11300 PR middle-end/56883
11301 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
11302 expand_omp_for_static_chunk): Use simple_p = true in
11303 force_gimple_operand_gsi calls when assigning to addressable decls.
11305 2013-04-09 Jeff Law <law@redhat.com>
11307 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
11308 when the boolean was created by converting a wider object which
11309 had a boolean range.
11311 2013-04-09 Richard Biener <rguenther@suse.de>
11313 * tree-vectorizer.h (slp_void_p): Remove.
11314 (slp_tree): Typedef before _slp_tree declaration.
11315 (struct _slp_tree): Use a vector of slp_tree as children.
11316 (vect_get_place_in_interleaving_chain): Remove.
11317 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
11319 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
11321 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
11322 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
11323 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
11324 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
11325 Use slp_node instead of slp_void_p and adjust.
11327 2013-04-09 Richard Biener <rguenther@suse.de>
11329 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
11330 work that is not necessary.
11332 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11334 PR tree-optimization/56854
11335 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
11336 forward into clobber stmts if it would change MEM_REF lhs into
11339 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
11341 * tree.c (type_hash_lookup, type_hash_add): Make static.
11342 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
11344 2013-04-09 Richard Biener <rguenther@suse.de>
11346 * tree.h (unsave_expr_now): Remove.
11347 * tree-inline.c (mark_local_for_remap_r): Remove.
11348 (unsave_expr_1): Likewise.
11349 (unsave_r): Likewise.
11350 (unsave_expr_now): Likewise.
11351 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
11352 (propagate_tree_value): Likewise.
11354 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
11356 * doc/rtl.texi (sequence): Rewrite documentation to match the
11357 current use of SEQUENCE rtl objects.
11358 * rtl.def (SEQUENCE): Likewise.
11360 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
11361 Update documentation.
11362 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
11363 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
11365 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
11367 2013-04-08 Teresa Johnson <tejohnson@google.com>
11369 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
11370 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
11372 (estimate_edge_size_and_time): Add comment to suggest using rounding
11374 (estimate_node_size_and_time): Ditto.
11375 (remap_edge_change_prob): Use helper rounding divide methods.
11376 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
11377 (gimple_mod_pow2_value_transform): Ditto.
11378 (gimple_mod_subtract_transform): Ditto.
11379 (gimple_ic_transform): Ditto.
11380 (gimple_stringops_transform): Ditto.
11381 * stmt.c (conditional_probability): Ditto.
11382 (emit_case_dispatch_table): Ditto.
11383 * lto-cgraph.c (merge_profile_summaries): Ditto.
11384 * tree-optimize.c (execute_fixup_cfg): Ditto.
11385 * cfgcleanup.c (try_forward_edges): Ditto.
11386 * cfgloopmanip.c (scale_loop_profile): Ditto.
11388 (duplicate_loop_to_header_edge): Ditto.
11389 (lv_adjust_loop_entry_edge): Ditto.
11390 * tree-vect-loop.c (vect_transform_loop): Ditto.
11391 * profile.c (compute_branch_probabilities): Ditto.
11392 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11393 * lto-streamer-in.c (input_cfg): Ditto.
11394 * gimple-streamer-in.c (input_bb): Ditto.
11395 * ipa-cp.c (update_profiling_info): Ditto.
11396 (update_specialized_profile): Ditto.
11397 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11398 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11400 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11401 (compute_trg_info): Ditto.
11402 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11403 (purge_dead_edges): Ditto.
11404 * loop-unswitch.c (unswitch_loop): Ditto.
11405 * cgraphclones.c (cgraph_clone_edge): Ditto.
11406 (cgraph_clone_node): Ditto.
11407 * tree-inline.c (copy_bb): Ditto.
11408 (copy_edges_for_bb): Ditto.
11409 (initialize_cfun): Ditto.
11410 (copy_cfg_body): Ditto.
11411 (expand_call_inline): Ditto.
11413 2013-04-08 Kai Tietz <ktietz@redhat.com>
11415 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11416 TARGET_CYGWIN64 by TARGET_64BIT.
11418 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11420 * config/epiphany/epiphany.md (GPR_1): New constant.
11421 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11422 * config/epiphany/epiphany.c (gen_compare_reg):
11423 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11424 is already in place.
11425 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11426 Don't require being called during rtl expansion; If y operlaps r0,
11428 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11429 (epiphany_expand_epilogue): Likewise.
11431 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11432 Don't use CC_FPmode for ORDERED / UNORDERED.
11433 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11435 * config/epiphany/constraints.md (CnL): New constraint.
11436 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11437 * config/epiphany/predicates.md (add_operand): Allow 1024.
11439 * config/epiphany/epiphany.md (logical_op): New code iterator.
11440 (op_mnc): New code attribute.
11441 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11442 (mov_f+1, mov_f+2): New peephole2 patterns.
11444 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11445 (cstoresi4): Also allow re-use of zero result when doing a NE
11446 comparison to a non-zero operand.
11447 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11449 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11450 Use gen_addsi3_i / gen_subsi3_i.
11452 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11456 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11457 contain anything but clobbers, at most one __builtin_stack_restore,
11458 optionally debug stmts and final resx, and if it has at least one
11459 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11460 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11461 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11462 which isn't defaut definition, remove them.
11463 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11464 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11465 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11466 with MEM_REF LHS with SSA_NAME address.
11468 2013-04-08 Jeff Law <law@redhat.com>
11470 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11472 2013-04-08 Richard Biener <rguenther@suse.de>
11474 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11476 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11477 determined vector type.
11478 (vect_analyze_data_refs): Likewise.
11479 (vect_get_new_vect_var): Adjust.
11480 (vect_create_destination_var): Preserve SSA name versions.
11481 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11482 not dump anything here.
11484 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11486 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11487 Add member lr_slot_known.
11488 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11490 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11491 Remove code that sets lr_slot_offset according to what a previous
11492 version of epiphany_emit_save_restore used to do.
11493 (epiphany_emit_save_restore): When doing an lr save or restore,
11494 set/verify lr_slot_known and lr_slot_offset.
11496 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11499 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11502 2013-04-08 Richard Biener <rguenther@suse.de>
11504 * alias.c (find_base_term): Fix thinko in previous change.
11506 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11508 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11509 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11510 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11511 if possible to compute val.
11512 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11513 For QImode integers don't require anything about precision. Use
11514 const_with_all_bytes_same to find out if the constant doesn't have
11515 repeated bytes in it.
11517 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11519 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11522 2013-04-08 Marek Polacek <polacek@redhat.com>
11524 PR rtl-optimization/48182
11525 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11528 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11531 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11532 nuses, make sure we have a label.
11534 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11537 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11538 (rs6000_emit_swdiv_low_precision): Remove.
11539 (rs6000_emit_swdiv): Rewrite to handle between one and four
11540 iterations of Newton-Raphson generally; modify required number of
11541 iterations for some cases.
11542 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11544 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11546 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11547 set-but-unused variable.
11549 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11550 basic blocks of released function bodies garbage-collectable.
11552 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11553 (struct rtl_opt_pass): Add TODO_df_finish.
11555 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11557 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11559 * config/arm/constraints.md (q): New constraint.
11560 * config/arm/ldrdstrd.md: New file.
11561 * config/arm/arm.md (ldrdstrd.md) New include.
11562 (arm_movdi): Use "q" instead of "r" constraint
11563 for double-word memory access.
11564 (movdf_soft_insn): Likewise.
11565 * config/arm/vfp.md (movdi_vfp): Likewise.
11566 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
11567 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
11568 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11569 (mem_ok_for_ldrd_strd): Likewise.
11570 (output_move_double): Update assertion.
11572 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11574 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11576 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11578 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11579 define_insn_and_split.
11580 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11582 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11584 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11585 define_insn_and_split.
11586 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
11587 (shiftsi3_compare): New pattern.
11588 (rrx): New pattern.
11589 * config/arm/unspecs.md (UNSPEC_RRX): New.
11591 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11593 * config/arm/arm.md (negdi_extendsidi): New pattern.
11594 (negdi_zero_extendsidi): Likewise.
11596 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11598 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11599 define_insn_and_split.
11600 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11601 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11603 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11605 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11606 define_insn_and_split.
11607 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11608 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11610 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11612 * config/arm/arm.md (subsi3_carryin): New pattern.
11613 (subsi3_carryin_const): Likewise.
11614 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11615 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11617 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11619 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11621 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11623 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
11624 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
11626 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11628 * config/arm/arm.c (arm_expand_builtin): Change fcode
11629 type to unsigned int.
11631 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11633 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11635 2013-04-04 Ian Lance Taylor <iant@google.com>
11637 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11640 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11642 PR middle-end/56729
11643 * df-scan.c (df_insn_delete): Disable failing assert.
11645 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11647 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11648 New function prototype.
11649 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11650 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11651 (arm_builtin_vectorized_function): New function.
11653 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11655 * config/arm/arm_neon_builtins.def: New file.
11656 * config/arm/arm.c (neon_builtin_data): Move contents to
11657 arm_neon_builtins.def.
11658 (enum arm_builtins): Include neon builtin definitions.
11659 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
11660 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
11662 2013-04-04 Marek Polacek <polacek@redhat.com>
11664 PR tree-optimization/48186
11665 * predict.c (maybe_hot_frequency_p): Return false if
11666 HOT_BB_FREQUENCY_FRACTION is 0.
11667 (cgraph_maybe_hot_edge_p): Likewise.
11669 2013-04-04 Richard Biener <rguenther@suse.de>
11671 PR tree-optimization/56826
11672 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11675 2013-04-04 Richard Biener <rguenther@suse.de>
11677 PR tree-optimization/56213
11678 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
11679 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
11681 2013-04-04 Richard Biener <rguenther@suse.de>
11683 PR tree-optimization/56837
11684 * tree-loop-distribution.c (classify_partition): For non-zero
11685 values require that the value has the same precision as its
11686 mode to be useful as memset value.
11688 2013-04-03 Nick Clifton <nickc@redhat.com>
11690 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
11691 (fmssf4): Use fmsf.s on E3V5 architectures.
11692 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11693 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11695 2013-04-03 Jeff Law <law@redhat.com>
11697 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11698 (lra-eliminations.o): Likewise.
11700 2013-04-03 Teresa Johnson <tejohnson@google.com>
11702 * gcov-io.c (compute_working_sets): Moved most of body of old
11703 compute_working_sets here from profile.c.
11704 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
11705 (gcov_working_set_t): Moved typedef here from basic-block.h
11706 (compute_working_set): Declare.
11707 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11708 (get_working_sets): Renamed from compute_working_set,
11709 replace most of body with call to new compute_working_sets.
11710 (get_exec_counts): Replace call to compute_working_sets
11711 to get_working_sets.
11712 * profile.h (get_working_sets): Renamed from compute_working_set.
11713 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
11714 to get_working_sets.
11715 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11716 * gcov-dump.c (dump_working_sets): New function.
11718 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11720 * hwint.c (sext_hwi, zext_hwi): New functions.
11721 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11722 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11723 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11724 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11725 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11726 (sext_hwi, zext_hwi): New functions.
11728 2013-04-03 Jeff Law <law@redhat.com>
11730 PR tree-optimization/56799
11731 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11732 back test for widening conversion erroneously dropped in prior change.
11734 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11737 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11738 instead of next_real_insn.
11740 2013-04-03 Marek Polacek <polacek@redhat.com>
11743 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
11745 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11748 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11750 (thumb1_output_casesi): Likewise.
11751 (thumb2_output_casesi): Likewise.
11753 2013-04-03 Richard Biener <rguenther@suse.de>
11755 PR tree-optimization/56817
11756 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11758 (tree_unroll_loops_completely_1): ... new function to manually
11759 walk the loop tree, properly defering outer loops of unrolled
11760 loops to later iterations.
11762 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11764 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11765 (vectorizable_load): Likewise.
11766 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11767 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11769 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11771 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11774 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11776 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11778 2013-04-03 Bin Cheng <bin.cheng@arm.com>
11780 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11782 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11784 PR tree-optimization/56790
11785 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11788 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11790 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11792 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11793 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11796 2013-04-03 Jakub Jelinek <jakub@redhat.com>
11799 * tree.h (force_folding_builtin_constant_p): New decl.
11800 * builtins.c (force_folding_builtin_constant_p): New variable.
11801 (fold_builtin_constant_p): Fold immediately also if
11802 force_folding_builtin_constant_p.
11804 2013-04-03 Richard Biener <rguenther@suse.de>
11806 PR tree-optimization/56812
11807 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11808 DRs of the same interleaving chain are independent.
11810 2013-04-02 Jason Merrill <jason@redhat.com>
11812 * gdbinit.in (pbb): Use debug fn.
11814 2013-04-02 Lawrence Crowl <crowl@google.com>
11816 * sese.h (struct ivtype_map_elt_s): Remove unused.
11817 (extern debug_ivtype_map): Remove unused.
11818 (extern eq_ivtype_map_elts): Remove unused.
11819 * sese.c (debug_ivtype_map): Removed unused.
11820 (debug_ivtype_map_1): Removed unused.
11821 (debug_ivtype_elt): Remove unused.
11822 (eq_ivtype_map_elts): Remove unused.
11825 2013-04-02 Kai Tietz <ktietz@redhat.com>
11828 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11829 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
11830 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11832 (legitimize_pe_coff_symbol): Likewise.
11833 (is_imported_p): New helper-function.
11834 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11835 for Windows x64 targets.
11836 (ix86_expand_prologue): Optimize for pe-coff targets.
11837 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11838 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11839 medium/large code-model.
11840 (legitimize_pic_address): Likewise.
11841 (legitimize_tls_address): Likewise.
11842 (ix86_expand_call): Likewise.
11843 (x86_output_mi_thunk): Likewise.
11844 (get_dllimport_decl): Add new beimport argument.
11845 (construct_plt_address): Don't assert for x64 pe-coff targets.
11846 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11848 (SYMBOL_FLAG_STUBVAR): New macro.
11849 (SYMBOL_REF_STUBVAR_P): Likewise.
11850 * config/i386/winnt.c (stub_list): New structure.
11851 (stub_head): New local variable.
11852 (i386_pe_record_stub): New function.
11853 (i386_pe_file_end): Emit refptr-stubs.
11855 2013-04-02 Jakub Jelinek <jakub@redhat.com>
11857 PR rtl-optimization/56745
11858 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11859 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11862 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11863 and both of them are MEM_REFs, just compare first argument for
11864 equality and attempt to deal even with differing offsets.
11867 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11868 of gimple_clobber_p to be MEM_REF.
11869 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11870 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11871 after gimplification.
11872 * asan.c (get_mem_ref_of_assignment): Don't instrument
11873 gimple_clobber_p stmts.
11874 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11875 gimple_clobber_p stmt if they have MEM_REF lhs and
11876 are dead because of another gimple_clobber_p stmt.
11877 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11878 gimple_clobber_p stmts like debug stmts.
11879 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11880 that refer to unused VAR_DECLs or uninitialized values.
11881 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11882 gimple_clobber_p stmts if they refer to removed parameters.
11883 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11886 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
11888 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11889 using SWI48 mode attribute.
11891 2013-04-02 Wei Mi <wmi@google.com>
11893 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11894 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11895 *<rotate_insn><mode>3_mask in i386.md.
11897 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11899 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11901 2013-04-02 Richard Biener <rguenther@suse.de>
11903 PR tree-optimization/56778
11904 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11905 Runtime alias tests are not supported for gather loads.
11906 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11907 stmts referenced from SSA operands before updating SSA form.
11909 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11910 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11912 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
11913 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11914 * config/arm/cortex-a53.md: New file.
11915 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11916 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11917 * config/arm/arm.c (arm_issue_rate): Likewise.
11918 * config/arm/arm-tune.md: Regenerate
11919 * config/arm/arm-tables.opt: Regenerate.
11920 * config/arm/arm-cores.def: Add cortex-a53.
11922 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11924 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11927 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11929 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11930 scalar load/store operations using B/H registers.
11931 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11933 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11935 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11937 * config/aarch64/aarch64.c
11938 (aarch64_simd_scalar_immediate_valid_for_move): New.
11939 * config/aarch64/aarch64-protos.h
11940 (aarch64_simd_scalar_immediate_valid_for_move): New.
11941 * config/aarch64/constraints.md (Dh, Dq): New.
11942 * config/aarch64/iterators.md (hq): New.
11944 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11946 * reorg.c (get_branch_condition): Deal with conditional returns.
11947 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11949 2013-04-01 Wei Mi <wmi@google.com>
11951 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11952 Truncate operand 2 using %b asm operand modifier.
11953 (*<shift_insn><mode>3_mask): Ditto.
11954 (*<rotate_insn><mode>3_mask): Ditto.
11956 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11958 PR middle-end/56798
11959 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11961 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11963 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11965 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11967 2013-03-30 Lawrence Crowl <crowl@google.com>
11969 * dse.c (clear_alias_sets): Remove never set.
11970 (disqualified_clear_alias_sets): Remove never set.
11971 (clear_alias_mode_pool): Remove never set.
11972 (dse_step0): Remove condition that is never true.
11973 (canon_address): Remove condition that is never true.
11974 (dse_step7): Remove condition that is never true.
11975 (rest_of_handle_dse): Remove condition that is never true.
11976 (rest_of_handle_dse::did_global): Remove never read from above.
11977 (dse_step2_spill): Remove never called from above.
11978 (dse_step5_spill): Remove never called from above.
11980 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11982 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11983 JUMP_TABLE_DATA changes.
11984 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11985 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11986 (Insns) <jump_table_data>: New entry.
11987 * doc/tm.texi: Regenerate.
11989 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11991 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
11992 for table jump at the end of a basic block using tablejump_p.
11993 * targhooks.c (default_invalid_within_doloop): Likewise.
11994 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
11995 target hook implementation that is identical to the default hook.
11996 (rs6000_invalid_within_doloop): Remove.
11998 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
11999 unused variable from tablejump_p call.
12001 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12002 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12003 (INSN_DELETED_P): Likewise.
12004 (emit_jump_table_data): New prototype.
12005 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12006 after 4th as unused.
12007 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12008 * sched-vis.c (print_insn): Likewise.
12009 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12010 insn for compatibility with back ends that use next_active_insn to
12011 identify jump table data.
12012 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12013 (remove_insn): Likewise.
12014 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12016 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12017 (emit_jump_table_data): New function.
12019 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12020 basic block, a JUMP_TABLE_DATA never is.
12021 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12022 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12023 off from code handling real insns.
12024 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12025 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12026 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12027 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12028 is not a NONDEBUG_INSN_P.
12029 * ira-costs.c (scan_one_insn): Likewise.
12030 * jump.c (mark_all_labels): Likewise.
12031 (mark_jump_label_1): Likewise.
12032 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12033 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12035 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12036 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12037 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12038 JUMP_TABLE_DATA_P insns.
12039 (calculate_elim_costs_all_insns): Likewise.
12040 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12041 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12042 (delete_output_reload): Code style fixups.
12043 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12044 insn flags on this non-insn.
12045 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12046 as scheduling barriers, for pre-change compatibility.
12047 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12048 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12050 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12051 redundant JUMP_TABLE_DATA_P test.
12052 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12053 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12054 (frv_for_each_packet): Likewise.
12055 * config/i386/i386.c (min_insn_size): Likewise.
12056 (ix86_avoid_jump_mispredicts): Likewise.
12057 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12058 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12059 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12060 (mips16_insn_length): Robustify.
12061 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12062 (mips16_split_long_branches): Likewise.
12063 * config/pa/pa.c (pa_combine_instructions): Likewise.
12064 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12065 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12066 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12067 as contributing to pool range lengths.
12068 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12069 Remove redundant JUMP_TABLE_DATA_P test.
12070 (sh_loop_align): Likewise.
12071 (split_branches): Likewise.
12072 (sh_insn_length_adjustment): Likewise.
12073 * config/spu/spu.c (get_branch_target): Likewise.
12075 2013-03-29 Jan Hubicka <jh@suse.cz>
12077 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12078 gcov streaming; stream hot bb threshold to ltrans.
12079 * predict.c (get_hot_bb_threshold): Break out from ....
12080 (maybe_hot_count_p): ... here.
12081 (set_hot_bb_threshold): New function.
12082 * lto-section-in.c (lto_section_name): Add profile.
12083 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12084 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12085 and data-streamer.h
12086 (histogram_entry): New structure.
12087 (histogram, histogram_pool): New global vars.
12088 (histogram_hash): New structure.
12089 (histogram_hash::hash): New method.
12090 (histogram_hash::equal): Likewise.
12091 (account_time_size): New function.
12092 (cmp_counts): New function.
12093 (dump_histogram): New function.
12094 (ipa_profile_generate_summary): New function.
12095 (ipa_profile_write_summary): New function.
12096 (ipa_profile_read_summary): New function.
12097 (ipa_profile): Decide on threshold.
12098 (pass_ipa_profile): Add ipa_profile_write_summary and
12099 ipa_profile_read_summary.
12100 * Makefile.in (ipa.o): Update dependencies.
12101 * lto-streamer.h (LTO_section_ipa_profile): New section.
12103 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12105 * tree.h (VAR_P): New.
12107 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12110 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12112 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12114 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12115 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12116 (control_flow_insn_p): Likewise.
12117 * cfgrtl.c (duplicate_insn_chain): Likewise.
12118 * final.c (get_attr_length_1): Likewise.
12119 (shorten_branches): Likewise.
12120 (final_scan_insn): Likewise.
12121 * function.c (instantiate_virtual_regs): Likewise.
12122 * gcse.c (insert_insn_end_basic_block): Likewise.
12123 * ira-costs.c (scan_one_insn): Likewise.
12124 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12125 * lra.c (check_rtl): Likewise.
12126 * reload1.c (elimination_costs_in_insn): Likewise.
12127 * reorg.c (follow_jumps): Likewise.
12129 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12130 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12131 (thumb_far_jump_used_p): Likewise.
12132 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12133 (workaround_speculation): Likewise.
12134 (add_sched_insns_for_speculation): Likewise.
12135 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
12136 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12137 (frv_for_each_packet): Likewise.
12138 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
12139 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
12140 (final_emit_insn_group_barriers): Likewise.
12141 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12142 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12143 (mips16_insn_length): Likewise.
12144 * config/pa/pa.c (pa_reorg): Likewise.
12145 (pa_combine_instructions): Likewise.
12146 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
12147 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
12148 (sh_reorg): Likewise.
12149 (split_branches): Likewise.
12150 * config/spu/spu.c (get_branch_target): Likewise.
12152 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
12155 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
12157 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
12158 Fix declaration name.
12160 2013-03-28 Lawrence Crowl <crowl@google.com>
12162 * graphds.h (struct graph.indicies): Remove unused.
12163 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
12164 (SCOP_ORIGINAL_PDDRS): Remove unused.
12165 * sese.h (extern insert_loop_close_phis): Removed unused.
12166 (extern insert_guard_phis): Removed unused.
12167 (extern ivtype_map_elt_info): Removed unused.
12168 (new_ivtype_map_elt): Removed unused.
12169 * sese.c (ivtype_map_elt_info): Removed unused.
12171 2013-03-28 Lawrence Crowl <crowl@google.com>
12173 * Makefile.in: Add several missing include dependences.
12175 (test-dump.o): New. This object is not added to any executable,
12176 but is present for ad-hoc testing.
12178 (debug (const bitmap_head_def &)): New.
12179 (debug (const bitmap_head_def *)): New.
12181 (extern debug (const bitmap_head_def &)): New.
12182 (extern debug (const bitmap_head_def *)): New.
12184 (debug (edge_def &)): New.
12185 (debug (edge_def *)): New.
12187 (debug (basic_block_def &)): New.
12188 (debug (basic_block_def *)): New.
12190 (dump_node (const_tree, int, FILE *)): Correct source file.
12192 (debug (die_struct &)): New.
12193 (debug (die_struct *)): New.
12195 (extern debug (die_struct &)): New.
12196 (extern debug (die_struct *)): New.
12197 * gimple-pretty-print.c
12198 (debug (gimple_statement_d &)): New.
12199 (debug (gimple_statement_d *)): New.
12200 * gimple-pretty-print.h
12201 (extern debug (gimple_statement_d &)): New.
12202 (extern debug (gimple_statement_d *)): New.
12204 (debug (ira_allocno_copy &)): New.
12205 (debug (ira_allocno_copy *)): New.
12206 (debug (ira_allocno &)): New.
12207 (debug (ira_allocno *)): New.
12209 (extern debug (ira_allocno_copy &)): New.
12210 (extern debug (ira_allocno_copy *)): New.
12211 (extern debug (ira_allocno &)): New.
12212 (extern debug (ira_allocno *)): New.
12214 (debug (live_range &)): New.
12215 (debug (live_range *)): New.
12217 (debug (lra_live_range &)): New.
12218 (debug (lra_live_range *)): New.
12220 (debug (lra_live_range &)): New.
12221 (debug (lra_live_range *)): New.
12223 (debug (omega_pb_d &)): New.
12224 (debug (omega_pb_d *)): New.
12226 (extern debug (omega_pb_d &)): New.
12227 (extern debug (omega_pb_d *)): New.
12229 (debug (const rtx_def &)): New.
12230 (debug (const rtx_def *)): New.
12232 (debug_tree (tree): Move within file.
12233 (debug_raw (const tree_node &)): New.
12234 (debug_raw (const tree_node *)): New.
12235 (dump_tree_via_hooks (const tree_node *, int)): New.
12236 (debug (const tree_node &)): New.
12237 (debug (const tree_node *)): New.
12238 (debug_verbose (const tree_node &)): New.
12239 (debug_verbose (const tree_node *)): New.
12240 (debug_head (const tree_node &)): New.
12241 (debug_head (const tree_node *)): New.
12242 (debug_body (const tree_node &)): New.
12243 (debug_body (const tree_node *)): New.
12244 (debug_vec_tree (tree): Move and reimplement in terms of dump.
12245 (debug (vec<tree, va_gc> &)): New.
12246 (debug (vec<tree, va_gc> *)): New.
12248 (extern debug (const rtx_def &)): New.
12249 (extern debug (const rtx_def *)): New.
12251 (debug_raw (simple_bitmap_def &)): New.
12252 (debug_raw (simple_bitmap_def *)): New.
12253 (debug (simple_bitmap_def &)): New.
12254 (debug (simple_bitmap_def *)): New.
12256 (extern debug (simple_bitmap_def &)): New.
12257 (extern debug (simple_bitmap_def *)): New.
12258 (extern debug_raw (simple_bitmap_def &)): New.
12259 (extern debug_raw (simple_bitmap_def *)): New.
12261 (debug (vinsn_def &)): New.
12262 (debug (vinsn_def *)): New.
12263 (debug_verbose (vinsn_def &)): New.
12264 (debug_verbose (vinsn_def *)): New.
12265 (debug (expr_def &)): New.
12266 (debug (expr_def *)): New.
12267 (debug_verbose (expr_def &)): New.
12268 (debug_verbose (expr_def *)): New.
12269 (debug (vec<rtx> &)): New.
12270 (debug (vec<rtx> *)): New.
12272 (extern debug (vinsn_def &)): New.
12273 (extern debug (vinsn_def *)): New.
12274 (extern debug_verbose (vinsn_def &)): New.
12275 (extern debug_verbose (vinsn_def *)): New.
12276 (extern debug (expr_def &)): New.
12277 (extern debug (expr_def *)): New.
12278 (extern debug_verbose (expr_def &)): New.
12279 (extern debug_verbose (expr_def *)): New.
12280 (extern debug (vec<rtx> &)): New.
12281 (extern debug (vec<rtx> *)): New.
12283 (_list_iter_cond_expr): Make inline instead of static.
12285 (debug (sreal &)): New.
12286 (debug (sreal *)): New.
12288 (extern debug (sreal &)): New.
12289 (extern debug (sreal *)): New.
12291 (extern debug_raw (const tree_node &)): New.
12292 (extern debug_raw (const tree_node *)): New.
12293 (extern debug (const tree_node &)): New.
12294 (extern debug (const tree_node *)): New.
12295 (extern debug_verbose (const tree_node &)): New.
12296 (extern debug_verbose (const tree_node *)): New.
12297 (extern debug_head (const tree_node &)): New.
12298 (extern debug_head (const tree_node *)): New.
12299 (extern debug_body (const tree_node &)): New.
12300 (extern debug_body (const tree_node *)): New.
12301 (extern debug (vec<tree, va_gc> &)): New.
12302 (extern debug (vec<tree, va_gc> *)): New.
12304 (debug (struct loop &)): New.
12305 (debug (struct loop *)): New.
12306 (debug_verbose (struct loop &)): New.
12307 (debug_verbose (struct loop *)): New.
12308 * tree-dump.c: Add header dependence.
12310 (extern debug (struct loop &)): New.
12311 (extern debug (struct loop *)): New.
12312 (extern debug_verbose (struct loop &)): New.
12313 (extern debug_verbose (struct loop *)): New.
12315 (debug (data_reference &)): New.
12316 (debug (data_reference *)): New.
12317 (debug (vec<data_reference_p> &)): New.
12318 (debug (vec<data_reference_p> *)): New.
12319 (debug (vec<ddr_p> &)): New.
12320 (debug (vec<ddr_p> *)): New.
12322 (extern debug (data_reference &)): New.
12323 (extern debug (data_reference *)): New.
12324 (extern debug (vec<data_reference_p> &)): New.
12325 (extern debug (vec<data_reference_p> *)): New.
12326 (extern debug (vec<ddr_p> &)): New.
12327 (extern debug (vec<ddr_p> *)): New.
12329 (debug (pt_solution &)): New.
12330 (debug (pt_solution *)): New.
12332 (extern debug (pt_solution &)): New.
12333 (extern debug (pt_solution *)): New.
12335 (debug (_var_map &)): New.
12336 (debug (_var_map *)): New.
12337 (debug (tree_live_info_d &)): New.
12338 (debug (tree_live_info_d *)): New.
12340 (extern debug (_var_map &)): New.
12341 (extern debug (_var_map *)): New.
12342 (extern debug (tree_live_info_d &)): New.
12343 (extern debug (tree_live_info_d *)): New.
12345 2013-03-28 Jan Hubicka <jh@suse.cz>
12347 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
12349 2013-03-28 Ian Bolton <ian.bolton@arm.com>
12351 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
12352 record only when desired or required.
12354 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
12356 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
12357 *vec_extractv2di_1_rex64. Use x64 isa attribute.
12359 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12361 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
12362 (*andsi3_compare0_uxtw): New pattern.
12363 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
12364 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
12366 2013-03-28 Jan Hubicka <jh@suse.cz>
12368 * data-streamer-in.c (streamer_read_gcov_count): New function.
12369 * gimple-streamer-out.c: Include value-prof.h.
12370 (output_gimple_stmt): Output histogram.
12371 (output_bb): Use streamer_write_gcov_count.
12372 * value-prof.c: Include data-streamer.h
12373 (dump_histogram_value): Add HIST_TYPE_MAX.
12374 (stream_out_histogram_value): New function.
12375 (stream_in_histogram_value): New function.
12376 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
12377 (stream_out_histogram_value, stream_in_histogram_value): Declare.
12378 * data-streamer-out.c (streamer_write_gcov_count): New function.
12379 (streamer_write_gcov_count_stream): New function.
12380 * lto-cgraph.c (lto_output_edge): Update counter streaming.
12381 (lto_output_node): Likewise.
12382 (input_node, input_edge): Likewise.
12383 * lto-streamer-out.c (output_cfg): Update streaming.
12384 * lto-streamer-in.c (input_cfg): Likewise.
12385 * data-streamer.h (streamer_write_gcov_count,
12386 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
12387 * gimple-streamer-in.c: Include value-prof.h
12388 (input_gimple_stmt): Input histograms.
12389 (input_bb): Update profile streaming.
12391 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12393 * genmodes.c (emit_max_int): New function.
12394 (emit_insn_modes_h): Added call to emit_max_function.
12395 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12397 * machmode.def: Fixed comment.
12399 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12401 * combine.c (try_combine): Removed useless assert.
12402 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12404 2013-03-28 Marek Polacek <polacek@redhat.com>
12405 Richard Biener <rguenther@suse.de>
12407 PR tree-optimization/56695
12408 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12409 build signed result of a vector comparison.
12410 * tree-cfg.c (verify_gimple_comparison): Check that a result
12411 of a vector comparison has signed type.
12413 2013-03-28 Richard Biener <rguenther@suse.de>
12415 PR tree-optimization/37021
12416 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12417 do not restrict gaps between groups.
12418 * tree-vect-stmts.c (vectorizable_load): Properly account for
12419 a gap between groups.
12421 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12423 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12424 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12427 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12429 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12430 * doc/extend.texi (Named Address Spaces): Ditto.
12431 (Variable Attributes): Ditto.
12433 2013-03-27 Kai Tietz <ktietz@redhat.com>
12435 * config.build: Add support for cygwin x64 target.
12436 * config.gcc: Likewise.
12437 * config.host: Likewise.
12438 * configure.ac: Likewise
12439 * configure: Regenerated.
12441 2013-03-27 Kai Tietz <ktietz@redhat.com>
12443 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12444 * config/i386/t-cygwin-w64: New file.
12445 * config/i386/cygwin-w64.h: New file.
12446 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12447 and add support for x64-cygwin target.
12448 (CPP_SPEC): Likewise.
12449 (CXX_WRAP_SPEC_LIST): Undefine before define.
12450 (LIBGCJ_SONAME): Use 15 as version.
12452 2013-03-27 Richard Biener <rguenther@suse.de>
12454 PR tree-optimization/56716
12455 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12456 dumping for ref nodes.
12458 2013-03-27 Martin Jambor <mjambor@suse.cz>
12460 PR tree-optimization/55334
12461 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12462 restricted pointers to arrays.
12464 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12466 * Makefile.in (.SUFFIXES): Add .cc.
12467 (.c.o): Apply same recipe for implicit rule .cc.o.
12469 2013-03-27 Richard Biener <rguenther@suse.de>
12471 PR tree-optimization/37021
12472 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12473 REALPART/IMAGPART_EXPRs around the supported refs.
12474 * tree-ssa-structalias.c (find_func_aliases): Assume that
12475 floating-point values are not used to transfer pointers.
12477 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12479 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12480 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12481 * doc/tm.texi: Regenerate.
12482 * targhooks.h (default_has_ifunc_p): New.
12483 * targhooks.c (default_has_ifunc_p): Ditto.
12484 * config/linux-protos.h: New file.
12485 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12486 this hook for linux which disables support of indirect functions in
12488 * config/linux-android.c: New file.
12489 * config/t-linux-android.c: Ditto.
12490 * config.gcc: Added new object file linux-android.o.
12491 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
12492 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12493 * varasm.c (do_assemble_alias): Likewise.
12494 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12495 doesn't support indirect functions.
12496 * configure: Regenerate.
12498 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12501 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12502 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12505 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12507 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12509 2013-03-27 Terry Guo <terry.guo@arm.com>
12511 * config/arm/arm-cores.def: Added core cortex-r7.
12512 * config/arm/arm-tune.md: Regenerated.
12513 * config/arm/arm-tables.opt: Regenerated.
12514 * doc/invoke.texi: Added entry for core cortex-r7.
12516 2013-03-27 Walter Lee <walt@tilera.com>
12518 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12519 double-decrement of next_scratch_regno.
12521 2013-03-27 Walter Lee <walt@tilera.com>
12523 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12525 (insn_v1mulus): Ditto.
12526 (insn_v2muls): Ditto.
12528 2013-03-27 Walter Lee <walt@tilera.com>
12530 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12531 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12533 2013-03-27 Walter Lee <walt@tilera.com>
12535 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
12536 (*sibcall_value): Ditto.
12538 2013-03-27 Walter Lee <walt@tilera.com>
12540 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12541 (insn_mnz_v8qi): ... this ...
12542 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12544 (insn_v<n>mnz): Replaced by ...
12545 (insn_v1mnz): ... this ...
12546 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12548 (insn_mz_<mode>): Replaced by ...
12549 (insn_mz_v8qi): ... this ...
12550 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12552 (insn_v<n>mz): Replaced by ...
12553 (insn_v1mz): ... this ...
12554 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12557 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12559 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12561 2013-03-26 Roland McGrath <mcgrathr@google.com>
12563 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12564 than fprintf with a non-constant, non-format string.
12566 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12568 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12569 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12570 operand 0 predicate.
12571 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12572 attribute. Use general_x64nomem_operand as operand 1 predicate.
12573 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12574 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12575 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12576 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12577 (mov<mode>_insv_1): Remove expander. Merge insn with
12578 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12579 Use general_x64nomem_operand as operand 1 predicate.
12580 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12581 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12582 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12583 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12584 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12585 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12586 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12587 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12588 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12589 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12590 (general_x64nomem_operand): Ditto.
12592 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
12594 * config/rtems.opt: Add -pthread option.
12596 2013-03-26 Richard Biener <rguenther@suse.de>
12598 * alias.c (find_base_term): Avoid redundant and not used recursion.
12599 (base_alias_check): Get the initial base term from the caller.
12600 (true_dependence_1): Compute and pass base terms to base_alias_check.
12601 (write_dependence_p): Likewise.
12602 (may_alias_p): Likewise.
12604 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12606 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12607 PC-relative load in SI modes and above only.
12609 2013-03-26 Xinyu Qi <xyqi@marvell.com>
12611 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
12612 * config/arm/iwmmxt.md (WCGR0): Update.
12613 (WCGR1, WCGR2, WCGR3): Likewise.
12615 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12617 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12618 Use x64 and nox64 isa attributes.
12620 2013-03-26 Richard Biener <rguenther@suse.de>
12622 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12623 alignment computations and rely on get_object_alignment_1
12624 for the !TYPE_P case.
12625 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12627 2013-03-26 Walter Lee <walt@tilera.com>
12629 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12630 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12632 2013-03-25 Jeff Law <law@redhat.com>
12634 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12635 check for INTEGRAL_TYPE_P that was missing due to checking in
12636 wrong version of prior patch.
12638 2013-03-25 Walter Lee <walt@tilera.com>
12640 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12641 TILEGX_INSN_SHUFFLEBYTES1.
12642 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12644 (tilegx_builtins): Ditto.
12645 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12647 2013-03-25 Walter Lee <walt@tilera.com>
12649 * config/tilegx/tilegx.md (floatsisf2): New pattern.
12650 (floatunssisf2): New pattern.
12651 (floatsidf2): New pattern.
12652 (floatunssidf2): New pattern.
12654 2013-03-25 Walter Lee <walt@tilera.com>
12656 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12657 tests for constraint J, K, N, P.
12659 2013-03-25 Walter Lee <walt@tilera.com>
12661 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12662 Use indirect/pcrel encoding.
12663 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12666 2013-03-25 Steve Ellcey <sellcey@mips.com>
12668 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12669 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12670 * config/mips/mips.c (mips_option_override): Set IMADD default.
12671 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12672 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12673 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12674 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12675 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12677 2013-03-25 Jeff Law <law@redhat.com>
12679 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12680 slightly to avoid creating and folding useless trees. Simplify
12681 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12683 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
12685 * config/i386/i386.md (*zero_extendsidi2): Merge with
12686 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12687 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12688 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12689 targets. Clarify comment.
12691 2013-03-25 Martin Jambor <mjambor@suse.cz>
12693 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12694 pass-through jump functions differently.
12695 (ipa_read_jump_function): Likewise. Also use setter functions to set
12698 2013-03-25 Martin Jambor <mjambor@suse.cz>
12700 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12701 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12703 (ipa_get_indirect_edge_target): New function.
12704 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12705 ipa_get_indirect_edge_target. Update all callers.
12706 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12707 ipa_get_indirect_edge_target_1 instead of calling
12708 ipa_get_indirect_edge_target.
12709 (create_specialized_node): Pass aggvlas to
12710 ipcp_discover_new_direct_edges.
12712 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12714 * config/arm/arm.md (f_sels, f_seld): New types.
12715 (*cmov<mode>): New pattern.
12716 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12719 2013-03-25 Kai Tietz <ktietz@redhat.com>
12721 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12722 POSIX-printf for mingw-hosted builds.
12724 2013-03-25 Richard Biener <rguenther@suse.de>
12726 PR middle-end/56694
12727 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12728 must-not-throw stmt location.
12730 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12732 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12733 Emit load-acquire versions when acq is true.
12734 (arm_emit_store_exclusive): Add rel parameter.
12735 Emit store-release versions when rel is true.
12736 (arm_split_compare_and_swap): Use acquire-release instructions
12738 of barriers when appropriate.
12739 (arm_split_atomic_op): Likewise.
12740 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12741 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12742 (VUNSPEC_SLX): Likewise.
12743 (VUNSPEC_LDA): Likewise.
12744 (VUNSPEC_STL): Likewise.
12745 * config/arm/sync.md (atomic_load<mode>): New pattern.
12746 (atomic_store<mode>): Likewise.
12747 (arm_load_acquire_exclusive<mode>): Likewise.
12748 (arm_load_acquire_exclusivesi): Likewise.
12749 (arm_load_acquire_exclusivedi): Likewise.
12750 (arm_store_release_exclusive<mode>): Likewise.
12752 2013-03-25 Catherine Moore <clm@codesourcery.com>
12754 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12755 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12756 * config/mip/predicates.md (lwsp_swsp_operand,
12757 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12758 sb16_operand, db4_operand, db7_operand, ib3_operand,
12759 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12760 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12761 andi16_operand): New predicates.
12762 * config/mips/mips.md (compression): New attribute.
12763 (enabled): New attribute.
12764 (length): Consider compression in computing length.
12765 (shift_compression): New code attribute.
12766 (*add<mode>3): New operands. Record compression.
12767 (sub<mode>3): Likewise.
12768 (one_cmpl<mode>2): Likewise.
12769 (*and<mode>3): Likewise.
12770 (*ior<mode>3): Likewise.
12771 (unnamed pattern for xor): Likewise.
12772 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12773 (*<optab><mode>3): Likewise.
12774 (*mov<mode>_internal: Likewise.
12775 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12776 (mips_unsigned_immediate_p): New.
12777 (umips_lwsp_swsp_address_p): New.
12778 (m16_based_address_p): New.
12779 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12780 (mips_unsigned_immediate_p): New prototype.
12781 (lwsp_swsp_address_p): New prototype.
12782 (m16_based_address_p): New prototype.
12783 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12784 (mips_signed_immediate_p): New function.
12785 (m16_based_address_p): New function.
12786 (lwsp_swsp_address_p): New function.
12787 (mips_print_operand_punctuation): Recognize short delay slot insns
12788 for microMIPS.add<mode>3"
12790 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12793 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12794 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12796 2013-03-25 Richard Biener <rguenther@suse.de>
12798 PR tree-optimization/56689
12799 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12802 2013-03-25 Richard Biener <rguenther@suse.de>
12804 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12806 (memory_references): Likewise.
12807 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12808 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12809 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12810 (gather_mem_refs_in_loops): Fold into ...
12811 (analyze_memory_references): ... this. Move initialization
12812 to tree_ssa_lim_initialize.
12813 (fill_always_executed_in): Rename to ...
12814 (fill_always_executed_in_1): ... this.
12815 (fill_always_executed_in): Move contains_call computation to
12816 this new function from ...
12817 (tree_ssa_lim_initialize): ... here.
12818 (tree_ssa_lim): Call fill_always_executed_in.
12820 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12822 * postreload.c (reload_combine): Fix code detecting returns.
12824 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12826 * function.c (emit_use_return_register_into_block): On cc0 targets,
12827 do not emit the sequence between cc0 setter and user.
12829 2013-03-25 Kai Tietz <ktietz@redhat.com>
12831 * config/i386/predicates.md (local_symbolic_operand): Interpret
12832 dll-imported symbols as none-local.
12834 2013-03-25 Richard Biener <rguenther@suse.de>
12836 * tree-ssa-loop-im.c (struct depend): Remove.
12837 (struct lim_aux_data): Make depends a vec of gimples.
12838 (free_lim_aux_data): Adjust.
12839 (add_dependency): Likewise.
12840 (set_level): Likewise.
12842 2013-03-25 Richard Biener <rguenther@suse.de>
12844 PR middle-end/56434
12845 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12846 the pointer returned by calls with ECF_MALLOC set.
12848 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12850 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12852 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12854 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
12855 using MMXMODE mode iterator.
12856 (*move<mode>_internal): Merge with *movv2sf_internal and
12857 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12859 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12861 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12862 (record_last_mem_set_info): Likewise.
12864 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12865 of XNEWVEC followed by memset.
12866 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12868 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12870 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12871 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12872 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12873 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12874 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12875 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12876 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12877 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12878 BARRIER_P instead of GET_CODE.
12880 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12882 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12883 inaccuracy in the probing code.
12885 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12886 (ctrapdi4): Likewise.
12888 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12890 * calls.c (expand_call): Add missing guard to code handling return
12891 of non-BLKmode structures in MSB.
12892 * function.c (expand_function_end): Likewise.
12894 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12896 * combine.c (try_combine): Adjust comment. Do not add the set of
12897 insn #0 if the destination indirectly is set or dies in insn #2.
12898 Tidy up code to distribute a new note.
12900 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12902 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12903 also for alternatives 16 and 17.
12905 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12907 * config/i386/sse.md (*mov<mode>_internal): Merge with
12908 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12909 Emit insn template depending on type attribute. Use
12910 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12911 movd instead of movq mnemonic for interunit moves. Rewrite mode
12912 attribute calculation. Remove unit attribute calculation.
12913 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12914 Set prefix_data16 attribute for DImode ssemov types.
12915 Use Ym instead of y for SSE-MMX conversion alternatives.
12916 Reorder operand constraints.
12918 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12920 * df.h (df_insn_delete): Adjust prototype.
12921 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12922 and let it decide whether mark the basic block dirty.
12923 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12924 * df-scan.c (df_insn_info_delete): New helper function, split
12925 off from df_insn_delete.
12926 (df_scan_free_bb_info): Use it.
12927 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12929 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12930 that the insn is actually an insn and it has a non-NULL basic block.
12931 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12933 2013-03-22 Richard Biener <rguenther@suse.de>
12935 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12937 (mem_ref_alloc): Do not allocate them.
12938 (refs_independent_p): Do not query or maintain a cache.
12940 2013-03-22 Richard Biener <rguenther@suse.de>
12942 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12943 (gather_mem_refs_in_loops): Do not compute it.
12944 (analyze_memory_references): Do not allocate it.
12945 (tree_ssa_lim_finalize): Do not free it.
12946 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12948 2013-03-22 Richard Biener <rguenther@suse.de>
12950 * is-a.h (as_a): Use gcc_checking_assert.
12952 2013-03-22 Ian Bolton <ian.bolton@arm.com>
12954 * config/aarch64/aarch64.c (aarch64_print_operand): New
12955 format specifier for printing a constant in hex.
12956 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12957 format specifier for printing second operand.
12959 2013-03-22 Richard Biener <rguenther@suse.de>
12961 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12963 (gather_mem_refs_in_loops): Perform store accumulation here.
12964 (create_vop_ref_mapping_loop): Remove.
12965 (create_vop_ref_mapping): Likewise.
12966 (analyze_memory_references): Initialize refs_stored_in_loop.
12967 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
12968 (record_indep_loop): Remove.
12969 (record_dep_loop): New function.
12970 (ref_indep_loop_p_1): Adjust to only walk over references
12971 in the loop, not its subloops.
12972 (ref_indep_loop_p): Rename to ...
12973 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12974 maintaining a more fine-grained cache.
12975 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12976 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12978 2013-03-22 Richard Biener <rguenther@suse.de>
12980 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12981 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12982 aggregate mem_ref_loc.
12983 (free_mem_ref_locs): Inline into ...
12984 (memref_free): ... this and adjust.
12985 (mem_ref_alloc): Adjust.
12986 (mem_ref_locs_alloc): Remove.
12987 (record_mem_ref_loc): Adjust.
12988 (get_all_locs_in_loop): Rewrite into ...
12989 (for_all_locs_in_loop): ... this iterator.
12990 (rewrite_mem_ref_loc): New functor.
12991 (rewrite_mem_refs): Use for_all_locs_in_loop.
12992 (sm_set_flag_if_changed): New functor.
12993 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
12994 (ref_always_accessed): New functor.
12995 (ref_always_accessed_p): Use for_all_locs_in_loop.
12997 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12999 * tree-pass.h (PROP_gimple_lvec): New.
13000 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13001 (init_optimization_passes): Move pass_lower_vector.
13002 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13004 (pass_lower_vector): Provide PROP_gimple_lvec.
13005 (pass_lower_vector_ssa): Likewise.
13006 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13008 2013-03-21 Mark Wielaard <mjw@redhat.com>
13010 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13012 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13014 * config/i386/i386.md (*movdi_internal): Disparage slightly
13015 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13016 conversion alternatives.
13018 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13020 PR middle-end/48087
13021 * diagnostic.def (DK_WERROR): New kind.
13022 * diagnostic.h (werrorcount): Define.
13023 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13024 promoted to DK_ERROR, increment DK_WERROR counter instead of
13026 * toplev.c (toplev_main): Call print_ignored_options even if
13027 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13028 even if just werrorcount is non-zero.
13031 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13033 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13034 (string_cst_pool_decl): New function.
13035 (optimize_one_addr_into_implicit_ptr): New function.
13036 (resolve_addr_in_expr): Optimize DWARF location expression
13037 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13038 which doesn't live in memory, but has DW_AT_location or
13039 DW_AT_const_value, or refers to a string literal, into
13040 DW_OP_GNU_implicit_pointer.
13041 (optimize_location_into_implicit_ptr): New function.
13042 (resolve_addr): If removing DW_AT_location of a variable because
13043 it was DW_OP_addr of address of the variable, but the variable doesn't
13044 live in memory, try to emit const value attribute for the initializer.
13046 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13048 * tree.h (VECTOR_TYPE_P): New macro.
13049 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13050 TYPE_MODE): Use it.
13051 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13052 VEC_COND_EXPR cannot be lvalues.
13053 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13055 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13057 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13058 Restrict the transformation to equal modes.
13060 2013-03-21 Richard Biener <rguenther@suse.de>
13062 PR tree-optimization/39326
13063 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13064 (MEM_ANALYZABLE): Adjust.
13065 (record_mem_ref_loc): Move bitmap ops ...
13066 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13067 unanalyzable refs, do not record locations for it.
13068 (analyze_memory_references): Allocate ref zero as shared
13070 (refs_independent_p): Do not test for unanalyzed mems here.
13071 (ref_indep_loop_p_1): Special-case disambiguation against
13072 the unanalyzed ref.
13073 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13075 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13077 * config/arm/arm-protos.h (tune_params): Add
13078 prefer_neon_for_64bits field.
13079 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13080 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13081 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13082 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13083 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13084 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13085 (arm_option_override): Handle -mneon-for-64bits new option.
13086 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13087 (prefer_neon_for_64bits): Declare new variable.
13088 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13089 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13090 (arch_enabled): Handle new arch types. Remove support for onlya8
13092 (one_cmpldi2): Use new arch names.
13093 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13094 * config/arm/arm.opt (mneon-for-64bits): Add option.
13095 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13096 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13097 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13099 * doc/invoke.texi (-mneon-for-64bits): Document.
13101 2013-03-21 Richard Biener <rguenther@suse.de>
13103 PR tree-optimization/39326
13104 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13105 (sort_bbs_in_loop_postorder_cmp): New function.
13106 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13108 2013-03-21 Richard Biener <rguenther@suse.de>
13110 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13111 (vect_insert_into_interleaving_chain): Likewise.
13112 (vect_drs_dependent_in_basic_block): Inline ...
13113 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13115 (vect_analyze_data_ref_dependence): ... here. Simplify.
13116 (vect_check_interleaving): Simplify.
13117 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13118 (vect_slp_analyze_data_ref_dependences): ... this new function.
13119 (dr_group_sort_cmp): New function.
13120 (vect_analyze_data_ref_accesses): Compute data-reference groups
13121 here instead of in vect_analyze_data_ref_dependence. Use
13122 a more efficient algorithm.
13123 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13124 vect_slp_analyze_data_ref_dependences. Call
13125 vect_analyze_data_ref_accesses earlier.
13126 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13127 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13128 (vect_slp_analyze_data_ref_dependences): New prototype.
13130 2013-03-21 Richard Biener <rguenther@suse.de>
13132 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13133 ref is stored in the loop.
13134 (find_refs_for_sm): Walk only over all stores.
13135 (store_motion_loop): Allocate from lim_bitmap_obstack.
13136 (store_motion): Likewise.
13138 2013-03-21 Richard Biener <rguenther@suse.de>
13140 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
13141 Update virtual SSA form.
13143 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13145 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
13146 * configure: Regenerate.
13147 * config.in: Regenerate.
13148 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
13149 if !HAVE_LD_EH_FRAME_CIEV3.
13151 2013-03-21 Richard Biener <rguenther@suse.de>
13153 * tree-cfg.c (verify_expr_no_block): New function.
13154 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
13155 nor DECL_VALUE_EXPR have locations with associated blocks.
13156 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
13157 (clear_unused_block_pointer): Remove code dealing with
13158 blocks in DECL_DEBUG_EXPR locations.
13160 2013-03-21 Richard Biener <rguenther@suse.de>
13162 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
13163 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
13164 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
13165 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
13166 instead of DECL_DEBUG_EXPR_IS_FROM.
13167 * gimplify.c (gimplify_modify_expr): Likewise.
13168 * tree-cfg.c (verify_expr_location_1): Likewise.
13169 * tree-complex.c (create_one_component_var): Likewise.
13170 * tree-sra.c (create_access_replacement): Likewise.
13171 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
13172 (clear_unused_block_pointer): Likewise.
13173 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
13174 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
13175 * var-tracking.c (var_debug_decl): Likewise.
13176 (track_expr_p): Likewise.
13177 * tree-inline.c (add_local_variables): Likewise. Set
13178 DECL_HAS_DEBUG_EXPR_P after copying it.
13179 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
13180 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
13182 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13185 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
13186 * configure: Regenerate.
13187 * config.in: Regenerate.
13188 * config/i386/i386.md (*movdf_internal): Use
13189 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13190 movd instead of movq mnemonic for interunit moves.
13191 (*movdi_internal): Ditto.
13193 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13195 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
13196 (abd<mode>_3): New pattern.
13197 (aba<mode>_3): New pattern.
13198 (fabd<mode>_3): New pattern.
13200 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13202 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
13203 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
13204 occurrence of REGISTER_PREFIX as its empty string.
13206 2013-03-20 Jeff Law <law@redhat.com>
13208 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
13209 addititional equivalences for equality comparisons between an SSA_NAME
13210 and a constant where the SSA_NAME was set from a widening conversion.
13212 2013-03-20 Walter Lee <walt@tilera.com>
13214 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
13216 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13218 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
13219 depending on type attribute.
13220 (*movti_internal): Ditto.
13221 (*movtf_internal): Ditto.
13222 (*movxf_internal): Ditto.
13223 (*movdf_internal): Ditto.
13224 (*movsf_internal): Ditto.
13226 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13228 * config/i386/i386.md (*movti_internal): Set prefix attribute to
13229 maybe_vex for sselog1 and ssemov types.
13230 (*movdi_internal): Reorder operand constraints.
13231 (*movsi_internal): Ditto. Set prefix attribute to
13232 maybe_vex for sselog1 and ssemov types.
13233 (*movtf_internal): Set prefix attribute to maybe_vex
13234 for sselog1 and ssemov types.
13235 (*movdf_internal): Ditto. Set prefix_data16 attribute for
13236 DImode ssemov types. Reorder operand constraints.
13237 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
13238 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
13239 attribute for SImode ssemov types. Reorder operand constraints.
13241 2013-03-20 Martin Jambor <mjambor@suse.cz>
13243 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
13244 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
13246 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
13248 * config/rs6000/predicates.md (indexed_address, update_address_mem
13249 update_indexed_address_mem): New predicates.
13250 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
13251 attribute for load/store instructions.
13252 * config/rs6000/dfp.md (movsd_store): Likewise.
13253 (movsd_load): Likewise.
13254 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
13255 (unnamed HI->DI extend define_insn): Likewise.
13256 (unnamed SI->DI extend define_insn): Likewise.
13257 (unnamed QI->SI extend define_insn): Likewise.
13258 (unnamed QI->HI extend define_insn): Likewise.
13259 (unnamed HI->SI extend define_insn): Likewise.
13260 (unnamed HI->SI extend define_insn): Likewise.
13261 (extendsfdf2_fpr): Likewise.
13262 (movsi_internal1): Likewise.
13263 (movsi_internal1_single): Likewise.
13264 (movhi_internal): Likewise.
13265 (movqi_internal): Likewise.
13266 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
13267 attribute for load/store instructions.
13268 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
13270 (mov<mode>_softfloat): Likewise.
13271 (mov<mode>_hardfloat32): Likewise.
13272 (mov<mode>_hardfloat64): Likewise.
13273 (mov<mode>_softfloat64): Likewise.
13274 (movdi_internal32): Likewise.
13275 (movdi_internal64): Likewise.
13276 (probe_stack_<mode>): Likewise.
13278 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
13280 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
13281 floating point, and decimal floating point to reload iterator.
13283 * config/rs6000/constraints.md (wl constraint): New constraints to
13284 return FLOAT_REGS if certain options are used to reduce the number
13285 of separate patterns that exist in the file.
13286 (wx constraint): Likewise.
13287 (wz constraint): Likewise.
13289 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13290 -mdebug=reg, print wg, wl, wx, and wz constraints.
13291 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13292 Initialize the reload functions for 64-bit binary/decimal floating
13294 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13295 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13296 create the buffer on the stack to overcome not having a 32-bit
13298 (rs6000_emit_move): Likewise.
13299 (rs6000_secondary_memory_needed_rtx): Likewise.
13300 (rs6000_alloc_sdmode_stack_slot): Likewise.
13301 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13302 via xxlxor, just like DFmode 0.0.
13304 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
13305 define as 1 if we are running on a power7 or newer.
13306 (enum r6000_reg_class_enum): Add new constraints.
13308 * config/rs6000/dfp.md (movsd): Delete, combine with binary
13309 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13310 with other moves by using conditional constraits (wg). Use LFIWZX
13311 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
13312 (movsd splitter): Likewise.
13313 (movsd_hardfloat): Likewise.
13314 (movsd_softfloat): Likewise.
13316 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
13317 binary and decimal floating point moves.
13318 (fmove_ok): New attributes to combine binary and decimal floating
13319 point moves, and to combine power6x (mfpgpr) moves along normal
13321 (real_value_to_target): Likewise.
13322 (f32_lr): Likewise.
13323 (f32_lm): Likewise.
13324 (f32_li): Likewise.
13325 (f32_sr): Likewise.
13326 (f32_sm): Likewise.
13327 (f32_si): Likewise.
13328 (movsf): Combine binary and decimal floating point moves. Combine
13329 power6x (mfpgpr) moves with other moves by using conditional
13330 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
13331 (mov<mode> for SFmode/SDmode); Likewise.
13332 (SFmode/SDmode splitters): Likewise.
13333 (movsf_hardfloat): Likewise.
13334 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
13335 (movsf_softfloat): Likewise.
13336 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
13338 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
13339 wx and wz constraints.
13341 * config/rs6000/constraints.md (wg constraint): New constraint to
13342 return FLOAT_REGS if -mmfpgpr (power6x) was used.
13344 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
13347 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13348 -mdebug=reg, print wg, wl, wx, and wz constraints.
13349 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13350 Initialize the reload functions for 64-bit binary/decimal floating
13352 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13353 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13354 create the buffer on the stack to overcome not having a 32-bit
13356 (rs6000_emit_move): Likewise.
13357 (rs6000_secondary_memory_needed_rtx): Likewise.
13358 (rs6000_alloc_sdmode_stack_slot): Likewise.
13359 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13360 via xxlxor, just like DFmode 0.0.
13362 * config/rs6000/dfp.md (movdd): Delete, combine with binary
13363 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13364 with other moves by using conditional constraits (wg). Use LFIWZX
13365 and STFIWX for loading SDmode on power7.
13366 (movdd splitters): Likewise.
13367 (movdd_hardfloat32): Likewise.
13368 (movdd_softfloat32): Likewise.
13369 (movdd_hardfloat64_mfpgpr): Likewise.
13370 (movdd_hardfloat64): Likewise.
13371 (movdd_softfloat64): Likewise.
13373 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
13374 64-bit binary and decimal floating point moves.
13375 (FMOVE64X): Likewise.
13376 (movdf): Combine 64-bit binary and decimal floating point moves.
13377 Combine power6x (mfpgpr) moves with other moves by using
13378 conditional constraits (wg).
13379 (mov<mode> for DFmode/DDmode): Likewise.
13380 (DFmode/DDmode splitters): Likewise.
13381 (movdf_hardfloat32): Likewise.
13382 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
13383 (movdf_softfloat32): Likewise.
13384 (movdf_hardfloat64_mfpgpr): Likewise.
13385 (movdf_hardfloat64): Likewise.
13386 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
13387 (movdf_softfloat64): Likewise.
13388 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13389 (reload_<mode>_load): Move to later in the file so they aren't in
13390 the middle of the floating point move insns.
13391 (reload_<mode>_store): Likewise.
13393 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13396 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13397 constraint if -mdebug=reg.
13398 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13399 Enable using dd reload support if needed.
13401 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13402 binary and decimal floating point moves in rs6000.md.
13403 (movtd_internal): Likewise.
13405 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13406 decimal floating point moves.
13408 (movtf_internal): Likewise.
13409 (mov<mode>_internal, TDmode/TFmode): Likewise.
13410 (movtf_softfloat): Likewise.
13411 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13413 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13414 movdi_internal64, using wg constraint for move direct operations.
13415 (movdi_internal64): Likewise.
13417 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13418 MODES_TIEABLE_P for selected modes. Print the numerical value of
13419 the various virtual registers. Use GPR/FPR first/last values,
13420 instead of hard coding the register numbers. Print which modes
13421 have reload functions registered.
13422 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13423 settings before/after setting cpu, target and subtarget settings.
13424 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13425 and for secondary reload failures in rs6000_secondary_reload_inner.
13426 (rs6000_secondary_reload_fail): Likewise.
13427 (rs6000_secondary_reload_inner): Likewise.
13429 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13430 macros for first/last GPR and FPR registers.
13431 (LAST_GPR_REGNO): Likewise.
13432 (FIRST_FPR_REGNO): Likewise.
13433 (LAST_FPR_REGNO): Likewise.
13435 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13436 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13437 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13438 (vcond<mode><mode>): Likewise.
13439 (vcondu<mode><mode>): Likewise.
13440 (vector_gtu<mode>): Likewise.
13441 (vector_gte<mode>): Likewise.
13442 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13443 to prevent the compiler from converting DImode operations to TImode.
13444 (ior<mode>3): Likewise.
13445 (and<mode>3): Likewise.
13446 (one_cmpl<mode>2): Likewise.
13447 (nor<mode>3): Likewise.
13448 (andc<mode>3): Likewise.
13450 * config/rs6000/constraints.md (wt constraint): New constraint
13451 that returns VSX_REGS if TImode is allowed in VSX registers.
13453 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13454 constant under VSX.
13456 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13457 similar to TImode, but it is restricted to being in the GPRs.
13459 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13460 TImode to occupy a single VSX register.
13462 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13463 -mvsx-timode for power7/power8.
13464 (power7 cpu): Likewise.
13465 (power8 cpu): Likewise.
13467 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13468 sure that TFmode/TDmode take up two registers if they are ever
13469 allowed in the upper VSX registers.
13470 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13472 (rs6000_init_hard_regno_mode_ok): Likewise.
13473 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13474 constraint. Print if LRA is turned on.
13475 (rs6000_option_override_internal): Give an error if -mvsx-timode
13476 and VSX is not enabled.
13477 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13478 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13479 to reg+offset addressing. Use PTImode when checking offset
13480 addresses for validity.
13481 (reg_offset_addressing_ok_p): Likewise.
13482 (rs6000_legitimate_offset_address_p): Likewise.
13483 (rs6000_legitimize_address): Likewise.
13484 (rs6000_legitimize_reload_address): Likewise.
13485 (rs6000_legitimate_address_p): Likewise.
13486 (rs6000_eliminate_indexed_memrefs): Likewise.
13487 (rs6000_emit_move): Likewise.
13488 (rs6000_secondary_reload): Likewise.
13489 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13490 reloads to fpr registers to continue to use reg+offset addressing,
13491 but 64-bit reloads to altivec registers need reg+reg addressing.
13492 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13493 it. Treat LO_SUM like a PLUS operation.
13494 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13495 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13496 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13497 registers to share a register with a smaller sized type, since VSX
13498 puts scalars in the upper 64-bits.
13499 (print_operand): Add support for PTImode.
13500 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13501 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13502 registers, but don't have arithmetic support.
13503 (rs6000_memory_move_cost): Add test for VSX.
13504 (rs6000_opt_masks): Add -mvsx-timode.
13506 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13509 (VSr): Use wt constraint for TImode.
13510 (VSv): Drop TImode support.
13511 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13512 (vsx_movti_64bit): Likewise.
13513 (vsx_movti_32bit): Likewise.
13514 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13515 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13516 one '?' on the appropriate output constraint. Do not allow TImode
13517 logical operations on 32-bit systems.
13518 (vsx_ior<mode>3): Likewise.
13519 (vsx_xor<mode>3): Likewise.
13520 (vsx_one_cmpl<mode>2): Likewise.
13521 (vsx_nor<mode>3): Likewise.
13522 (vsx_andc<mode>3): Likewise.
13523 (vsx_concat_<mode>): Likewise.
13524 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13526 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13527 OPTION_MASK_VSX_TIMODE.
13528 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13529 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13531 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13532 (TI2 iterator): New iterator for TImode, PTImode.
13533 (wd mode attribute): Add values for vector types.
13534 (movti_string): Replace TI move operations with operations for TImode
13535 and PTImode. Add support for TImode being allowed in VSX registers.
13536 (mov<mode>_string, TImode/PTImode): Likewise.
13537 (movti_ppc64): Likewise.
13538 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13539 (TI mode splitters): Likewise.
13541 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13544 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
13546 PR tree-optimization/56355
13547 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13548 Also handle integers with undefined overflow.
13550 2013-03-20 Catherine Moore <clm@codesourcery.com>
13551 Maciej W. Rozycki <macro@codesourcery.com>
13552 Tom de Vries <tom@codesourcery.com>
13553 Nathan Sidwell <nathan@codesourcery.com>
13554 Iain Sandoe <iain@codesourcery.com>
13555 Nathan Froyd <froydnj@codesourcery.com>
13556 Chao-ying Fu <fu@mips.com>
13558 * doc/extend.texi: (micromips, nomicromips, nocompression):
13559 Document new function attributes.
13560 * doc/invoke.texi (minterlink-compressed, mmicromips,
13561 m14k, m14ke, m14kec): Document new options.
13562 (minterlink-mips16): Update documentation.
13563 * doc/md.texi (ZC, ZD): Document new constraints.
13564 * configure.ac (gcc_cv_as_micromips): Check if linker
13565 supports the .set micromips directive.
13566 * configure: Regenerate.
13567 * config.in: Regenerate.
13568 * config/mips/mips-tables.opt: Regenerate.
13569 * config/mips/micromips.md: New file.
13570 * constraints.md (ZC, ZD): New constraints.
13571 * config/mips/predicates.md (movep_src_register): New predicate.
13572 (movep_src_operand): New predicate.
13573 (non_volatile_mem_operand): New predicate.
13574 * config/mips/mips.md (multimem): New type.
13575 (length): Differentiate between 17-bit and 18-bit branch offsets.
13576 (MOVEP1, MOVEP2): New mode iterator.
13577 (mov_<load>l): Use ZC constraint.
13578 (mov_<load>r): Likewise.
13579 (mov_<store>l): Likewise.
13580 (mov_<store>r): Likewise.
13581 (*branch_equality<mode>_inverted): Add microMIPS support.
13582 (*branch_equality<mode>): Likewise.
13583 (*jump_absolute): Likewise.
13584 (indirect_jump_<mode>): Likewise.
13585 (tablejump_<mode>): Likewise.
13586 (<optab>_internal): Likewise.
13587 (sibcall_internal): Likewise.
13588 (sibcall_value_internal): Likewise.
13589 (prefetch): Use constraint ZD.
13590 * config/mips/mips.opt (minterlink-compressed): New option.
13591 (minterlink-mips16): Now an alias for minterlink-compressed.
13592 (mmicromips): New option.
13593 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13594 (compare_and_swap_12): Likewise.
13595 (sync_add<mode>): Likewise.
13596 (sync_<optab>_12): Likewise.
13597 (sync_old_<optab>_12): Likewise.
13598 (sync_new_<optab>_12): Likewise.
13599 (sync_nand_12): Likewise.
13600 (sync_old_nand_12): Likewise.
13601 (sync_new_nand_12): Likewise.
13602 (sync_sub<mode>): Likewise.
13603 (sync_old_add<mode>): Likewise.
13604 (sync_old_sub<mode>): Likewise.
13605 (sync_new_add<mode>): Likewise.
13606 (sync_new_sub<mode>): Likewise.
13607 (sync_<optab><mode>): Likewise.
13608 (sync_old_<optab><mode>): Likewise.
13609 (sync_new_<optab><mode>): Likewise.
13610 (sync_nand<mode>): Likewise.
13611 (sync_old_nand<mode>): Likewise.
13612 (sync_new_nand<mode>): Likewise.
13613 (sync_lock_test_and_set<mode>): Likewise.
13614 (test_and_set_12): Likewise.
13615 (atomic_compare_and_swap<mode>): Likewise.
13616 (atomic_exchange<mode>_llsc): Likewise.
13617 (atomic_fetch_add<mode>_llsc): Likewise.
13618 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13619 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13620 (umips_save_restore_pattern_p): Likewise.
13621 (umips_load_store_pair_p): Likewise.
13622 (umips_output_load_store_pair): Likewise.
13623 (umips_movep_target_p): Likewise.
13624 (umips_12bit_offset_address_p): Likewise.
13625 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13626 (mips_base_mips16): Rename this...
13627 (mips_base_compression_flags): ...to this. Update all uses.
13628 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13629 (mips_mips16_decl_p): Delete.
13630 (mips_nomips16_decl_p): Delete.
13631 (mips_get_compress_on_flags): New function.
13632 (mips_get_compress_off_flags): New function.
13633 (mips_get_compress_mode): New function.
13634 (mips_get_compress_on_name): New function.
13635 (mips_get_compress_off_name): New function.
13636 (mips_insert_attributes): Support multiple compression types.
13637 (mips_merge_decl_attributes): Likewise.
13638 (umips_12bit_offset_address_p): New function.
13639 (mips_start_function_definition): Emit .set micromips directive.
13640 (mips_call_may_need_jalx_p): New function.
13641 (mips_function_ok_for_sibcall): Add microMIPS support.
13642 (mips_print_operand_punctuation): Support short delay slots and
13644 (umips_swm_mask, umips_swm_encoding): New.
13645 (umips_build_save_restore): New function.
13646 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13647 (was_mips16_p): Remove.
13648 (old_compression_mode): New.
13649 (mips_set_compression_mode): New function.
13650 (mips_set_current_function): Add microMIPS support.
13651 (mips_option_override): Likewise.
13652 (umips_save_restore_pattern_p): New function.
13653 (umips_output_save_restore): New function.
13654 (umips_load_store_pair_p_1): New function.
13655 (umips_load_store_pair_p): New function.
13656 (umips_output_load_store_pair_1): New function.
13657 (umips_output_load_store_pair): New function.
13658 (umips_movep_target_p) New function.
13659 (mips_prepare_pch_save): Add microMIPS support.
13660 * config/mips/mips.h (TARGET_COMPRESSION): New.
13661 (TARGET_CPU_CPP_BUILTINS): Update macro
13662 to use new compression flags and to support microMIPS.
13663 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13664 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13665 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13666 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13667 (ASM_SPEC): Support mmicromips and mno-micromips.
13668 (M16STORE_REG_P): New macro.
13669 (MIPS_CALL): Support TARGET_MICROMIPS.
13670 (MICROMIPS_J): New macro.
13671 (mips_base_mips16): Rename this...
13672 (mips_base_compression_flags): ...to this.
13673 (UMIPS_12BIT_OFFSET_P): New macro.
13674 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13675 (MULTILIB_DIRNAMES): Likewise.
13676 2013-03-20 Richard Biener <rguenther@suse.de>
13678 PR tree-optimization/56661
13679 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13680 the result does not have to be distinct.
13682 2013-03-20 Richard Biener <rguenther@suse.de>
13684 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13687 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13688 Steven Bosscher <steven@gcc.gnu.org>
13690 PR rtl-optimization/56605
13691 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13693 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13696 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13697 that require movd instead of movq.
13699 2013-03-20 Richard Biener <rguenther@suse.de>
13701 * tree-ssa-structalias.c (struct variable_info): Add pointer
13702 to the first field of an aggregate with sub-vars. Make
13703 this and the pointer to the next subfield its ID.
13704 (vi_next): New function.
13705 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13706 storedanything_id, integer_id): Increment by one.
13707 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13708 get_call_clobber_vi): Adjust.
13709 (solution_set_expand): Simplify and speedup.
13710 (solution_set_add): Inline into ...
13711 (set_union_with_increment): ... this. Adjust accordingly.
13712 (do_sd_constraint): Likewise.
13713 (do_ds_constraint): Likewise.
13714 (do_complex_constraint): Simplify.
13715 (build_pred_graph): Adjust.
13716 (solve_graph): Likewise. Simplify and speedup.
13717 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13718 get_constraint_for_component_ref, get_constraint_for_1,
13719 first_vi_for_offset, first_or_preceding_vi_for_offset,
13720 create_function_info_for, create_variable_info_for_1,
13721 create_variable_info_for, intra_create_variable_infos): Adjust.
13722 (init_base_vars): Push NULL for ID zero.
13723 (compute_points_to_sets): Adjust.
13725 2013-03-20 Richard Biener <rguenther@suse.de>
13727 * cfgloop.c (verify_loop_structure): Streamline and avoid
13728 ICEing on corrupt loop tree.
13729 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13732 2013-03-20 Richard Biener <rguenther@suse.de>
13734 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13735 check whether an SSA update is needed.
13737 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13739 * config/mips/constraints.md (T): Rename to...
13743 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13744 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13746 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13748 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13749 (*subsi3_carryin_uxtw): Likewise.
13751 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13753 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13754 (*rorsi3_insn_uxtw): Likewise.
13756 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13758 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13759 (*extrsi5_insn_uxtw): Likewise.
13761 2013-03-19 Richard Biener <rguenther@suse.de>
13763 PR tree-optimization/56273
13764 * passes.c (init_optimization_passes): Move second VRP after DOM.
13766 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
13768 * config/i386/i386.md (*movti_internal): Merge from
13769 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13770 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13771 nox64 isa attributes.
13773 2013-03-18 Richard Biener <rguenther@suse.de>
13775 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13777 (merge_node_constraints): Likewise.
13778 (build_succ_graph): Likewise.
13779 (valid_graph_edge): Inline into single caller.
13780 (unify_nodes): Likewise. Use bitmap_set_bit return value
13782 (scc_visit): Fix formatting and variable use.
13783 (do_sd_constraint): Use gcc_checking_assert.
13784 (do_ds_constraint): Likewise.
13785 (do_complex_constraint): Likewise.
13786 (condense_visit): Likewise. Cleanup.
13787 (dump_pred_graph): New function.
13788 (perform_var_substitution): Dump the pred-graph before
13789 variable substitution.
13790 (find_equivalent_node): Use gcc_checking_assert.
13791 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13793 2013-03-18 Richard Biener <rguenther@suse.de>
13795 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13796 Remove cond_expr_stmt_list argument and do not gimplify the
13798 (vect_loop_versioning): Adjust.
13799 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13800 Cleanup to use less temporaries.
13801 (vect_create_data_ref_ptr): Cleanup.
13803 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13805 PR tree-optimization/56635
13806 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13807 require types_compatible_p types.
13809 2013-03-18 Nick Clifton <nickc@redhat.com>
13811 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13812 spurious backslash.
13814 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13815 Add missing line to comment describing function.
13817 2013-03-18 Richard Biener <rguenther@suse.de>
13819 PR tree-optimization/56210
13820 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13821 Handle string / character search functions.
13822 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13824 2013-03-18 Richard Biener <rguenther@suse.de>
13826 PR middle-end/56483
13827 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13828 and implement properly.
13829 * gimple.h (gimple_cond_single_var_p): Remove.
13831 2013-03-18 Richard Biener <rguenther@suse.de>
13833 * tree-data-ref.h (find_data_references_in_loop): Declare.
13834 * tree-data-ref.c (get_references_in_stmt): Use a stack
13835 vector pre-allocated in the callers.
13836 (find_data_references_in_stmt): Adjust.
13837 (graphite_find_data_references_in_stmt): Likewise.
13838 (create_rdg_vertices): Likewise.
13839 (find_data_references_in_loop): Export.
13840 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13841 Compute dependences here...
13842 (vect_analyze_data_refs): ...not here. When we encounter
13843 a non-vectorizable data reference in basic-block vectorization
13844 truncate the data reference vector. Do not bother to
13845 fixup data-dependence information for gather loads.
13846 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13847 of data references, as reported.
13849 2013-03-18 Richard Biener <rguenther@suse.de>
13851 PR tree-optimization/3713
13852 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13853 has_constants and expr.
13854 (stmt_has_constants): Properly valueize SSA names when deciding
13855 whether the stmt has constants.
13857 2013-03-18 Richard Biener <rguenther@suse.de>
13859 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13860 whole function when there is nothing to do.
13861 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13862 * tree-vectorizer.c (vectorize_loops): Update virtual and
13863 loop-closed SSA once.
13864 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13866 2013-03-18 Richard Biener <rguenther@suse.de>
13868 PR middle-end/56113
13869 * domwalk.c (bb_postorder): New global static.
13870 (cmp_bb_postorder): New function.
13871 (walk_dominator_tree): Replace scheme imposing an order for
13872 visiting dominator sons by one sorting them at the time they
13873 are pushed on the stack.
13875 2013-03-18 Richard Biener <rguenther@suse.de>
13877 PR tree-optimization/39326
13878 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13879 (struct mem_ref): Replace mem member with ao_ref typed member.
13880 (MEM_ANALYZABLE): Adjust.
13881 (memref_eq): Likewise.
13882 (mem_ref_alloc): Likewise.
13883 (gather_mem_refs_stmt): Likewise.
13884 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13885 (execute_sm_if_changed_flag_set): Adjust.
13886 (execute_sm): Likewise.
13887 (ref_always_accessed_p): Likewise.
13888 (refs_independent_p): Likewise.
13889 (can_sm_ref_p): Likewise.
13891 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13894 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13895 return 1 even for !unsignedp.
13897 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13899 * config/i386/i386.md (isa): Add x64 and nox64.
13900 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13901 (*pushtf): Enable *roF alternative for x64 isa only.
13902 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13903 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13904 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13905 (*movtf_internal): Merge from *movtf_internal_rex64 and
13906 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13907 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13908 nox64 isa attributes.
13909 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13910 nox64 isa attributes.
13911 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13913 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13915 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13917 2013-03-17 Jakub Jelinek <jakub@redhat.com>
13920 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13921 class names. Remove trailing comma after "ALL_REGS".
13923 2013-03-16 Jan Hubicka <jh@suse.cz>
13925 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13926 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13927 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13928 of cgraph_get_create_node.
13929 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13931 2013-03-16 Jason Merrill <jason@redhat.com>
13934 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13935 with DW_AT_default_value.
13937 2013-03-16 Jakub Jelinek <jakub@redhat.com>
13939 * BASE-VER: Set to 4.9.0.
13941 2013-03-14 Andi Kleen <ak@linux.intel.com>
13944 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13945 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13946 Document _x* TSX intrinsics.
13948 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13949 David Holsgrove <david.holsgrove@xilinx.com>
13951 * configure.ac: Add MicroBlaze TLS support detection.
13952 * configure: Regenerate.
13953 * config/microblaze/microblaze-protos.h
13954 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13955 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13956 * config/microblaze/microblaze.c (microblaze_address_type): Add
13957 ADDRESS_TLS and tls_reloc address types.
13958 (microblaze_address_info): Add tls_reloc.
13959 (TARGET_HAVE_TLS): Define.
13960 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13961 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
13962 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13963 load_tls_operand, microblaze_call_tls_get_addr,
13964 microblaze_legitimize_tls_address): New functions.
13965 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13966 (get_base_reg): Use microblaze_tls_symbol_p.
13967 (microblaze_classify_address): Handle TLS.
13968 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13969 label_mentioned_p and microblaze_tls_referenced_p.
13970 (microblaze_legitimize_address): Handle TLS.
13971 (microblaze_address_insns): Handle ADDRESS_TLS.
13972 (pic_address_needs_scratch): Handle TLS.
13973 (print_operand_address): Handle TLS.
13974 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13975 (microblaze_expand_move): Handle TLS.
13976 (microblaze_legitimate_constant_p): Check
13977 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
13978 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13979 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13980 (PIC_OFFSET_TABLE_REGNUM): Set.
13981 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13982 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13983 (addsi3, movsi_internal2, movdf_internal): Update constraints
13984 * config/microblaze/predicates.md (arith_plus_operand): Define
13985 (move_operand): Redefine as move_src_operand,
13986 check microblaze_tls_referenced_p.
13988 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13990 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13991 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13993 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13995 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
13998 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14000 PR tree-optimization/53265
14001 * common.opt (Waggressive-loop-optimizations): New option.
14002 * tree-ssa-loop-niter.c: Include tree-pass.h.
14003 (do_warn_aggressive_loop_optimizations): New function.
14004 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14005 if number_of_latch_executions returned constant.
14006 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14007 early. If number_of_latch_executions returned constant, set
14008 nb_iterations_upper_bound back to it.
14009 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14011 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14012 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14014 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14015 (MULTILIB_OSDIRNAMES): Set.
14016 * genmultilib: If defaultosdirname doesn't start with :: , set
14017 defaultosdirname2 instead, clear it and emit two . multilib_raw
14018 entries instead of just one.
14020 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14022 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14023 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14024 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14025 (SUBTARGET_OVERRIDE_OPTIONS): New.
14027 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14030 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14031 (musermode): Convert to Var(TARGET_USERMODE).
14032 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14033 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14034 * config/sh/sh.c (sh_option_override): Use
14035 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14036 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14038 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14040 (udivsi3_i4_single, divsi3_i4_single): Use
14041 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14043 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14045 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14048 2013-03-13 Richard Biener <rguenther@suse.de>
14050 PR tree-optimization/56608
14051 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14052 calls when vectorizing basic-blocks.
14054 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14057 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14060 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14062 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14064 2013-03-11 Jan Hubicka <jh@suse.cz>
14067 * lto-streamer-out.c (output_symbol_p): Skip references from
14068 constructors of external variables.
14070 2013-03-11 Jan Hubicka <jh@suse.cz>
14072 PR middle-end/56571
14073 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14075 * emit-rtl.c (verify_rtx_sharing): Likewise.
14076 (copy_insn_1): Likewise.
14077 * rtl.c (copy_rtx): Likewise.
14079 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14082 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14083 output_operand_lossage message.
14085 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14088 * arm.c (shift_op): Validate RTL pattern on the fly.
14089 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14092 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14095 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14096 new call patterns that clobber %fr12.
14097 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14098 split and postreload patterns.
14099 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14100 registers %fr12 and %fr12R as call used.
14102 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14104 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14105 (canon_address, record_store, replace_read, check_mem_read_rtx,
14106 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14107 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14108 rest_of_handle_dse): Likewise.
14110 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14112 PR middle-end/56524
14113 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14115 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14116 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14117 (save_optabs_if_changed): Replace with...
14118 (init_tree_optimization_optabs): ...this.
14119 * optabs.c (save_optabs_if_changed): Rename to...
14120 (init_tree_optimization_optabs): ...this. Take the optimization node
14121 as argument. Do nothing if the base optabs are already correct.
14122 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14123 to recompute optabs.
14124 * function.h (function): Remove optabs field.
14125 * function.c (invoke_set_current_function_hook): Call
14126 init_tree_optimization_optabs. Use the result to initialize
14129 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14131 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14132 if GTMA_HAS_NO_INSTRUMENTATION.
14133 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14134 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14135 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
14136 * gimple-pretty-print.c (dump_gimple_transaction): Handle
14137 GTMA_HAS_NO_INSTRUMENTATION.
14139 2013-03-08 Jakub Jelinek <jakub@redhat.com>
14141 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
14144 2013-03-08 Marek Polacek <polacek@redhat.com>
14145 Jakub Jelinek <jakub@redhat.com>
14147 PR tree-optimization/56478
14148 * predict.c (is_comparison_with_loop_invariant_p): Change the
14149 type of loop_step to tree.
14150 (predict_loops): Adjust.
14151 (predict_iv_comparison): Perform the computations on double_ints.
14153 2013-03-08 Richard Biener <rguenther@suse.de>
14155 PR tree-optimization/56570
14156 * tree-cfg.c (verify_expr_location_1): Verify locations for
14158 * tree-sra.c (create_access_replacement): Strip locations
14159 from DECL_DEBUG_EXPRs.
14161 2013-03-08 Richard Biener <rguenther@suse.de>
14163 * tree-inline.c (expand_call_inline): Do not associate
14164 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
14165 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
14167 2013-03-08 Richard Biener <rguenther@suse.de>
14169 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
14170 or block changes with -Og. Fix for location / block encoding
14171 changes and PHI arguments with locations.
14173 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
14175 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
14177 (struct output_info): Likewise.
14178 (register_overhead): Remove bad gcc_assert.
14179 (bitmap_find_bit): If there is only a single bitmap element, do not
14180 count a miss as a search.
14181 (print_statistics): Update for counter type changes.
14182 (dump_bitmap_statistics): Likewise. Print headers such that they
14183 are properly lined up with the printed counters.
14185 2013-03-07 Jakub Jelinek <jakub@redhat.com>
14187 PR tree-optimization/56559
14188 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
14189 check that it has only a single use.
14191 2013-03-07 Richard Biener <rguenther@suse.de>
14193 * doc/invoke.texi (fwhole-program): Discourage use in combination
14196 2013-03-06 Jakub Jelinek <jakub@redhat.com>
14198 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
14200 PR tree-optimization/56539
14201 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
14202 instead of GSI_CONTINUE_LINKING as last argument to
14203 force_gimple_operand_gsi. Adjust function comment.
14205 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
14208 PR middle-end/56548
14209 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
14210 promoted mode, convert the result back to the original mode.
14212 2013-03-06 Richard Biener <rguenther@suse.de>
14214 PR middle-end/56294
14215 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
14216 (insert_updated_phi_nodes_compare_uids): New function.
14217 (update_ssa): Sort symbols_to_rename after UID before
14218 traversing it to insert PHI nodes.
14220 2013-03-06 Richard Biener <rguenther@suse.de>
14222 PR middle-end/50494
14223 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
14224 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
14227 2013-02-13 Richard Biener <rguenther@suse.de>
14230 * varasm.c (output_constant_def_1): Get the decl representing
14231 the constant as argument.
14232 (output_constant_def): Wrap output_constant_def_1.
14233 (make_decl_rtl): Use output_constant_def_1 with the decl
14234 representing the constant.
14235 (build_constant_desc): Optionally re-use a decl already
14236 representing the constant.
14237 (tree_output_constant_def): Adjust.
14239 2013-03-06 Joey Ye <joey.ye@arm.com>
14242 * gcc.c (convert_white_space): New function.
14243 (main): Handles white space in function name.
14245 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
14248 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
14249 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
14250 to SH_DIV_CALL_TABLE for TARGET_SH2.
14251 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
14253 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
14254 call-table options.
14256 2013-03-05 Sterling Augustine <saugustine@google.com>
14257 Cary Coutant <ccoutant@google.com>
14260 * dwarf2out.c (resolve_addr): Don't call
14261 remove_loc_list_addr_table_entries a second time for the same
14264 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14267 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
14268 (avoid_complex_debug_insns): New function.
14269 (expand_debug_locations): Call it.
14271 PR rtl-optimization/56484
14272 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
14273 lifetimes of hard registers on small register class machines.
14275 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14277 * config/microblaze/microblaze-protos.h: Rename
14278 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
14279 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
14281 (microblaze_fast_interrupt_function_p): New function.
14282 (microblaze_is_interrupt_handler): Rename to
14283 microblaze_is_interrupt_variant and add fast_interrupt check.
14284 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
14285 (save_restore_insns): Likewise.
14286 (compute_frame_size): Likewise.
14287 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
14288 (microblaze_globalize_label): Likewise.
14289 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
14290 * config/microblaze/microblaze.md: Use wrapper
14291 microblaze_is_interrupt_variant.
14293 2013-03-05 Kai Tietz <ktietz@redhat.com>
14295 * sdbout.c (sdbout_one_type): Switch to current function's section
14296 supporting cold/hot.
14298 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14300 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
14303 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14305 PR middle-end/56461
14306 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
14307 if VALGRIND_GET_VBITS is defined, temporarily make object
14308 memory all defined, and restore previous valgrind addressability
14309 and definability afterwards. Free this_object at the end.
14311 PR middle-end/56461
14312 * lra.c (lra): Call lra_clear_live_ranges if live_p,
14313 right before calling lra_create_live_ranges, also call it
14314 when clearing live_p. Only call lra_clear_live_ranges
14315 at the end if live_p.
14317 PR middle-end/56461
14318 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
14320 2013-03-05 Richard Biener <rguenther@suse.de>
14322 PR tree-optimization/56521
14323 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
14326 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
14329 * except.h (remove_unreachable_eh_regions): New prototype.
14330 * except.c (remove_eh_handler_splicer): New function, split out
14331 of remove_eh_handler.
14332 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
14333 warning about running it on many EH regions one at a time.
14334 (remove_unreachable_eh_regions_worker): New function, walk the
14335 EH tree in depth-first order and remove non-marked regions.
14336 (remove_unreachable_eh_regions): New function.
14337 * tree-eh.c (mark_reachable_handlers): New function, split out
14338 from remove_unreachable_handlers.
14339 (remove_unreachable_handlers): Use mark_reachable_handlers and
14340 remove_unreachable_eh_regions.
14341 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
14342 and remove_unreachable_eh_regions.
14344 2013-03-05 Richard Biener <rguenther@suse.de>
14346 PR middle-end/56525
14347 * loop-init.c (fix_loop_structure): Remove loops in two stages,
14348 not freeing them until the end.
14350 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14352 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
14354 2013-03-05 Richard Biener <rguenther@suse.de>
14356 PR tree-optimization/56270
14357 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
14358 of loads after scheduling an SLP instance.
14360 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14362 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
14364 (check_gcc_parallelize): Run guality.exp as a separate job from
14365 vect.exp with unsorted.exp and $(dg_target_exps) separately from
14366 struct-layout-1.exp with stackalign.exp.
14368 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
14370 PR middle-end/56461
14371 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
14372 load_index sbitmap even if some bit in it isn't set.
14374 PR middle-end/56461
14375 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
14376 (discover_iteration_bound_by_body_walk): Change queues to
14377 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
14378 spelling in comment. Call safe_push on queues[bound_index] directly.
14379 Release queues[queue_index] in every iteration unconditionally.
14380 Release bounds vector.
14382 PR middle-end/56461
14383 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
14384 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
14385 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14388 2013-03-05 Richard Biener <rguenther@suse.de>
14391 * tree-inline.c (remap_blocks_to_null): New function.
14392 (expand_call_inline): When expanding a call stmt without
14393 an associated block inline remap all callee blocks to NULL.
14395 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14397 PR rtl-optimization/56494
14398 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14399 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14400 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14402 PR middle-end/56461
14403 * sel-sched-ir.c (free_sched_pools): Release
14404 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14405 if succs_info_pool.max_top isn't -1.
14408 * opts.c (opts_obstack, opts_concat): Moved to...
14409 * opts-common.c (opts_obstack, opts_concat): ... here.
14411 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14413 PR middle-end/56461
14414 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14416 2013-03-04 Martin Jambor <mjambor@suse.cz>
14418 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14419 all appropriate places.
14421 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14423 PR tree-optimization/56424
14424 * ipa-split.c (split_function): Do not set the RSO flag if result is
14425 not by reference and its type is a register type.
14427 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14429 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14430 (microblaze_legitimate_pic_operand): Likewise
14431 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14432 new function microblaze_legitimate_pic_operand
14433 * config/microblaze/microblaze-protos.h
14434 (microblaze_legitimate_pic_operand): Declare.
14436 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14438 * config/microblaze/predicates.md (call_insn_simple_operand):
14439 New predicate for supported rtx code types.
14440 * config/microblaze/microblaze.md (call_internal1): Use
14441 call_insn_simple_operand predicate.
14443 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14445 PR middle-end/56461
14446 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14447 partitions.ordered_remove.
14449 PR middle-end/56461
14450 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14451 vec_oprnds0.create (1) for modifier == NONE.
14453 PR middle-end/56461
14454 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14455 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14456 vec_oprnds1 right before pushing anything to it for
14459 PR middle-end/56461
14460 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14461 set nbbs to 0 instead of having separate code path.
14462 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14463 instead of false as last argument if returning NULL.
14465 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14467 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14468 the attribute is now called "target" instead of "option".
14469 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14470 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14471 attribute/pragma name for TARGET_OPTION_VALID_P and
14472 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14473 * doc/tm.texi: Regenerated.
14475 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14477 * config/microblaze/microblaze.c:
14478 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14479 * config/microblaze/microblaze.h: Add -mxl-reorder to
14481 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14482 instructions emitted if TARGET_REORDER.
14483 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14484 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14487 2013-03-01 Xinliang David Li <davidxl@google.com>
14489 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14492 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14494 PR middle-end/56461
14495 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14496 vector even when returning true. Fix up function comment formatting.
14498 PR middle-end/56461
14499 * ira-build.c (ira_loop_nodes_count): New variable.
14500 (create_loop_tree_nodes): Initialize it.
14501 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14503 PR middle-end/56461
14504 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14505 method on dr_chain and result_chain.
14506 * tree-vect-stmts.c (vectorizable_store): Only call
14507 result_chain.create if j == 0.
14509 PR middle-end/56461
14510 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14511 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14512 before overwriting it.
14514 2013-03-01 Tobias Burnus <burnus@net-b.de>
14516 * doc/extended.texi (C Extensions): Change order in @menu
14518 (Other MIPS Built-in Functions): Move last MIPS entry before
14519 "picoChip Built-in Functions".
14520 (SH Built-in Functions): Move after RX Built-in Functions.
14521 * doc/gcc.texi (Introduction): Change order in @menu
14523 * doc/md.texi (Constraints): Ditto.
14524 * gty.texi (Type Information): Ditto.
14525 (User-provided marking routines for template types): Make
14527 * doc/invoke.texi (AArch64 Options): Move before
14528 "Adapteva Epiphany Options".
14530 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14531 Jakub Jelinek <jakub@redhat.com>
14534 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14535 no_address_safety_analysis attribute.
14536 * doc/extend.texi (no_address_safety_attribute): Rename to
14537 no_sanitize_address attribute, mention no_address_safety_analysis
14538 attribute as deprecated alias.
14540 2013-02-28 Jakub Jelinek <jakub@redhat.com>
14542 PR middle-end/56461
14543 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14544 type to vec<vec<tree> > *.
14545 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14546 to be vec<tree> instead of vec<tree> *, set vec_defs
14547 to vNULL and call vec_defs.create (number_of_vects), adjust other
14549 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14550 vectorizable_condition): Adjust vect_get_slp_defs callers.
14552 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14554 * config/aarch64/aarch64.c
14555 (aarch64_float_const_representable): Remove unused variable.
14557 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14559 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14561 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14563 * config/aarch64/aarch64-builtins.c
14564 (aarch64_init_simd_builtins): Make static.
14566 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14568 * config/aarch64/aarch64.c
14569 (aarch64_simd_make_constant): Make static.
14571 2013-02-28 Martin Jambor <mjambor@suse.cz>
14573 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14574 with no initialization to the RHS of debug statements.
14576 2013-02-28 Martin Jambor <mjambor@suse.cz>
14578 PR tree-optimization/56294
14579 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14581 (get_access_replacement): Do not call create_access_replacement.
14582 Assert a replacement exists.
14583 (get_repl_default_def_ssa_name): Create the replacement declaration
14586 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14588 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14589 final_end_function.
14591 2013-02-28 Marek Polacek <polacek@redhat.com>
14593 PR rtl-optimization/56466
14594 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14595 if we're changing a loop.
14596 (peel_loops_completely): Likewise.
14598 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14601 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14603 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14606 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14607 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14608 INTX_FTYPE_FX, FX_FTYPE_INTX.
14609 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14611 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14613 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14614 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14615 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14616 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14617 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14618 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14619 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14620 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14621 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14622 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14623 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14624 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14625 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14626 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14627 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14628 (avrxmega6): Increase max flash segments from 5 to 6.
14629 * config/avr/t-multilib: Regenerate.
14630 * config/avr/avr-tables.opt: Regenerate.
14631 * doc/avr-mmcu.texi: Regenerate.
14633 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14635 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14636 (avr_device_to_arch): Rename to avr_device_to_ld.
14637 (avr_device_to_as): New prototype.
14638 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14639 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14640 * config/avr/driver-avr.c (avr_device_to_as): New.
14641 (avr_device_to_arch): Rename to avr_device_to_ld.
14643 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14645 PR middle-end/56461
14646 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14647 method on dr_chain and result_chain.
14649 PR middle-end/56461
14650 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14651 pointer_set_destroy on not_executed_last_iteration.
14653 PR middle-end/56461
14654 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
14656 PR middle-end/56461
14657 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14658 FOR_EACH_DEFINED_FUNCTION when freeing state.
14660 PR middle-end/56461
14661 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14663 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14666 PR middle-end/56461
14667 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14668 known_aggs[i].items and release known_aggs vector.
14670 PR middle-end/56461
14671 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14673 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14675 * config/microblaze/microblaze.c (microblaze_emit_compare):
14676 Use xor for EQ/NE comparisions.
14677 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14678 (cbranchsf4): Adjust operator to comparison_operator.
14680 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14682 PR middle-end/56461
14683 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14685 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14686 vec_safe_push, always update *slot.
14687 (redirect_edge_var_map_clear): Use vec_free.
14688 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14689 (free_var_map_entry): Use vec_free.
14690 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14691 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14693 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
14695 PR middle-end/45472
14696 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14697 when the may_trap_p bit of the exprs being merged differs.
14698 Reorder tests for speculativeness in the logical and operator.
14700 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14702 * incpath.c (add_standard_paths): Use reconcat instead of concat
14703 where appropriate and avoid leaking memory.
14705 * opts.h: Include obstack.h.
14706 (opts_concat): New prototype.
14707 (opts_obstack): New declaration.
14708 * opts.c (opts_concat): New function.
14709 (opts_obstack): New variable.
14710 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14711 (finish_options): Use opts_concat instead of concat
14712 and XOBNEWVEC instead of XNEWVEC.
14713 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14714 generate_option): Likewise.
14715 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14716 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14719 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14720 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14722 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14724 PR middle-end/56461
14725 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14727 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14729 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14730 (arm_block_move_unaligned_straight): Likewise.
14731 (arm_adjust_block_mem): Likewise.
14733 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14736 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14737 temp, cond and label.
14738 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14741 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14742 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14745 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14746 of prologue/epilogue functions.
14749 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14750 Remove unused variable cfa_offset.
14751 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14754 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14755 type promotion to unsigned.
14758 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14759 for HOST_WIDE_INT of 32 bit / same size as int.
14760 (arm_block_move_unaligned_straight): Likewise.
14761 (arm_adjust_block_mem): Likewise.
14764 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14767 2013-02-26 Marek Polacek <polacek@redhat.com>
14769 PR tree-optimization/56426
14770 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
14772 2013-02-26 Richard Biener <rguenther@suse.de>
14775 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14776 unused variable loops.
14778 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14780 PR tree-optimization/56448
14781 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14782 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14783 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14784 later operands of the references, or even first operand for
14785 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14787 PR tree-optimization/56443
14788 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14789 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14790 to type_for_mode langhook.
14792 2013-02-25 Matt Turner <mattst88@gmail.com>
14794 * doc/invoke.texi: Document r4700.
14796 2013-02-25 Richard Biener <rguenther@suse.de>
14798 PR tree-optimization/56175
14799 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14801 (simplify_bitwise_binary): ... here. Also guard the conversion
14802 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14804 2013-02-25 Catherine Moore <clm@codesourcery.com>
14807 2013-02-24 Catherine Moore <clm@codesourcery.com>
14808 Maciej W. Rozycki <macro@codesourcery.com>
14809 Tom de Vries <tom@codesourcery.com>
14810 Nathan Sidwell <nathan@codesourcery.com>
14811 Iain Sandoe <iain@codesourcery.com>
14812 Nathan Froyd <froydnj@codesourcery.com>
14813 Chao-ying Fu <fu@mips.com>
14815 * doc/extend.texi: (micromips, nomicromips, nocompression):
14816 Document new function attributes.
14817 * doc/invoke.texi (minterlink-compressed, mmicromips,
14818 m14k, m14ke, m14kec): Document new options.
14819 (minterlink-mips16): Update documentation.
14820 * doc/md.texi (ZC, ZD): Document new constraints.
14821 * configure.ac (gcc_cv_as_micromips): Check if linker
14822 supports the .set micromips directive.
14823 * configure: Regenerate.
14824 * config.in: Regenerate.
14825 * config/mips/mips-tables.opt: Regenerate.
14826 * config/mips/micromips.md: New file.
14827 * constraints.md (ZC, AD): New constraints.
14828 * config/mips/predicates.md (movep_src_register): New predicate.
14829 (movep_src_operand): New predicate.
14830 (non_volatile_mem_operand): New predicate.
14831 * config/mips/mips.md (multimem): New type.
14832 (length): Differentiate between 17-bit and 18-bit branch offsets.
14833 (MOVEP1, MOVEP2): New mode iterator.
14834 (mov_<load>l): Use ZC constraint.
14835 (mov_<load>r): Likewise.
14836 (mov_<store>l): Likewise.
14837 (mov_<store>r): Likewise.
14838 (*branch_equality<mode>_inverted): Add microMIPS support.
14839 (*branch_equality<mode>): Likewise.
14840 (*jump_absolute): Likewise.
14841 (indirect_jump_<mode>): Likewise.
14842 (tablejump_<mode>): Likewise.
14843 (<optab>_internal): Likewise.
14844 (sibcall_internal): Likewise.
14845 (sibcall_value_internal): Likewise.
14846 (prefetch): Use constraint ZD.
14847 * config/mips/mips.opt (minterlink-compressed): New option.
14848 (minterlink-mips16): Now an alias for minterlink-compressed.
14849 (mmicromips): New option.
14850 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14851 (compare_and_swap_12): Likewise.
14852 (sync_add<mode>): Likewise.
14853 (sync_<optab>_12): Likewise.
14854 (sync_old_<optab>_12): Likewise.
14855 (sync_new_<optab>_12): Likewise.
14856 (sync_nand_12): Likewise.
14857 (sync_old_nand_12): Likewise.
14858 (sync_new_nand_12): Likewise.
14859 (sync_sub<mode>): Likewise.
14860 (sync_old_add<mode>): Likewise.
14861 (sync_old_sub<mode>): Likewise.
14862 (sync_new_add<mode>): Likewise.
14863 (sync_new_sub<mode>): Likewise.
14864 (sync_<optab><mode>): Likewise.
14865 (sync_old_<optab><mode>): Likewise.
14866 (sync_new_<optab><mode>): Likewise.
14867 (sync_nand<mode>): Likewise.
14868 (sync_old_nand<mode>): Likewise.
14869 (sync_new_nand<mode>): Likewise.
14870 (sync_lock_test_and_set<mode>): Likewise.
14871 (test_and_set_12): Likewise.
14872 (atomic_compare_and_swap<mode>): Likewise.
14873 (atomic_exchange<mode>_llsc): Likewise.
14874 (atomic_fetch_add<mode>_llsc): Likewise.
14875 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14876 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14877 (umips_save_restore_pattern_p): Likewise.
14878 (umips_load_store_pair_p): Likewise.
14879 (umips_output_load_store_pair): Likewise.
14880 (umips_movep_target_p): Likewise.
14881 (umips_12bit_offset_address_p): Likewise.
14882 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14883 (mips_base_mips16): Rename this...
14884 (mips_base_compression_flags): ...to this. Update all uses.
14885 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14886 (mips_mips16_decl_p): Delete.
14887 (mips_nomips16_decl_p): Delete.
14888 (mips_get_compress_on_flags): New function.
14889 (mips_get_compress_off_flags): New function.
14890 (mips_get_compress_mode): New function.
14891 (mips_get_compress_on_name): New function.
14892 (mips_get_compress_off_name): New function.
14893 (mips_insert_attributes): Support multiple compression types.
14894 (mips_merge_decl_attributes): Likewise.
14895 (umips_12bit_offset_address_p): New function.
14896 (mips_start_function_definition): Emit .set micromips directive.
14897 (mips_call_may_need_jalx_p): New function.
14898 (mips_function_ok_for_sibcall): Add microMIPS support.
14899 (mips_print_operand_punctuation): Support short delay slots and
14901 (umips_swm_mask, umips_swm_encoding): New.
14902 (umips_build_save_restore): New function.
14903 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14904 (was_mips16_p): Remove.
14905 (old_compression_mode): New.
14906 (mips_set_compression_mode): New function.
14907 (mips_set_current_function): Add microMIPS support.
14908 (mips_option_override): Likewise.
14909 (umips_save_restore_pattern_p): New function.
14910 (umips_output_save_restore): New function.
14911 (umips_load_store_pair_p_1): New function.
14912 (umips_load_store_pair_p): New function.
14913 (umips_output_load_store_pair_1): New function.
14914 (umips_output_load_store_pair): New function.
14915 (umips_movep_target_p) New function.
14916 (mips_prepare_pch_save): Add microMIPS support.
14917 * config/mips/mips.h (TARGET_COMPRESSION): New.
14918 (TARGET_CPU_CPP_BUILTINS): Update macro
14919 to use new compression flags and to support microMIPS.
14920 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14921 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14922 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14923 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14924 (ASM_SPEC): Support mmicromips and mno-micromips.
14925 (M16STORE_REG_P): New macro.
14926 (MIPS_CALL): Support TARGET_MICROMIPS.
14927 (MICROMIPS_J): New macro.
14928 (mips_base_mips16): Rename this...
14929 (mips_base_compression_flags): ...to this.
14930 (UMIPS_12BIT_OFFSET_P): New macro.
14931 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14932 (MULTILIB_DIRNAMES): Likewise.
14934 2013-02-25 Tom de Vries <tom@codesourcery.com>
14936 PR rtl-optimization/56131
14937 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14938 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14939 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14941 2013-02-25 Tobias Burnus <burnus@net-b.de>
14943 * doc/invoke.texi (-fsanitize=): Move from optimization
14944 to debugging options.
14946 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14948 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14950 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14951 Alexander Monakov <amonakov@ispras.ru>
14953 PR middle-end/56077
14954 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14955 flush pending lists also on non-jumps. Adjust comment.
14957 2013-02-24 Catherine Moore <clm@codesourcery.com>
14958 Maciej W. Rozycki <macro@codesourcery.com>
14959 Tom de Vries <tom@codesourcery.com>
14960 Nathan Sidwell <nathan@codesourcery.com>
14961 Iain Sandoe <iain@codesourcery.com>
14962 Nathan Froyd <froydnj@codesourcery.com>
14963 Chao-ying Fu <fu@mips.com>
14965 * doc/extend.texi: (micromips, nomicromips, nocompression):
14966 Document new function attributes.
14967 * doc/invoke.texi (minterlink-compressed, mmicromips,
14968 m14k, m14ke, m14kec): Document new options.
14969 (minterlink-mips16): Update documentation.
14970 * doc/md.texi (ZC, ZD): Document new constraints.
14971 * configure.ac (gcc_cv_as_micromips): Check if linker
14972 supports the .set micromips directive.
14973 * configure: Regenerate.
14974 * config.in: Regenerate.
14975 * config/mips/mips-tables.opt: Regenerate.
14976 * config/mips/micromips.md: New file.
14977 * constraints.md (ZC, AD): New constraints.
14978 * config/mips/predicates.md (movep_src_register): New predicate.
14979 (movep_src_operand): New predicate.
14980 (non_volatile_mem_operand): New predicate.
14981 * config/mips/mips.md (multimem): New type.
14982 (length): Differentiate between 17-bit and 18-bit branch offsets.
14983 (MOVEP1, MOVEP2): New mode iterator.
14984 (mov_<load>l): Use ZC constraint.
14985 (mov_<load>r): Likewise.
14986 (mov_<store>l): Likewise.
14987 (mov_<store>r): Likewise.
14988 (*branch_equality<mode>_inverted): Add microMIPS support.
14989 (*branch_equality<mode>): Likewise.
14990 (*jump_absolute): Likewise.
14991 (indirect_jump_<mode>): Likewise.
14992 (tablejump_<mode>): Likewise.
14993 (<optab>_internal): Likewise.
14994 (sibcall_internal): Likewise.
14995 (sibcall_value_internal): Likewise.
14996 (prefetch): Use constraint ZD.
14997 * config/mips/mips.opt (minterlink-compressed): New option.
14998 (minterlink-mips16): Now an alias for minterlink-compressed.
14999 (mmicromips): New option.
15000 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15001 (compare_and_swap_12): Likewise.
15002 (sync_add<mode>): Likewise.
15003 (sync_<optab>_12): Likewise.
15004 (sync_old_<optab>_12): Likewise.
15005 (sync_new_<optab>_12): Likewise.
15006 (sync_nand_12): Likewise.
15007 (sync_old_nand_12): Likewise.
15008 (sync_new_nand_12): Likewise.
15009 (sync_sub<mode>): Likewise.
15010 (sync_old_add<mode>): Likewise.
15011 (sync_old_sub<mode>): Likewise.
15012 (sync_new_add<mode>): Likewise.
15013 (sync_new_sub<mode>): Likewise.
15014 (sync_<optab><mode>): Likewise.
15015 (sync_old_<optab><mode>): Likewise.
15016 (sync_new_<optab><mode>): Likewise.
15017 (sync_nand<mode>): Likewise.
15018 (sync_old_nand<mode>): Likewise.
15019 (sync_new_nand<mode>): Likewise.
15020 (sync_lock_test_and_set<mode>): Likewise.
15021 (test_and_set_12): Likewise.
15022 (atomic_compare_and_swap<mode>): Likewise.
15023 (atomic_exchange<mode>_llsc): Likewise.
15024 (atomic_fetch_add<mode>_llsc): Likewise.
15025 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15026 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15027 (umips_save_restore_pattern_p): Likewise.
15028 (umips_load_store_pair_p): Likewise.
15029 (umips_output_load_store_pair): Likewise.
15030 (umips_movep_target_p): Likewise.
15031 (umips_12bit_offset_address_p): Likewise.
15032 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15033 (mips_base_mips16): Rename this...
15034 (mips_base_compression_flags): ...to this. Update all uses.
15035 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15036 (mips_mips16_decl_p): Delete.
15037 (mips_nomips16_decl_p): Delete.
15038 (mips_get_compress_on_flags): New function.
15039 (mips_get_compress_off_flags): New function.
15040 (mips_get_compress_mode): New function.
15041 (mips_get_compress_on_name): New function.
15042 (mips_get_compress_off_name): New function.
15043 (mips_insert_attributes): Support multiple compression types.
15044 (mips_merge_decl_attributes): Likewise.
15045 (umips_12bit_offset_address_p): New function.
15046 (mips_start_function_definition): Emit .set micromips directive.
15047 (mips_call_may_need_jalx_p): New function.
15048 (mips_function_ok_for_sibcall): Add microMIPS support.
15049 (mips_print_operand_punctuation): Support short delay slots and
15051 (umips_swm_mask, umips_swm_encoding): New.
15052 (umips_build_save_restore): New function.
15053 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15054 (was_mips16_p): Remove.
15055 (old_compression_mode): New.
15056 (mips_set_compression_mode): New function.
15057 (mips_set_current_function): Add microMIPS support.
15058 (mips_option_override): Likewise.
15059 (umips_save_restore_pattern_p): New function.
15060 (umips_output_save_restore): New function.
15061 (umips_load_store_pair_p_1): New function.
15062 (umips_load_store_pair_p): New function.
15063 (umips_output_load_store_pair_1): New function.
15064 (umips_output_load_store_pair): New function.
15065 (umips_movep_target_p) New function.
15066 (mips_prepare_pch_save): Add microMIPS support.
15067 * config/mips/mips.h (TARGET_COMPRESSION): New.
15068 (TARGET_CPU_CPP_BUILTINS): Update macro
15069 to use new compression flags and to support microMIPS.
15070 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15071 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15072 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15073 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15074 (ASM_SPEC): Support mmicromips and mno-micromips.
15075 (M16STORE_REG_P): New macro.
15076 (MIPS_CALL): Support TARGET_MICROMIPS.
15077 (MICROMIPS_J): New macro.
15078 (mips_base_mips16): Rename this...
15079 (mips_base_compression_flags): ...to this.
15080 (UMIPS_12BIT_OFFSET_P): New macro.
15081 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15082 (MULTILIB_DIRNAMES): Likewise.
15084 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15087 * target-globals.c (save_target_globals): For init_reg_sets and
15088 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15090 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15092 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15093 * config/aarch64/t-aarch64
15094 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15096 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15098 PR inline-asm/56148
15099 * lra-constraints.c (process_alt_operands): Reload operand
15100 conflicting with earlier clobber only if no more other conflicting
15103 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15106 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15107 if not linking a shared library.
15109 2013-02-22 Seth LaForge <sethml@google.com>
15111 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15113 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15115 * config/arm/arm.md (split for extendsidi): Update condition.
15116 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15117 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15118 (qhs_zextenddi_cstr): Likewise.
15120 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15122 PR middle-end/56420
15123 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15124 avoid signed wrapping.
15125 (expand_mult): Handle properly multiplication by
15126 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15127 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15128 in the compiler if coeff is HOST_WIDE_INT_MIN.
15129 (expand_divmod): Don't make ext_op1 static, change it's type to
15130 uhwi. Avoid undefined behavior in -INTVAL (op1).
15132 PR rtl-optimization/50339
15133 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15135 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
15136 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
15137 into splitting_ashiftrt field.
15138 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
15140 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
15143 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
15145 PR middle-end/56108
15146 * trans-mem.c (execute_tm_mark): Do not expand transactions that
15147 are sure to go irrevocable.
15149 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
15151 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
15152 scalars are valid operands.
15154 2013-02-21 Martin Jambor <mjambor@suse.cz>
15156 PR tree-optimization/56310
15157 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
15158 only matching indices and non-negative final offsets.
15159 (intersect_aggregates_with_edge): Pass src_idx to
15160 agg_replacements_to_vector. Pass src_idx insstead of index to
15161 intersect_with_agg_replacements.
15163 2013-02-21 Martin Jambor <mjambor@suse.cz>
15165 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
15166 instead of hard-wired defaults.
15168 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
15170 * doc/invoke.texi (MIPS Options): Update documentation of the
15171 floating-point multiply-accumulate instruction restrictions.
15173 2013-02-21 Kostya Serebryany <kcc@google.com>
15175 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
15176 asan_shadow_offset on x86_64 linux.
15178 2013-02-21 Richard Biener <rguenther@suse.de>
15180 PR tree-optimization/56415
15182 2013-02-11 Richard Biener <rguenther@suse.de>
15184 PR tree-optimization/56273
15185 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15188 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15191 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
15194 PR inline-asm/56405
15195 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
15196 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
15198 2013-02-20 Jan Hubicka <jh@suse.cz>
15200 PR tree-optimization/56265
15201 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
15202 when target is referenced for first time.
15204 2013-02-20 Richard Biener <rguenther@suse.de>
15206 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
15207 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
15208 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
15209 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
15210 not return anything.
15211 (rename_ssa_copies): Do not remove unused locals.
15212 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
15213 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
15214 * passes.c (execute_function_todo): Do not schedule unused locals
15215 removal if cleanup_tree_cfg did something.
15216 * tree-ssa-live.c (remove_unused_locals): Dump statistics
15217 about the number of removed locals.
15219 2013-02-20 Richard Biener <rguenther@suse.de>
15221 PR tree-optimization/56398
15222 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
15224 2013-02-20 Martin Jambor <mjambor@suse.cz>
15226 PR tree-optimization/55334
15227 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
15228 restricted pointers to arrays.
15230 2013-02-20 Richard Biener <rguenther@suse.de>
15231 Jakub Jelinek <jakub@redhat.com>
15233 PR tree-optimization/56396
15234 * tree-ssa-ccp.c (n_const_val): New static variable.
15235 (get_value): Return NULL for SSA names we don't have a lattice
15237 (ccp_initialize): Initialize n_const_val.
15238 * tree-ssa-copy.c (n_copy_of): New static variable.
15239 (init_copy_prop): Initialize n_copy_of.
15240 (get_value): Return NULL_TREE for SSA names we don't have a
15243 2013-02-20 Martin Jambor <mjambor@suse.cz>
15245 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
15247 2013-02-20 Richard Biener <rguenther@suse.de>
15249 * genpreds.c (write_lookup_constraint): Do not compare first
15250 letter of the constraint again.
15252 2013-02-20 Richard Biener <rguenther@suse.de>
15254 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
15256 (get_use_iv_cost): Terminate hashtable walk when coming across
15259 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
15261 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
15262 reassociation for avx2 targets.
15264 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15266 * config/microblaze/microblaze.c: microblaze_has_clz = 0
15267 Add version check for v8.10.a to enable microblaze_has_clz
15268 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
15269 version and TARGET_PATTERN_COMPARE check
15270 * config/microblaze/microblaze.md: New clzsi2 instruction
15272 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15274 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
15275 function before branching.
15277 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15279 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
15281 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
15283 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15285 PR middle-end/55889
15286 * sel-sched.c: Include ira.h.
15287 (implicit_clobber_conflict_p): New function.
15288 (moveup_expr): Use it.
15289 * Makefile.in (sel-sched.o): Depend on ira.h.
15291 2013-02-19 Richard Biener <rguenther@suse.de>
15293 PR tree-optimization/56384
15294 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
15295 (vn_hash_type): Split out from ...
15296 (vn_hash_constant_with_type): ... here.
15297 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
15298 (vn_phi_eq): Compare types from vn_phi_s structure.
15299 (vn_phi_lookup): Populate vn_phi_s type.
15300 (vn_phi_insert): Likewise.
15302 2013-02-19 Jakub Jelinek <jakub@redhat.com>
15304 PR tree-optimization/56350
15305 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
15306 if haven't found reduction or nested cycle operand, rather than
15307 asserting we must find it.
15309 PR tree-optimization/56381
15310 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
15313 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
15314 Jakub Jelinek <jakub@redhat.com>
15317 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
15318 (swap_optab_enable): Same.
15319 (init_all_optabs): Use argument instead of global.
15320 * tree.h (struct tree_optimization_option): New field target_optabs.
15321 * expr.h (init_all_optabs): Add argument to prototype.
15322 (TREE_OPTIMIZATION_OPTABS): New.
15323 (save_optabs_if_changed): Protoize.
15324 * optabs.h: Declare this_fn_optabs.
15325 * optabs.c (save_optabs_if_changed): New.
15326 Declare this_fn_optabs.
15327 (init_optabs): Add argument to init_all_optabs() call.
15328 * function.c (invoke_set_current_function_hook): Handle per
15330 * function.h (struct function): New field optabs.
15331 * config/mips/mips.c (mips_set_mips16_mode): Handle when
15332 optimization_current_node has changed.
15333 * target-globals.h (save_target_globals_default_opts): Protoize.
15334 * target-globals.c (save_target_globals_default_opts): New.
15336 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15339 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
15340 registers %fr12 and %fr12R as call used.
15343 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
15344 and HImode, require all displacements to be an integer multiple of
15346 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
15347 only allow QImode and HImode when reload is in progress and strict is
15348 true. Likewise for symbolic addresses. Use base14_operand to check
15349 displacements in REG+BASE addresses.
15351 2013-02-18 Richard Biener <rguenther@suse.de>
15353 PR tree-optimization/56366
15354 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15355 handle sign-conversion of outer-loop initial induction value.
15357 2013-02-18 Richard Biener <rguenther@suse.de>
15359 PR middle-end/56349
15360 * cfghooks.c (merge_blocks): If we merge a latch into another
15361 block adjust references to it.
15362 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
15363 (verify_loop_structure): Verify that a recorded latch is in fact
15366 2013-02-18 Richard Biener <rguenther@suse.de>
15368 PR tree-optimization/56321
15369 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
15370 order SSA name release and virtual operand unlinking.
15372 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15374 * config/microblaze/microblaze.md (save_stack_block): Define.
15375 (restore_stack_block): Likewise.
15377 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15379 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
15380 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
15381 * config/microblaze/microblaze.c (microblaze_option_override):
15382 Bail out early for PIC modes when target does not support PIC.
15384 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15386 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
15387 Replace with a microblaze version.
15388 (microblaze_trampoline_init): Adapt for microblaze.
15389 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15392 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15393 Dodji Seketeli <dodji@redhat.com>
15396 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15397 (instrument_mem_region_access): Do not forget to always put
15398 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15399 0) statement, even for cases where either 'base' or 'base + len'
15400 are not instrumented -- because they have been previously
15401 instrumented. Simplify the logic by putting all the statements
15402 instrument 'base + len' inside a sequence, and then insert that
15403 sequence right before the current insertion point. Then, to
15404 instrument 'base + len', just get an iterator on that statement.
15405 And do not forget to update the pointer to iterator the function
15406 received as argument.
15408 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15410 PR rtl-optimization/56348
15411 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15413 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15415 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15416 (clean_graph_dump_file): Pass base to start_graph_dump.
15418 2013-02-14 Richard Henderson <rth@redhat.com>
15421 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15423 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15425 * collect2-aix.h: Define F_LOADONLY.
15427 2013-02-14 Richard Biener <rguenther@suse.de>
15430 * varasm.c (output_constant_def_1): Get the decl representing
15431 the constant as argument.
15432 (output_constant_def): Wrap output_constant_def_1.
15433 (make_decl_rtl): Use output_constant_def_1 with the decl
15434 representing the constant.
15435 (build_constant_desc): Optionally re-use a decl already
15436 representing the constant.
15437 (tree_output_constant_def): Adjust.
15439 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15442 * asan.c (instrument_builtin_call): Really put the length of the
15443 second source argument into src1_len.
15445 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15447 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15448 argument. If it is false, don't create edge from then_bb to
15450 (insert_if_then_before_iter): Pass true to it.
15451 (build_check_stmt): Pass false to it.
15452 (transform_statements): Flush hash table only on extended basic
15453 block boundaries, rather than at the beginning of every bb.
15454 Don't flush hash table on nonfreeing_call_p calls.
15455 * tree-flow.h (nonfreeing_call_p): New prototype.
15456 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15458 2013-02-13 David S. Miller <davem@davemloft.net>
15460 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15462 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15465 * ira.c (max_regno_before_ira): Move from ...
15467 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15470 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15472 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15474 2013-02-13 Richard Biener <rguenther@suse.de>
15477 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15478 globals in MEM_REFs.
15480 2013-02-13 Richard Biener <rguenther@suse.de>
15482 * loop-init.c (loop_optimizer_init): Clear loop state when
15483 re-initializing preserved loops.
15484 * loop-unswitch.c (unswitch_single_loop): Return whether
15485 we unswitched the loop. Do not verify loop state here.
15486 (unswitch_loops): When we unswitched a loop discover new loops.
15488 2013-02-13 Kostya Serebryany <kcc@google.com>
15490 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15492 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15494 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15496 Avoid instrumenting duplicated memory access in the same basic block
15497 * Makefile.in (asan.o): Add new dependency on hash-table.h
15498 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15499 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15500 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15501 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15502 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15503 (get_mem_ref_of_assignment): New functions.
15504 (get_mem_refs_of_builtin_call): Extract from
15505 instrument_builtin_call and tweak a little bit to make it fit with
15507 (instrument_builtin_call): Use the new
15508 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15509 of is_gimple_builtin_call.
15510 (instrument_derefs, instrument_mem_region_access): Insert the
15511 instrumented memory reference into the hash table.
15512 (maybe_instrument_assignment): Renamed instrument_assignment into
15513 this, and change it to advance the iterator when instrumentation
15514 actually happened and return true in that case. This makes it
15515 homogeneous with maybe_instrument_assignment, and thus give a
15516 chance to callers to be more 'regular'.
15517 (transform_statements): Clear the memory reference hash table
15518 whenever we enter a new BB, when we cross a function call, or when
15519 we are done transforming statements. Use
15520 maybe_instrument_assignment instead of instrumentation. No more
15521 need to special case maybe_instrument_assignment and advance the
15522 iterator after calling it; it's now handled just like
15523 maybe_instrument_call. Update comment.
15525 2013-02-13 Richard Biener <rguenther@suse.de>
15527 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15528 Fix loop discovery code.
15530 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15532 PR inline-asm/56148
15533 * lra-constraints.c (process_alt_operands): Match early clobber
15534 operand with itself. Check conflicts with earlyclobber only if
15535 the operand is not reloaded. Prefer to reload conflicting operand
15536 if earlyclobber and matching operands are the same.
15538 2013-02-12 Richard Biener <rguenther@suse.de>
15541 * lto-streamer-out.c (write_symbol): Do not output symbols
15542 for hard register variables.
15544 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
15547 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15548 (umulsidi3_insn, mulsidi3_insn): New insns.
15550 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
15552 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15553 (struct tune_params): Add vec_costs field.
15554 * config/arm/arm.c (arm_builtin_vectorization_cost)
15555 (arm_add_stmt_cost): New functions.
15556 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15557 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15558 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15559 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15560 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15561 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15562 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15564 2013-02-12 Richard Biener <rguenther@suse.de>
15567 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15568 decls again if possible.
15570 2013-02-12 Richard Biener <rguenther@suse.de>
15572 PR middle-end/56288
15573 * tree-ssa.c (verify_ssa_name): Fix check, move
15574 SSA_NAME_IN_FREE_LIST check up.
15576 2013-02-12 Jakub Jelinek <jakub@redhat.com>
15577 Steven Bosscher <steven@gcc.gnu.org>
15579 PR rtl-optimization/56151
15580 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15581 equal to op0 or op1, and last_insn pattern is CODE operation
15582 with MEM dest and one of the operands matches that MEM.
15584 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15586 * doc/extend.texi: Document Function Multiversioning and "default"
15587 parameter string to target attribute.
15588 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15589 target attribute parameter is "default".
15590 (ix86_compare_version_priority): Remove checks for target attribute.
15591 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15592 Remove check for target attribute equal to NULL. Add assert.
15593 (ix86_generate_version_dispatcher_body): Change error to sorry.
15595 2013-02-11 Iain Sandoe <iain@codesourcery.com>
15596 Jack Howarth <howarth@bromo.med.uc.edu>
15597 Patrick Marlier <patrick.marlier@gmail.com>
15600 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15601 define ENDFILE_SPEC as TM_DESTRUCTOR.
15602 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15604 2013-02-11 Alexander Potapenko <glider@google.com>
15605 Jack Howarth <howarth@bromo.med.uc.edu>
15606 Jakub Jelinek <jakub@redhat.com>
15609 * config/darwin.c (cdtor_record): Rename ctor_record.
15610 (sort_cdtor_records): Rename sort_ctor_records.
15611 (finalize_dtors): New routine to sort destructors by
15612 priority before use in assemble_integer.
15613 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15615 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
15617 PR rtl-optimization/56275
15618 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15619 offset is non-negative and less than cmode size before
15620 calling simplify_subreg.
15622 2013-02-11 Richard Biener <rguenther@suse.de>
15624 PR tree-optimization/56264
15625 * cfgloop.h (fix_loop_structure): Adjust prototype.
15626 * loop-init.c (fix_loop_structure): Return the number of
15627 newly discovered loops.
15628 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15629 are discovered, do a full loop-closed SSA rewrite.
15631 2013-02-11 Richard Biener <rguenther@suse.de>
15633 PR tree-optimization/56273
15634 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15636 (check_array_ref): Fix missing newline in dumps.
15637 (search_for_addr_array): Likewise.
15639 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
15641 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15643 2013-02-09 Jakub Jelinek <jakub@redhat.com>
15646 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15648 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15650 PR rtl-optimization/56246
15651 * lra-constraints.c (simplify_operand_subreg): Try to reuse
15653 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15654 constraints are satisfied.
15656 2013-02-08 Jeff Law <law@redhat.com>
15659 * emit-rtl.c (reg_is_parm_p): New function.
15660 * regs.h (reg_is_parm_p): New prototype.
15661 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15662 callee-clobbered registers.
15664 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15667 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15668 If there is no implicit builtin declaration, just return NULL.
15670 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15672 * config/i386/sse.md (FMAMODEM): New mode iterator.
15673 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15674 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15676 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15678 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15679 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15680 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15682 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15684 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15685 (microblaze*-*-elf): Likewise.
15686 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15688 * config/microblaze/microblaze-c.c: Add builtin defines for
15689 _LITTLE_ENDIAN and _BIG_ENDIAN.
15690 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15691 add to TARGET_DEFAULT flags.
15692 Expand ASM_SPEC and LINK_SPEC.
15693 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15694 * config/microblaze/microblaze.md: Update extendsidi2 and
15695 movdi_internal instructions to use low-order / high-order reg
15697 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15698 options and inversemask / mask of LITTLE_ENDIAN.
15699 * config/microblaze/t-microblaze: Expand multilib options to
15700 include mlittle-endian (le) and update exceptions patterns.
15702 2013-02-08 Jakub Jelinek <jakub@redhat.com>
15704 PR rtl-optimization/56195
15705 * lra-constraints.c (get_reload_reg): Don't reuse regs
15706 if they have smaller mode than requested, if they have
15707 wider mode than requested, try to return a SUBREG.
15709 PR tree-optimization/56250
15710 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15711 if type is unsigned and code isn't MULT_EXPR.
15713 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15715 PR tree-optimization/56064
15716 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15717 bits according to mode.
15718 * fixed-value.h (fixed_from_double_int)
15719 (const_fixed_from_double_int): Adjust comments.
15721 2013-02-08 Richard Biener <rguenther@suse.de>
15724 * lto-streamer.h (struct data_in): Remove current_file, current_line
15725 and current_col members.
15726 * lto-streamer-out.c (lto_output_location): Stream changed bits
15727 en-block for efficiency.
15728 * lto-streamer-in.c (clear_line_info): Remove.
15729 (lto_input_location): Cache current file, line and column
15730 globally via local statics. Read changed bits en-block.
15731 (input_function): Do not call clear_line_info.
15732 (lto_read_body): Likewise.
15733 (lto_input_toplevel_asms): Likewise.
15735 2013-02-08 Michael Matz <matz@suse.de>
15737 PR tree-optimization/52448
15738 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15739 (nt_call_phase): New static.
15740 (add_or_mark_expr): Only mark accesses with newer phase than any
15742 (nonfreeing_call_p): New.
15743 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15744 (nt_fini_block): Keep blocks marked as visited.
15745 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15747 2013-02-08 Richard Biener <rguenther@suse.de>
15749 * ira.c (ira): Free broken dominator information.
15751 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15753 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15755 2013-02-08 Marek Polacek <polacek@redhat.com>
15757 * cfgloop.c (verify_loop_structure): Add more checking of headers.
15759 2013-02-08 Richard Biener <rguenther@suse.de>
15761 PR middle-end/56181
15762 * cfgloop.h (flow_loops_find): Adjust.
15763 (bb_loop_header_p): Declare.
15764 * cfgloop.c (bb_loop_header_p): New function split out from ...
15765 (flow_loops_find): ... here. Adjust function signature,
15766 support incremental loop structure update.
15767 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15768 * cfgloopmanip.c (fix_loop_structure): Move ...
15769 * loop-init.c (fix_loop_structure): ... here.
15770 (apply_loop_flags): Split out from ...
15771 (loop_optimizer_init): ... here.
15772 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15773 in incremental mode, only remove dead loops here.
15775 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15778 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15779 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15780 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15781 (*round<mode>3.libgcc): New insns for fixed-modes.
15782 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15783 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15784 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15785 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15786 implementations. Define to __builtin_avr_absFX,
15787 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15788 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15789 __builtin_avr_countlsFX, respectively.
15790 * config/avr/avr-c.c (target.h): Include it.
15791 (enum avr_builtin_id): New enum.
15792 (avr_resolve_overloaded_builtin): New static function.
15793 (avr_register_target_pragmas): Use it to set
15794 targetm.resolve_overloaded_builtin.
15795 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15796 tree nodes used by DEF_BUILTIN.
15797 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15798 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15799 <AVR_BUILTIN_xxBITS>: Same.
15801 2013-02-08 Richard Biener <rguenther@suse.de>
15803 * cfgloop.c (verify_loop_structure): Properly handle
15804 a loop exiting to another loop header.
15805 * ira-int.h (ira_loops): Remove.
15806 * ira.c (ira_loops): Remove.
15807 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15808 (do_reload): Use loop_optimizer_finalize.
15809 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15810 number_of_loops to access the loop tree.
15811 (more_one_region_p): Likewise.
15812 (finish_loop_tree_nodes): Likewise.
15813 (rebuild_regno_allocno_maps): Likewise.
15814 (mark_loops_for_removal): Likewise.
15815 (mark_all_loops_for_removal): Likewise.
15816 (remove_unnecessary_regions): Likewise.
15817 (ira_build): Likewise.
15818 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15820 2013-02-08 Richard Biener <rguenther@suse.de>
15822 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15823 * ipa-pure-const.c (analyze_function): Avoid calling
15824 mark_irreducible_loops twice.
15825 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
15827 2013-02-07 David S. Miller <davem@davemloft.net>
15829 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15831 * var-tracking.c (vt_add_function_parameter): Test the presence of
15832 HAVE_window_save properly and do not remap argument registers when
15833 we have a leaf function.
15835 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
15838 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15840 * config/i386/i386.c (ix86_print_operand): Ditto.
15842 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15844 * lra-constraints.c (process_alt_operands): Fix recently added comment.
15846 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15848 PR rtl-optimization/56225
15849 * lra-constraints.c (process_alt_operands): Check that reload hard
15850 reg can hold value for strict_low_part.
15852 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15855 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15856 dwarf2out_end_function.
15857 (in_first_function_p, maybe_at_text_label_p,
15858 first_loclabel_num_not_at_text_label): New variables.
15859 (dwarf2out_var_location): In the first function find out
15860 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
15861 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15864 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15866 PR rtl-optimization/56178
15867 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15868 SUBREG of a register. Tidy up related block of code.
15869 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15870 note if the source is a register or a SUBREG of a register.
15872 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15875 * config/rs6000/rs6000.md (ptrm): New mode attr.
15876 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15877 call_value_indirect_aix<pttrsize>,
15878 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15881 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15883 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15884 if -bnortl. Convert to strcmp and strncmp.
15886 2013-02-07 Alan Modra <amodra@gmail.com>
15889 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15890 addresses won't wrap when offsetting.
15891 (rs6000_secondary_reload): Provide secondary reloads needed for
15892 wrapping LO_SUM addresses.
15894 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15896 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15897 MACH, just __MACH__.
15899 2013-02-06 Richard Biener <rguenther@suse.de>
15901 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15902 instead of calling fix_loop_structure.
15904 2013-02-06 Jakub Jelinek <jakub@redhat.com>
15906 PR middle-end/56217
15907 * omp-low.c (use_pointer_for_field): Return false if
15908 lower_send_shared_vars doesn't generate any copy-out code.
15910 2013-02-06 Tom de Vries <tom@codesourcery.com>
15912 PR rtl-optimization/56131
15913 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15914 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15915 of the label is NULL. Add comment.
15917 2013-02-05 Jakub Jelinek <jakub@redhat.com>
15919 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15922 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15923 (STATIC_LIBTSAN_LIBS): Likewise.
15924 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15925 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15926 is defined, don't add anything else beyond that.
15927 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15928 (LINK_COMMAND_SPEC): Use them.
15930 PR tree-optimization/56205
15931 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15932 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15933 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15935 2013-02-05 Richard Biener <rguenther@suse.de>
15937 PR tree-optimization/53342
15938 PR tree-optimization/53185
15939 * tree-vectorizer.h (vect_check_strided_load): Remove.
15940 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15941 not disallow peeling for vectorized strided loads.
15942 (vect_check_strided_load): Make static and simplify.
15943 (vect_analyze_data_refs): Adjust.
15944 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15945 correctly when vectorizing strided loads.
15947 2013-02-05 Richard Biener <rguenther@suse.de>
15949 * doc/install.texi: Refer to ISL, not PPL.
15951 2013-02-05 Jan Hubicka <jh@suse.cz>
15953 PR tree-optimization/55789
15954 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15956 2013-02-05 Jan Hubicka <jh@suse.cz>
15958 PR tree-optimization/55789
15959 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15960 the dead call anyway.
15962 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15965 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15967 2013-02-04 Alexander Potapenko <glider@google.com>
15968 Jack Howarth <howarth@bromo.med.uc.edu>
15969 Jakub Jelinek <jakub@redhat.com>
15972 * config/darwin.c (sort_ctor_records): Stabilized qsort
15973 on constructor priority by using original position.
15974 (finalize_ctors): New routine to sort constructors by
15975 priority before use in assemble_integer.
15976 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15978 2013-02-04 Jakub Jelinek <jakub@redhat.com>
15981 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15982 about visibility on artificial decls.
15983 * config/sol2.c (solaris_assemble_visibility): Likewise.
15985 2013-02-04 Kai Tietz <ktietz@redhat.com>
15988 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15989 argument and improve checking of return-argument types for 16-byte
15991 (ix86_function_value_1): Add additional valtype argument on call
15992 of function_value_64.
15993 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
15994 handling infunction_value_64 function.
15996 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15998 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16000 2013-02-04 Richard Biener <rguenther@suse.de>
16002 PR tree-optimization/56188
16003 * tree-ssa-structalias.c (label_visit): Consider case with
16004 initially non-empty points-to set.
16005 (perform_var_substitution): Dump node mapping and clean up.
16007 2013-02-04 Richard Guenther <rguenther@suse.de>
16010 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16011 node prevail as last resort.
16012 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16013 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16015 2013-02-04 Richard Biener <rguenther@suse.de>
16017 PR tree-optimization/56113
16018 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16020 (equiv_class_lookup_or_add): ... this.
16021 (label_visit): Adjust and fix error in previous patch.
16022 (perform_var_substitution): Adjust.
16024 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16026 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16027 * config/sh/sh4-300.md: Likewise.
16028 * config/sh/sh4a.md: Likewise.
16029 * config/sh/constraints.md: Likewise.
16030 * config/sh/sh.md: Likewise.
16031 * config/sh/netbsd-elf.h: Likewise.
16032 * config/sh/predicates.md: Likewise.
16033 * config/sh/sh-protos.h: Likewise.
16034 * config/sh/ushmedia.h: Likewise.
16035 * config/sh/linux.h: Likewise.
16036 * config/sh/sh.c: Likewise.
16037 * config/sh/superh.h: Likewise.
16038 * config/sh/elf.h: Likewise.
16039 * config/sh/sh4.md: Likewise.
16040 * config/sh/sh.h: Likewise.
16042 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16044 * config/pa/constraints.md: Adjust unused letters. Change "T"
16045 constraint to match_test floating_point_store_memory_operand().
16046 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16047 (base14_operand): New.
16048 (floating_point_store_memory_operand): New.
16049 (integer_store_memory_operand): Revise to use base14_operand and
16050 reg_plus_base_memory_operand.
16051 (move_dest_operand): Allow symbolic_memory_operands.
16052 (symbolic_memory_operand): Check for LO_SOM.
16053 (symbolic_operand): Change default case to break.
16054 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16055 CONST_DOUBLE values to be reloaded by putting them into memory when
16056 the destination is a floating point register.
16057 (movdf): Remove code to handle CONST_DOUBLE.
16059 (reload_indf_r1): New.
16060 (reload_insf_r1): New.
16061 Consistently use "Q" and "T" constraints with integer and floating
16062 point move instructions, respectively.
16063 (movdi): Remove FAIL.
16064 Change predicate for source operand unamed DImode move from
16065 general_operand to move_src_operand.
16066 (umulsidi3): Change predicate for destination operand to
16068 Likewise for similar unamed patterns.
16069 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16070 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16071 (hppa_legitimize_address): Simplify mask calculation.
16072 (pa_emit_move_sequence): Revised handling of secondary reloads from
16073 REG+D addresses for floating point loads and stores. Directly handle
16074 loading CONST0_RTX (mode) to a floating point register.
16075 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16076 to floating point registers. Don't restrict secondary reloads to
16077 floating point registers to integer modes. Revise some comments and
16079 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16080 (pa_legitimate_address_p): New.
16081 (pa_legitimize_reload_address): New.
16082 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16083 (STRICT_REG_OK_FOR_BASE_P): New.
16084 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16085 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16087 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16088 Andrew Dixie <andrewd@gentrack.com>
16090 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16093 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16095 * expmed.c (extract_bit_field_1): Pass the full width of the
16096 structure to get_best_reg_extraction_insn.
16098 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16101 * configure.ac (use_cxa_atexit): Add AIX.
16102 * configure: Regenerate.
16104 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16106 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16109 * final.c (need_profile_function): New variable.
16110 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16111 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16112 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16113 notes, targetm.asm_out.function_prologue doesn't emit anything,
16114 HAVE_prologue and profiler should be emitted before prologue,
16115 set need_profile_function instead of emitting it.
16116 (final_scan_insn): If need_profile_function, emit
16117 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16118 NOTE_INSN_FUNCTION_BEG note.
16120 2013-02-01 Richard Henderson <rth@redhat.com>
16122 * config/rs6000/rs6000.md (smulditi3): New.
16125 * config/alpha/alpha.md (umulditi3): New.
16127 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16129 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16130 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16132 2013-02-01 Richard Biener <rguenther@suse.de>
16134 PR tree-optimization/56113
16135 * tree-ssa-structalias.c (label_visit): Reduce work for
16136 single-predecessor nodes.
16138 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
16140 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
16141 range isn't testing for zero.
16143 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
16145 PR middle-end/56113
16146 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
16148 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
16149 Nick Clifton <nickc@redhat.com>
16151 * config/v850/constraints.md (Q): Define as a memory constraint.
16152 * config/v850/predicates.md (label_ref_operand): New predicate.
16153 (e3v5_shift_operand): New predicate.
16154 (ior_operator): New predicate.
16155 * config/v850/t-v850: Add e3v5 multilib.
16156 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
16157 (v850_gen_movdi): Prototype.
16158 * config/v850/v850.c: Add support for e3v5 architecture.
16159 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
16161 (construct_save_jarl): Add e3v5 long JARL support.
16162 (v850_adjust_insn_length): New function. Adjust length of call
16163 insns when using e3v5 instructions.
16164 (v850_gen_movdi): New function: Generate instructions to move a
16166 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
16167 (CPP_SPEC): Define __v850e3v5__ as appropriate.
16168 (TARGET_USE_FPU): Enable for e3v5.
16169 (CONST_OK_FOR_W): New macro.
16170 (ADJUST_INSN_LENGTH): Define.
16171 * config/v850/v850.md (UNSPEC_LOOP): Define.
16172 (attr cpu): Add v850e3v5.
16173 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
16174 (movdi): New pattern.
16175 (movdi_internal): New pattern.
16176 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
16177 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
16178 (cstoresf4): Likewise.
16179 (cstoredf4): Likewise.
16180 (insv): New pattern.
16181 (rotlso3_a): New pattern.
16182 (rotlsi3_b): New pattern
16183 (rotlsi3_v850e3v5): New pattern.
16184 (doloop_begin): New pattern.
16185 (fix_loop_counter): New pattern.
16186 (doloop_end): New pattern.
16187 (branch_normal): Add e3v5 long branch support.
16188 (branch_invert): Likewise.
16189 (branch_z_normal): Likewise.
16190 (branch_z_invert): Likewise.
16191 (branch_nz_normal): Likewise.
16192 (branch_nz_invert): Likewise.
16193 (call_internal_short): Add e3v5 register-indirect JARL support.
16194 (call_internal_long): Likewise.
16195 (call_value_internal_short): Likewise.
16196 (call_value_internal_long): Likewise.
16197 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
16198 (mloop): New option.
16199 * config.gcc: Add support for configuring v840e3v5 target.
16200 * doc/invoke.texi: Document new v850 specific command line options.
16202 2013-01-31 Paul Koning <ni1d@arrl.net>
16206 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
16207 children if parent is a class.
16208 (prune_unused_types_prune): Don't add DW_AT_declaration.
16210 2013-01-31 Richard Biener <rguenther@suse.de>
16212 PR tree-optimization/56157
16213 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
16214 match up operand with SLP child.
16216 2013-01-31 Jason Merrill <jason@redhat.com>
16219 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
16220 parameters the first time.
16221 (gen_scheduled_generic_parms_dies): Check completeness here.
16223 2013-01-31 Richard Biener <rguenther@suse.de>
16225 PR middle-end/53073
16226 * common.opt (faggressive-loop-optimizations): New flag,
16227 enabled by default.
16228 * doc/invoke.texi (faggressive-loop-optimizations): Document.
16229 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
16230 infer_loop_bounds_from_undefined by it.
16232 2013-01-31 Richard Biener <rguenther@suse.de>
16234 PR tree-optimization/56150
16235 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
16236 visit virtual operands.
16237 (find_uses_to_rename_bb): Likewise.
16239 2013-01-31 Richard Biener <rguenther@suse.de>
16241 PR tree-optimization/56150
16242 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
16243 mixed store non-store stmts.
16245 2013-01-30 Jakub Jelinek <jakub@redhat.com>
16248 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
16249 LIBASAN_EARLY_SPEC is defined.
16250 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
16251 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
16253 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
16256 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
16257 invalid args instead of ICEing on it.
16258 (ix86_valid_target_attribute_tree): Return error_mark_node if
16259 ix86_valid_target_attribute_inner_p failed.
16260 (ix86_valid_target_attribute_p): Return false only if
16261 ix86_valid_target_attribute_tree returned error_mark_node. Allow
16262 target("default") attribute.
16263 (sorted_attr_string): Change argument from const char * to tree,
16264 merge in all target attribute arguments rather than just one.
16265 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
16266 instead of free. Avoid using strcat.
16267 (ix86_mangle_function_version_assembler_name): Mangle
16268 target("default") as if no target attribute is present. Adjust
16269 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
16270 instead of xmalloc and XDELETEVEC instead of free.
16271 (ix86_function_versions): Don't return true if one of the decls
16272 doesn't have target attribute. If they don't and one of the decls
16273 is DECL_FUNCTION_VERSIONED, report an error. Adjust
16274 sorted_attr_string caller. Use XDELETEVEC instead of free.
16275 (ix86_supports_function_versions): Remove.
16276 (make_name): Fix up formatting.
16277 (make_dispatcher_decl): Remove resolver_name and its initialization.
16278 Avoid leaking memory.
16279 (is_function_default_version): Return true if there is
16280 target("default") attribute rather than no target attribute at all.
16281 (make_resolver_func): Avoid leaking memory.
16282 (ix86_generate_version_dispatcher_body): Likewise.
16283 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
16284 * target.def (supports_function_versions): Remove.
16285 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
16286 * doc/tm.texi: Regenerated.
16288 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
16290 PR rtl-optimization/56144
16291 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
16292 for values with side effects.
16294 2013-01-30 Richard Biener <rguenther@suse.de>
16296 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
16297 (sparseset_pop): Likewise.
16298 * cfganal.c (compute_idf): Likewise. Increase work-stack size
16299 to be able to use quick_push in the worker loop.
16301 2013-01-30 Marek Polacek <polacek@redhat.com>
16303 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
16305 2013-01-30 Richard Biener <rguenther@suse.de>
16308 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
16310 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
16312 PR tree-optimization/56064
16313 * fixed-value.c (fixed_from_double_int): New function.
16314 * fixed-value.h (fixed_from_double_int): New prototype.
16315 (const_fixed_from_double_int): New static inline function.
16316 * fold-const.c (native_interpret_fixed): New static function.
16317 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
16318 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
16319 (native_encode_fixed): New static function.
16320 (native_encode_expr) <FIXED_CST>: Use it.
16321 (native_interpret_int): Move double_int worker code to...
16322 * double-int.c (double_int::from_buffer): ...this new static method.
16323 * double-int.h (double_int::from_buffer): Prototype it.
16325 2013-01-30 Richard Biener <rguenther@suse.de>
16327 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
16328 New pointer-map and obstack.
16329 (init_alias_vars): Allocate pointer-map and obstack.
16330 (delete_points_to_sets): Free them.
16331 (find_what_var_points_to): Cache result.
16332 (find_what_p_points_to): Adjust for changed interface of
16333 find_what_var_points_to.
16334 (compute_points_to_sets): Likewise.
16335 (ipa_pta_execute): Likewise.
16337 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16339 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
16340 * configure: Regenerate.
16341 * config.in: Regenerate.
16342 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
16343 #nobits/#progbits if supported.
16345 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
16348 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
16349 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
16350 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
16352 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16354 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
16355 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
16357 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16359 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
16360 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
16362 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16364 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
16366 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
16367 * config/arm/cortex-a7.md: New bypasses using
16368 arm_mac_accumulator_is_result.
16370 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16372 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
16373 (cortex_a7_neon_mla): Likewise.
16374 (cortex_a7_fpfmad): New reservation.
16375 (cortex_a7_fpmacs): Use ffmas and update required units.
16376 (cortex_a7_fpmuld): Update required units and latency.
16377 (cortex_a7_fpmacd): Likewise.
16378 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
16379 (cortex_a7_neon). Likewise.
16380 (bypass) Update participating units.
16382 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16384 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
16385 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
16387 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16388 (vfp_fmul): Use ffmad.
16389 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16390 (cortex_r4_fmacd): Use ffmad.
16391 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16392 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16393 (cortex_a9_fmacd): Use ffmad.
16394 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16395 (cortex_a8_vfp_macd): Use ffmad.
16396 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16397 (cortex_a5_fpmacd): Use ffmad.
16398 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16399 (cortex_a15_vfp_macd): Use ffmad.
16400 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16402 2013-01-29 Jason Merrill <jason@redhat.com>
16405 * varasm.c (default_assemble_visibility): Don't warn about
16406 visibility on artificial decls.
16408 2013-01-29 Richard Biener <rguenther@suse.de>
16410 PR tree-optimization/56113
16411 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16413 (label_visit): Free duplicate bitmaps and record the leader instead.
16414 (perform_var_substitution): Adjust.
16416 2013-01-29 Richard Biener <rguenther@suse.de>
16418 PR tree-optimization/55270
16419 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16420 the CFG, schedule loops for fixup.
16422 2013-01-29 Nick Clifton <nickc@redhat.com>
16424 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16427 2013-01-28 Leif Ekblad <leif@rdos.net>
16429 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16430 * config/i386/i386.h (TARGET_RDOS): New macro.
16431 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16432 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16433 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16434 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16435 DEFAULT_LARGE_SECTION_THRESHOLD.
16436 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16437 * config/i386/rdos.h: New file.
16438 * config/i386/rdos64.h: New file.
16440 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16443 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16446 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16448 PR rtl-optimization/56117
16449 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16450 call cselib_lookup_from_insn on the MEM before calling
16451 add_insn_mem_dependence.
16453 2013-01-28 Richard Biener <rguenther@suse.de>
16455 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16456 to a stmt that didn't have one.
16457 (copy_phis_for_bb): Likewise for PHI arguments.
16458 (copy_debug_stmt): Likewise for debug stmts.
16460 2013-01-28 Richard Biener <rguenther@suse.de>
16462 PR tree-optimization/56034
16463 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16464 (partition_builtin_p): Adjust.
16465 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16466 it is the last partition.
16467 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16468 up the vertex for the definition.
16469 (classify_partition): Classify whether a partition is a
16470 PKIND_REDUCTION, thus has uses outside of the loop.
16471 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16472 Merge all PKIND_REDUCTION partitions into the last partition.
16473 (tree_loop_distribution): Seed partitions from reductions as well.
16475 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16477 PR tree-optimization/56125
16478 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16479 pow(x,c) into sqrt(x) * powi(x, n/2) or
16480 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16481 optimizing for size.
16482 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16483 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16486 PR tree-optimization/56094
16487 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16488 to UNKNOWN_LOCATION while gimplifying expr.
16490 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16493 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16494 operand 0 in movabs insn template for -masm=intel asm alternative.
16495 (*movabs<mode>_2): Ditto for operand 1.
16497 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16500 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16503 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16505 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16508 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16510 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16511 Undef to avoid warning.
16513 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16515 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16516 * configure: Regenerate.
16518 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16520 PR tree-optimization/56098
16521 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16522 for stmts with volatile ops.
16523 (cond_store_replacement): Don't optimize if assign has volatile ops.
16524 (cond_if_else_store_replacement_1): Don't optimize if either
16525 then_assign or else_assign have volatile ops.
16526 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16529 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16531 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16533 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16535 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16536 missing ':' in asm example.
16538 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16540 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16541 entries into lane and laneq entries.
16542 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16543 Remove AdvSIMD scalar modes.
16544 (aarch64_sq<r>dmulh_laneq<mode>): New.
16545 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16547 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16548 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16549 * config/aarch64/iterators.md (VCOND): New.
16552 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16555 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16556 Add NULL LIBNAME argument to existing definitions.
16557 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16558 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16559 * config/avr/avr.c (DEF_BUILTIN): Same.
16560 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16561 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16562 implementation is available (DECL_ASSEMBLER_NAME is set).
16563 (avr_fold_absfx): New static function.
16564 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16565 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16566 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16567 AVR_BUILTIN_ABSLLK.
16568 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16569 (abshk, absk, abslk, absllk): Provide as static inline functions.
16571 2013-01-25 Marek Polacek <polacek@redhat.com>
16573 PR tree-optimization/56035
16574 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16576 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
16578 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16579 (*movtf_internal_rex64): Add (!o,C) alternative
16580 (*movxf_internal_rex64): Ditto.
16581 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16583 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
16585 * doc/invoke.texi: fix typo.
16586 * doc/objc.texi: fix typo.
16588 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16590 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16591 for the first two alternatives.
16593 2013-01-24 Diego Novillo <dnovillo@google.com>
16595 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
16596 (ggc-zone.o): Remove.
16597 * configure.ac: Remove option --with-gc.
16598 * configure: Re-generate.
16599 * doc/install.texi: Remove documentation for --with-gc.
16600 * gengtype.c (write_enum_defn): Remove. Update all users.
16601 (write_Types_process_field): Remove generation of gt_e_* argument.
16602 (output_type_enum): Remove. Update all users.
16603 (write_enum_defn): Remove. Update all users.
16604 (enum alloc_zone): Remove. Update all users.
16605 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
16606 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16607 Update all callers.
16608 (struct ptr_data): Remove field TYPE. Update all users.
16609 (gt_pch_note_object): Remove argument TYPE. Update all users.
16610 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16612 * ggc-none.c (ggc_alloc_typed_stat): Remove.
16613 (struct alloc_zone): Remove.
16614 (ggc_internal_alloc_zone_stat): Remove.
16615 (ggc_internal_cleared_alloc_zone_stat): Remove.
16616 * ggc-page.c (ggc_alloc_typed_stat): Remove.
16617 (ggc_pch_count_object): Remove last argument. Update all users.
16618 (ggc_pch_alloc_object): Remove last argument. Update all users.
16619 (struct alloc_zone): Remove.
16620 * ggc-zone.c: Remove.
16621 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
16622 (struct alloc_zone): Remove.
16623 (ggc_alloc_typed_stat): Remove.
16624 (ggc_alloc_typed): Remove.
16625 (ggc_splay_alloc): Remove first argument.
16626 (rtl_zone): Remove. Update all users.
16627 (tree_zone): Remove. Update all users.
16628 (tree_id_zone): Remove. Update all users.
16629 (ggc_internal_zone_alloc_stat): Remove. Update all users.
16630 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
16631 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
16632 * tree-ssanames.c: Remove references to zone allocator in comments.
16634 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
16636 * config/avr/avr.c (avr_out_fract): Make register numbers that
16637 might be outside of source operand signed.
16639 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
16641 * config/i386/constraints.md (Yf): New constraint.
16642 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16643 of f constraint to conditionaly disable x87 register preferences.
16644 (*movdf_internal): Ditto.
16645 (*movsf_internal): Ditto.
16647 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16649 PR inline-asm/55934
16650 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16651 that have operands with impossible constraints.
16652 Add a FIXME for a speed-up opportunity.
16653 * lra-constraints.c (process_alt_operands): Verify that a class
16654 selected from constraints on asms is valid for the operand mode.
16655 (curr_insn_transform): Remove incorrect comment.
16657 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
16659 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16660 TOC operand is a valid symbol ref in the constant pool.
16662 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16664 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
16666 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
16669 * config/avr/stdfix.h: New file.
16670 * t-avr (stdfix-gcc.h): New rule to build it.
16671 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16673 2013-01-23 Kostya Serebryany <kcc@google.com>
16675 * config/darwin.h: remove dependency on
16676 CoreFoundation (asan on Mac OS).
16678 2013-01-23 Jakub Jelinek <jakub@redhat.com>
16681 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16682 instead of cmpdi_operand for first comparison operand.
16683 Don't assert that comparison operands aren't both constants.
16685 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16687 * doc/install.texi (Downloading the Source): Update references to
16688 downloading separate components.
16690 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16692 * doc/extend.texi (__int128): Improve grammar.
16694 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
16697 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16698 alternative to (o,r).
16699 (*movdi_internal_rex64): Remove (!o,n) alternative.
16700 (DImode immediate->memory splitter): Remove.
16701 (DImode immediate->memory peephole2): Remove.
16702 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16703 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16704 alternative to (!o,*r).
16705 (*movtf_internal_sse): New pattern.
16706 (*movxf_internal_rex64): New pattern.
16707 (*movxf_internal): Disable for TARGET_64BIT.
16708 (*movdf_internal_rex64): Remove (!o,F) alternative.
16710 2013-01-22 Jakub Jelinek <jakub@redhat.com>
16712 PR middle-end/56074
16713 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16714 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16715 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16716 stmt locations where LOCATION_LOCUS of the stmt location is
16717 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16720 * config/i386/i386.md (UNSPEC_STOS): New.
16721 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16722 *strsetqi_1): Add UNSPEC_STOS.
16724 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16727 * doc/invoke.texi: Remove left over -Wsynth example.
16729 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16731 PR tree-optimization/56051
16732 * fold-const.c (fold_binary_loc): Don't fold
16733 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16734 a narrowing conversion, or widening conversion from signed
16737 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16739 PR rtl-optimization/56023
16740 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16741 dependent on debug instruction.
16743 2013-01-21 Martin Jambor <mjambor@suse.cz>
16745 PR middle-end/56022
16746 * function.c (allocate_struct_function): Call
16747 invoke_set_current_function_hook earlier.
16749 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16751 * reload1.c (init_reload): Only initialize reload_obstack
16752 during the first call.
16754 2013-01-21 Marek Polacek <polacek@redhat.com>
16756 * cfgloop.c (verify_loop_structure): Fix up grammar.
16758 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16760 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16761 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16763 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16766 * config/arm/marvell-pj4.md: Update copyright year.
16767 Fix up use of alu to alu_reg and simple_alu_imm.
16769 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16771 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16773 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16776 * lra-constraints.c (curr_insn_transform): Don't reuse original
16777 insn for secondary memory move when memory mode should be different.
16779 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16781 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16782 atomic_storedi_1): New patterns.
16784 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16786 btver2 pipeline descriptions.
16787 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16789 * config/i386/i386.md (btver2_decode): New type attributes.
16790 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
16792 * config/i386/btver2.md: New file describing btver2 pipelines.
16794 2013-01-19 Andrew Pinski <apinski@cavium.com>
16796 PR tree-optimization/52631
16797 * tree-ssa-sccvn (visit_use): Before looking up the original
16798 statement, try looking up the simplified expression.
16800 2013-01-19 Anthony Green <green@moxielogic.com>
16802 * config/moxie/moxie.c (moxie_expand_prologue): Set
16803 current_function_static_stack_size.
16805 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16807 PR tree-optimization/56029
16808 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16809 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16811 2013-01-18 Sharad Singhai <singhai@google.com>
16813 PR tree-optimization/55995
16814 * dumpfile.c (dump_loc): Print location only if available.
16815 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16817 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16820 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16821 secondary memory move.
16822 (inherit_reload_reg): Use rclass instead of cl for
16823 check_secondary_memory_needed_p.
16825 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16827 PR middle-end/56015
16828 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
16829 the case where writing real complex part of target modifies op1.
16831 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16833 * config/aarch64/aarch64-simd.md
16834 (aarch64_vcond_internal<mode>): Handle unordered cases.
16835 * config/aarch64/iterators.md (v_cmp_result): New.
16837 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16838 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16840 * config/arm/marvell-pj4.md: New file.
16841 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16842 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16843 (generic_vfp): Likewise.
16844 * config/arm/arm-cores.def: Add marvell-pj4.
16845 * config/arm/arm-tune.md: Regenerate.
16846 * config/arm/arm-tables.opt: Regenerate.
16847 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16848 * doc/invoke.texi: Document marvell-pj4.
16850 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16852 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16854 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16859 * var-tracking.c (negative_power_of_two_p): New.
16860 (global_get_addr_cache, local_get_addr_cache): New.
16861 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16862 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16864 (vt_stack_offset_p): Remove.
16865 (vt_canon_true_dep): Always canonicalize loc's address.
16866 (clobber_overlapping_mems): Make sure we have a MEM.
16867 (local_get_addr_clear_given_value): New.
16868 (val_reset): Clear local cached entries.
16869 (compute_bb_dataflow): Create and release the local cache.
16870 Disable duplicate MEMs clobbering.
16871 (emit_notes_in_bb): Clobber MEMs likewise.
16872 (vt_emit_notes): Create and release the local cache.
16873 (vt_initialize, vt_finalize): Create and release the global
16874 cache, respectively.
16875 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
16877 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16879 PR libmudflap/53359
16880 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16881 not found in the symtab.
16883 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16886 PR rtl-optimization/55547
16887 PR rtl-optimization/53827
16890 * alias.c (offset_overlap_p): New, factored out of...
16891 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16892 the conservative special case for symbolic constants. Don't
16893 adjust zero sizes on alignment.
16895 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16897 PR rtl-optimization/52573
16898 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16899 REG_UNUSED for the same register.
16901 2013-01-17 Richard Biener <rguenther@suse.de>
16902 Marek Polacek <polacek@redhat.com>
16904 PR rtl-optimization/55833
16905 * loop-unswitch.c (unswitch_loops): Move loop verification...
16906 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16907 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16908 Set it to true when we're removing a loop from hierarchy tree in
16909 an irreducible region.
16910 (fix_bb_placements): Adjust caller.
16911 (fix_loop_placements): Likewise.
16913 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
16915 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16916 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16917 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16918 Remove ID. Adjust comments.
16919 * config/avr/avr-c.c (avr_builtin_name): Remove.
16920 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16921 * config/avr/avr.c (avr_tolower): New static function.
16922 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16923 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16924 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16927 2013-01-17 Jan Hubicka <jh@suse.cz>
16929 PR tree-optimization/55273
16930 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16932 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
16935 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16936 store through atomic_store<mode>_1.
16937 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16939 2013-01-17 Martin Jambor <mjambor@suse.cz>
16941 PR tree-optimizations/55264
16942 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16943 for virtual methods.
16944 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16945 virtual methods before inlining is over.
16946 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16948 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16951 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16953 PR rtl-optimization/56005
16954 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16955 pending reads for prefetch.
16957 2013-01-16 Ian Bolton <ian.bolton@arm.com>
16959 * config/aarch64/aarch64.md
16960 (*cstoresi_neg_uxtw): New pattern.
16961 (*cmovsi_insn_uxtw): New pattern.
16962 (*<optab>si3_uxtw): New pattern.
16963 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16964 (*<optab>si3_insn_uxtw): New pattern.
16965 (*bswapsi2_uxtw): New pattern.
16967 2013-01-16 Richard Biener <rguenther@suse.de>
16969 * tree-inline.c (tree_function_versioning): Remove set but
16970 never used variable.
16972 2013-01-16 Richard Biener <rguenther@suse.de>
16974 PR tree-optimization/55964
16975 * tree-flow.h (rename_variables_in_loop): Remove.
16976 (rename_variables_in_bb): Likewise.
16977 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16978 (copy_loop_before): Adjust and delete update-ssa status.
16979 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
16980 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
16981 (rename_variables_in_loop): Remove.
16982 (slpeel_update_phis_for_duplicate_loop): Likewise.
16983 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16984 use available cfg machinery instead of duplicating it.
16985 Update PHI nodes and perform poor-mans SSA update here.
16986 (slpeel_tree_peel_loop_to_edge): Adjust.
16988 2013-01-16 Richard Biener <rguenther@suse.de>
16990 PR tree-optimization/54767
16991 PR tree-optimization/53465
16992 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
16993 (vrp_visit_phi_node): For PHI arguments coming via backedges
16994 drop all symbolical range information.
16995 (execute_vrp): Compute backedges.
16997 2013-01-16 Richard Biener <rguenther@suse.de>
16999 * doc/install.texi: Update CLooG and ISL requirements to
17002 2013-01-16 Christian Bruel <christian.bruel@st.com>
17005 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17006 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17007 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17009 2013-01-16 DJ Delorie <dj@redhat.com>
17011 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17012 (UNSPECV_SP_SWITCH_E): New.
17013 (sp_switch_1): Change to an unspec.
17014 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17017 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17019 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17020 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17021 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17022 (expand_mem_thread_fence): Ditto.
17023 (expand_mem_signal_fence): Ditto.
17024 (expand_atomic_load): Ditto.
17025 (expand_atomic_store): Ditto.
17027 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17029 PR rtl-optimization/55547
17030 PR rtl-optimization/53827
17033 * alias.c (memrefs_conflict_p): Set sizes to negative after
17036 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17039 * function.c (thread_prologue_and_epilogue_insns): Always
17040 add crtl->drap_reg to set_up_by_prologue.set, even if
17041 stack_realign_drap is false.
17043 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17045 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17046 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17047 *call): Fix indention.
17049 2013-01-15 Tom de Vries <tom@codesourcery.com>
17052 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17055 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17057 PR rtl-optimization/55153
17058 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17060 2013-01-15 Martin Jambor <mjambor@suse.cz>
17062 PR tree-optimization/55920
17063 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17064 accesses as grp_to_be_debug_replaced.
17066 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17068 PR tree-optimization/55920
17069 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17070 there is non-useless type conversion needed from debug rhs to lhs,
17071 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17073 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17074 Mikael Pettersson <mikpe@it.uu.se>
17077 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17079 (ASM_OUTPUT_CASE_LABEL): Remove.
17080 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17081 * final.c (shorten_branches): Update alignment of labels before
17082 jump tables if CASE_VECTOR_SHORTEN_MODE.
17084 2013-01-15 Richard Biener <rguenther@suse.de>
17087 * system.h: Do not include gmp.h for building host tools.
17089 2013-01-15 Richard Biener <rguenther@suse.de>
17091 PR middle-end/55882
17092 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17093 account for bitpos when computing alignment.
17095 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17097 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17098 (ix86_target_macros_internal): Likewise.
17100 * config/i386/i386.c (m_CORE2I7): Removed.
17101 (m_CORE_HASWELL): New macro.
17102 (m_CORE_ALL): Likewise.
17103 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17104 (initial_ix86_arch_features): Likewise.
17105 (processor_target_table): Initializations for Core avx2.
17106 (cpu_names): New names "core-avx2".
17107 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17108 PROCESSOR_CORE_HASWELL.
17109 (ix86_issue_rate): New case.
17110 (ia32_multipass_dfa_lookahead): Likewise.
17111 (ix86_sched_init_global): Likewise.
17113 * config/i386/i386.h (TARGET_HASWELL): New macro.
17114 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17115 (processor_type): New PROCESSOR_HASWELL.
17117 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17119 PR tree-optimization/55955
17120 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17121 *SHIFT_EXPR and *ROTATE_EXPR codes.
17123 PR tree-optimization/48766
17124 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17125 -ftrapv disable -fwrapv.
17127 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17130 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17131 etc. to 1 and not to __flash.
17132 Use LL suffix for __INT24_MAX__ with -mint8.
17133 Use ULL suffix for __UINT24_MAX__ with -mint8.
17135 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17137 * config/avr/avr-arch.h
17138 (struct base_arch_s): Use typedef avr_arch_t instead.
17139 (struct arch_info_s): Use typedef avr_arch_info_t instead.
17140 (struct mcu_type_s): Use typedef avr_mcu_t instead.
17141 * config/avr/avr.c: Same.
17142 * config/avr/avr-devices.c: Same.
17143 * config/avr/driver-avr.c: Same.
17144 * config/avr/gen-avr-mmcu-texi.c: Same.
17145 * config/avr/avr-mcus.def: Adjust comment.
17147 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
17149 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
17150 * config/aarch64/iterators.md (VALLDI): New.
17152 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
17153 Andi Kleen <ak@linux.intel.com>
17156 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
17157 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
17160 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17162 * config/avr/avr-stdint.h: Remove trailing blanks.
17163 * config/avr/avr-log.h: Same.
17164 * config/avr/avr-arch.h: Same.
17165 * config/avr/avr-devices.c: Same.
17166 * config/avr/avr-dimode.md: Same.
17167 * config/avr/predicates.md: Same.
17168 * config/avr/avr-c.c: Same. And fix typo.
17170 * config/avr/avr-protos.h: Same. And:
17171 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
17172 (init_cumulative_args): Rename to avr_init_cumulative_args.
17173 (expand_prologue): Rename to avr_expand_prologue.
17174 (expand_epilogue): Rename to avr_expand_epilogue.
17175 (adjust_insn_length): Rename to avr_adjust_insn_length.
17176 (notice_update_cc): Rename to avr_notice_update_cc.
17177 (final_prescan_insn): Rename to avr_final_prescan_insn.
17178 * config/avr/avr.c: Same.
17179 * config/avr/avr.h: Same.
17180 * config/avr/avr.md: Remove trailing blanks.
17181 (prologue): Use avr_expand_prologue.
17182 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
17184 2013-01-14 Richard Biener <rguenther@suse.de>
17186 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
17187 verify_location, collect_subblocks): New functions.
17188 (verify_gimple_in_cfg): Verify that locations only reference
17189 BLOCKs in the functions BLOCK tree.
17191 2013-01-14 Richard Biener <rguenther@suse.de>
17193 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
17195 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
17197 (insert_out_of_ssa_copy_on_edge): Likewise.
17198 (rewrite_close_phi_out_of_ssa): Likewise.
17199 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
17201 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
17202 propagated constants.
17203 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
17206 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17208 * config/avr/avr-modes.def: Add GPL copyright notice.
17210 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
17212 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
17213 MEMMODEL_MASK to determine memory model.
17214 (atomic_store<mode>): Ditto from operands[2].
17215 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
17217 2013-01-13 Jakub Jelinek <jakub@redhat.com>
17220 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
17221 (fold_gimple_assign): Don't call unshare_expr here.
17222 (fold_ctor_reference): Call unshare_expr.
17224 2013-01-13 Terry Guo <terry.guo@arm.com>
17226 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
17227 * doc/fragments.texi: Document MULTILIB_REUSE.
17228 * gcc.c (multilib_reuse): New internal spec.
17229 (set_multilib_dir): Also search multilib from multilib_reuse.
17230 * genmultilib (tmpmultilib3): Refactor code.
17231 (tmpmultilib4): Ditto.
17232 (multilib_reuse): New multilib argument.
17234 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
17236 * Makefile.in: Update copyright.
17238 2013-01-12 Tom de Vries <tom@codesourcery.com>
17240 PR middle-end/55890
17241 * calls.c (expand_call): Check if arg_nr is valid.
17243 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17245 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
17246 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
17247 documentation. Add missing '__' in front of
17248 __builtin_ia32_packssdw256.
17250 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17253 * config/s390/s390.c (s390_preferred_reload_class): Do not return
17254 NO_REGS for larl operands.
17255 (s390_reload_larl_operand): Use s390_load_address instead of
17258 2013-01-11 Richard Biener <rguenther@suse.de>
17260 * tree-cfg.c (verify_node_sharing_1): Split out from ...
17261 (verify_node_sharing): ... here.
17262 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
17264 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
17266 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
17267 Substitute TREECHECKING.
17268 * configure: Regenerate.
17269 * Makefile.in (TREECHECKING): New.
17271 2013-01-11 Richard Guenther <rguenther@suse.de>
17273 PR tree-optimization/44061
17274 * tree-vrp.c (extract_range_basic): Compute zero as
17275 value-range for __builtin_constant_p of function parameters.
17277 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
17279 Update copyright years.
17281 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
17283 PR rtl-optimization/55672
17284 * lra-eliminations.c (mark_not_eliminable): Permit addition with
17285 const to be eliminable.
17287 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
17289 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
17290 * configure: Regenerate.
17292 2013-01-10 Richard Biener <rguenther@suse.de>
17294 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
17296 2013-01-10 Richard Biener <rguenther@suse.de>
17299 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
17300 locations for virtual PHI arguments.
17301 (rewrite_update_phi_arguments): Likewise.
17303 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
17305 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
17308 2013-01-10 Jakub Jelinek <jakub@redhat.com>
17310 PR tree-optimization/55921
17311 * tree-complex.c (expand_complex_asm): New function.
17312 (expand_complex_operations_1): Call it for GIMPLE_ASM.
17314 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17317 * config/s390/s390.c (s390_symref_operand_p)
17318 (s390_loadrelative_operand_p): Merge the two functions.
17319 (s390_check_qrst_address, print_operand_address): Add parameters
17320 to s390_loadrelative_operand_p invokation.
17321 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
17322 (s390_reload_larl_operand, s390_secondary_reload): Use
17323 s390_loadrelative_operand_p instead of s390_symref_operand_p.
17324 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
17326 2013-01-09 Mike Stump <mikestump@comcast.net>
17328 * dse.c (record_store): Remove unnecessary assert.
17330 2013-01-09 Jan Hubicka <jh@suse.cz>
17332 PR tree-optimization/55569
17333 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
17334 * cfgloop.h (scale_loop_profile): Likewise.
17336 2013-01-09 Jan Hubicka <jh@suse.cz>
17339 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
17341 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
17343 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
17345 PR middle-end/55114
17346 * expr.h (maybe_emit_group_store): Declare.
17347 * expr.c (maybe_emit_group_store): New function.
17348 * builtins.c (expand_builtin_int_roundingfn): Call it.
17349 (expand_builtin_int_roundingfn_2): Likewise.
17351 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
17353 PR rtl-optimization/55829
17354 * lra-constraints.c (match_reload): Add code for absent output.
17355 (curr_insn_transform): Add code for reloads of matched inputs
17358 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
17360 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
17361 attribute of movddup insn to DF.
17362 (*vec_interleave_lowv2df): Ditto.
17363 (vec_dupv2df): Ditto.
17365 2013-01-09 Jan Hubicka <jh@suse.cz>
17367 PR tree-optimiation/55875
17368 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
17369 EVERY_ITERATION parameter.
17370 (number_of_iterations_exit): Check if exit is executed every iteration.
17371 (idx_infer_loop_bounds): Similarly here.
17372 (n_of_executions_at_most): Simplify
17373 to only test for cases where statement is dominated by the
17374 particular bound; handle correctly the "postdominance" test.
17375 (scev_probably_wraps_p): Use max loop iterations info
17376 as a global bound first.
17378 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
17379 Nick Clifton <nickc@redhat.com>
17381 * config/v850/v850.md (cbranchsf4): New pattern.
17382 (cstoresf4): New pattern.
17383 (cbranchdf4): New pattern.
17384 (cstoredf4): New pattern.
17385 (movsicc): Disallow floating point comparisons.
17386 (cmpsf_le_insn): Fix order of operators.
17387 (cmpsf_lt_insn): Likewise.
17388 (cmpsf_eq_insn): Likewise.
17389 (cmpdf_le_insn): Likewise.
17390 (cmpdf_lt_insn): Likewise.
17391 (cmpdf_eq_insn): Likewise.
17392 (cmpsf_ge_insn): Use LE comparison.
17393 (cmpdf_ge_insn): Likewise.
17394 (cmpsf_gt_insn): Use LT comparison.
17395 (cmpdf_gt_insn): Likewise.
17396 (cmpsf_ne_insn): Delete pattern.
17397 (cmpdf_ne_insn): Delete pattern.
17398 * config/v850/v850.c (v850_gen_float_compare): Use
17399 gen_cmpdf_eq_insn for NE comparison.
17400 (v850_float_z_comparison_operator)
17401 (v850_float_nz_comparison_operator): Move from here ...
17402 * config/v850/predicates.md: ... to here. Move GT and GE
17403 comparisons into v850_float_z_comparison_operator.
17404 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17406 (v850_float_nz_comparison_operator): Likewise.
17408 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17410 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17411 with calls to gen_insvsi/gen_insvdi.
17413 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17415 * config/i386/i386.c (initial_ix86_tune_features): Set up
17416 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17418 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17419 Jakub Jelinek <jakub@redhat.com>
17421 PR tree-optimization/48189
17422 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17423 If nitercst is 0, don't predict the exit edge.
17425 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17427 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17428 in asm_fprintf with reg_names.
17429 (aarch64_print_operand_address): Likewise.
17430 (aarch64_return_addr): Likewise.
17431 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17433 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17435 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17436 (INT_U6_BITS): Likewise.
17437 * config/pa/predicates.md (uint6_operand): New predicate.
17438 (shift5_operand, shift6_operand): Likewise.
17439 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17441 (lshrdi3): Use shift6_operand.
17442 (shrpsi4, shrpdi4): New insn patterns.
17443 (extzv): Delete expander.
17444 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17445 predicates in unamed zero extract patterns. Tighten common constraint.
17446 (extv): Delete expander.
17447 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17448 predicates in unamed sign extract patterns. Tighten common constraint.
17449 (insv): Delete expander.
17450 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17451 predicates in unamed insert patterns. Tighten common constraint.
17452 Change uint32_operand predicate to uint6_operand predicate in unamed
17453 DImode pattern to insert constant values of type 1...1xxxx.
17455 2013-01-04 Jan Hubicka <jh@suse.cz>
17457 PR tree-optimization/55823
17458 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17461 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17462 Uros Bizjak <ubizjak@gmail.com>
17464 PR rtl-optimization/55845
17465 * df-problems.c (can_move_insns_across): Stop scanning at
17466 volatile_insn_p source instruction or give up if
17467 across_from .. across_to range contains any volatile_insn_p
17470 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17472 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17473 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17475 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17476 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17478 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17481 * asan.c (asan_clear_shadow): New function.
17482 (asan_emit_stack_protection): Use it.
17484 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17486 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17487 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17488 with tab instead of space.
17490 2013-01-08 Nick Clifton <nickc@redhat.com>
17492 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17493 register bank 0 at the start of an interrupt handler.
17494 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17497 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17499 * config/aarch64/aarch64-simd.md
17500 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17501 (aarch64_simd_bsl): Likewise.
17502 (aarch64_vcond_internal<mode>): Likewise.
17503 (vcond<mode><mode>): Likewise.
17504 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17505 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17507 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17509 * config/aarch64/aarch64-builtins.c
17510 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17512 2013-01-08 Martin Jambor <mjambor@suse.cz>
17515 * tree-sra.c (analyze_access_subtree): Return true also after
17516 potentially creating a debug-only replacement.
17518 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17520 PR middle-end/55890
17521 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17523 PR tree-optimization/54120
17524 * tree-vrp.c (range_fits_type_p): Don't allow
17525 src_precision < precision from signed vr to unsigned_p
17526 if vr->min or vr->max is negative.
17527 (simplify_float_conversion_using_ranges): Test can_float_p
17528 against CODE_FOR_nothing.
17530 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17531 Richard Biener <rguenther@suse.de>
17533 PR middle-end/55851
17534 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17535 types instead of just INTEGER_TYPE types.
17537 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
17539 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17542 2013-01-07 Steve Ellcey <sellcey@mips.com>
17545 * config/mips/mips.opt: Change mad to mmad to match documentation.
17547 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17550 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17551 .progmemx.data now.
17553 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17556 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17557 (avr_addrspace_t): Add .section_name field.
17558 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17560 (avr_addrspace): Same. Initialize .section_name. Remove last
17561 NULL entry. Put __memx into .progmemx.data.
17562 (progmem_section_prefix): Remove.
17563 (avr_asm_init_sections): No need to initialize progmem_section.
17564 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17565 section name prefix.
17566 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17567 retrieve the progmem section.
17568 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17569 boundary to run over avr_addrspace[].
17570 (avr_register_target_pragmas): Ditto.
17572 2013-01-06 Jakub Jelinek <jakub@redhat.com>
17574 * varasm.c (output_constant_def_contents): For asan_protect_global
17575 protected strings, adjust DECL_ALIGN if needed, before testing for
17577 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17578 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17580 (output_object_block): For asan protected decls, emit asan padding
17581 after their contents.
17582 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17583 (asan_finish_file): Test it here instead.
17585 2013-01-07 Nick Clifton <nickc@redhat.com>
17586 Matthias Klose <doko@debian.org>
17587 Doug Kwan <dougkwan@google.com>
17588 H.J. Lu <hongjiu.lu@intel.com>
17591 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17593 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17595 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17597 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
17599 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17601 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17604 * doc/install.texi (Cross-Compiler-Specific Options): Document
17607 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17609 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17610 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17611 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17612 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17613 vqmovun_high_s64): Fix source operand number and update copyright.
17615 2013-01-07 Richard Biener <rguenther@suse.de>
17617 PR middle-end/55890
17618 * gimple.h (gimple_call_builtin_p): New overload.
17619 * gimple.c (validate_call): New function.
17620 (gimple_call_builtin_p): Likewise.
17621 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17622 Use gimple_call_builtin_p.
17623 (find_func_clobbers): Likewise.
17624 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17625 (strlen_optimize_stmt): Likewise.
17627 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17629 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17630 (vld1q_dup_*): Likewise.
17631 (vld1_*): Likewise.
17632 (vld1q_*): Likewise.
17633 (vld1_lane_*): Likewise.
17634 (vld1q_lane_*): Likewise.
17636 2013-01-07 Richard Biener <rguenther@suse.de>
17638 * lto-streamer.h (LTO_minor_version): Bump to 2.
17640 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17642 * config/aarch64/aarch64-protos.h
17643 (aarch64_const_double_zero_rtx_p): Rename to...
17644 (aarch64_float_const_zero_rtx_p): ...this.
17645 (aarch64_float_const_representable_p): New.
17646 (aarch64_output_simd_mov_immediate): Likewise.
17647 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17648 move immediate case.
17649 * config/aarch64/aarch64.c
17650 (aarch64_const_double_zero_rtx_p): Rename to...
17651 (aarch64_float_const_zero_rtx_p): ...this.
17652 (aarch64_print_operand): Allow printing of new constants.
17653 (aarch64_valid_floating_const): New.
17654 (aarch64_legitimate_constant_p): Check for valid floating-point
17656 (aarch64_simd_valid_immediate): Likewise.
17657 (aarch64_vect_float_const_representable_p): New.
17658 (aarch64_float_const_representable_p): Likewise.
17659 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17660 (aarch64_output_simd_mov_immediate): New.
17661 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17662 (*movdf_aarch64): Likewise.
17663 * config/aarch64/constraints.md (Ufc): New.
17664 (Y): call aarch64_float_const_zero_rtx.
17665 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17667 2013-01-07 Richard Biener <rguenther@suse.de>
17669 PR tree-optimization/55888
17670 PR tree-optimization/55862
17671 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17672 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17673 not if it is contained therein.
17675 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17677 * config/avr/t-avr: Typo.
17679 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17682 * config/avr/t-avr: Don't automatically rebuild
17683 $(srcdir)/config/avr/t-multilib
17684 $(srcdir)/config/avr/avr-tables.opt
17685 $(srcdir)/doc/avr-mmcu.texi
17686 (avr-mcus): New phony target to build them on request.
17687 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17688 * avr/avr-mcus.def: Adjust comments.
17690 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
17692 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17694 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17696 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17698 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17700 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17702 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
17704 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17705 to generate profiling.
17706 * config/rs6000/aix64.h (LIB_SPEC): Same.
17708 2013-01-04 Andrew Pinski <apinski@cavium.com>
17710 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17712 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17714 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
17716 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17718 (ix86_expand_move): Ditto.
17719 (ix86_zero_extend_to_Pmode): Ditto.
17720 (ix86_expand_call): Ditto.
17721 (ix86_expand_special_args_builtin): Ditto.
17722 (ix86_expand_builtin): Ditto.
17724 2013-01-04 Richard Biener <rguenther@suse.de>
17726 PR tree-optimization/55862
17727 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17728 translating them through PHI nodes.
17730 2013-01-04 Martin Jambor <mjambor@suse.cz>
17732 PR tree-optimization/55755
17733 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17734 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17736 2013-01-04 Marek Polacek <polacek@redhat.com>
17738 PR middle-end/55859
17739 * opts.c (default_options_optimization): Clarify error message.
17741 2013-01-04 Richard Biener <rguenther@suse.de>
17743 PR middle-end/55863
17744 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17747 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17750 * config/pa/pa.md (movsi): Revert previous change.
17751 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17754 2013-01-03 Richard Henderson <rth@redhat.com>
17756 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17757 after eliminating TLS symbols.
17759 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
17762 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17763 * graphite-poly.c (debug_gmp_value): Likewise.
17765 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
17768 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17769 selected code model, define __code_mode_small__, __code_model_medium__,
17770 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17771 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17772 xchg temporary register with %k. Declare temporary register as
17774 [__x86_64__]: For medium and large code models, preserve %rbx register.
17776 2013-01-03 Richard Biener <rguenther@suse.de>
17778 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
17779 (dump_subscript): Adjust.
17780 (finalize_ddr_dependent): Do not dump redundant info.
17781 (analyze_siv_subscript): Adjust.
17782 (subscript_dependence_tester): Likewise.
17783 (compute_affine_dependence): Likewise.
17785 2013-01-03 Richard Biener <rguenther@suse.de>
17788 2013-01-03 Richard Biener <rguenther@suse.de>
17790 PR tree-optimization/55857
17791 * tree-vect-stmts.c (vectorizable_load): Do not setup
17792 re-alignment for invariant loads.
17794 2013-01-02 Richard Biener <rguenther@suse.de>
17796 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17797 invariant load do not generate a vector load from the scalar location.
17799 2013-01-03 Richard Biener <rguenther@suse.de>
17801 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17802 for not vectorizing.
17803 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17804 not build INDIRECT_REFs, call get_name once only.
17805 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17806 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17808 2013-01-03 Richard Biener <rguenther@suse.de>
17810 PR tree-optimization/55857
17811 * tree-vect-stmts.c (vectorizable_load): Do not setup
17812 re-alignment for invariant loads.
17814 2013-01-03 Richard Biener <rguenther@suse.de>
17817 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17818 prefer a built-in decl.
17820 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17822 * gcc.c (process_command): Update copyright notice dates.
17823 * gcov.c (print_version): Likewise.
17824 * gcov-dump.c (print_version): Likewise.
17826 PR rtl-optimization/55838
17827 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17828 iv0.step, iv1.step and step.
17830 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17831 Marc Glisse <marc.glisse@inria.fr>
17833 PR tree-optimization/55832
17834 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17835 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17836 integer_{one,zero}_node.
17838 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17841 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17842 * var-tracking.c (reverse_op): Don't add reverse ops to
17843 VALUEs that have already
17844 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
17846 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17848 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17850 2013-01-02 Teresa Johnson <tejohnson@google.com>
17852 * dumpfile.c (dump_loc): Print filename with location.
17853 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17854 new location_t parameter to emit complete unroll message with
17855 new dump framework.
17856 (canonicalize_loop_induction_variables): Compute loops location
17857 and pass to try_unroll_loop_completely.
17858 * loop-unroll.c (report_unroll_peel): New function.
17859 (peel_loops_completely): Use new dump format with location
17860 for main dumpfile message, and invoke report_unroll_peel on success.
17861 (decide_unrolling_and_peeling): Ditto.
17862 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17863 by report_unroll_peel.
17864 (decide_peel_completely): Ditto.
17865 (decide_unroll_constant_iterations): Ditto.
17866 (decide_unroll_runtime_iterations): Ditto.
17867 (decide_peel_simple): Ditto.
17868 (decide_unroll_stupid): Ditto.
17869 * cfgloop.c (get_loop_location): New function.
17870 * cfgloop.h (get_loop_location): Declare.
17872 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17874 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17877 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17879 PR middle-end/55198
17880 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17881 BLKmode objects when EXPAND_MEMORY is specified.
17883 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17885 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17887 (fold_builtin_cpu): Do not share cpu model decls across statements.
17889 2013-01-02 Jason Merrill <jason@redhat.com>
17892 * tree.c (build_array_type_1): Revert earlier change.
17894 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17896 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17898 * config/aarch64/aarch64-tune.md: Re-generate.
17900 2013-01-02 Richard Biener <rguenther@suse.de>
17902 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17903 invariant load do not generate a vector load from the scalar location.
17905 2013-01-02 Richard Biener <rguenther@suse.de>
17908 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17909 * configure: Regenerate.
17911 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17913 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17914 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17915 (expand_builtin_int_roundingfn_2): Keep the original target around
17916 for the fallback case.
17918 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17920 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17921 to be clear for sign changes.
17923 2013-01-01 Jan Hubicka <jh@suse.cz>
17925 * ipa-inline-analysis.c: Fix formatting.
17927 2013-01-01 Jakub Jelinek <jakub@redhat.com>
17929 PR tree-optimization/55831
17930 * tree-vect-loop.c (get_initial_def_for_induction): Use
17931 gsi_after_labels instead of gsi_start_bb.
17933 Copyright (C) 2013 Free Software Foundation, Inc.
17935 Copying and distribution of this file, with or without modification,
17936 are permitted in any medium without royalty provided the copyright
17937 notice and this notice are preserved.