1 2013-08-10 Jan Hubicka <jh@suse.cz>
3 Workaround binutils PR14342
4 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5 (gimple_init_edge_profiler): Likewise.
6 (gimple_gen_ic_func_profiler): Likewise.
8 2013-08-09 Jan Hubicka <jh@suse.cz>
10 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12 2013-08-09 Xinliang David Li <davidxl@google.com>
14 * config/i386/stringop.def: New file.
15 * config/i386/stringop.opt: New file.
16 * config/i386/i386-opts.h: Include stringopt.def.
17 * config/i386/i386.opt: Include stringopt.opt.
18 * config/i386/i386.c (ix86_option_override_internal):
19 Override default size based stringop inline strategies
21 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
24 2013-08-09 Jan Hubicka <jh@suse.cz>
26 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
28 2013-08-09 Jan Hubicka <jh@suse.cz>
30 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
32 (dump_cgraph_node): Dump profile-id.
33 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
34 and common_target_probability.
35 * lto-cgraph.c (lto_output_edge): Stream common targets.
36 (lto_output_node): Stream profile ids.
37 (input_node): Stream profile ids.
38 (input_edge): Stream common targets.
39 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
40 * ipa.c: Include value-prof.h
41 (ipa_profile_generate_summary): Turn indirect call statement histograms
43 (ipa_profile): Turn common targets into speculative edges.
45 2013-08-09 Jan Hubicka <jh@suse.cz>
47 * cgraph.h (cgraph_node): Add profile_id.
48 * value-prof.c (cgraph_node_map): Turn into pointer_map.
49 (init_node_map): Rewrite to handle hashes increas of incremental
51 (del_node_map): Update.
52 (find_func_by_funcdef_no): Replace by ...
53 (find_func_by_profile_id): ... this one.
54 (gimple_ic_transform): Do not remove useful histograms when
55 speculation is not done; dump info when indirect call removal
57 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
58 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
59 (__gcov_indirect_call_profiler_v2): .. this one.
60 * profile.h (init_node_map): Update.
61 * coverage.c (coverage_compute_profile_id): New function.
62 * coverage.h (coverage_compute_profile_id): Declare.
63 * tree-profile.c (init_ic_make_global_vars): Make
64 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
65 (gimple_init_edge_profiler): Update prototype of
66 __gcov_indirect_call_profiler.
67 (gimple_gen_ic_func_profiler): Simplify.
68 (tree_profiling): Use init_node_map
70 2013-08-09 Jan Hubicka <jh@suse.cz>
72 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only non-speculative
74 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
75 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
76 (cgraph_set_call_stmt): Likewise.
77 (cgraph_create_edge_1): Fix release checking compilatoin;
79 (cgraph_free_edge): Free indirect info.
80 (cgraph_turn_edge_to_speculative): New function.
81 (cgraph_speculative_call_info): New function.
82 (cgraph_make_edge_direct): Return direct edge; handle speculation.
83 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative
85 (dump_cgraph_node): Dump speculation.
86 (verify_edge_count_and_frequency): Accept speculative edges.
87 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
88 (verify_cgraph_node): Handle speculation.
89 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
90 (cgraph_set_call_stmt): Update prototype.
91 (cgraph_make_edge_direct): Update prototype.
92 (cgraph_speculative_call_info): Declare.
93 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
94 to change; update call of ipa_find_references.
95 * ipa-ref.c (ipa_record_reference): Fix return value; clear
96 lto_stmt_uid and speculative flags.
97 (ipa_dump_references): Dump speculation.
98 (ipa_clone_references): Clone speculative flag.
99 (ipa_clone_referring): Likewise.
100 (ipa_clone_ref): New function.
101 (ipa_find_reference): Look into lto_stmt_uids
102 (ipa_clear_stmts_in_references): Do not clear speculative calls.
103 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
104 (ipa_find_reference): Update declaration.
105 (ipa_clone_ref): Declare.
106 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
107 stream speculative flag.
108 (lto_output_ref): Stream statements uids and speculation.
109 (input_ref): Likewise.
110 (input_edge): Stream speuclation.
111 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
112 (cgraph_set_call_stmt_including_clones): Handle speculation.
113 * ipa-inline.c (heap_edge_removal_hook): New function.
114 (inline_small_functions): Register it.
115 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
116 also initialize refs.
117 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
119 (try_make_edge_direct_simple_call): Likewise.
120 (try_make_edge_direct_simple_call): Likewise.
121 (update_indirect_edges_after_inlining): Likewise.
122 (remove_described_reference): Look proper lto_stmt_uid.
123 (propagate_controlled_uses): Likewise.
124 (propagate_controlled_uses): Liekwise.
125 * tree-inline.c (copy_bb): Copy speculative edges.
126 (redirect_all_calls): New function.
127 (copy_cfg_body): Do redirection after loop info
129 (delete_unreachable_blocks_update_callgraph): Updadte
132 2013-08-09 Jan Hubicka <jh@suse.cz>
134 * lto-streamer-out.c (output_function): Renumber PHIs.
135 * lto-streamer-in.c (input_function): Likewise.
137 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
139 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
140 (get_lane_unsigned): Likewise.
141 (dup_lane_scalar): Likewise.
142 (get_lane): enable for VALL.
143 * config/aarch64/aarch64-simd.md
144 (aarch64_dup_lane_scalar<mode>): Remove.
145 (aarch64_get_lane_signed<mode>): Likewise.
146 (aarch64_get_lane_unsigned<mode>): Likewise.
147 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
148 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
149 (aarch64_get_lane<mode>): Enable for all vector modes.
150 (aarch64_get_lanedi): Remove misleading constraints.
151 * config/aarch64/arm_neon.h
152 (__aarch64_vget_lane_any): Define.
153 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
154 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
155 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
156 * config/aarch64/iterators.md (VDQQH): New.
160 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
162 * configure.ac: Add GAS check for LEON instructions on SPARC.
163 * configure: Regenerate.
164 * config.in: Likewise.
165 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
167 * config/sparc/sparc.opt (LEON, LEON3): New masks.
168 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
170 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
171 (AS_LEON_FLAG): New macro.
172 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
173 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
174 Deal with LEON and LEON3 for the memory model.
175 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
176 (atomic_compare_and_swap<mode>_1): Likewise.
177 (*atomic_compare_and_swap<mode>_1): Likewise.
179 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
181 * config/arm/neon.md (vcond): Fix floating-point vector
182 comparisons against 0.
184 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
186 * lra-constraints.c (emit_spill_move): Remove assert.
187 (process_alt_operands): Add more debugging
188 output. Increase reject for spilling into memory. Decrease
189 reject for reloading scratch.
190 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
192 2013-08-08 Steve Ellcey <sellcey@mips.com>
194 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
195 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
196 micromips incompatible. Add nan2008.
197 (MULTILIB_DIRNAMES): Add nan2008.
198 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
199 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
200 and micromips incompatible. Add nan2008.
201 (MULTILIB_DIRNAMES): Add nan2008.
202 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
204 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
206 PR rtl-optimization/58079
207 * combine.c (combine_simplify_rtx): Avoid using SUBST if
208 simplify_comparison has widened a comparison with an integer.
210 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
212 * config/arm/neon.md (movmisalign<mode>): Disable when we
213 don't allow unaligned accesses.
214 (*movmisalign<mode>_neon_store): Likewise.
215 (*movmisalign<mode>_neon_load): Likewise.
216 (*movmisalign<mode>_neon_store): Likewise.
217 (*movmisalign<mode>_neon_load): Likewise.
219 2013-08-08 Jan Hubicka <jh@suse.cz>
221 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
222 (make_pass_rebuild_cgraph_edges): Also clear references.
223 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
224 * ipa-inline-transform.c (inline_transform): Remove all references
226 * cgraphunit.c (expand_function): Remove all references after expansion.
227 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
228 (ipa_find_reference): Rewrite to iterator.
229 (remove_stmt_references): Likewise.
230 (ipa_clear_stmts_in_references): New function.
231 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
232 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
234 * ipa-split.c (split_function): Remove references in split function.
236 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
239 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
240 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
242 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
245 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
246 suggested mode for the assignment isn't compatible with the
249 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
252 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
254 2013-08-07 Xinliang David Li <davidxl@google.com>
256 * config/i386/i386.opt: New option -mtune-ctrl=.
257 * config/i386/x86-tune.def: New file.
258 * config/i386/i386.h: include x86-tune.def.
259 * config/i386/i386.c (ix86_option_override_internal):
260 Parsing -mtune-ctrl= option and set tune features.
262 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
265 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
266 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
269 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
271 * rtl.h (update_alignments): Declare.
272 * final.c (grow_label_align): New function extracted from...
273 (shorten_branches): ...here. Call it.
274 (update_alignments): New function.
275 * reorg.c (sibling_labels): New variable.
276 (get_label_before): Add SIBLING parameter. If it is non-zero, push
277 the new label along with it onto the sibling_labels vector.
278 (fill_simple_delay_slots): Adjust call to get_label_before.
279 (fill_slots_from_thread): Likewise.
280 (relax_delay_slots): Likewise.
281 (make_return_insns): Likewise.
282 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
284 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
286 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
287 document its semantics.
288 (diagnostic_report_diagnostic): Adjust accordingly.
290 2013-08-07 David Malcolm <dmalcolm@redhat.com>
292 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
293 (sparc_option_override): ...and port to new C++ pass API.
294 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
296 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
298 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
300 2013-08-06 Caroline Tice <cmtice@google.com>
302 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
303 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
304 * tree-pass.h: Add pass_vtable_verify.
305 * varasm.c (assemble_variable): Add code to properly set the comdat
306 section and name for the .vtable_map_vars section.
307 (assemble_vtyv_preinit_initializer): New function.
308 (default_sectin_type_flags): Make sure .vtable_map_vars section has
310 * output.h: Add function decl for assemble_vtv_preinit_initializer.
311 * vtable-verify.c: New file.
312 * vtable-verify.h: New file.
313 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
314 initialiation levels.
315 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
316 * passes.def: Insert pass_vtable_verify.
317 * aclocal.m4: Reorder includes.
318 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
319 -fvtv-counts options.
320 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
321 as appropriate, if -fvtable-verify=... is used.
322 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
323 -fvtable-verify=... is used.
324 * Makefile.in (OBJS): Add vtable-verify.o to list.
325 (vtable-verify.o): Add new build rule.
326 (GTFILES): Add vtable-verify.c to list.
327 * common.opt (fvtable-verify=): New flag.
328 (vtv_priority): Values for fvtable-verify= flag.
329 (fvtv-counts): New flag.
330 (fvtv-debug): New flag.
331 * tree.h (save_vtable_map_decl): New extern function decl.
333 2013-08-07 David Malcolm <dmalcolm@redhat.com>
335 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
336 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
337 (pass_data_rl78_devirt): ...new pass_data instance and...
338 (make_pass_rl78_devirt): ...new function.
339 (rl78_asm_file_start): Port pass registration to new C++ API.
341 2013-08-07 David Malcolm <dmalcolm@redhat.com>
343 * coretypes.h (rtl_opt_pass): Add.
345 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
346 (epiphany_init): Port to new C++ pass API.
347 (epiphany_optimize_mode_switching): Likewise.
348 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
349 (pass_manager::get_pass_mode_switching): New.
350 (pass_manager::get_pass_peephole2): New.
351 * mode-switching.c (pass_mode_switching): Add clone method.
352 * recog.c (pass_peephole2): Add clone method.
353 (pass_split_all_insns): Add clone method.
355 2013-08-06 David Malcolm <dmalcolm@redhat.com>
357 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
358 (mips_option_override): ...here, porting to new C++ API for passes.
360 2013-08-06 Jan Hubicka <jh@suse.cz>
362 * cgraph.c (cgraph_get_body): New function based on lto.c
364 * cgraph.h (cgraph_get_body): Declare.
365 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
367 * cgraphunit.c (expand_function): Get body prior expanding.
368 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
369 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
371 * passes.c (do_per_function_toporder): Get body.
372 * tree-inline.c (expand_call_inline): Get body prior inlining it.
373 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
375 2013-08-06 Martin Jambor <mjambor@suse.cz>
378 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
379 is not re-finalized. Rename second parameter to no_collect.
381 2013-08-06 Martin Jambor <mjambor@suse.cz>
384 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
385 MEM_REF has proper alignment information.
387 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
390 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
392 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
393 rtx (*) (rtx, ...) with insn_gen_fn.
394 * genoutput.c (output_insn_data): Cast gen_? function pointers to
395 insn_gen_fn::stored_funcptr. Add initializer braces.
397 2013-08-05 David Malcolm <dmalcolm@redhat.com>
399 Rewrite how instances of passes are cloned to remove assumptions
400 about their sizes (thus allowing pass subclasses to have
401 additional data fields, albeit non-GC-managed ones at this point).
403 * passes.c (make_pass_instance): Now that passes have clone
404 methods, rewrite this function to eliminate XNEW and memcpy
405 calls that used hardcoded sizes. Since this function no longer
406 creates pass instances, rename it to...
407 (add_pass_instance): ...this. Document the old way that passes were
408 numbered and flagged, and rework this function to continue using it.
409 (next_pass_1): Add an initial_pass argument for use by
411 (position_pass): When adding multiple instances of a pass, use
412 the pass's clone method, rather than relying on the XNEW/memcpy
413 within the former make_pass_instance (now add_pass_instance).
414 (pass_manager::pass_manager): When invoking next_pass_1, also supply
415 the initial instance of the current pass within the pass manager.
417 2013-08-05 David Malcolm <dmalcolm@redhat.com>
419 This is the automated part of the conversion of passes from C
420 structs to C++ classes.
422 Patch autogenerated by refactor_passes.py from
423 https://github.com/davidmalcolm/gcc-refactoring-scripts
424 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
426 * asan.c (pass_asan): Convert from a global struct to a subclass of
427 gimple_opt_pass along with...
428 (pass_data_asan): ...new pass_data instance and...
429 (make_pass_asan): ...new function.
430 (pass_asan_O0): Convert from a global struct to a subclass of
431 gimple_opt_pass along with...
432 (pass_data_asan_O0): ...new pass_data instance and...
433 (make_pass_asan_O0): ...new function.
434 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
435 subclass of rtl_opt_pass along with...
436 (pass_data_inc_dec): ...new pass_data instance and...
437 (make_pass_inc_dec): ...new function.
438 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
439 a subclass of rtl_opt_pass along with...
440 (pass_data_reorder_blocks): ...new pass_data instance and...
441 (make_pass_reorder_blocks): ...new function.
442 (pass_duplicate_computed_gotos): Convert from a global struct to a
443 subclass of rtl_opt_pass along with...
444 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
445 (make_pass_duplicate_computed_gotos): ...new function.
446 (pass_partition_blocks): Convert from a global struct to a subclass of
447 rtl_opt_pass along with...
448 (pass_data_partition_blocks): ...new pass_data instance and...
449 (make_pass_partition_blocks): ...new function.
450 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
451 struct to a subclass of rtl_opt_pass along with...
452 (pass_data_branch_target_load_optimize1): ...new pass_data instance
454 (make_pass_branch_target_load_optimize1): ...new function.
455 (pass_branch_target_load_optimize2): Convert from a global struct to a
456 subclass of rtl_opt_pass along with...
457 (pass_data_branch_target_load_optimize2): ...new pass_data instance
459 (make_pass_branch_target_load_optimize2): ...new function.
460 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
461 of rtl_opt_pass along with...
462 (pass_data_jump): ...new pass_data instance and...
463 (make_pass_jump): ...new function.
464 (pass_jump2): Convert from a global struct to a subclass of
465 rtl_opt_pass along with...
466 (pass_data_jump2): ...new pass_data instance and...
467 (make_pass_jump2): ...new function.
468 * cfgexpand.c (pass_expand): Convert from a global struct to a
469 subclass of rtl_opt_pass along with...
470 (pass_data_expand): ...new pass_data instance and...
471 (make_pass_expand): ...new function.
472 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
473 of rtl_opt_pass along with...
474 (pass_data_free_cfg): ...new pass_data instance and...
475 (make_pass_free_cfg): ...new function.
476 (pass_into_cfg_layout_mode): Convert from a global struct to a
477 subclass of rtl_opt_pass along with...
478 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
479 (make_pass_into_cfg_layout_mode): ...new function.
480 (pass_outof_cfg_layout_mode): Convert from a global struct to a
481 subclass of rtl_opt_pass along with...
482 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
483 (make_pass_outof_cfg_layout_mode): ...new function.
484 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
485 struct to a subclass of gimple_opt_pass along with...
486 (pass_data_build_cgraph_edges): ...new pass_data instance and...
487 (make_pass_build_cgraph_edges): ...new function.
488 (pass_rebuild_cgraph_edges): Convert from a global struct to a
489 subclass of gimple_opt_pass along with...
490 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
491 (make_pass_rebuild_cgraph_edges): ...new function.
492 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
493 subclass of gimple_opt_pass along with...
494 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
496 (make_pass_remove_cgraph_callee_edges): ...new function.
497 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
498 struct to a subclass of rtl_opt_pass along with...
499 (pass_data_stack_adjustments): ...new pass_data instance and...
500 (make_pass_stack_adjustments): ...new function.
501 * combine.c (pass_combine): Convert from a global struct to a subclass
502 of rtl_opt_pass along with...
503 (pass_data_combine): ...new pass_data instance and...
504 (make_pass_combine): ...new function.
505 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
506 global struct to a subclass of rtl_opt_pass along with...
507 (pass_data_compare_elim_after_reload): ...new pass_data instance
509 (make_pass_compare_elim_after_reload): ...new function.
510 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
511 of rtl_opt_pass along with...
512 (pass_data_rtl_cprop): ...new pass_data instance and...
513 (make_pass_rtl_cprop): ...new function.
514 * cse.c (pass_cse): Convert from a global struct to a subclass of
515 rtl_opt_pass along with...
516 (pass_data_cse): ...new pass_data instance and...
517 (make_pass_cse): ...new function.
518 (pass_cse2): Convert from a global struct to a subclass of
519 rtl_opt_pass along with...
520 (pass_data_cse2): ...new pass_data instance and...
521 (make_pass_cse2): ...new function.
522 (pass_cse_after_global_opts): Convert from a global struct to a
523 subclass of rtl_opt_pass along with...
524 (pass_data_cse_after_global_opts): ...new pass_data instance and...
525 (make_pass_cse_after_global_opts): ...new function.
526 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
527 of rtl_opt_pass along with...
528 (pass_data_ud_rtl_dce): ...new pass_data instance and...
529 (make_pass_ud_rtl_dce): ...new function.
530 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
531 rtl_opt_pass along with...
532 (pass_data_fast_rtl_dce): ...new pass_data instance and...
533 (make_pass_fast_rtl_dce): ...new function.
534 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
535 a subclass of rtl_opt_pass along with...
536 (pass_data_df_initialize_opt): ...new pass_data instance and...
537 (make_pass_df_initialize_opt): ...new function.
538 (pass_df_initialize_no_opt): Convert from a global struct to a
539 subclass of rtl_opt_pass along with...
540 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
541 (make_pass_df_initialize_no_opt): ...new function.
542 (pass_df_finish): Convert from a global struct to a subclass of
543 rtl_opt_pass along with...
544 (pass_data_df_finish): ...new pass_data instance and...
545 (make_pass_df_finish): ...new function.
546 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
547 rtl_opt_pass along with...
548 (pass_data_rtl_dse1): ...new pass_data instance and...
549 (make_pass_rtl_dse1): ...new function.
550 (pass_rtl_dse2): Convert from a global struct to a subclass of
551 rtl_opt_pass along with...
552 (pass_data_rtl_dse2): ...new pass_data instance and...
553 (make_pass_rtl_dse2): ...new function.
554 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
555 subclass of rtl_opt_pass along with...
556 (pass_data_dwarf2_frame): ...new pass_data instance and...
557 (make_pass_dwarf2_frame): ...new function.
558 * except.c (pass_set_nothrow_function_flags): Convert from a global
559 struct to a subclass of rtl_opt_pass along with...
560 (pass_data_set_nothrow_function_flags): ...new pass_data instance
562 (make_pass_set_nothrow_function_flags): ...new function.
563 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
564 subclass of rtl_opt_pass along with...
565 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
567 (make_pass_convert_to_eh_region_ranges): ...new function.
568 * final.c (pass_compute_alignments): Convert from a global struct to a
569 subclass of rtl_opt_pass along with...
570 (pass_data_compute_alignments): ...new pass_data instance and...
571 (make_pass_compute_alignments): ...new function.
572 (pass_final): Convert from a global struct to a subclass of
573 rtl_opt_pass along with...
574 (pass_data_final): ...new pass_data instance and...
575 (make_pass_final): ...new function.
576 (pass_shorten_branches): Convert from a global struct to a subclass of
577 rtl_opt_pass along with...
578 (pass_data_shorten_branches): ...new pass_data instance and...
579 (make_pass_shorten_branches): ...new function.
580 (pass_clean_state): Convert from a global struct to a subclass of
581 rtl_opt_pass along with...
582 (pass_data_clean_state): ...new pass_data instance and...
583 (make_pass_clean_state): ...new function.
584 * function.c (pass_instantiate_virtual_regs): Convert from a global
585 struct to a subclass of rtl_opt_pass along with...
586 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
587 (make_pass_instantiate_virtual_regs): ...new function.
588 (pass_leaf_regs): Convert from a global struct to a subclass of
589 rtl_opt_pass along with...
590 (pass_data_leaf_regs): ...new pass_data instance and...
591 (make_pass_leaf_regs): ...new function.
592 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
593 subclass of rtl_opt_pass along with...
594 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
596 (make_pass_thread_prologue_and_epilogue): ...new function.
597 (pass_match_asm_constraints): Convert from a global struct to a
598 subclass of rtl_opt_pass along with...
599 (pass_data_match_asm_constraints): ...new pass_data instance and...
600 (make_pass_match_asm_constraints): ...new function.
601 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
602 subclass of rtl_opt_pass along with...
603 (pass_data_rtl_fwprop): ...new pass_data instance and...
604 (make_pass_rtl_fwprop): ...new function.
605 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
606 rtl_opt_pass along with...
607 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
608 (make_pass_rtl_fwprop_addr): ...new function.
609 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
610 rtl_opt_pass along with...
611 (pass_data_rtl_pre): ...new pass_data instance and...
612 (make_pass_rtl_pre): ...new function.
613 (pass_rtl_hoist): Convert from a global struct to a subclass of
614 rtl_opt_pass along with...
615 (pass_data_rtl_hoist): ...new pass_data instance and...
616 (make_pass_rtl_hoist): ...new function.
617 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
618 subclass of gimple_opt_pass along with...
619 (pass_data_lower_cf): ...new pass_data instance and...
620 (make_pass_lower_cf): ...new function.
621 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
622 from a global struct to a subclass of gimple_opt_pass along with...
623 (pass_data_strength_reduction): ...new pass_data instance and...
624 (make_pass_strength_reduction): ...new function.
625 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
626 of rtl_opt_pass along with...
627 (pass_data_rtl_ifcvt): ...new pass_data instance and...
628 (make_pass_rtl_ifcvt): ...new function.
629 (pass_if_after_combine): Convert from a global struct to a subclass of
630 rtl_opt_pass along with...
631 (pass_data_if_after_combine): ...new pass_data instance and...
632 (make_pass_if_after_combine): ...new function.
633 (pass_if_after_reload): Convert from a global struct to a subclass of
634 rtl_opt_pass along with...
635 (pass_data_if_after_reload): ...new pass_data instance and...
636 (make_pass_if_after_reload): ...new function.
637 * init-regs.c (pass_initialize_regs): Convert from a global struct to
638 a subclass of rtl_opt_pass along with...
639 (pass_data_initialize_regs): ...new pass_data instance and...
640 (make_pass_initialize_regs): ...new function.
641 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
642 of ipa_opt_pass_d along with...
643 (pass_data_ipa_cp): ...new pass_data instance and...
644 (make_pass_ipa_cp): ...new function.
645 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
646 global struct to a subclass of gimple_opt_pass along with...
647 (pass_data_inline_parameters): ...new pass_data instance and...
648 (make_pass_inline_parameters): ...new function.
649 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
650 subclass of gimple_opt_pass along with...
651 (pass_data_early_inline): ...new pass_data instance and...
652 (make_pass_early_inline): ...new function.
653 (pass_ipa_inline): Convert from a global struct to a subclass of
654 ipa_opt_pass_d along with...
655 (pass_data_ipa_inline): ...new pass_data instance and...
656 (make_pass_ipa_inline): ...new function.
657 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
658 struct to a subclass of gimple_opt_pass along with...
659 (pass_data_local_pure_const): ...new pass_data instance and...
660 (make_pass_local_pure_const): ...new function.
661 (pass_ipa_pure_const): Convert from a global struct to a subclass of
662 ipa_opt_pass_d along with...
663 (pass_data_ipa_pure_const): ...new pass_data instance and...
664 (make_pass_ipa_pure_const): ...new function.
665 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
666 to a subclass of ipa_opt_pass_d along with...
667 (pass_data_ipa_reference): ...new pass_data instance and...
668 (make_pass_ipa_reference): ...new function.
669 * ipa-split.c (pass_split_functions): Convert from a global struct to
670 a subclass of gimple_opt_pass along with...
671 (pass_data_split_functions): ...new pass_data instance and...
672 (make_pass_split_functions): ...new function.
673 (pass_feedback_split_functions): Convert from a global struct to a
674 subclass of gimple_opt_pass along with...
675 (pass_data_feedback_split_functions): ...new pass_data instance and...
676 (make_pass_feedback_split_functions): ...new function.
677 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
678 global struct to a subclass of simple_ipa_opt_pass along with...
679 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
681 (make_pass_ipa_function_and_variable_visibility): ...new function.
682 (pass_ipa_free_inline_summary): Convert from a global struct to a
683 subclass of simple_ipa_opt_pass along with...
684 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
685 (make_pass_ipa_free_inline_summary): ...new function.
686 (pass_ipa_whole_program_visibility): Convert from a global struct to a
687 subclass of ipa_opt_pass_d along with...
688 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
690 (make_pass_ipa_whole_program_visibility): ...new function.
691 (pass_ipa_profile): Convert from a global struct to a subclass of
692 ipa_opt_pass_d along with...
693 (pass_data_ipa_profile): ...new pass_data instance and...
694 (make_pass_ipa_profile): ...new function.
695 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
696 ipa_opt_pass_d along with...
697 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
698 (make_pass_ipa_cdtor_merge): ...new function.
699 * ira.c (pass_ira): Convert from a global struct to a subclass of
700 rtl_opt_pass along with...
701 (pass_data_ira): ...new pass_data instance and...
702 (make_pass_ira): ...new function.
703 (pass_reload): Convert from a global struct to a subclass of
704 rtl_opt_pass along with...
705 (pass_data_reload): ...new pass_data instance and...
706 (make_pass_reload): ...new function.
707 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
708 subclass of rtl_opt_pass along with...
709 (pass_data_cleanup_barriers): ...new pass_data instance and...
710 (make_pass_cleanup_barriers): ...new function.
711 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
712 of rtl_opt_pass along with...
713 (pass_data_loop2): ...new pass_data instance and...
714 (make_pass_loop2): ...new function.
715 (pass_rtl_loop_init): Convert from a global struct to a subclass of
716 rtl_opt_pass along with...
717 (pass_data_rtl_loop_init): ...new pass_data instance and...
718 (make_pass_rtl_loop_init): ...new function.
719 (pass_rtl_loop_done): Convert from a global struct to a subclass of
720 rtl_opt_pass along with...
721 (pass_data_rtl_loop_done): ...new pass_data instance and...
722 (make_pass_rtl_loop_done): ...new function.
723 (pass_rtl_move_loop_invariants): Convert from a global struct to a
724 subclass of rtl_opt_pass along with...
725 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
726 (make_pass_rtl_move_loop_invariants): ...new function.
727 (pass_rtl_unswitch): Convert from a global struct to a subclass of
728 rtl_opt_pass along with...
729 (pass_data_rtl_unswitch): ...new pass_data instance and...
730 (make_pass_rtl_unswitch): ...new function.
731 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
732 subclass of rtl_opt_pass along with...
733 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
735 (make_pass_rtl_unroll_and_peel_loops): ...new function.
736 (pass_rtl_doloop): Convert from a global struct to a subclass of
737 rtl_opt_pass along with...
738 (pass_data_rtl_doloop): ...new pass_data instance and...
739 (make_pass_rtl_doloop): ...new function.
740 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
741 a subclass of rtl_opt_pass along with...
742 (pass_data_lower_subreg): ...new pass_data instance and...
743 (make_pass_lower_subreg): ...new function.
744 (pass_lower_subreg2): Convert from a global struct to a subclass of
745 rtl_opt_pass along with...
746 (pass_data_lower_subreg2): ...new pass_data instance and...
747 (make_pass_lower_subreg2): ...new function.
748 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
749 struct to a subclass of ipa_opt_pass_d along with...
750 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
751 (make_pass_ipa_lto_gimple_out): ...new function.
752 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
753 of ipa_opt_pass_d along with...
754 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
755 (make_pass_ipa_lto_finish_out): ...new function.
756 * mode-switching.c (pass_mode_switching): Convert from a global struct
757 to a subclass of rtl_opt_pass along with...
758 (pass_data_mode_switching): ...new pass_data instance and...
759 (make_pass_mode_switching): ...new function.
760 * modulo-sched.c (pass_sms): Convert from a global struct to a
761 subclass of rtl_opt_pass along with...
762 (pass_data_sms): ...new pass_data instance and...
763 (make_pass_sms): ...new function.
764 * omp-low.c (pass_expand_omp): Convert from a global struct to a
765 subclass of gimple_opt_pass along with...
766 (pass_data_expand_omp): ...new pass_data instance and...
767 (make_pass_expand_omp): ...new function.
768 (pass_lower_omp): Convert from a global struct to a subclass of
769 gimple_opt_pass along with...
770 (pass_data_lower_omp): ...new pass_data instance and...
771 (make_pass_lower_omp): ...new function.
772 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
773 of gimple_opt_pass along with...
774 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
775 (make_pass_diagnose_omp_blocks): ...new function.
776 * passes.c (pass_early_local_passes): Convert from a global struct to
777 a subclass of simple_ipa_opt_pass along with...
778 (pass_data_early_local_passes): ...new pass_data instance and...
779 (make_pass_early_local_passes): ...new function.
780 (pass_all_early_optimizations): Convert from a global struct to a
781 subclass of gimple_opt_pass along with...
782 (pass_data_all_early_optimizations): ...new pass_data instance and...
783 (make_pass_all_early_optimizations): ...new function.
784 (pass_all_optimizations): Convert from a global struct to a subclass
785 of gimple_opt_pass along with...
786 (pass_data_all_optimizations): ...new pass_data instance and...
787 (make_pass_all_optimizations): ...new function.
788 (pass_all_optimizations_g): Convert from a global struct to a subclass
789 of gimple_opt_pass along with...
790 (pass_data_all_optimizations_g): ...new pass_data instance and...
791 (make_pass_all_optimizations_g): ...new function.
792 (pass_rest_of_compilation): Convert from a global struct to a subclass
793 of rtl_opt_pass along with...
794 (pass_data_rest_of_compilation): ...new pass_data instance and...
795 (make_pass_rest_of_compilation): ...new function.
796 (pass_postreload): Convert from a global struct to a subclass of
797 rtl_opt_pass along with...
798 (pass_data_postreload): ...new pass_data instance and...
799 (make_pass_postreload): ...new function.
800 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
801 subclass of rtl_opt_pass along with...
802 (pass_data_gcse2): ...new pass_data instance and...
803 (make_pass_gcse2): ...new function.
804 * postreload.c (pass_postreload_cse): Convert from a global struct to
805 a subclass of rtl_opt_pass along with...
806 (pass_data_postreload_cse): ...new pass_data instance and...
807 (make_pass_postreload_cse): ...new function.
808 * predict.c (pass_profile): Convert from a global struct to a subclass
809 of gimple_opt_pass along with...
810 (pass_data_profile): ...new pass_data instance and...
811 (make_pass_profile): ...new function.
812 (pass_strip_predict_hints): Convert from a global struct to a subclass
813 of gimple_opt_pass along with...
814 (pass_data_strip_predict_hints): ...new pass_data instance and...
815 (make_pass_strip_predict_hints): ...new function.
816 * recog.c (pass_peephole2): Convert from a global struct to a subclass
817 of rtl_opt_pass along with...
818 (pass_data_peephole2): ...new pass_data instance and...
819 (make_pass_peephole2): ...new function.
820 (pass_split_all_insns): Convert from a global struct to a subclass of
821 rtl_opt_pass along with...
822 (pass_data_split_all_insns): ...new pass_data instance and...
823 (make_pass_split_all_insns): ...new function.
824 (pass_split_after_reload): Convert from a global struct to a subclass
825 of rtl_opt_pass along with...
826 (pass_data_split_after_reload): ...new pass_data instance and...
827 (make_pass_split_after_reload): ...new function.
828 (pass_split_before_regstack): Convert from a global struct to a
829 subclass of rtl_opt_pass along with...
830 (pass_data_split_before_regstack): ...new pass_data instance and...
831 (make_pass_split_before_regstack): ...new function.
832 (pass_split_before_sched2): Convert from a global struct to a subclass
833 of rtl_opt_pass along with...
834 (pass_data_split_before_sched2): ...new pass_data instance and...
835 (make_pass_split_before_sched2): ...new function.
836 (pass_split_for_shorten_branches): Convert from a global struct to a
837 subclass of rtl_opt_pass along with...
838 (pass_data_split_for_shorten_branches): ...new pass_data instance
840 (make_pass_split_for_shorten_branches): ...new function.
841 * ree.c (pass_ree): Convert from a global struct to a subclass of
842 rtl_opt_pass along with...
843 (pass_data_ree): ...new pass_data instance and...
844 (make_pass_ree): ...new function.
845 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
846 subclass of rtl_opt_pass along with...
847 (pass_data_stack_regs): ...new pass_data instance and...
848 (make_pass_stack_regs): ...new function.
849 (pass_stack_regs_run): Convert from a global struct to a subclass of
850 rtl_opt_pass along with...
851 (pass_data_stack_regs_run): ...new pass_data instance and...
852 (make_pass_stack_regs_run): ...new function.
853 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
854 subclass of rtl_opt_pass along with...
855 (pass_data_cprop_hardreg): ...new pass_data instance and...
856 (make_pass_cprop_hardreg): ...new function.
857 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
858 subclass of rtl_opt_pass along with...
859 (pass_data_reginfo_init): ...new pass_data instance and...
860 (make_pass_reginfo_init): ...new function.
861 * regmove.c (pass_regmove): Convert from a global struct to a subclass
862 of rtl_opt_pass along with...
863 (pass_data_regmove): ...new pass_data instance and...
864 (make_pass_regmove): ...new function.
865 * regrename.c (pass_regrename): Convert from a global struct to a
866 subclass of rtl_opt_pass along with...
867 (pass_data_regrename): ...new pass_data instance and...
868 (make_pass_regrename): ...new function.
869 * reorg.c (pass_delay_slots): Convert from a global struct to a
870 subclass of rtl_opt_pass along with...
871 (pass_data_delay_slots): ...new pass_data instance and...
872 (make_pass_delay_slots): ...new function.
873 (pass_machine_reorg): Convert from a global struct to a subclass of
874 rtl_opt_pass along with...
875 (pass_data_machine_reorg): ...new pass_data instance and...
876 (make_pass_machine_reorg): ...new function.
877 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
878 of rtl_opt_pass along with...
879 (pass_data_sched): ...new pass_data instance and...
880 (make_pass_sched): ...new function.
881 (pass_sched2): Convert from a global struct to a subclass of
882 rtl_opt_pass along with...
883 (pass_data_sched2): ...new pass_data instance and...
884 (make_pass_sched2): ...new function.
885 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
886 to a subclass of rtl_opt_pass along with...
887 (pass_data_stack_ptr_mod): ...new pass_data instance and...
888 (make_pass_stack_ptr_mod): ...new function.
889 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
890 to a subclass of rtl_opt_pass along with...
891 (pass_data_rtl_store_motion): ...new pass_data instance and...
892 (make_pass_rtl_store_motion): ...new function.
893 * tracer.c (pass_tracer): Convert from a global struct to a subclass
894 of gimple_opt_pass along with...
895 (pass_data_tracer): ...new pass_data instance and...
896 (make_pass_tracer): ...new function.
897 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
898 to a subclass of gimple_opt_pass along with...
899 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
900 (make_pass_diagnose_tm_blocks): ...new function.
901 (pass_lower_tm): Convert from a global struct to a subclass of
902 gimple_opt_pass along with...
903 (pass_data_lower_tm): ...new pass_data instance and...
904 (make_pass_lower_tm): ...new function.
905 (pass_tm_init): Convert from a global struct to a subclass of
906 gimple_opt_pass along with...
907 (pass_data_tm_init): ...new pass_data instance and...
908 (make_pass_tm_init): ...new function.
909 (pass_tm_mark): Convert from a global struct to a subclass of
910 gimple_opt_pass along with...
911 (pass_data_tm_mark): ...new pass_data instance and...
912 (make_pass_tm_mark): ...new function.
913 (pass_tm_edges): Convert from a global struct to a subclass of
914 gimple_opt_pass along with...
915 (pass_data_tm_edges): ...new pass_data instance and...
916 (make_pass_tm_edges): ...new function.
917 (pass_tm_memopt): Convert from a global struct to a subclass of
918 gimple_opt_pass along with...
919 (pass_data_tm_memopt): ...new pass_data instance and...
920 (make_pass_tm_memopt): ...new function.
921 (pass_ipa_tm): Convert from a global struct to a subclass of
922 simple_ipa_opt_pass along with...
923 (pass_data_ipa_tm): ...new pass_data instance and...
924 (make_pass_ipa_tm): ...new function.
925 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
926 subclass of gimple_opt_pass along with...
927 (pass_data_call_cdce): ...new pass_data instance and...
928 (make_pass_call_cdce): ...new function.
929 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
930 subclass of gimple_opt_pass along with...
931 (pass_data_build_cfg): ...new pass_data instance and...
932 (make_pass_build_cfg): ...new function.
933 (pass_split_crit_edges): Convert from a global struct to a subclass of
934 gimple_opt_pass along with...
935 (pass_data_split_crit_edges): ...new pass_data instance and...
936 (make_pass_split_crit_edges): ...new function.
937 (pass_warn_function_return): Convert from a global struct to a
938 subclass of gimple_opt_pass along with...
939 (pass_data_warn_function_return): ...new pass_data instance and...
940 (make_pass_warn_function_return): ...new function.
941 (pass_warn_function_noreturn): Convert from a global struct to a
942 subclass of gimple_opt_pass along with...
943 (pass_data_warn_function_noreturn): ...new pass_data instance and...
944 (make_pass_warn_function_noreturn): ...new function.
945 (pass_warn_unused_result): Convert from a global struct to a subclass
946 of gimple_opt_pass along with...
947 (pass_data_warn_unused_result): ...new pass_data instance and...
948 (make_pass_warn_unused_result): ...new function.
949 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
950 a subclass of gimple_opt_pass along with...
951 (pass_data_merge_phi): ...new pass_data instance and...
952 (make_pass_merge_phi): ...new function.
953 * tree-complex.c (pass_lower_complex): Convert from a global struct to
954 a subclass of gimple_opt_pass along with...
955 (pass_data_lower_complex): ...new pass_data instance and...
956 (make_pass_lower_complex): ...new function.
957 (pass_lower_complex_O0): Convert from a global struct to a subclass of
958 gimple_opt_pass along with...
959 (pass_data_lower_complex_O0): ...new pass_data instance and...
960 (make_pass_lower_complex_O0): ...new function.
961 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
962 subclass of gimple_opt_pass along with...
963 (pass_data_lower_eh): ...new pass_data instance and...
964 (make_pass_lower_eh): ...new function.
965 (pass_refactor_eh): Convert from a global struct to a subclass of
966 gimple_opt_pass along with...
967 (pass_data_refactor_eh): ...new pass_data instance and...
968 (make_pass_refactor_eh): ...new function.
969 (pass_lower_resx): Convert from a global struct to a subclass of
970 gimple_opt_pass along with...
971 (pass_data_lower_resx): ...new pass_data instance and...
972 (make_pass_lower_resx): ...new function.
973 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
974 of gimple_opt_pass along with...
975 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
976 (make_pass_lower_eh_dispatch): ...new function.
977 (pass_cleanup_eh): Convert from a global struct to a subclass of
978 gimple_opt_pass along with...
979 (pass_data_cleanup_eh): ...new pass_data instance and...
980 (make_pass_cleanup_eh): ...new function.
981 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
982 to a subclass of simple_ipa_opt_pass along with...
983 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
984 (make_pass_ipa_lower_emutls): ...new function.
985 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
986 a subclass of gimple_opt_pass along with...
987 (pass_data_if_conversion): ...new pass_data instance and...
988 (make_pass_if_conversion): ...new function.
989 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
990 subclass of gimple_opt_pass along with...
991 (pass_data_build_ssa): ...new pass_data instance and...
992 (make_pass_build_ssa): ...new function.
993 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
994 global struct to a subclass of gimple_opt_pass along with...
995 (pass_data_loop_distribution): ...new pass_data instance and...
996 (make_pass_loop_distribution): ...new function.
997 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
998 subclass of gimple_opt_pass along with...
999 (pass_data_mudflap_1): ...new pass_data instance and...
1000 (make_pass_mudflap_1): ...new function.
1001 (pass_mudflap_2): Convert from a global struct to a subclass of
1002 gimple_opt_pass along with...
1003 (pass_data_mudflap_2): ...new pass_data instance and...
1004 (make_pass_mudflap_2): ...new function.
1005 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1006 subclass of gimple_opt_pass along with...
1007 (pass_data_mudflap_1): ...new pass_data instance and...
1008 (make_pass_mudflap_1): ...new function.
1009 (pass_mudflap_2): Convert from a global struct to a subclass of
1010 gimple_opt_pass along with...
1011 (pass_data_mudflap_2): ...new pass_data instance and...
1012 (make_pass_mudflap_2): ...new function.
1013 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1014 gimple_opt_pass along with...
1015 (pass_data_nrv): ...new pass_data instance and...
1016 (make_pass_nrv): ...new function.
1017 (pass_return_slot): Convert from a global struct to a subclass of
1018 gimple_opt_pass along with...
1019 (pass_data_return_slot): ...new pass_data instance and...
1020 (make_pass_return_slot): ...new function.
1021 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1022 to a subclass of gimple_opt_pass along with...
1023 (pass_data_object_sizes): ...new pass_data instance and...
1024 (make_pass_object_sizes): ...new function.
1025 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1026 global struct to a subclass of gimple_opt_pass along with...
1027 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1029 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1030 (pass_fixup_cfg): Convert from a global struct to a subclass of
1031 gimple_opt_pass along with...
1032 (pass_data_fixup_cfg): ...new pass_data instance and...
1033 (make_pass_fixup_cfg): ...new function.
1034 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1035 (make_pass_mudflap_1): ...new function.
1036 (pass_mudflap_2): Replace declaration with that of...
1037 (make_pass_mudflap_2): ...new function.
1038 (pass_asan): Replace declaration with that of...
1039 (make_pass_asan): ...new function.
1040 (pass_asan_O0): Replace declaration with that of...
1041 (make_pass_asan_O0): ...new function.
1042 (pass_tsan): Replace declaration with that of...
1043 (make_pass_tsan): ...new function.
1044 (pass_tsan_O0): Replace declaration with that of...
1045 (make_pass_tsan_O0): ...new function.
1046 (pass_lower_cf): Replace declaration with that of...
1047 (make_pass_lower_cf): ...new function.
1048 (pass_refactor_eh): Replace declaration with that of...
1049 (make_pass_refactor_eh): ...new function.
1050 (pass_lower_eh): Replace declaration with that of...
1051 (make_pass_lower_eh): ...new function.
1052 (pass_lower_eh_dispatch): Replace declaration with that of...
1053 (make_pass_lower_eh_dispatch): ...new function.
1054 (pass_lower_resx): Replace declaration with that of...
1055 (make_pass_lower_resx): ...new function.
1056 (pass_build_cfg): Replace declaration with that of...
1057 (make_pass_build_cfg): ...new function.
1058 (pass_early_tree_profile): Replace declaration with that of...
1059 (make_pass_early_tree_profile): ...new function.
1060 (pass_cleanup_eh): Replace declaration with that of...
1061 (make_pass_cleanup_eh): ...new function.
1062 (pass_sra): Replace declaration with that of...
1063 (make_pass_sra): ...new function.
1064 (pass_sra_early): Replace declaration with that of...
1065 (make_pass_sra_early): ...new function.
1066 (pass_early_ipa_sra): Replace declaration with that of...
1067 (make_pass_early_ipa_sra): ...new function.
1068 (pass_tail_recursion): Replace declaration with that of...
1069 (make_pass_tail_recursion): ...new function.
1070 (pass_tail_calls): Replace declaration with that of...
1071 (make_pass_tail_calls): ...new function.
1072 (pass_tree_loop): Replace declaration with that of...
1073 (make_pass_tree_loop): ...new function.
1074 (pass_tree_loop_init): Replace declaration with that of...
1075 (make_pass_tree_loop_init): ...new function.
1076 (pass_lim): Replace declaration with that of...
1077 (make_pass_lim): ...new function.
1078 (pass_tree_unswitch): Replace declaration with that of...
1079 (make_pass_tree_unswitch): ...new function.
1080 (pass_predcom): Replace declaration with that of...
1081 (make_pass_predcom): ...new function.
1082 (pass_iv_canon): Replace declaration with that of...
1083 (make_pass_iv_canon): ...new function.
1084 (pass_scev_cprop): Replace declaration with that of...
1085 (make_pass_scev_cprop): ...new function.
1086 (pass_empty_loop): Replace declaration with that of...
1087 (make_pass_empty_loop): ...new function.
1088 (pass_record_bounds): Replace declaration with that of...
1089 (make_pass_record_bounds): ...new function.
1090 (pass_graphite): Replace declaration with that of...
1091 (make_pass_graphite): ...new function.
1092 (pass_graphite_transforms): Replace declaration with that of...
1093 (make_pass_graphite_transforms): ...new function.
1094 (pass_if_conversion): Replace declaration with that of...
1095 (make_pass_if_conversion): ...new function.
1096 (pass_loop_distribution): Replace declaration with that of...
1097 (make_pass_loop_distribution): ...new function.
1098 (pass_vectorize): Replace declaration with that of...
1099 (make_pass_vectorize): ...new function.
1100 (pass_slp_vectorize): Replace declaration with that of...
1101 (make_pass_slp_vectorize): ...new function.
1102 (pass_complete_unroll): Replace declaration with that of...
1103 (make_pass_complete_unroll): ...new function.
1104 (pass_complete_unrolli): Replace declaration with that of...
1105 (make_pass_complete_unrolli): ...new function.
1106 (pass_parallelize_loops): Replace declaration with that of...
1107 (make_pass_parallelize_loops): ...new function.
1108 (pass_loop_prefetch): Replace declaration with that of...
1109 (make_pass_loop_prefetch): ...new function.
1110 (pass_iv_optimize): Replace declaration with that of...
1111 (make_pass_iv_optimize): ...new function.
1112 (pass_tree_loop_done): Replace declaration with that of...
1113 (make_pass_tree_loop_done): ...new function.
1114 (pass_ch): Replace declaration with that of...
1115 (make_pass_ch): ...new function.
1116 (pass_ccp): Replace declaration with that of...
1117 (make_pass_ccp): ...new function.
1118 (pass_phi_only_cprop): Replace declaration with that of...
1119 (make_pass_phi_only_cprop): ...new function.
1120 (pass_build_ssa): Replace declaration with that of...
1121 (make_pass_build_ssa): ...new function.
1122 (pass_build_alias): Replace declaration with that of...
1123 (make_pass_build_alias): ...new function.
1124 (pass_build_ealias): Replace declaration with that of...
1125 (make_pass_build_ealias): ...new function.
1126 (pass_dominator): Replace declaration with that of...
1127 (make_pass_dominator): ...new function.
1128 (pass_dce): Replace declaration with that of...
1129 (make_pass_dce): ...new function.
1130 (pass_dce_loop): Replace declaration with that of...
1131 (make_pass_dce_loop): ...new function.
1132 (pass_cd_dce): Replace declaration with that of...
1133 (make_pass_cd_dce): ...new function.
1134 (pass_call_cdce): Replace declaration with that of...
1135 (make_pass_call_cdce): ...new function.
1136 (pass_merge_phi): Replace declaration with that of...
1137 (make_pass_merge_phi): ...new function.
1138 (pass_split_crit_edges): Replace declaration with that of...
1139 (make_pass_split_crit_edges): ...new function.
1140 (pass_pre): Replace declaration with that of...
1141 (make_pass_pre): ...new function.
1142 (pass_profile): Replace declaration with that of...
1143 (make_pass_profile): ...new function.
1144 (pass_strip_predict_hints): Replace declaration with that of...
1145 (make_pass_strip_predict_hints): ...new function.
1146 (pass_lower_complex_O0): Replace declaration with that of...
1147 (make_pass_lower_complex_O0): ...new function.
1148 (pass_lower_complex): Replace declaration with that of...
1149 (make_pass_lower_complex): ...new function.
1150 (pass_lower_vector): Replace declaration with that of...
1151 (make_pass_lower_vector): ...new function.
1152 (pass_lower_vector_ssa): Replace declaration with that of...
1153 (make_pass_lower_vector_ssa): ...new function.
1154 (pass_lower_omp): Replace declaration with that of...
1155 (make_pass_lower_omp): ...new function.
1156 (pass_diagnose_omp_blocks): Replace declaration with that of...
1157 (make_pass_diagnose_omp_blocks): ...new function.
1158 (pass_expand_omp): Replace declaration with that of...
1159 (make_pass_expand_omp): ...new function.
1160 (pass_expand_omp_ssa): Replace declaration with that of...
1161 (make_pass_expand_omp_ssa): ...new function.
1162 (pass_object_sizes): Replace declaration with that of...
1163 (make_pass_object_sizes): ...new function.
1164 (pass_strlen): Replace declaration with that of...
1165 (make_pass_strlen): ...new function.
1166 (pass_fold_builtins): Replace declaration with that of...
1167 (make_pass_fold_builtins): ...new function.
1168 (pass_stdarg): Replace declaration with that of...
1169 (make_pass_stdarg): ...new function.
1170 (pass_early_warn_uninitialized): Replace declaration with that of...
1171 (make_pass_early_warn_uninitialized): ...new function.
1172 (pass_late_warn_uninitialized): Replace declaration with that of...
1173 (make_pass_late_warn_uninitialized): ...new function.
1174 (pass_cse_reciprocals): Replace declaration with that of...
1175 (make_pass_cse_reciprocals): ...new function.
1176 (pass_cse_sincos): Replace declaration with that of...
1177 (make_pass_cse_sincos): ...new function.
1178 (pass_optimize_bswap): Replace declaration with that of...
1179 (make_pass_optimize_bswap): ...new function.
1180 (pass_optimize_widening_mul): Replace declaration with that of...
1181 (make_pass_optimize_widening_mul): ...new function.
1182 (pass_warn_function_return): Replace declaration with that of...
1183 (make_pass_warn_function_return): ...new function.
1184 (pass_warn_function_noreturn): Replace declaration with that of...
1185 (make_pass_warn_function_noreturn): ...new function.
1186 (pass_cselim): Replace declaration with that of...
1187 (make_pass_cselim): ...new function.
1188 (pass_phiopt): Replace declaration with that of...
1189 (make_pass_phiopt): ...new function.
1190 (pass_forwprop): Replace declaration with that of...
1191 (make_pass_forwprop): ...new function.
1192 (pass_phiprop): Replace declaration with that of...
1193 (make_pass_phiprop): ...new function.
1194 (pass_tree_ifcombine): Replace declaration with that of...
1195 (make_pass_tree_ifcombine): ...new function.
1196 (pass_dse): Replace declaration with that of...
1197 (make_pass_dse): ...new function.
1198 (pass_nrv): Replace declaration with that of...
1199 (make_pass_nrv): ...new function.
1200 (pass_rename_ssa_copies): Replace declaration with that of...
1201 (make_pass_rename_ssa_copies): ...new function.
1202 (pass_sink_code): Replace declaration with that of...
1203 (make_pass_sink_code): ...new function.
1204 (pass_fre): Replace declaration with that of...
1205 (make_pass_fre): ...new function.
1206 (pass_check_data_deps): Replace declaration with that of...
1207 (make_pass_check_data_deps): ...new function.
1208 (pass_copy_prop): Replace declaration with that of...
1209 (make_pass_copy_prop): ...new function.
1210 (pass_vrp): Replace declaration with that of...
1211 (make_pass_vrp): ...new function.
1212 (pass_uncprop): Replace declaration with that of...
1213 (make_pass_uncprop): ...new function.
1214 (pass_return_slot): Replace declaration with that of...
1215 (make_pass_return_slot): ...new function.
1216 (pass_reassoc): Replace declaration with that of...
1217 (make_pass_reassoc): ...new function.
1218 (pass_rebuild_cgraph_edges): Replace declaration with that of...
1219 (make_pass_rebuild_cgraph_edges): ...new function.
1220 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
1221 (make_pass_remove_cgraph_callee_edges): ...new function.
1222 (pass_build_cgraph_edges): Replace declaration with that of...
1223 (make_pass_build_cgraph_edges): ...new function.
1224 (pass_local_pure_const): Replace declaration with that of...
1225 (make_pass_local_pure_const): ...new function.
1226 (pass_tracer): Replace declaration with that of...
1227 (make_pass_tracer): ...new function.
1228 (pass_warn_unused_result): Replace declaration with that of...
1229 (make_pass_warn_unused_result): ...new function.
1230 (pass_diagnose_tm_blocks): Replace declaration with that of...
1231 (make_pass_diagnose_tm_blocks): ...new function.
1232 (pass_lower_tm): Replace declaration with that of...
1233 (make_pass_lower_tm): ...new function.
1234 (pass_tm_init): Replace declaration with that of...
1235 (make_pass_tm_init): ...new function.
1236 (pass_tm_mark): Replace declaration with that of...
1237 (make_pass_tm_mark): ...new function.
1238 (pass_tm_memopt): Replace declaration with that of...
1239 (make_pass_tm_memopt): ...new function.
1240 (pass_tm_edges): Replace declaration with that of...
1241 (make_pass_tm_edges): ...new function.
1242 (pass_split_functions): Replace declaration with that of...
1243 (make_pass_split_functions): ...new function.
1244 (pass_feedback_split_functions): Replace declaration with that of...
1245 (make_pass_feedback_split_functions): ...new function.
1246 (pass_strength_reduction): Replace declaration with that of...
1247 (make_pass_strength_reduction): ...new function.
1248 (pass_ipa_lower_emutls): Replace declaration with that of...
1249 (make_pass_ipa_lower_emutls): ...new function.
1250 (pass_ipa_function_and_variable_visibility): Replace declaration with
1252 (make_pass_ipa_function_and_variable_visibility): ...new function.
1253 (pass_ipa_tree_profile): Replace declaration with that of...
1254 (make_pass_ipa_tree_profile): ...new function.
1255 (pass_early_local_passes): Replace declaration with that of...
1256 (make_pass_early_local_passes): ...new function.
1257 (pass_ipa_whole_program_visibility): Replace declaration with that
1259 (make_pass_ipa_whole_program_visibility): ...new function.
1260 (pass_ipa_lto_gimple_out): Replace declaration with that of...
1261 (make_pass_ipa_lto_gimple_out): ...new function.
1262 (pass_ipa_increase_alignment): Replace declaration with that of...
1263 (make_pass_ipa_increase_alignment): ...new function.
1264 (pass_ipa_inline): Replace declaration with that of...
1265 (make_pass_ipa_inline): ...new function.
1266 (pass_ipa_free_lang_data): Replace declaration with that of...
1267 (make_pass_ipa_free_lang_data): ...new function.
1268 (pass_ipa_free_inline_summary): Replace declaration with that of...
1269 (make_pass_ipa_free_inline_summary): ...new function.
1270 (pass_ipa_cp): Replace declaration with that of...
1271 (make_pass_ipa_cp): ...new function.
1272 (pass_ipa_reference): Replace declaration with that of...
1273 (make_pass_ipa_reference): ...new function.
1274 (pass_ipa_pure_const): Replace declaration with that of...
1275 (make_pass_ipa_pure_const): ...new function.
1276 (pass_ipa_pta): Replace declaration with that of...
1277 (make_pass_ipa_pta): ...new function.
1278 (pass_ipa_lto_finish_out): Replace declaration with that of...
1279 (make_pass_ipa_lto_finish_out): ...new function.
1280 (pass_ipa_tm): Replace declaration with that of...
1281 (make_pass_ipa_tm): ...new function.
1282 (pass_ipa_profile): Replace declaration with that of...
1283 (make_pass_ipa_profile): ...new function.
1284 (pass_ipa_cdtor_merge): Replace declaration with that of...
1285 (make_pass_ipa_cdtor_merge): ...new function.
1286 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
1288 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1289 (pass_init_datastructures): Replace declaration with that of...
1290 (make_pass_init_datastructures): ...new function.
1291 (pass_fixup_cfg): Replace declaration with that of...
1292 (make_pass_fixup_cfg): ...new function.
1293 (pass_expand): Replace declaration with that of...
1294 (make_pass_expand): ...new function.
1295 (pass_instantiate_virtual_regs): Replace declaration with that of...
1296 (make_pass_instantiate_virtual_regs): ...new function.
1297 (pass_rtl_fwprop): Replace declaration with that of...
1298 (make_pass_rtl_fwprop): ...new function.
1299 (pass_rtl_fwprop_addr): Replace declaration with that of...
1300 (make_pass_rtl_fwprop_addr): ...new function.
1301 (pass_jump): Replace declaration with that of...
1302 (make_pass_jump): ...new function.
1303 (pass_jump2): Replace declaration with that of...
1304 (make_pass_jump2): ...new function.
1305 (pass_lower_subreg): Replace declaration with that of...
1306 (make_pass_lower_subreg): ...new function.
1307 (pass_cse): Replace declaration with that of...
1308 (make_pass_cse): ...new function.
1309 (pass_fast_rtl_dce): Replace declaration with that of...
1310 (make_pass_fast_rtl_dce): ...new function.
1311 (pass_ud_rtl_dce): Replace declaration with that of...
1312 (make_pass_ud_rtl_dce): ...new function.
1313 (pass_rtl_dce): Replace declaration with that of...
1314 (make_pass_rtl_dce): ...new function.
1315 (pass_rtl_dse1): Replace declaration with that of...
1316 (make_pass_rtl_dse1): ...new function.
1317 (pass_rtl_dse2): Replace declaration with that of...
1318 (make_pass_rtl_dse2): ...new function.
1319 (pass_rtl_dse3): Replace declaration with that of...
1320 (make_pass_rtl_dse3): ...new function.
1321 (pass_rtl_cprop): Replace declaration with that of...
1322 (make_pass_rtl_cprop): ...new function.
1323 (pass_rtl_pre): Replace declaration with that of...
1324 (make_pass_rtl_pre): ...new function.
1325 (pass_rtl_hoist): Replace declaration with that of...
1326 (make_pass_rtl_hoist): ...new function.
1327 (pass_rtl_store_motion): Replace declaration with that of...
1328 (make_pass_rtl_store_motion): ...new function.
1329 (pass_cse_after_global_opts): Replace declaration with that of...
1330 (make_pass_cse_after_global_opts): ...new function.
1331 (pass_rtl_ifcvt): Replace declaration with that of...
1332 (make_pass_rtl_ifcvt): ...new function.
1333 (pass_into_cfg_layout_mode): Replace declaration with that of...
1334 (make_pass_into_cfg_layout_mode): ...new function.
1335 (pass_outof_cfg_layout_mode): Replace declaration with that of...
1336 (make_pass_outof_cfg_layout_mode): ...new function.
1337 (pass_loop2): Replace declaration with that of...
1338 (make_pass_loop2): ...new function.
1339 (pass_rtl_loop_init): Replace declaration with that of...
1340 (make_pass_rtl_loop_init): ...new function.
1341 (pass_rtl_move_loop_invariants): Replace declaration with that of...
1342 (make_pass_rtl_move_loop_invariants): ...new function.
1343 (pass_rtl_unswitch): Replace declaration with that of...
1344 (make_pass_rtl_unswitch): ...new function.
1345 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
1346 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1347 (pass_rtl_doloop): Replace declaration with that of...
1348 (make_pass_rtl_doloop): ...new function.
1349 (pass_rtl_loop_done): Replace declaration with that of...
1350 (make_pass_rtl_loop_done): ...new function.
1351 (pass_web): Replace declaration with that of...
1352 (make_pass_web): ...new function.
1353 (pass_cse2): Replace declaration with that of...
1354 (make_pass_cse2): ...new function.
1355 (pass_df_initialize_opt): Replace declaration with that of...
1356 (make_pass_df_initialize_opt): ...new function.
1357 (pass_df_initialize_no_opt): Replace declaration with that of...
1358 (make_pass_df_initialize_no_opt): ...new function.
1359 (pass_reginfo_init): Replace declaration with that of...
1360 (make_pass_reginfo_init): ...new function.
1361 (pass_inc_dec): Replace declaration with that of...
1362 (make_pass_inc_dec): ...new function.
1363 (pass_stack_ptr_mod): Replace declaration with that of...
1364 (make_pass_stack_ptr_mod): ...new function.
1365 (pass_initialize_regs): Replace declaration with that of...
1366 (make_pass_initialize_regs): ...new function.
1367 (pass_combine): Replace declaration with that of...
1368 (make_pass_combine): ...new function.
1369 (pass_if_after_combine): Replace declaration with that of...
1370 (make_pass_if_after_combine): ...new function.
1371 (pass_ree): Replace declaration with that of...
1372 (make_pass_ree): ...new function.
1373 (pass_partition_blocks): Replace declaration with that of...
1374 (make_pass_partition_blocks): ...new function.
1375 (pass_match_asm_constraints): Replace declaration with that of...
1376 (make_pass_match_asm_constraints): ...new function.
1377 (pass_regmove): Replace declaration with that of...
1378 (make_pass_regmove): ...new function.
1379 (pass_split_all_insns): Replace declaration with that of...
1380 (make_pass_split_all_insns): ...new function.
1381 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1382 (make_pass_fast_rtl_byte_dce): ...new function.
1383 (pass_lower_subreg2): Replace declaration with that of...
1384 (make_pass_lower_subreg2): ...new function.
1385 (pass_mode_switching): Replace declaration with that of...
1386 (make_pass_mode_switching): ...new function.
1387 (pass_sms): Replace declaration with that of...
1388 (make_pass_sms): ...new function.
1389 (pass_sched): Replace declaration with that of...
1390 (make_pass_sched): ...new function.
1391 (pass_ira): Replace declaration with that of...
1392 (make_pass_ira): ...new function.
1393 (pass_reload): Replace declaration with that of...
1394 (make_pass_reload): ...new function.
1395 (pass_clean_state): Replace declaration with that of...
1396 (make_pass_clean_state): ...new function.
1397 (pass_branch_prob): Replace declaration with that of...
1398 (make_pass_branch_prob): ...new function.
1399 (pass_value_profile_transformations): Replace declaration with that
1401 (make_pass_value_profile_transformations): ...new function.
1402 (pass_postreload_cse): Replace declaration with that of...
1403 (make_pass_postreload_cse): ...new function.
1404 (pass_gcse2): Replace declaration with that of...
1405 (make_pass_gcse2): ...new function.
1406 (pass_split_after_reload): Replace declaration with that of...
1407 (make_pass_split_after_reload): ...new function.
1408 (pass_branch_target_load_optimize1): Replace declaration with that
1410 (make_pass_branch_target_load_optimize1): ...new function.
1411 (pass_thread_prologue_and_epilogue): Replace declaration with that
1413 (make_pass_thread_prologue_and_epilogue): ...new function.
1414 (pass_stack_adjustments): Replace declaration with that of...
1415 (make_pass_stack_adjustments): ...new function.
1416 (pass_peephole2): Replace declaration with that of...
1417 (make_pass_peephole2): ...new function.
1418 (pass_if_after_reload): Replace declaration with that of...
1419 (make_pass_if_after_reload): ...new function.
1420 (pass_regrename): Replace declaration with that of...
1421 (make_pass_regrename): ...new function.
1422 (pass_cprop_hardreg): Replace declaration with that of...
1423 (make_pass_cprop_hardreg): ...new function.
1424 (pass_reorder_blocks): Replace declaration with that of...
1425 (make_pass_reorder_blocks): ...new function.
1426 (pass_branch_target_load_optimize2): Replace declaration with that
1428 (make_pass_branch_target_load_optimize2): ...new function.
1429 (pass_leaf_regs): Replace declaration with that of...
1430 (make_pass_leaf_regs): ...new function.
1431 (pass_split_before_sched2): Replace declaration with that of...
1432 (make_pass_split_before_sched2): ...new function.
1433 (pass_compare_elim_after_reload): Replace declaration with that of...
1434 (make_pass_compare_elim_after_reload): ...new function.
1435 (pass_sched2): Replace declaration with that of...
1436 (make_pass_sched2): ...new function.
1437 (pass_stack_regs): Replace declaration with that of...
1438 (make_pass_stack_regs): ...new function.
1439 (pass_stack_regs_run): Replace declaration with that of...
1440 (make_pass_stack_regs_run): ...new function.
1441 (pass_df_finish): Replace declaration with that of...
1442 (make_pass_df_finish): ...new function.
1443 (pass_compute_alignments): Replace declaration with that of...
1444 (make_pass_compute_alignments): ...new function.
1445 (pass_duplicate_computed_gotos): Replace declaration with that of...
1446 (make_pass_duplicate_computed_gotos): ...new function.
1447 (pass_variable_tracking): Replace declaration with that of...
1448 (make_pass_variable_tracking): ...new function.
1449 (pass_free_cfg): Replace declaration with that of...
1450 (make_pass_free_cfg): ...new function.
1451 (pass_machine_reorg): Replace declaration with that of...
1452 (make_pass_machine_reorg): ...new function.
1453 (pass_cleanup_barriers): Replace declaration with that of...
1454 (make_pass_cleanup_barriers): ...new function.
1455 (pass_delay_slots): Replace declaration with that of...
1456 (make_pass_delay_slots): ...new function.
1457 (pass_split_for_shorten_branches): Replace declaration with that of...
1458 (make_pass_split_for_shorten_branches): ...new function.
1459 (pass_split_before_regstack): Replace declaration with that of...
1460 (make_pass_split_before_regstack): ...new function.
1461 (pass_convert_to_eh_region_ranges): Replace declaration with that
1463 (make_pass_convert_to_eh_region_ranges): ...new function.
1464 (pass_shorten_branches): Replace declaration with that of...
1465 (make_pass_shorten_branches): ...new function.
1466 (pass_set_nothrow_function_flags): Replace declaration with that of...
1467 (make_pass_set_nothrow_function_flags): ...new function.
1468 (pass_dwarf2_frame): Replace declaration with that of...
1469 (make_pass_dwarf2_frame): ...new function.
1470 (pass_final): Replace declaration with that of...
1471 (make_pass_final): ...new function.
1472 (pass_rtl_seqabstr): Replace declaration with that of...
1473 (make_pass_rtl_seqabstr): ...new function.
1474 (pass_release_ssa_names): Replace declaration with that of...
1475 (make_pass_release_ssa_names): ...new function.
1476 (pass_early_inline): Replace declaration with that of...
1477 (make_pass_early_inline): ...new function.
1478 (pass_inline_parameters): Replace declaration with that of...
1479 (make_pass_inline_parameters): ...new function.
1480 (pass_update_address_taken): Replace declaration with that of...
1481 (make_pass_update_address_taken): ...new function.
1482 (pass_convert_switch): Replace declaration with that of...
1483 (make_pass_convert_switch): ...new function.
1484 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1485 to a subclass of simple_ipa_opt_pass along with...
1486 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1487 (make_pass_ipa_tree_profile): ...new function.
1488 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1489 subclass of gimple_opt_pass along with...
1490 (pass_data_sra_early): ...new pass_data instance and...
1491 (make_pass_sra_early): ...new function.
1492 (pass_sra): Convert from a global struct to a subclass of
1493 gimple_opt_pass along with...
1494 (pass_data_sra): ...new pass_data instance and...
1495 (make_pass_sra): ...new function.
1496 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1497 gimple_opt_pass along with...
1498 (pass_data_early_ipa_sra): ...new pass_data instance and...
1499 (make_pass_early_ipa_sra): ...new function.
1500 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1501 subclass of gimple_opt_pass along with...
1502 (pass_data_ccp): ...new pass_data instance and...
1503 (make_pass_ccp): ...new function.
1504 (pass_fold_builtins): Convert from a global struct to a subclass of
1505 gimple_opt_pass along with...
1506 (pass_data_fold_builtins): ...new pass_data instance and...
1507 (make_pass_fold_builtins): ...new function.
1508 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1509 subclass of gimple_opt_pass along with...
1510 (pass_data_copy_prop): ...new pass_data instance and...
1511 (make_pass_copy_prop): ...new function.
1512 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1513 global struct to a subclass of gimple_opt_pass along with...
1514 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1515 (make_pass_rename_ssa_copies): ...new function.
1516 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1517 subclass of gimple_opt_pass along with...
1518 (pass_data_dce): ...new pass_data instance and...
1519 (make_pass_dce): ...new function.
1520 (pass_dce_loop): Convert from a global struct to a subclass of
1521 gimple_opt_pass along with...
1522 (pass_data_dce_loop): ...new pass_data instance and...
1523 (make_pass_dce_loop): ...new function.
1524 (pass_cd_dce): Convert from a global struct to a subclass of
1525 gimple_opt_pass along with...
1526 (pass_data_cd_dce): ...new pass_data instance and...
1527 (make_pass_cd_dce): ...new function.
1528 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1529 subclass of gimple_opt_pass along with...
1530 (pass_data_dominator): ...new pass_data instance and...
1531 (make_pass_dominator): ...new function.
1532 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1533 gimple_opt_pass along with...
1534 (pass_data_phi_only_cprop): ...new pass_data instance and...
1535 (make_pass_phi_only_cprop): ...new function.
1536 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1537 subclass of gimple_opt_pass along with...
1538 (pass_data_dse): ...new pass_data instance and...
1539 (make_pass_dse): ...new function.
1540 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1541 a subclass of gimple_opt_pass along with...
1542 (pass_data_forwprop): ...new pass_data instance and...
1543 (make_pass_forwprop): ...new function.
1544 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1545 struct to a subclass of gimple_opt_pass along with...
1546 (pass_data_tree_ifcombine): ...new pass_data instance and...
1547 (make_pass_tree_ifcombine): ...new function.
1548 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1549 subclass of gimple_opt_pass along with...
1550 (pass_data_ch): ...new pass_data instance and...
1551 (make_pass_ch): ...new function.
1552 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1553 subclass of gimple_opt_pass along with...
1554 (pass_data_tree_loop): ...new pass_data instance and...
1555 (make_pass_tree_loop): ...new function.
1556 (pass_tree_loop_init): Convert from a global struct to a subclass of
1557 gimple_opt_pass along with...
1558 (pass_data_tree_loop_init): ...new pass_data instance and...
1559 (make_pass_tree_loop_init): ...new function.
1560 (pass_lim): Convert from a global struct to a subclass of
1561 gimple_opt_pass along with...
1562 (pass_data_lim): ...new pass_data instance and...
1563 (make_pass_lim): ...new function.
1564 (pass_tree_unswitch): Convert from a global struct to a subclass of
1565 gimple_opt_pass along with...
1566 (pass_data_tree_unswitch): ...new pass_data instance and...
1567 (make_pass_tree_unswitch): ...new function.
1568 (pass_predcom): Convert from a global struct to a subclass of
1569 gimple_opt_pass along with...
1570 (pass_data_predcom): ...new pass_data instance and...
1571 (make_pass_predcom): ...new function.
1572 (pass_vectorize): Convert from a global struct to a subclass of
1573 gimple_opt_pass along with...
1574 (pass_data_vectorize): ...new pass_data instance and...
1575 (make_pass_vectorize): ...new function.
1576 (pass_graphite): Convert from a global struct to a subclass of
1577 gimple_opt_pass along with...
1578 (pass_data_graphite): ...new pass_data instance and...
1579 (make_pass_graphite): ...new function.
1580 (pass_graphite_transforms): Convert from a global struct to a subclass
1581 of gimple_opt_pass along with...
1582 (pass_data_graphite_transforms): ...new pass_data instance and...
1583 (make_pass_graphite_transforms): ...new function.
1584 (pass_check_data_deps): Convert from a global struct to a subclass of
1585 gimple_opt_pass along with...
1586 (pass_data_check_data_deps): ...new pass_data instance and...
1587 (make_pass_check_data_deps): ...new function.
1588 (pass_iv_canon): Convert from a global struct to a subclass of
1589 gimple_opt_pass along with...
1590 (pass_data_iv_canon): ...new pass_data instance and...
1591 (make_pass_iv_canon): ...new function.
1592 (pass_scev_cprop): Convert from a global struct to a subclass of
1593 gimple_opt_pass along with...
1594 (pass_data_scev_cprop): ...new pass_data instance and...
1595 (make_pass_scev_cprop): ...new function.
1596 (pass_record_bounds): Convert from a global struct to a subclass of
1597 gimple_opt_pass along with...
1598 (pass_data_record_bounds): ...new pass_data instance and...
1599 (make_pass_record_bounds): ...new function.
1600 (pass_complete_unroll): Convert from a global struct to a subclass of
1601 gimple_opt_pass along with...
1602 (pass_data_complete_unroll): ...new pass_data instance and...
1603 (make_pass_complete_unroll): ...new function.
1604 (pass_complete_unrolli): Convert from a global struct to a subclass of
1605 gimple_opt_pass along with...
1606 (pass_data_complete_unrolli): ...new pass_data instance and...
1607 (make_pass_complete_unrolli): ...new function.
1608 (pass_parallelize_loops): Convert from a global struct to a subclass
1609 of gimple_opt_pass along with...
1610 (pass_data_parallelize_loops): ...new pass_data instance and...
1611 (make_pass_parallelize_loops): ...new function.
1612 (pass_loop_prefetch): Convert from a global struct to a subclass of
1613 gimple_opt_pass along with...
1614 (pass_data_loop_prefetch): ...new pass_data instance and...
1615 (make_pass_loop_prefetch): ...new function.
1616 (pass_iv_optimize): Convert from a global struct to a subclass of
1617 gimple_opt_pass along with...
1618 (pass_data_iv_optimize): ...new pass_data instance and...
1619 (make_pass_iv_optimize): ...new function.
1620 (pass_tree_loop_done): Convert from a global struct to a subclass of
1621 gimple_opt_pass along with...
1622 (pass_data_tree_loop_done): ...new pass_data instance and...
1623 (make_pass_tree_loop_done): ...new function.
1624 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1625 struct to a subclass of gimple_opt_pass along with...
1626 (pass_data_cse_reciprocals): ...new pass_data instance and...
1627 (make_pass_cse_reciprocals): ...new function.
1628 (pass_cse_sincos): Convert from a global struct to a subclass of
1629 gimple_opt_pass along with...
1630 (pass_data_cse_sincos): ...new pass_data instance and...
1631 (make_pass_cse_sincos): ...new function.
1632 (pass_optimize_bswap): Convert from a global struct to a subclass of
1633 gimple_opt_pass along with...
1634 (pass_data_optimize_bswap): ...new pass_data instance and...
1635 (make_pass_optimize_bswap): ...new function.
1636 (pass_optimize_widening_mul): Convert from a global struct to a
1637 subclass of gimple_opt_pass along with...
1638 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1639 (make_pass_optimize_widening_mul): ...new function.
1640 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1641 subclass of gimple_opt_pass along with...
1642 (pass_data_phiopt): ...new pass_data instance and...
1643 (make_pass_phiopt): ...new function.
1644 (pass_cselim): Convert from a global struct to a subclass of
1645 gimple_opt_pass along with...
1646 (pass_data_cselim): ...new pass_data instance and...
1647 (make_pass_cselim): ...new function.
1648 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1649 subclass of gimple_opt_pass along with...
1650 (pass_data_phiprop): ...new pass_data instance and...
1651 (make_pass_phiprop): ...new function.
1652 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1653 subclass of gimple_opt_pass along with...
1654 (pass_data_pre): ...new pass_data instance and...
1655 (make_pass_pre): ...new function.
1656 (pass_fre): Convert from a global struct to a subclass of
1657 gimple_opt_pass along with...
1658 (pass_data_fre): ...new pass_data instance and...
1659 (make_pass_fre): ...new function.
1660 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1661 subclass of gimple_opt_pass along with...
1662 (pass_data_reassoc): ...new pass_data instance and...
1663 (make_pass_reassoc): ...new function.
1664 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1665 subclass of gimple_opt_pass along with...
1666 (pass_data_sink_code): ...new pass_data instance and...
1667 (make_pass_sink_code): ...new function.
1668 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1669 subclass of gimple_opt_pass along with...
1670 (pass_data_strlen): ...new pass_data instance and...
1671 (make_pass_strlen): ...new function.
1672 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1673 struct to a subclass of gimple_opt_pass along with...
1674 (pass_data_build_alias): ...new pass_data instance and...
1675 (make_pass_build_alias): ...new function.
1676 (pass_build_ealias): Convert from a global struct to a subclass of
1677 gimple_opt_pass along with...
1678 (pass_data_build_ealias): ...new pass_data instance and...
1679 (make_pass_build_ealias): ...new function.
1680 (pass_ipa_pta): Convert from a global struct to a subclass of
1681 simple_ipa_opt_pass along with...
1682 (pass_data_ipa_pta): ...new pass_data instance and...
1683 (make_pass_ipa_pta): ...new function.
1684 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1685 subclass of gimple_opt_pass along with...
1686 (pass_data_uncprop): ...new pass_data instance and...
1687 (make_pass_uncprop): ...new function.
1688 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1689 global struct to a subclass of gimple_opt_pass along with...
1690 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1691 (make_pass_late_warn_uninitialized): ...new function.
1692 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1693 to a subclass of gimple_opt_pass along with...
1694 (pass_data_init_datastructures): ...new pass_data instance and...
1695 (make_pass_init_datastructures): ...new function.
1696 (pass_early_warn_uninitialized): Convert from a global struct to a
1697 subclass of gimple_opt_pass along with...
1698 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1699 (make_pass_early_warn_uninitialized): ...new function.
1700 (pass_update_address_taken): Convert from a global struct to a
1701 subclass of gimple_opt_pass along with...
1702 (pass_data_update_address_taken): ...new pass_data instance and...
1703 (make_pass_update_address_taken): ...new function.
1704 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1705 struct to a subclass of gimple_opt_pass along with...
1706 (pass_data_release_ssa_names): ...new pass_data instance and...
1707 (make_pass_release_ssa_names): ...new function.
1708 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1709 subclass of gimple_opt_pass along with...
1710 (pass_data_stdarg): ...new pass_data instance and...
1711 (make_pass_stdarg): ...new function.
1712 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1713 global struct to a subclass of gimple_opt_pass along with...
1714 (pass_data_convert_switch): ...new pass_data instance and...
1715 (make_pass_convert_switch): ...new function.
1716 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1717 to a subclass of gimple_opt_pass along with...
1718 (pass_data_tail_recursion): ...new pass_data instance and...
1719 (make_pass_tail_recursion): ...new function.
1720 (pass_tail_calls): Convert from a global struct to a subclass of
1721 gimple_opt_pass along with...
1722 (pass_data_tail_calls): ...new pass_data instance and...
1723 (make_pass_tail_calls): ...new function.
1724 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1725 struct to a subclass of gimple_opt_pass along with...
1726 (pass_data_lower_vector): ...new pass_data instance and...
1727 (make_pass_lower_vector): ...new function.
1728 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1729 gimple_opt_pass along with...
1730 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1731 (make_pass_lower_vector_ssa): ...new function.
1732 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1733 to a subclass of gimple_opt_pass along with...
1734 (pass_data_slp_vectorize): ...new pass_data instance and...
1735 (make_pass_slp_vectorize): ...new function.
1736 (pass_ipa_increase_alignment): Convert from a global struct to a
1737 subclass of simple_ipa_opt_pass along with...
1738 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1739 (make_pass_ipa_increase_alignment): ...new function.
1740 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1741 gimple_opt_pass along with...
1742 (pass_data_vrp): ...new pass_data instance and...
1743 (make_pass_vrp): ...new function.
1744 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1745 subclass of simple_ipa_opt_pass along with...
1746 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1747 (make_pass_ipa_free_lang_data): ...new function.
1748 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1749 gimple_opt_pass along with...
1750 (pass_data_tsan): ...new pass_data instance and...
1751 (make_pass_tsan): ...new function.
1752 (pass_tsan_O0): Convert from a global struct to a subclass of
1753 gimple_opt_pass along with...
1754 (pass_data_tsan_O0): ...new pass_data instance and...
1755 (make_pass_tsan_O0): ...new function.
1756 * var-tracking.c (pass_variable_tracking): Convert from a global
1757 struct to a subclass of rtl_opt_pass along with...
1758 (pass_data_variable_tracking): ...new pass_data instance and...
1759 (make_pass_variable_tracking): ...new function.
1760 * web.c (pass_web): Convert from a global struct to a subclass of
1761 rtl_opt_pass along with...
1762 (pass_data_web): ...new pass_data instance and...
1763 (make_pass_web): ...new function.
1764 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1765 declaration with that of...
1766 (make_pass_mode_switch_use): ...new function.
1767 (pass_resolve_sw_modes): Replace declaration with that of...
1768 (make_pass_resolve_sw_modes): ...new function.
1769 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1770 from a global struct to a subclass of rtl_opt_pass along with...
1771 (pass_data_mode_switch_use): ...new pass_data instance and...
1772 (make_pass_mode_switch_use): ...new function.
1773 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1774 from a global struct to a subclass of rtl_opt_pass along with...
1775 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1776 (make_pass_resolve_sw_modes): ...new function.
1777 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1778 struct to a subclass of rtl_opt_pass along with...
1779 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1780 (make_pass_insert_vzeroupper): ...new function.
1781 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1782 global struct to a subclass of rtl_opt_pass along with...
1783 (pass_data_work_around_errata): ...new pass_data instance and...
1784 (make_pass_work_around_errata): ...new function.
1785 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1786 struct to a subclass of rtl_opt_pass along with...
1787 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1788 (make_pass_mips_machine_reorg2): ...new function.
1790 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1792 * passes.c (pass_manager::operator new): New.
1794 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1796 Handwritten part of conversion of passes to C++ classes.
1798 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1799 (toplev.o): Add dep on PASS_MANAGER_H.
1800 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1801 of early local pases to reflect this moving from a global to a
1802 member of gcc::pass_manager.
1803 (cgraph_add_new_function): Likewise.
1804 * lto-cgraph.c (lto_output_node): Update for conversion of
1805 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1806 * passes.c (opt_pass::clone): New.
1807 (opt_pass::gate): New.
1808 (opt_pass::execute): New.
1809 (opt_pass::opt_pass): New.
1810 (pass_manager::execute_early_local_passes): New.
1811 (pass_manager::execute_pass_mode_switching): new.
1812 (finish_optimization_passes): Convert to...
1813 (pass_manager::finish_optimization_passes): ...this.
1814 (finish_optimization_passes): Update for conversion of passes to
1816 (register_dump_files_1): Use has_gate since we cannot portably
1817 check a vtable entry against NULL.
1818 (dump_one_pass): Likewise.
1819 (ipa_write_summaries_2): Likewise.
1820 (ipa_write_optimization_summaries_1): Likewise.
1821 (ipa_read_summaries_1): Likewise.
1822 (ipa_read_optimization_summaries_1): Likewise.
1823 (execute_ipa_stmt_fixups): Likewise.
1824 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1825 pass-creation functions rather than wiring up globals, and
1826 storing the results in fields of pass_manager generated using
1828 (pass_manager::dump_profile_report): Update for conversion of
1829 passes to C++ classes.
1830 (pass_manager::execute_ipa_summary_passes): Likewise.
1831 (execute_one_ipa_transform_pass): Likewise.
1832 (execute_one_pass): Use has_gate and has_execute since we cannot
1833 portably check a vtable entry against NULL.
1834 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1835 (pass_manager): Use pass-instances.def to add fields for the
1836 various pass instances.
1837 * toplev.c (finalize): Update for move of
1838 finish_optimization_passes to a method of gcc::pass_manager.
1839 * toplev.h (finish_optimization_passes): Move to method of class
1841 * tree-pass.h (struct pass_data): New.
1842 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
1843 (opt_pass::gate): Convert to virtual function.
1844 (opt_pass::~opt_pass): New.
1845 (opt_pass::clone): New.
1846 (opt_pass::execute): Convert to virtual function.
1847 (opt_pass::opt_pass): New.
1848 (opt_pass::ctxt_): new.
1849 (gimple_opt_pass): Convert to subclass of opt_pass.
1850 (gimple_opt_pass::gimple_opt_pass): New.
1851 (rtl_opt_pass): Convert to subclass of opt_pass.
1852 (rtl_opt_pass::rtl_opt_pass): New.
1853 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1854 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1855 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1856 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1857 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1858 invocation of pass_mode_switching to reflect this moving from a
1859 global to a member of gcc::pass_manager.
1860 (ix86_option_override): Rework how pass_insert_vzeroupper is
1861 added to the pass_manager to reflect autogenerated changes.
1862 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
1864 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
1866 PR rtl-optimization/57708
1867 * recog.c (peep2_find_free_register): Validate all regs in a
1870 2013-08-05 Jan Hubicka <jh@suse.cz>
1873 * cgraph.c (verify_cgraph_node): Accept local flags from other
1875 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1876 (function_and_variable_visibility): Likewise.
1877 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1879 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1881 * graph.c (init_graph_slim_pretty_print): Remove.
1882 (print_graph_cfg): Do not call it. Use local pretty printer.
1883 (start_graph_dump): Likewise.
1885 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1887 * gimple-pretty-print.c (buffer): Remove.
1888 (initialized): Likewise.
1889 (maybe_init_pretty_print): Likewise.
1890 (print_gimple_stmt): Do not call it. Use non-static local
1891 pretty_printer variable.
1892 (print_gimple_expr): Likewise.
1893 (print_gimple_seq): Likewise.
1894 (gimple_dump_bb): Likewise.
1896 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1898 * asan.c (asan_pp): Remove.
1899 (asan_pp_initialized): Likewise.
1900 (asan_pp_initialize): Likewise.
1901 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1902 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1903 (asan_add_global): Likewise.
1905 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1907 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1908 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1909 * pretty-print.c (pp_formatted_text_data): Likewise.
1910 (pp_indent): Rename from pp_base_indent.
1911 (pp_format): Rename from pp_base_format.
1912 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1913 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1914 (pp_flush): Rename from pp_base_flush.
1915 (pp_set_line_maximum_length): Rename from
1916 pp_base_set_line_maximum_length.
1917 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1918 (pp_set_prefix): Rename from pp_base_set_prefix.
1919 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1920 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1921 (pp_append_text): Rename from pp_base_append_text.
1922 (pp_formatted_text): Rename from pp_base_formatted_text.
1923 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1924 (pp_remaining_character_count_for_line): Rename from
1925 pp_base_remaining_character_count_for_line.
1926 (pp_newline): Rename from pp_base_newline.
1927 (pp_character): Rename from pp_base_character.
1928 (pp_string): Rename from pp_base_string.
1929 (pp_maybe_space): Rename from pp_base_maybe_space.
1930 * asan.c (asan_pp_string): Adjust.
1931 (asan_emit_stack_protection): Likewise.
1932 (asan_add_global): Likewise.
1933 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1934 * tree-mudflap.c (mf_varname_tree): Likewise.
1935 * tree-pretty-print.c (pp_tree_identifier): Rename from
1936 pp_base_tree_identifier.
1937 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1938 Declare as function.
1940 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1942 * pretty-print.h (pp_bar_bar): New.
1943 (pp_ampersand_ampersand): Likewise.
1944 (pp_less_equal): Likewise.
1945 (pp_greater_equal): Likewise.
1946 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1947 printer functions instead of pp_string or operators and punctuators.
1948 (dump_gimple_call): Likewise.
1949 (dump_gimple_omp_for): Likewise.
1950 (dump_gimple_transaction): Likewise.
1951 (dump_gimple_phi): Likewise.
1952 (pp_gimple_stmt_1): Likewise.
1953 * sched-vis.c (print_insn): Likewise.
1954 * tree-mudflap.c (mf_varname_tree): Likewise.
1955 * tree-pretty-print.c (dump_block_node): Likewise.
1956 (dump_generic_node): Likewise.
1958 2013-08-02 Jan Hubicka <jh@suse.cz>
1960 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1962 * lto-streamer-out.c (tree_is_indexable): Results decls and
1963 parm decls are not indexable.
1964 (DFS_write_tree_body): Do not follow args and results.
1965 (hash_tree): Likewise.
1966 (output_functions): Rearrange so struct function is needed
1967 only when real body is output; be able to also ouptut abstract
1968 functions; output DECL_ARGUMENTS and DECL_RESULT.
1969 (lto_output): When not in WPA, ale store abstract functions.
1970 (write_symbol): Do not care about RESULT_DECL.
1971 (output_symbol_p): Handle correctly sbtract decls.
1972 * lto-streamer-in.c (input_function): Rearrange so struct
1973 function can be NULL at entry; allow streaming of
1974 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
1975 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
1976 sanity check during LTO.
1977 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
1978 RESULT_DECl and DECL_ARGUMENTS.
1979 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
1982 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1984 * pretty-print.h (pp_underscore): New.
1986 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
1987 printer functions instead of pp_character.
1988 (dump_binary_rhs): Likewise.
1989 (dump_ternary_rhs): Likewise.
1990 (dump_gimple_call_args): Likewise.
1991 (pp_points_to_solution): Likewise.
1992 (dump_gimple_call): Likewise.
1993 (dump_gimple_switch): Likewise.
1994 (dump_gimple_cond): Likewise.
1995 (dump_gimple_bind): Likewise.
1996 (dump_gimple_try): Likewise.
1997 (dump_gimple_omp_for): Likewise.
1998 (dump_gimple_omp_continue): Likewise.
1999 (dump_gimple_omp_single): Likewise.
2000 (dump_gimple_omp_sections): Likewise.
2001 (dump_gimple_omp_block): Likewise.
2002 (dump_gimple_omp_critical): Likewise.
2003 (dump_gimple_transaction): Likewise.
2004 (dump_gimple_asm): Likewise.
2005 (dump_gimple_phi): Likewise.
2006 (dump_gimple_omp_parallel): Likewise.
2007 (dump_gimple_omp_task): Likewise.
2008 (dump_gimple_omp_atomic_load): Likewise.
2009 (dump_gimple_omp_atomic_store): Likewise.
2010 (dump_gimple_mem_ops): Likewise.
2011 (pp_gimple_stmt_1): Likewise.
2012 (pp_cfg_jump): Likewise.
2013 (dump_implicit_edges): Likewise.
2014 (gimple_dump_bb_for_graph): Likewise.
2015 * graph.c (draw_cfg_node): Likewise.
2016 * langhooks.c (lhd_print_error_function): Likewise.
2017 * sched-vis.c (print_exp): Likewise.
2018 (print_value): Likewise.
2019 (print_pattern): Likewise.
2020 (print_insn): Likewise.
2021 (rtl_dump_bb_for_graph): Likewise.
2022 * tree-pretty-print.c (dump_function_declaration): Likewise.
2023 (dump_array_domain): Likewise.
2024 (dump_omp_clause): Likewise.
2025 (dump_location): Likewise.
2026 (dump_generic_node): Likewise.
2027 (print_struct_decl): Likewise.
2028 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2030 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2032 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2033 candidate table when replacing a candidate statement.
2034 (replace_rhs_if_not_dup): Likewise.
2035 (replace_one_candidate): Likewise.
2037 2013-08-02 Jan Hubicka <jh@suse.cz>
2038 Martin Liska <marxin.liska@gmail.com>
2040 * cgraphunit.c (add_new_function): Fix logic when adding from
2042 (assemble_thunk): Rename to ...
2043 (expand_thunk); .. this one; export; get it working with
2044 general functions; make produced gimple valid.
2045 * cgraph.h (expand_thunk): Declare.
2047 2013-08-02 Jan Hubicka <jh@suse.cz>
2049 * ipa-cp.c (gather_context_independent_values): Use
2050 ipa_get_param_move_cost.
2051 (get_replacement_map): Remove PARAM; move parameter folding
2053 (create_specialized_node): Update.
2054 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2055 assert that we have gimple body; update move_cost.
2056 (count_formal_params): Assert that we have gimple body.
2057 (ipa_dump_param): New function.
2058 (ipa_alloc_node_params): Break out from ...
2059 (ipa_initialize_node_params): ... here.
2060 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2061 (ipa_write_node_info): Stream move costs.
2062 (ipa_read_node_info): Read move costs.
2063 (ipa_update_after_lto_read): Do not recompute node params.
2064 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2065 (ipa_get_param): Check we are not in WPA.
2066 (ipa_get_param_move_cost): New.
2067 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2068 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2069 parm numbers to be present.
2071 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2073 PR rtl-optimization/58048
2074 * lra-constraints.c (process_alt_operands): Don't check asm
2075 operand on register.
2077 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2079 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2080 the implied StoreLoad barrier for atomic operations if before.
2082 2013-08-02 Jan Hubicka <jh@suse.cz>
2083 Martin Liska <marxin.liska@gmail.com>
2085 * cgraph.c (cgraph_function_body_availability): Do not check
2087 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2088 symtab_node_availability): Declare.
2089 * ipa.c (can_replace_by_local_alias): New.
2090 (function_and_variable_visibility): Use it.
2091 * symtab.c (symtab_for_node_and_aliases,
2092 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2094 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2096 PR rtl-optimization/57963
2097 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2098 (lra_constraints): Use them.
2100 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2102 * config/arm/types.md (define_attr "type"): Add "load_acq"
2104 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2106 (cortex_a53_store1): Likewise.
2108 2013-08-01 Jan Hubicka <jh@suse.cz>
2110 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
2111 partitions are not needed.
2113 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
2115 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
2116 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
2117 MAYBE_NON_Q_CLASS_P where appropriate.
2119 2013-08-01 Jan Hubicka <jh@suse.cz>
2121 * cgraph.h (release_function_body): Declare.
2122 * tree.c (free_lang_data_in_decl): Free, parameters and return values
2123 of unused delcarations.
2125 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2127 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
2128 RTL form when subtracting a constant.
2130 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2132 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
2133 Generate canonical plus rtx with negated immediate instead of minus
2135 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
2137 2013-08-01 Jan Hubicka <jh@suse.cz>
2139 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
2140 (cgraph_release_function_body): Likewise.
2141 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2142 * cgraph.h (cgrpah_node): Rename abstract_and_needed
2143 to used_as_abstract_origin.
2144 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
2145 symbols used as abstract origins.
2146 * cgraphunit.c (analyze_functions): Update.
2147 * ipa.c (symtab_remove_unreachable_nodes): Recompute
2148 used_as_abstract_origin.
2149 * tree-inline.c (tree_function_versioning): Update
2150 used_as_abstract_origin; be ready for DECL_RESULT and
2151 DECL_ARGUMENTS to be NULL.
2153 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
2154 for abstract functions.
2155 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
2158 2013-08-01 Jan Hubicka <jh@suse.cz>
2160 * profile.c (compute_value_histograms): Fix thinko.
2162 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2164 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
2165 aarch-common-protos.h to extra_headers.
2166 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2167 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
2168 * config/aarch64/t-aarch64 (aarch-common.o): Define.
2170 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2172 * config/aarch64/aarch64.md (define_attr "type"): Delete.
2173 Include "../arm/types.md". Define "type" attribute for all patterns.
2174 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
2177 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
2179 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
2180 to support power8 load fusion.
2181 (fusion_gpr_mem_load): Likewise.
2183 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
2185 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
2186 declarations for power8 load fusion.
2187 (emit_fusion_gpr_load): Likewise.
2189 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
2190 tuning for power8, turn on fusion mode by default. Turn on sign
2191 extending fusion mode if normal fusion mode is on, and we are at
2193 (fusion_gpr_load_p): New function, return true if we can fuse an
2194 addis instruction with a dependent load to a GPR.
2195 (emit_fusion_gpr_load): Emit the instructions for power8 load
2198 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
2199 (VSX load fusion peepholes): New peepholes to fuse together an
2200 addi instruction with a VSX load instruction.
2202 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
2203 peepholes to fuse an addis instruction with a load to a GPR base
2204 register. If we are supporting sign extending fusions, convert
2205 sign extending loads to zero extending loads and add an explicit
2208 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2210 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
2211 aarch-common-protos.h to extra_headers.
2212 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2213 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
2214 (arm_early_store_addr_dep): Likewise.
2215 (arm_no_early_alu_shift_dep): Likewise.
2216 (arm_no_early_alu_shift_value_dep): Likewise.
2217 (arm_no_early_mul_dep): Likewise.
2218 (arm_no_early_store_addr_dep): Likewise.
2219 (arm_mac_accumulator_is_mul_result): Likewise.
2220 (arm_mac_accumulator_is_result): Likewise.
2221 * config/arm/aarch-common.c: ... here. New file.
2222 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
2224 (arm_early_store_addr_dep): Likewise.
2225 (arm_no_early_alu_shift_dep): Likewise.
2226 (arm_no_early_alu_shift_value_dep): Likewise.
2227 (arm_no_early_mul_dep): Likewise.
2228 (arm_no_early_store_addr_dep): Likewise.
2229 (arm_mac_accumulator_is_mul_result): Likewise.
2230 (arm_mac_accumulator_is_result): Likewise.
2231 * config/arm/aarch-common-protos.h: ... here. New file.
2232 * config/arm/t-arm (aarch-common.o): Define.
2234 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2236 * config/arm/arm.md: Include new file "types.md".
2237 (define_attr "type"): Move from here to ...
2238 (define_attr "mul32"): Likewise.
2239 (define_attr "mul64"): Likewise.
2240 * config/arm/types.md: ... here. New file.
2242 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
2244 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
2245 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
2247 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2249 * gen-pass-instances.awk: Fix offset of substr().
2251 2013-07-31 David Malcolm <dmalcolm@redhat.com>
2253 * Makefile.in (pass-instances.def): New.
2254 (passes.o): Replace dependency on passes.def with one on
2257 * gen-pass-instances.awk: New.
2259 * passes.c (pass_manager::pass_manager): Use pass-instances.def
2260 rather than passes.def, updating local definition of NEXT_PASS
2261 macro to add an extra NUM parameter (currently unused).
2263 2013-07-30 David Malcolm <dmalcolm@redhat.com>
2265 * Makefile.in (PASS_MANAGER_H): New.
2266 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
2267 (passes.o): Likewise.
2268 (statistics.o): Likewise.
2269 (cgraphunit.o): Likewise.
2270 (context.o): Depend on PASS_MANAGER_H.
2272 * pass_manager.h: New.
2274 * cgraphunit.c (cgraph_add_new_function): Update for moves
2275 of globals to fields of pass_manager.
2276 (analyze_function): Likewise.
2277 (expand_function): Likewise.
2278 (ipa_passes): Likewise.
2279 (compile): Likewise.
2281 * context.c (context::context): New.
2282 * context.h (context::context): New.
2283 (context::get_passes): New.
2284 (context::passes_): New.
2286 * lto-cgraph.c (input_node): Update for moves of globals to
2287 fields of pass_manager.
2289 * passes.c (all_passes): Remove, in favor of a field of the
2290 same name within the new class pass_manager.
2291 (all_small_ipa_passes): Likewise.
2292 (all_lowering_passes): Likewise.
2293 (all_regular_ipa_passes): Likewise.
2294 (all_late_ipa_passes): Likewise.
2295 (all_lto_gen_passes): Likewise.
2296 (passes_by_id): Likewise.
2297 (passes_by_id_size): Likewise.
2298 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2299 the new class pass_manager.
2300 (set_pass_for_id): Convert to...
2301 (pass_manager::set_pass_for_id): ...method.
2302 (get_pass_for_id): Convert to...
2303 (pass_manager::get_pass_for_id): ...method.
2304 (register_one_dump_file): Move body of implementation into...
2305 (pass_manager::register_one_dump_file): ...here.
2306 (register_dump_files_1): Convert to...
2307 (pass_manager::register_dump_files_1): ...method.
2308 (register_dump_files): Convert to...
2309 (pass_manager::register_dump_files): ...method.
2310 (create_pass_tab): Update for moves of globals to fields of
2312 (dump_passes): Move body of implementation into...
2313 (pass_manager::dump_passes): ...here.
2314 (register_pass): Move body of implementation into...
2315 (pass_manager::register_pass): ...here.
2316 (init_optimization_passes): Convert into...
2317 (pass_manager::pass_manager): ...constructor for new
2318 pass_manager class, and initialize the pass_lists array.
2319 (check_profile_consistency): Update for moves of globals to
2320 fields of pass_manager.
2321 (dump_profile_report): Move body of implementation into...
2322 (pass_manager::dump_profile_report): ...here.
2323 (ipa_write_summaries_1): Update for moves of pass lists from
2324 being globals to fields of pass_manager.
2325 (ipa_write_optimization_summaries): Likewise.
2326 (ipa_read_summaries): Likewise.
2327 (ipa_read_optimization_summaries): Likewise.
2328 (execute_all_ipa_stmt_fixups): Likewise.
2330 * statistics.c (statistics_fini): Update for moves of globals to
2331 fields of pass_manager.
2333 * toplev.c (general_init): Replace call to
2334 init_optimization_passes with construction of the pass_manager
2337 * tree-pass.h (all_passes): Remove, in favor of a field of the
2338 same name within the new class pass_manager.
2339 (all_small_ipa_passes): Likewise.
2340 (all_lowering_passes): Likewise.
2341 (all_regular_ipa_passes): Likewise.
2342 (all_lto_gen_passes): Likewise.
2343 (all_late_ipa_passes): Likewise.
2344 (passes_by_id): Likewise.
2345 (passes_by_id_size): Likewise.
2346 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2347 the new class pass_manager.
2348 (get_pass_for_id): Remove.
2350 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2352 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
2355 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2357 * arm.md (mulhi3): New expand pattern.
2359 2013-07-30 Jan Hubicka <jh@suse.cz>
2360 Martin Liska <marxin.liska@gmail.com>
2362 * profile.c (compute_value_histograms): Do not ICE when
2363 there is mismatch only on some counters.
2365 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2367 PR rtl-optimization/57637
2368 * function.c (move_insn_for_shrink_wrap): Also check the
2369 GEN set of the LIVE problem for the liveness analysis
2370 if it exists, otherwise give up.
2372 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2374 PR tree-optimization/57993
2375 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2376 replaced statement in the candidate table.
2377 (phi_add_costs): Return infinite cost when the hidden basis does
2378 not dominate all phis on which the candidate is dependent.
2379 (replace_one_candidate): Record replaced statement in the
2382 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2384 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2385 (ashlv2si3): New expander.
2386 (*ashlv2si3_i): New define_insn_and_split.
2387 * predicates.md (float_operation): Allow patterns with three
2390 PR rtl-optimization/58021
2391 * mode-switching.c (create_pre_exit): Always split off preceding
2392 insns if we are not at the basic block head.
2394 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2396 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2397 (UCLIBC_DYNAMIC_LINKER): New macro.
2398 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2400 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2401 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2403 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2404 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2405 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2406 for SF and DF modes. Use ieee_quad_format for TF mode.
2407 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2408 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2409 (mips_option_override): Handle `-mnan=legacy'.
2410 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2411 `-mabs=2008' and `-mnan=2008'.
2412 (OPTION_DEFAULT_SPECS): Add "nan" default.
2413 (ASM_SPEC): Handle `-mnan='.
2414 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2415 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2416 comment accordingly.
2417 (neg<mode>2): Likewise.
2418 * config/mips/mips.opt (mabs, mnan): New options.
2419 * doc/install.texi (Configuration): Document `--with-nan=' option.
2420 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2422 (MIPS Options): Document them.
2423 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2424 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2425 * configure: Regenerate.
2426 * config.in: Regenerate.
2428 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2430 * config/i386/i386.md (float post-reload splitters): Do not check
2431 for subregs of SSE registers.
2433 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2434 H.J. Lu <hongjiu.lu@intel.com>
2438 * config/i386/i386.md (post-reload splitter
2439 to avoid partial SSE reg dependency stalls): New pattern.
2441 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2443 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2445 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2447 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2448 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2450 2013-07-26 Cary Coutant <ccoutant@google.com>
2452 * dwarf2out.c (die_checksum_ordered): Don't include template
2453 instantiations in signature.
2454 (is_template_parameter): New function.
2455 (is_template_instantiation): New function.
2456 (generate_skeleton_bottom_up): Don't include template instantiations
2458 (generate_skeleton): Likewise.
2459 (break_out_comdat_types): Move recursive call to break out nested
2461 (prune_unused_types_mark_generic_parms_dies): Call
2462 is_template_parameter.
2464 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2466 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2467 uses vector registers.
2468 * config/aarch64/iterators.md: Add attributes rtn and vas.
2470 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2471 Richard Earnshaw <richard.earnshaw@arm.com>
2473 * combine.c (simplify_comparison): Re-canonicalize operands
2475 * config/arm/arm.md (movcond_addsi): New splitter.
2477 2013-07-25 Sterling Augustine <saugustine@google.com>
2479 * dwarf2out.c (size_of_pubnames): Move code to...
2480 (include_pubname_in_output): ...here. New.
2481 (want_pubnames): Rearrange.
2482 (output_pubnames): Call include_pubname_in_output. Move assertion.
2484 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2486 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2488 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2491 * doc/extend.texi: Remove obsolete builtins. Fix
2492 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2494 2013-07-25 Jan Hubicka <jh@suse.cz>
2496 * cgraph.c (release_function_body): Break out from ...
2497 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2499 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2500 old_tree in the map.
2501 (create_specialized_node): Update.
2502 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2504 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2505 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2506 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2507 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2510 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2512 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2513 addsi3_carryin_alt2_<optab>): Correct output template.
2515 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2517 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2518 Adjust for arm_restrict_it.
2519 Remove trailing whitespace.
2521 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2523 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2524 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2526 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2528 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2530 PR rtl-optimization/57960
2531 * lra-constraints.c (process_alt_operands): Use the right mode
2532 when checking strict_low.
2534 2013-07-25 Jan Hubicka <jh@suse.cz>
2536 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2537 * cgraph.c (cgraph_remove_node): Do not release function body
2538 when in cgraph streaming.
2539 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
2540 in other partitions are not considered reachable; fix handling of
2543 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2545 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2547 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2552 * config/arm/arm.md ("*sibcall_insn): Replace use of
2553 Ss with US. Adjust output for v5 and v4t.
2554 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
2556 * config/arm/constraints.md ("Ss"): Rename to US.
2558 2013-07-25 Terry Guo <terry.guo@arm.com>
2560 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2561 shift_add/shift_sub0/shift_sub1 RTXs.
2563 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2564 Anton Blanchard <anton@au1.ibm.com>
2566 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2567 (altivec_vpks<VI_char>ss): Likewise.
2568 (altivec_vpks<VI_char>us): Likewise.
2569 (altivec_vpku<VI_char>us): Likewise.
2570 (altivec_vpku<VI_char>um): Likewise.
2572 2013-07-24 David Malcolm <dmalcolm@redhat.com>
2574 Introduce context class.
2576 * Makefile.in (CONTEXT_H): New.
2577 (OBJS): Add context.o.
2578 (toplev.o): Add CONTEXT_H to dependencies.
2581 * toplev.c (general_init): Create the singleton gcc::context instance.
2587 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2589 PR rtl-optimization/57968
2590 * mode-switching.c (create_pre_exit): Allow instructions that
2591 don't set a return register to need a non-exit mode.
2593 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2594 Anton Blanchard <anton@au1.ibm.com>
2596 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2597 operands to vperm for little endian.
2598 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2599 of lvsl to create the control mask for a vperm for little endian.
2601 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2602 Anton Blanchard <anton@au1.ibm.com>
2604 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2605 two operands for little-endian.
2607 2013-07-23 Steve Ellcey <sellcey@mips.com>
2609 * config/mips/mips.c (mips_case_values_threshold): New.
2610 (TARGET_CASE_VALUES_THRESHOLD): Define.
2612 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2613 Anton Blanchard <anton@au1.ibm.com>
2615 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2616 selection of field for vector splat in little endian mode.
2618 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2620 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2621 expanders to rs6000.md.
2622 (ior<mode>3): Likewise.
2623 (and<mode>3): Likewise.
2624 (one_cmpl<mode>2): Likewise.
2625 (nor<mode>3): Likewise.
2626 (andc<mode>3): Likewise.
2627 (eqv<mode>3): Likewise.
2628 (nand<mode>3): Likewise.
2629 (orc<mode>3): Likewise.
2631 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2634 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2635 to split multi-word logical operations.
2636 (rs6000_split_logical_di): Likewise.
2637 (rs6000_split_logical): Likewise.
2639 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2640 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2641 and allow TImode operations in 32-bit.
2642 (vsx_and<mode>3_64bit): Likewise.
2643 (vsx_ior<mode>3_32bit): Likewise.
2644 (vsx_ior<mode>3_64bit): Likewise.
2645 (vsx_xor<mode>3_32bit): Likewise.
2646 (vsx_xor<mode>3_64bit): Likewise.
2647 (vsx_one_cmpl<mode>2_32bit): Likewise.
2648 (vsx_one_cmpl<mode>2_64bit): Likewise.
2649 (vsx_nor<mode>3_32bit): Likewise.
2650 (vsx_nor<mode>3_64bit): Likewise.
2651 (vsx_andc<mode>3_32bit): Likewise.
2652 (vsx_andc<mode>3_64bit): Likewise.
2653 (vsx_eqv<mode>3_32bit): Likewise.
2654 (vsx_eqv<mode>3_64bit): Likewise.
2655 (vsx_nand<mode>3_32bit): Likewise.
2656 (vsx_nand<mode>3_64bit): Likewise.
2657 (vsx_orc<mode>3_32bit): Likewise.
2658 (vsx_orc<mode>3_64bit): Likewise.
2660 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2661 logical types in GPRs.
2663 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2664 logical insns to rs6000.md, and allow TImode operations in
2666 (altivec_ior<mode>3): Likewise.
2667 (altivec_xor<mode>3): Likewise.
2668 (altivec_one_cmpl<mode>2): Likewise.
2669 (altivec_nor<mode>3): Likewise.
2670 (altivec_andc<mode>3): Likewise.
2672 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2673 attributes for moving the 128-bit logical operations into
2675 (BOOL_REGS_OUTPUT): Likewise.
2676 (BOOL_REGS_OP1): Likewise.
2677 (BOOL_REGS_OP2): Likewise.
2678 (BOOL_REGS_UNARY): Likewise.
2679 (BOOL_REGS_AND_CR0): Likewise.
2680 (one_cmpl<mode>2): Add support for DI logical operations on
2681 32-bit, splitting the operations to 32-bit.
2685 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2686 changes to combine the 32/64-bit code, allow logical operations on
2687 TI mode in 32-bit, and to use similar match_operator patterns like
2688 scalar mode uses. Combine the Altivec and VSX code for logical
2689 operations, and move it here.
2690 (ior<mode>3, 128-bit types): Likewise.
2691 (xor<mode>3, 128-bit types): Likewise.
2692 (one_cmpl<mode>3, 128-bit types): Likewise.
2693 (nor<mode>3, 128-bit types): Likewise.
2694 (andc<mode>3, 128-bit types): Likewise.
2695 (eqv<mode>3, 128-bit types): Likewise.
2696 (nand<mode>3, 128-bit types): Likewise.
2697 (orc<mode>3, 128-bit types): Likewise.
2698 (and<mode>3_internal): Likewise.
2699 (bool<mode>3_internal): Likewise.
2700 (boolc<mode>3_internal1): Likewise.
2701 (boolc<mode>3_internal2): Likewise.
2702 (boolcc<mode>3_internal1): Likewise.
2703 (boolcc<mode>3_internal2): Likewise.
2704 (eqv<mode>3_internal1): Likewise.
2705 (eqv<mode>3_internal2): Likewise.
2706 (one_cmpl1<mode>3_internal): Likewise.
2708 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2710 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2711 Rename flag_stack_usage to flag_stack_usage_info.
2713 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2715 * config/microblaze/sync.md: New file.
2716 * config/microblaze/microblaze.md: Include sync.md
2717 * config/microblaze/microblaze.c: Add print_operand 'y'.
2718 * config/microblaze/constraints.md: Add memory_contraint
2719 'Q' which is a single register.
2721 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2723 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2725 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2727 * reload.c (find_reloads): Exit loop once we find this operand
2728 cannot be reloaded somehow for this alternative.
2730 * reload.c (find_reloads): Exit loop once we find a hard register.
2732 * rtlanal.c (computed_jump_p): Exit loop once we find label
2735 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2737 * cfgloopmanip.c (remove_path): Exit loop after setting
2740 * gensupport.c (subst_dup): Avoid loop if code is not
2741 MATCH_DUP nor MATCH_OP_DUP.
2743 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2745 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2747 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2749 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2750 true for SP_REGNUM if mode == ptr_mode.
2751 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2752 with value R0_REGNUM + 31.
2754 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2756 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2757 pad pointer-typed argument downward.
2759 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2761 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2762 and __ILP32__ when the ILP32 model is in use.
2764 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2766 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2767 (aarch64_load_symref_appropriately): In the case of
2768 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2769 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2770 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2771 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2772 if 'mode' doesn't equal to 'ptr_mode'.
2773 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2774 'vcall_offset'; change to call aarch64_emit_move differently depending
2775 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2776 to calculate the upper bound of 'vcall_offset'.
2777 (aarch64_cannot_force_const_mem): Change to also return true if
2779 (aarch64_legitimize_reload_address): In the case of large
2780 displacements, add new local variable 'xmode' and an assertion
2781 based on it; change to use 'xmode' to generate the new rtx and
2783 (aarch64_asm_trampoline_template): Change to generate the template
2784 differently depending on TARGET_ILP32 or not; change to use
2785 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2786 (aarch64_trampoline_size): Removed.
2787 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2788 and replace immediate literals with it. Change to use 'ptr_mode'
2789 instead of 'DImode' and call convert_memory_address if the mode
2790 of 'fnaddr' doesn't equal to 'ptr_mode'.
2791 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2793 (aarch64_elf_asm_destructor): Likewise.
2794 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2795 on TARGET_ILP32 instead of aarch64_trampoline_size.
2796 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2797 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2798 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2799 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2800 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2801 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2802 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2803 depending on the value of 'mode'.
2804 (add_losym_<mode>): New.
2805 (ldr_got_small_<mode>): New, based on ldr_got_small.
2806 (ldr_got_small): Remove.
2807 (ldr_got_small_sidi): New.
2808 * config/aarch64/iterators.md (P): New.
2809 (PTR): Change to 'ptr_mode' in the condition.
2811 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2813 * config.gcc (aarch64*-*-*): Support --with-abi.
2814 (aarch64*-*-elf): Support --with-multilib-list.
2815 (aarch64*-*-linux*): Likewise.
2816 (supported_defaults): Add abi to aarch64*-*-*.
2817 * configure.ac: Mention AArch64 for --with-multilib-list.
2818 * configure: Re-generated.
2819 * config/aarch64/biarchilp32.h: New file.
2820 * config/aarch64/biarchlp64.h: New file.
2821 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2823 (MULTILIB_DEFAULTS): Ditto.
2824 (DRIVER_SELF_SPECS): Ditto.
2825 (ASM_SPEC): Update to also substitute -mabi.
2826 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2827 file whose name depends on -mabi= and -mbig-endian.
2828 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2830 (POINTER_SIZE): New define.
2831 (POINTERS_EXTEND_UNSIGNED): Ditto.
2832 (enum aarch64_abi_type): New enumeration tag.
2833 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2834 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2835 (TARGET_ILP32): New define.
2836 * config/aarch64/aarch64.opt (mabi): New.
2838 (ilp32, lp64): New values for -mabi.
2839 * config/aarch64/t-aarch64 (comma): New define.
2840 (MULTILIB_OPTIONS): Ditto.
2841 (MULTILIB_DIRNAMES): Ditto.
2842 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2843 * doc/invoke.texi: Document -mabi for AArch64.
2845 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
2847 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2849 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2850 Anton Blanchard <anton@au1.ibm.com>
2852 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2853 endianness when selecting field to splat.
2855 2013-07-22 Eric Christopher <echristo@gmail.com>
2857 * dwarf2out.c (die_odr_checksum): New function to use
2858 CHECKSUM_ macros and ULEB128 for DIE tag.
2859 (generate_type_signature): Use.
2861 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2863 * config.gcc (sparc*-*-*): Accept leon3 processor.
2864 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2865 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2866 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2867 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2868 (mfix-ut699): Adjust comment.
2869 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2870 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2871 (CPP_CPU_SPEC): Likewise.
2872 (ASM_CPU_SPEC): Likewise.
2873 * config/sparc/sparc.c (leon3_cost): New constant.
2874 (sparc_option_override): Add leon3 support.
2875 (mem_ref): New function.
2876 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2877 (sparc_do_work_around_errata): Look into the instruction in the delay
2878 slot and adjust accordingly. Add fix for the data cache nullify issues
2879 of the UT699. Change insertion position for the NOP.
2880 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2881 (leon3_load): New reservation.
2882 (leon_store): Bump latency to 2.
2883 (grfpu): New automaton.
2884 (grfpu_alu): New unit.
2885 (grfpu_ds): Likewise.
2886 (leon_fp_alu): Adjust.
2887 (leon_fp_mult): Delete.
2888 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2889 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2890 * config/sparc/sparc.md (cpu): Add leon3.
2891 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2893 (atomic_test_and_set): Likewise.
2896 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2898 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2899 default for R5900 targets.
2900 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2901 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2902 * config/mips/mips.c (mips_option_override): Report an error for
2903 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2904 for -march=r5900 -mhard-float.
2906 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2908 * df-problems.c (can_move_insns_across): Exit loop once we
2909 find a non-fixed, non-global register.
2911 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2914 * omega.c (omega_eliminate_red): Break after setting red_found.
2915 (omega_problem_has_red_equations): Similarly after setting found.
2916 (omega_query_variable): Similarly after setting coupled.
2918 2013-07-22 Marek Polacek <polacek@redhat.com>
2920 * gimplify.c: Don't include gimple.h twice.
2922 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2924 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2925 instead of TARGET_THUMB1.
2926 (Pz): New constraint.
2927 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2929 (compare_negsi_si): Likewise.
2930 (compare_addsi2_op0): Likewise.
2931 (compare_addsi2_op1): Likewise.
2932 (addsi3_carryin_<optab>): Likewise.
2933 (addsi3_carryin_alt2_<optab>): Likewise.
2934 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2935 for arm_restrict_it.
2936 (subsi3_carryin): Likewise.
2937 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2938 (minmax_arithsi): Disable for arm_restrict_it.
2939 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2940 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2941 (satsi_<SAT:code>_shift): Likewise.
2942 (arm_shiftsi3): Add alternative for 16-bit encoding.
2943 (arm32_movhf): Disable for arm_restrict_it.
2944 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2945 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2947 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2949 * config/arm/arm.md (attribute "insn"): Delete.
2950 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2951 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2952 (not_shiftsi): Update for attribute change.
2953 (not_shiftsi_compare0): Likewise.
2954 (not_shiftsi_compare0_scratch): Likewise.
2955 (arm_one_cmplsi2): Likewise.
2956 (thumb1_one_cmplsi2): Likewise.
2957 (notsi_compare0): Likewise.
2958 (notsi_compare0_scratch): Likewise.
2959 (thumb1_movdi_insn): Likewise.
2960 (arm_movsi_insn): Likewise.
2961 (movhi_insn_arch4): Likewise.
2962 (movhi_bytes): Likewise.
2963 (arm_movqi_insn): Likewise.
2964 (thumb1_movqi_insn): Likewise.
2965 (arm32_movhf): Likewise.
2966 (thumb1_movhf): Likewise.
2967 (arm_movsf_soft_insn): Likewise.
2968 (thumb1_movsf_insn): Likewise.
2969 (thumb_movdf_insn): Likewise.
2970 (movsicc_insn): Likewise.
2971 (movsfcc_soft_insn): Likewise.
2972 (and_scc): Likewise.
2973 (cond_move): Likewise.
2974 (if_move_not): Likewise.
2975 (if_not_move): Likewise.
2976 (if_shift_move): Likewise.
2977 (if_move_shift): Likewise.
2978 (if_shift_shift): Likewise.
2979 (if_not_arith): Likewise.
2980 (if_arith_not): Likewise.
2981 (cond_move_not): Likewise.
2982 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
2983 (neon_mov<mode>): Likewise.
2984 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
2985 (thumb2_movsi_vfp): Likewise.
2986 (movsf_vfp): Likewise.
2987 (thumb2_movsf_vfp): Likewise.
2988 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
2990 (cortexa7_older_only): Likewise.
2991 (cortexa7_younger): Likewise.
2992 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
2993 (1020alu_shift_op): Likewise.
2994 (1020alu_shift_reg_op): Likewise.
2995 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
2996 (alu_shift_op): Likewise.
2997 (alu_shift_reg_op): Likewise.
2998 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
2999 (11_alu_shift_op): Likewise.
3000 (11_alu_shift_reg_op): Likewise.
3001 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3002 (9_alu_shift_reg_op): Likewise.
3003 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3005 (cortex_a15_alu_shift): Likewise.
3006 (cortex_a15_alu_shift_reg): Likewise.
3007 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3008 (cortex_a5_alu_shift): Likewise.
3009 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3011 (cortex_a53_alu_shift): Likewise.
3012 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3014 (cortex_a7_alu_reg): Likewise.
3015 (cortex_a7_alu_shift): Likewise.
3016 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3017 (cortex_a8_alu_shift): Likewise.
3018 (cortex_a8_alu_shift_reg): Likewise.
3019 (cortex_a8_mov): Likewise.
3020 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3021 (cortex_a9_dp_shift): Likewise.
3022 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3023 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3024 (cortex_r4_mov): Likewise.
3025 (cortex_r4_alu_shift): Likewise.
3026 (cortex_r4_alu_shift_reg): Likewise.
3027 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3028 (526_alu_shift_op): Likewise.
3029 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3030 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3031 (626te_alu_shift_op): Likewise.
3032 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3033 (726te_alu_op): Likewise.
3034 (726te_alu_shift_op): Likewise.
3035 (726te_alu_shift_reg_op): Likewise.
3036 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3037 (mp626_alu_shift_op): Likewise.
3038 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3039 (pj4_alu_e1_conds): Likewise.
3040 (pj4_alu): Likewise.
3041 (pj4_alu_conds): Likewise.
3042 (pj4_shift): Likewise.
3043 (pj4_shift_conds): Likewise.
3044 (pj4_alu_shift): Likewise.
3045 (pj4_alu_shift_conds): Likewise.
3047 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3049 * config/arm/predicates.md (shiftable_operator_strict_it):
3051 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3052 Disable cond_exec version for arm_restrict_it.
3053 (thumb2_smaxsi3): Convert to generate cond_exec.
3054 (thumb2_sminsi3): Likewise.
3055 (thumb32_umaxsi3): Likewise.
3056 (thumb2_uminsi3): Likewise.
3057 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3058 (thumb2_neg_abssi2): Likewise.
3059 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3060 (thumb2_movsicc_insn): Adjust alternatives.
3061 (thumb2_mov_negscc): Disable for arm_restrict_it.
3062 (thumb2_mov_negscc_strict_it): New pattern.
3063 (thumb2_mov_notscc_strict_it): New pattern.
3064 (thumb2_mov_notscc): Disable for arm_restrict_it.
3065 (thumb2_ior_scc): Likewise.
3066 (thumb2_ior_scc_strict_it): New pattern.
3067 (thumb2_cond_move): Adjust for arm_restrict_it.
3068 (thumb2_cond_arith): Disable for arm_restrict_it.
3069 (thumb2_cond_arith_strict_it): New pattern.
3070 (thumb2_cond_sub): Adjust for arm_restrict_it.
3071 (thumb2_movcond): Likewise.
3072 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3073 (thumb2_zero_extendhisi2_v6): Likewise.
3074 (thumb2_zero_extendqisi2_v6): Likewise.
3075 (orsi_notsi_si): Likewise.
3076 (orsi_not_shiftsi_si): Likewise.
3078 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3080 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3081 instruction sequence is 1 byte shorter.
3083 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3085 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3086 it is not needed after split.
3088 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3091 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3092 second label for nonlocal goto receivers. Don't output pic base labels
3093 unless we're producing PIC; mark that action unreachable().
3094 (ix86_save_reg): If the function contains a nonlocal label, save the
3096 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3097 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3098 (update_pic_label_number_if_needed): New.
3099 (machopic_output_function_base_name): Adjust for nonlocal receiver
3101 (machopic_should_output_picbase_label): New.
3102 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3103 (nonlocal_goto_receiver): New insn and split.
3105 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3107 * config/aarch64/aarch64-builtins.c
3108 (aarch64_fold_builtin): Fold abs in all modes.
3109 * config/aarch64/aarch64-simd-builtins.def
3110 (abs): Enable for all modes.
3111 * config/aarch64/arm_neon.h
3112 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
3113 (vabs_f64): Add missing intrinsic.
3115 2013-07-19 Ian Bolton <ian.bolton@arm.com>
3117 * config/aarch64/arm_neon.h (vabs_s64): New function
3119 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
3122 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
3123 * config/avr/avr.md (adjust_len): Add `round'.
3124 * config/avr/avr-protos.h (avr_out_round): New prototype.
3125 (avr_out_plus): Add `out_label' argument.
3126 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
3127 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
3128 Handle the case where `insn' is just a pattern.
3129 (avr_out_bitop): Handle the case where `insn' is just a pattern.
3130 (avr_out_round): New function.
3131 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
3133 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
3135 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3136 Add check for flag_stack_usage to handle -fstack-usage support
3138 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
3140 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
3141 interaction for new Power8 flags and VSX.
3143 2013-07-18 Sriraman Tallam <tmsriram@google.com>
3146 * tree-inline.c (expand_call_inline): Emit errors during
3147 early_inlining only if optimization is not turned on.
3149 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3153 * passes.c (init_optimization_passes): Move the construction of
3154 the pass hierarchy into a new passes.def file.
3156 * Makefile.in (passes.o): Add dependency on passes.def.
3158 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3160 * passes.c (init_optimization_passes): Introduce macros for
3161 constructing the tree of passes (INSERT_PASSES_AFTER,
3162 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
3163 TERMINATE_PASS_LIST).
3165 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
3166 Wei Mi <wmi@google.com>
3168 PR rtl-optimization/57878
3169 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
3171 (reload_pseudo_compare_func): Check nregs first for reload
3174 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3176 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
3178 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
3180 * read-rtl.c (validate_const_int): Once an invalid character is
3181 seen, quit the loop.
3183 * gengtype.c (write_roots): Similarly once we find the "deletable"
3184 or "if_marked" option.
3186 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3188 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
3189 "xtab" and "sat". Move value "clz" from here to ...
3190 (attriubte "type"): ... here.
3191 (satsi_<SAT:code>): Delete "insn" attribute.
3192 (satsi_<SAT:code>_shift): Likewise.
3193 (arm_zero_extendqisi2addsi): Likewise.
3194 (arm_extendqisi2addsi): Likewise.
3195 (clzsi2): Update for attribute changes.
3196 (rbitsi2): Likewise.
3197 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
3199 (arm_usatsihi): Likewise.
3200 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3202 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3204 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
3205 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
3206 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
3207 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
3208 in alphabetical order.
3209 (attribute "core_cycles"): Update for attribute changes.
3210 (arm_addsi3): Likewise.
3211 (addsi3_compare0): Likewise.
3212 (addsi3_compare0_scratch): Likewise.
3213 (addsi3_compare_op1): Likewise.
3214 (addsi3_compare_op2): Likewise.
3215 (compare_addsi2_op0): Likewise.
3216 (compare_addsi2_op1): Likewise.
3217 (addsi3_carryin_shift_<optab>): Likewise.
3218 (subsi3_carryin_shift): Likewise.
3219 (rsbsi3_carryin_shift): Likewise.
3220 (arm_subsi3_insn): Likewise.
3221 (subsi3_compare0): Likewise.
3222 (subsi3_compare): Likewise.
3223 (arm_andsi3_insn): Likewise.
3224 (thumb1_andsi3_insn): Likewise.
3225 (andsi3_compare0): Likewise.
3226 (andsi3_compare0_scratch): Likewise.
3227 (zeroextractsi_compare0_scratch
3228 (andsi_not_shiftsi_si): Likewise.
3229 (iorsi3_insn): Likewise.
3230 (iorsi3_compare0): Likewise.
3231 (iorsi3_compare0_scratch): Likewise.
3232 (arm_xorsi3): Likewise.
3233 (thumb1_xorsi3_insn): Likewise.
3234 (xorsi3_compare0): Likewise.
3235 (xorsi3_compare0_scratch): Likewise.
3236 (satsi_<SAT:code>_shift): Likewise.
3238 (arm_shiftsi3): Likewise.
3239 (shiftsi3_compare0): Likewise.
3240 (not_shiftsi): Likewise.
3241 (not_shiftsi_compare0): Likewise.
3242 (not_shiftsi_compare0_scratch): Likewise.
3243 (arm_one_cmplsi2): Likewise.
3244 (thumb_one_complsi2): Likewise.
3245 (notsi_compare0): Likewise.
3246 (notsi_compare0_scratch): Likewise.
3247 (thumb1_zero_extendhisi2): Likewise.
3248 (arm_zero_extendhisi2): Likewise.
3249 (arm_zero_extendhisi2_v6): Likewise.
3250 (arm_zero_extendhisi2addsi): Likewise.
3251 (thumb1_zero_extendqisi2): Likewise.
3252 (thumb1_zero_extendqisi2_v6): Likewise.
3253 (arm_zero_extendqisi2): Likewise.
3254 (arm_zero_extendqisi2_v6): Likewise.
3255 (arm_zero_extendqisi2addsi): Likewise.
3256 (thumb1_extendhisi2): Likewise.
3257 (arm_extendhisi2): Likewise.
3258 (arm_extendhisi2_v6): Likewise.
3259 (arm_extendqisi): Likewise.
3260 (arm_extendqisi_v6): Likewise.
3261 (arm_extendqisi2addsi): Likewise.
3262 (thumb1_extendqisi2): Likewise.
3263 (thumb1_movdi_insn): Likewise.
3264 (arm_movsi_insn): Likewise.
3265 (movsi_compare0): Likewise.
3266 (movhi_insn_arch4): Likewise.
3267 (movhi_bytes): Likewise.
3268 (arm_movqi_insn): Likewise.
3269 (thumb1_movqi_insn): Likewise.
3270 (arm32_movhf): Likewise.
3271 (thumb1_movhf): Likewise.
3272 (arm_movsf_soft_insn): Likewise.
3273 (thumb1_movsf_insn): Likewise.
3274 (movdf_soft_insn): Likewise.
3275 (thumb_movdf_insn): Likewise.
3276 (arm_cmpsi_insn): Likewise.
3277 (cmpsi_shiftsi): Likewise.
3278 (cmpsi_shiftsi_swp): Likewise.
3279 (arm_cmpsi_negshiftsi_si): Likewise.
3280 (movsicc_insn): Likewise.
3281 (movsfcc_soft_insn): Likewise.
3282 (arith_shiftsi): Likewise.
3283 (arith_shiftsi_compare0
3284 (arith_shiftsi_compare0_scratch
3285 (sub_shiftsi): Likewise.
3286 (sub_shiftsi_compare0
3287 (sub_shiftsi_compare0_scratch
3288 (and_scc): Likewise.
3289 (cond_move): Likewise.
3290 (if_plus_move): Likewise.
3291 (if_move_plus): Likewise.
3292 (if_move_not): Likewise.
3293 (if_not_move): Likewise.
3294 (if_shift_move): Likewise.
3295 (if_move_shift): Likewise.
3296 (if_shift_shift): Likewise.
3297 (if_not_arith): Likewise.
3298 (if_arith_not): Likewise.
3299 (cond_move_not): Likewise.
3300 (thumb1_ashlsi3): Set type attribute.
3301 (thumb1_ashrsi3): Likewise.
3302 (thumb1_lshrsi3): Likewise.
3303 (thumb1_rotrsi3): Likewise.
3304 (shiftsi3_compare0_scratch): Likewise.
3305 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
3306 (neon_mov<mode>): Likewise.
3307 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
3309 (thumb2_movsi_insn): Likewise.
3310 (thumb2_cmpsi_neg_shiftsi): Likewise.
3311 (thumb2_extendqisi_v6): Likewise.
3312 (thumb2_zero_extendhisi2_v6): Likewise.
3313 (thumb2_zero_extendqisi2_v6): Likewise.
3314 (thumb2_shiftsi3_short): Likewise.
3315 (thumb2_addsi3_compare0_scratch): Likewise.
3316 (orsi_not_shiftsi_si): Likewise.
3317 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3318 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
3320 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3321 (1020alu_shift_op): Likewise.
3322 (1020alu_shift_reg_op): Likewise.
3323 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3324 (alu_shift_op): Likewise.
3325 (alu_shift_reg_op): Likewise.
3326 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3327 (11_alu_shift_op): Likewise.
3328 (11_alu_shift_reg_op): Likewise.
3329 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3330 (9_alu_shift_reg_op): Likewise.
3331 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3333 (cortex_a15_alu_shift): Likewise.
3334 (cortex_a15_alu_shift_reg): Likewise.
3335 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
3337 (cortex_a5_alu_shift): Likewise.
3338 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
3340 (cortex_a53_alu_shift): Likewise.
3341 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3343 (cortex_a7_alu_reg): Likewise.
3344 (cortex_a7_alu_shift): Likewise.
3345 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
3347 (cortex_a8_alu_shift): Likewise.
3348 (cortex_a8_alu_shift_reg): Likewise.
3349 (cortex_a8_mov): Likewise.
3350 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3351 (cortex_a9_dp_shift): Likewise.
3352 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
3354 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
3356 (cortex_r4_mov): Likewise.
3357 (cortex_r4_alu_shift): Likewise.
3358 (cortex_r4_alu_shift_reg): Likewise.
3359 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3360 (526_alu_shift_op): Likewise.
3361 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3362 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3363 (626te_alu_shift_op): Likewise.
3364 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3365 (726te_alu_op): Likewise.
3366 (726te_alu_shift_op): Likewise.
3367 (726te_alu_shift_reg_op): Likewise.
3368 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3369 (mp626_alu_shift_op): Likewise.
3370 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3371 (pj4_alu_e1_conds): Likewise.
3372 (pj4_alu): Likewise.
3373 (pj4_alu_conds): Likewise.
3374 (pj4_shift): Likewise.
3375 (pj4_shift_conds): Likewise.
3376 (pj4_alu_shift): Likewise.
3377 (pj4_alu_shift_conds): Likewise.
3378 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3380 (cortexa7_older_only): Likewise.
3381 (cortexa7_younger): Likewise.
3383 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3385 * ipa-pure-const.c (generate_summary): Rename to...
3386 (pure_const_generate_summary): ... this.
3388 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3390 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3392 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3395 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3398 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3400 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3401 enabled without -march=zEC12.
3402 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3405 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3407 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3408 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3409 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3410 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3411 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3412 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3413 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3415 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3416 and ISA_HAS_NMADD3_NMSUB3.
3417 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3418 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3419 (nmsub4<mode>, nmsub3<mode>): Likewise.
3420 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3422 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3424 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3425 TARGET_MIPS5400 checking.
3427 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3428 Peter Bergner <bergner@vnet.ibm.com>
3430 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3431 registers in the comment.
3432 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3433 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3434 rather than FIRST_PSEUDO_REGISTERS.
3436 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3438 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3439 enable extra ISA flags with TARGET_HTM.
3441 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3443 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3446 2013-07-15 Cong Hou <congh@google.com>
3448 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3449 in compare function for sorting.
3451 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3453 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3454 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3455 * config/rs6000/rs6000.opt: Add -mhtm option.
3456 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3457 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3458 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3459 __HTM__ if the HTM instructions are available.
3460 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3461 htm_spr_reg_operand): New define_predicates.
3462 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3463 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3465 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3466 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3467 HTM builtin functions.
3468 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3469 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3470 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3471 (rs6000_builtin_mask_calculate): Likewise.
3472 (rs6000_option_override_internal): Likewise.
3473 (bdesc_htm): Add new HTM builtin support.
3474 (htm_spr_num): New function.
3475 (htm_spr_regno): Likewise.
3476 (rs6000_htm_spr_icode): Likewise.
3477 (htm_expand_builtin): Likewise.
3478 (htm_init_builtins): Likewise.
3479 (rs6000_expand_builtin): Add support for HTM builtin functions.
3480 (rs6000_init_builtins): Likewise.
3481 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3483 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3484 (TARGET_HTM, MASK_HTM): Define macros.
3485 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3486 (FIXED_REGISTERS): Likewise.
3487 (CALL_USED_REGISTERS): Likewise.
3488 (CALL_REALLY_USED_REGISTERS): Likewise.
3489 (REG_ALLOC_ORDER): Likewise.
3490 (enum reg_class): Likewise.
3491 (REG_CLASS_NAMES): Likewise.
3492 (REG_CLASS_CONTENTS): Likewise.
3493 (REGISTER_NAMES): Likewise.
3494 (ADDITIONAL_REGISTER_NAMES): Likewise.
3495 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3496 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3497 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3498 * config/rs6000/htm.md: New file.
3499 * config/rs6000/htmintrin.h: New file.
3500 * config/rs6000/htmxlintrin.h: New file.
3502 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3504 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3505 Define SYMBOL_TINY_GOT, update comment.
3506 * config/aarch64/aarch64.c
3507 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3508 (aarch64_expand_mov_immediate): Likewise.
3509 (aarch64_print_operand): Likewise.
3510 (aarch64_classify_symbol): Likewise.
3511 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3512 (ldr_got_tiny): Define.
3514 2013-07-13 Tobias Grosser <tobias@grosser.es>
3516 PR tree-optimization/54094
3517 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3518 scheduling dimension for the parallelism check from the polyhedral
3519 information in the AST.
3520 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3523 2013-07-13 Jason Merrill <jason@redhat.com>
3525 * print-tree.c (debug_vec_tree): Use debug_raw.
3526 (debug_raw (vec<tree, va_gc> &)): New.
3527 (debug_raw (vec<tree, va_gc> *)): New.
3528 * tree.h: Declare them.
3530 2013-07-13 Bin Cheng <bin.cheng@arm.com>
3532 * ifcvt.c (ifcvt_after_combine): New static variable.
3533 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3535 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3536 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3537 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3539 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3541 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3543 2013-07-12 Michael Matz <matz@suse.de>
3546 * convert.c (convert_to_real): Reject non-float inner types.
3548 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3550 * config/aarch64/aarch64-protos.h
3551 (aarch64_simd_immediate_valid_for_move): Remove.
3552 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3553 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3555 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3557 2013-07-11 Steve Ellcey <sellcey@mips.com>
3559 * config/mips/mips.c (mips_conditional_register_usage): Do not
3560 use t[0-7] registers in MIPS16 mode when optimizing for size.
3562 2013-07-11 Sriraman Tallam <tmsriram@google.com>
3564 * config/i386/i386.c (dispatch_function_versions): Fix array
3565 indexing of function_version_info to match actual_versions.
3567 2013-07-11 Teresa Johnson <tejohnson@google.com>
3569 * vec.h (struct va_gc): Move release out-of-line.
3570 (va_gc::release): Call ggc_free on released vec.
3572 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3574 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3575 Require GOT register as additional operand in UNSPEC.
3576 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3577 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3578 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3579 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3580 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3581 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3582 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3584 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
3587 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3588 name seen by assembler/linker rather if available.
3590 2013-07-11 Andreas Schwab <schwab@suse.de>
3592 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3594 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3596 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
3598 2013-07-10 Joseph Myers <joseph@codesourcery.com>
3600 * doc/tm.texi.in: Move hook documentation to ....
3601 * target.def: ... here.
3603 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3605 * doc/tm.texi: Regenerate.
3607 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3610 * doc/invoke.texi: Document -Wconditionally-supported.
3612 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3615 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3618 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3621 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3622 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3623 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3624 Remove duplicate devices.
3625 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3626 * config/avr/t-multilib: Regenerate.
3627 * config/avr/avr-tables.opt: Regenerate.
3628 * doc/avr-mmcu.texi: Regenerate.
3630 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3633 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3635 2013-07-10 Graham Stott <graham.stott@btinternet.com>
3637 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3638 the cost of MULT when optimizing for size.
3640 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3642 * config/cr16/cr16-protos.h: Don't include target.h.
3644 2013-07-09 Joseph Myers <joseph@codesourcery.com>
3646 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3647 adjust register size for TDmode and TFmode for VSX registers.
3649 2013-07-08 Kai Tietz <ktietz@redhat.com>
3652 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3653 hook_bool_const_tree_true.
3655 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3657 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3658 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3659 * config/s390/s390.md: Define FPR*_REGNUM constants.
3660 Fix FPR2_REGNUM constant (18 -> 17).
3661 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3662 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3663 ("trunc<BFP:mode><DFP_ALL:mode>2")
3664 ("trunc<DFP_ALL:mode><BFP:mode>2")
3665 ("*extend<BFP:mode><DFP_ALL:mode>2")
3666 ("*extend<DFP_ALL:mode><BFP:mode>2")
3667 ("extend<BFP:mode><DFP_ALL:mode>2")
3668 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3671 2013-07-08 Graham Stott <graham.stott@btinternet.com>
3673 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
3675 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3677 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3678 and cfun_fpr_bit_p to cfun_fpr_save_p.
3679 (s390_frame_area, s390_register_info, s390_frame_info)
3680 (s390_emit_prologue, s390_emit_epilogue)
3681 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3683 * config/s390/s390.h: Define *_REGNUM macros for floating point
3686 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3688 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3690 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3692 PR rtl-optimization/57786
3693 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3694 and break out of the loop when it is set to false.
3696 2013-07-08 Jakub Jelinek <jakub@redhat.com>
3699 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3700 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3701 (const_int 63)) 0)).
3702 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3703 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3704 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3706 PR rtl-optimization/57829
3707 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3708 mask bits outside of mode are just sign-extension from mode to HWI.
3710 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3712 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3713 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3714 adjust_address instead of change_address to keep info about alignment.
3715 (emit_strmov): Remove.
3716 (emit_memmov): New function.
3717 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3718 (expand_movmem_epilogue): Likewise and return updated rtx for
3720 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3721 destination and source.
3722 (decide_alignment): Refactor, handle vector_loop.
3723 (ix86_expand_movmem): Likewise.
3724 (ix86_expand_setmem): Likewise.
3725 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3727 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
3729 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3730 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3732 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3734 * config/i386/sse.md (sse_movlhps): Change alternative 3
3735 of operand 2 to "m".
3737 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3740 * config/i386/sse.md (iptr): New mode attribute.
3741 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3742 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3743 (<sse>_vmmul<mode>3): Ditto.
3744 (<sse>_vmdiv<mode>3): Ditto.
3745 (sse_vmrcpv4sf2): Ditto.
3746 (<sse>_vmsqrt<mode>2): Ditto.
3747 (sse_vmrsqrtv4sf2): Ditto.
3748 (<sse>_vm<code><mode>3): Ditto.
3749 (avx_vmcmp<mode>3): Ditto.
3750 (<sse>_vmmaskcmp<mode>3): Ditto.
3751 (<sse>_comi): Ditto.
3752 (<sse>_ucomi): Ditto.
3753 (*xop_vmfrcz_<mode>): Ditto.
3754 (*fmai_fmadd_<mode>): Ditto.
3755 (*fmai_fmsub_<mode>): Ditto.
3756 (*fmai_fnmadd_<mode>): Ditto.
3757 (*fmai_fnmsub_<mode>): Ditto.
3758 (*fma4i_vmfmadd_<mode>): Ditto.
3759 (*fma4i_vmfmsub_<mode>): Ditto.
3760 (*fma4i_vmfnmadd_<mode>): Ditto.
3761 (*fma4i_vmfnmsub_<mode>): Ditto.
3762 (*xop_vmfrcz_<mode>): Ditto.
3763 (sse_cvtps2pi): Ditto.
3764 (sse_cvttps2pi): Ditto.
3765 (sse_cvtss2si): Ditto.
3766 (sse_cvtss2si_2): Ditto.
3767 (sse_cvtss2siq_2): Ditto.
3768 (sse_cvttss2si): Ditto.
3769 (sse_cvttss2siq): Ditto.
3770 (sse_cvtsd2si): Ditto.
3771 (sse_cvtsd2si_2): Ditto.
3772 (sse_cvtsd2siq_2): Ditto.
3773 (sse_cvttsd2si): Ditto.
3774 (sse_cvttsd2siq): Ditto.
3775 (sse_cvtsd2ss): Ditto.
3776 (sse_cvtss2sd): Ditto.
3777 (avx2_pbroadcast<mode>): Ditto.
3778 (avx2_pbroadcast<mode>_1): Ditto.
3779 (*avx_vperm_broadcast_v4sf): Ditto.
3781 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3782 (sse_movlhps): Ditto.
3783 (sse_storehps): Ditto.
3784 (sse_loadhps): Ditto.
3785 (sse_storelps): Ditto.
3786 (sse_loadlps): Ditto.
3787 (*vec_concatv4sf): Ditto.
3788 (*vec_interleave_highv2df): Ditto.
3789 (*vec_interleave_lowv2df): Ditto.
3790 (*vec_extractv2df_1_sse): Ditto.
3791 (*vec_extractv2df_0_sse): Ditto.
3792 (sse2_storelpd): Ditto.
3793 (sse2_loadlpd): Ditto.
3794 (sse2_movsd): Ditto.
3795 (*vec_concatv4si): Ditto.
3796 (vec_concatv2di): Ditto.
3798 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3799 for Intel asm dialect.
3800 (mmx_punpcklwd): Ditto.
3801 (mmx_punpckldq): Ditto.
3803 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3804 for intel assembler dialect.
3806 2013-07-06 Jakub Jelinek <jakub@redhat.com>
3809 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3810 for BUILT_IN_C{LZ,LRSB}*.
3811 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3812 * tree-vrp.c (extract_range_basic): Handle
3813 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3814 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3815 fall thru to code calling set_value*.
3816 * builtins.c (expand_builtin): Remove *IMAX cases.
3817 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3818 if width is bigger than 2*HWI.
3820 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3822 PR rtl-optimization/55342
3823 * lra-int.h (lra_subreg_reload_pseudos): New.
3824 * lra.c: Add undoing optional reloads to the block diagram.
3825 (lra_subreg_reload_pseudos): New.
3826 (lra_optional_reload_pseudos): Change comments.
3827 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3828 lra_optional_reload_pseudos after undo transformations.
3829 * lra-assigns.c (pseudo_prefix_title): New.
3830 (lra_setup_reg_renumber): Use it.
3831 (spill_for): Ditto. Check subreg reload pseudos too.
3832 (assign_by_spills): Consider subreg reload pseudos too.
3833 * lra-constraints.c (simplify_operand_subreg): Use
3834 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3835 (curr_insn_transform): Recognize and do optional reloads.
3836 (undo_optional_reloads): New.
3837 (lra_undo_inheritance): Call undo_optional_reloads.
3839 2013-07-05 Thomas Quinot <quinot@adacore.com>
3841 * tree-complex.c (expand_complex_operations_1): Fix typo.
3843 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3845 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3846 (tune_params): New member 'const vec_costs'.
3847 * config/aarch64/aarch64.c (generic_vector_cost): New.
3848 (generic_tunings): New member 'generic_vector_cost'.
3849 (aarch64_builtin_vectorization_cost): New.
3850 (aarch64_add_stmt_cost): New.
3851 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3852 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3854 2013-07-03 Jakub Jelinek <jakub@redhat.com>
3857 * config/i386/predicates.md (vsib_address_operand): Disallow
3858 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3860 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3863 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3864 expand_builtin_setjmp_receiver.
3865 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3866 with NULL for the label parameter.
3867 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3868 the frame-pointer. Adjust comments.
3869 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3870 only if LABEL is non-NULL.
3872 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3874 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3875 (ARM_ABI_AAPCS64): Ditto.
3877 (ARM_DEFAULT_ABI): Ditto.
3879 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3881 * config/aarch64/aarch64-builtins.c
3882 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3883 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3885 * config/aarch64/aarch64-simd.md
3886 (aarch64_ld1<VALL:mode>): New.
3887 (aarch64_st1<VALL:mode>): Likewise.
3888 * config/aarch64/arm_neon.h
3889 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3891 2013-07-02 Sriraman Tallam <tmsriram@google.com>
3893 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3895 (ix86_option_override_internal):Turn on all -mavx target flags by
3896 default as they are dependent on AVX anyway.
3898 2013-07-02 Cary Coutant <ccoutant@google.com>
3900 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3902 (loc_checksum_ordered): Likewise.
3903 (hash_loc_operands): Remove inline keyword.
3905 2013-07-02 Jakub Jelinek <jakub@redhat.com>
3907 PR tree-optimization/57741
3908 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3909 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3910 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3911 Allow REAL_CST step_exprs if flag_associative_math.
3912 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3914 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3916 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
3918 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3920 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3922 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3924 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3926 (iorsi3_insn): Likewise.
3927 (arm_xorsi3): Likewise.
3929 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3931 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3933 (attribute "type"): ... here, and prefix with "wmmx_".
3934 (attribute "core_cycles"): Update for attribute changes.
3935 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3936 (tbcstv4hi): Likewise.
3937 (tbcstv2si): Likewise.
3938 (iwmmxt_iordi3): Likewise.
3939 (iwmmxt_xordi3): Likewise.
3940 (iwmmxt_anddi3): Likewise.
3941 (iwmmxt_nanddi3): Likewise.
3942 (iwmmxt_arm_movdi): Likewise.
3943 (iwmmxt_movsi_insn): Likewise.
3944 (mov<mode>_internal): Likewise.
3945 (and<mode>3_iwmmxt): Likewise.
3946 (ior<mode>3_iwmmxt): Likewise.
3947 (xor<mode>3_iwmmxt): Likewise.
3948 (add<mode>3_iwmmxt): Likewise.
3949 (ssaddv8qi3): Likewise.
3950 (ssaddv4hi3): Likewise.
3951 (ssaddv2si3): Likewise.
3952 (usaddv8qi3): Likewise.
3953 (usaddv4hi3): Likewise.
3954 (usaddv2si3): Likewise.
3955 (sub<mode>3_iwmmxt): Likewise.
3956 (sssubv8qi3): Likewise.
3957 (sssubv4hi3): Likewise.
3958 (sssubv2si3): Likewise.
3959 (ussubv8qi3): Likewise.
3960 (ussubv4hi3): Likewise.
3961 (ussubv2si3): Likewise.
3962 (mulv4hi3_iwmmxt): Likewise.
3963 (smulv4hi3_highpart): Likewise.
3964 (umulv4hi3_highpart): Likewise.
3965 (iwmmxt_wmacs): Likewise.
3966 (iwmmxt_wmacsz): Likewise.
3967 (iwmmxt_wmacu): Likewise.
3968 (iwmmxt_wmacuz): Likewise.
3969 (iwmmxt_clrdi): Likewise.
3970 (iwmmxt_clrv8qi): Likewise.
3971 (iwmmxt_clr4hi): Likewise.
3972 (iwmmxt_clr2si): Likewise.
3973 (iwmmxt_uavgrndv8qi3): Likewise.
3974 (iwmmxt_uavgrndv4hi3): Likewise.
3975 (iwmmxt_uavgv8qi3): Likewise.
3976 (iwmmxt_uavgv4hi3): Likewise.
3977 (iwmmxt_tinsrb): Likewise.
3978 (iwmmxt_tinsrh): Likewise.
3979 (iwmmxt_tinsrw): Likewise.
3980 (iwmmxt_textrmub): Likewise.
3981 (iwmmxt_textrmsb): Likewise.
3982 (iwmmxt_textrmuh): Likewise.
3983 (iwmmxt_textrmsh): Likewise.
3984 (iwmmxt_textrmw): Likewise.
3985 (iwmxxt_wshufh): Likewise.
3986 (eqv8qi3): Likewise.
3987 (eqv4hi3): Likewise.
3988 (eqv2si3): Likewise.
3989 (gtuv8qi3): Likewise.
3990 (gtuv4hi3): Likewise.
3991 (gtuv2si3): Likewise.
3992 (gtv8qi3): Likewise.
3993 (gtv4hi3): Likewise.
3994 (gtv2si3): Likewise.
3995 (smax<mode>3_iwmmxt): Likewise.
3996 (umax<mode>3_iwmmxt): Likewise.
3997 (smin<mode>3_iwmmxt): Likewise.
3998 (umin<mode>3_iwmmxt): Likewise.
3999 (iwmmxt_wpackhss): Likewise.
4000 (iwmmxt_wpackwss): Likewise.
4001 (iwmmxt_wpackdss): Likewise.
4002 (iwmmxt_wpackhus): Likewise.
4003 (iwmmxt_wpackwus): Likewise.
4004 (iwmmxt_wpackdus): Likewise.
4005 (iwmmxt_wunpckihb): Likewise.
4006 (iwmmxt_wunpckihh): Likewise.
4007 (iwmmxt_wunpckihw): Likewise.
4008 (iwmmxt_wunpckilb): Likewise.
4009 (iwmmxt_wunpckilh): Likewise.
4010 (iwmmxt_wunpckilw): Likewise.
4011 (iwmmxt_wunpckehub): Likewise.
4012 (iwmmxt_wunpckehuh): Likewise.
4013 (iwmmxt_wunpckehuw): Likewise.
4014 (iwmmxt_wunpckehsb): Likewise.
4015 (iwmmxt_wunpckehsh): Likewise.
4016 (iwmmxt_wunpckehsw): Likewise.
4017 (iwmmxt_wunpckelub): Likewise.
4018 (iwmmxt_wunpckeluh): Likewise.
4019 (iwmmxt_wunpckeluw): Likewise.
4020 (iwmmxt_wunpckelsb): Likewise.
4021 (iwmmxt_wunpckelsh): Likewise.
4022 (iwmmxt_wunpckelsw): Likewise.
4023 (ror<mode>3): Likewise.
4024 (ashr<mode>3_iwmmxt): Likewise.
4025 (lshr<mode>3_iwmmxt): Likewise.
4026 (ashl<mode>3_iwmmxt): Likewise.
4027 (ror<mode>3_di): Likewise.
4028 (ashr<mode>3_di): Likewise.
4029 (lshr<mode>3_di): Likewise.
4030 (ashl<mode>3_di): Likewise.
4031 (iwmmxt_wmadds): Likewise.
4032 (iwmmxt_wmaddu): Likewise.
4033 (iwmmxt_tmia): Likewise.
4034 (iwmmxt_tmiaph): Likewise.
4035 (iwmmxt_tmiabb): Likewise.
4036 (iwmmxt_tmiatb): Likewise.
4037 (iwmmxt_tmiabt): Likewise.
4038 (iwmmxt_tmiatt): Likewise.
4039 (iwmmxt_tmovmskb): Likewise.
4040 (iwmmxt_tmovmskh): Likewise.
4041 (iwmmxt_tmovmskw): Likewise.
4042 (iwmmxt_waccb): Likewise.
4043 (iwmmxt_wacch): Likewise.
4044 (iwmmxt_waccw): Likewise.
4045 (iwmmxt_waligni): Likewise.
4046 (iwmmxt_walignr): Likewise.
4047 (iwmmxt_walignr0): Likewise.
4048 (iwmmxt_walignr1): Likewise.
4049 (iwmmxt_walignr2): Likewise.
4050 (iwmmxt_walignr3): Likewise.
4051 (iwmmxt_wsadb): Likewise.
4052 (iwmmxt_wsadh): Likewise.
4053 (iwmmxt_wsadbz): Likewise.
4054 (iwmmxt_wsadhz): Likewise.
4055 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4056 (iwmmxt_wabsdiffb): Likewise.
4057 (iwmmxt_wabsdiffh): Likewise.
4058 (iwmmxt_wabsdiffw): Likewise.
4059 (iwmmxt_waddsubhx): Likewise
4060 (iwmmxt_wsubaddhx): Likewise.
4061 (addc<mode>3): Likewise.
4062 (iwmmxt_avg4): Likewise.
4063 (iwmmxt_avg4r): Likewise.
4064 (iwmmxt_wmaddsx): Likewise.
4065 (iwmmxt_wmaddux): Likewise.
4066 (iwmmxt_wmaddsn): Likewise.
4067 (iwmmxt_wmaddun): Likewise.
4068 (iwmmxt_wmulwsm): Likewise.
4069 (iwmmxt_wmulwum): Likewise.
4070 (iwmmxt_wmulsmr): Likewise.
4071 (iwmmxt_wmulumr): Likewise.
4072 (iwmmxt_wmulwsmr): Likewise.
4073 (iwmmxt_wmulwumr): Likewise.
4074 (iwmmxt_wmulwl): Likewise.
4075 (iwmmxt_wqmulm): Likewise.
4076 (iwmmxt_wqmulwm): Likewise.
4077 (iwmmxt_wqmulmr): Likewise.
4078 (iwmmxt_wqmulwmr): Likewise.
4079 (iwmmxt_waddbhusm): Likewise.
4080 (iwmmxt_waddbhusl): Likewise.
4081 (iwmmxt_wqmiabb): Likewise.
4082 (iwmmxt_wqmiabt): Likewise.
4083 (iwmmxt_wqmiatb): Likewise.
4084 (iwmmxt_wqmiatt): Likewise.
4085 (iwmmxt_wqmiabbn): Likewise.
4086 (iwmmxt_wqmiabtn): Likewise.
4087 (iwmmxt_wqmiatbn): Likewise.
4088 (iwmmxt_wqmiattn): Likewise.
4089 (iwmmxt_wmiabb): Likewise.
4090 (iwmmxt_wmiabt): Likewise.
4091 (iwmmxt_wmiatb): Likewise.
4092 (iwmmxt_wmiatt): Likewise.
4093 (iwmmxt_wmiabbn): Likewise.
4094 (iwmmxt_wmiabtn): Likewise.
4095 (iwmmxt_wmiatbn): Likewise.
4096 (iwmmxt_wmiattn): Likewise.
4097 (iwmmxt_wmiawbb): Likewise.
4098 (iwmmxt_wmiawbt): Likewise.
4099 (iwmmxt_wmiawtb): Likewise.
4100 (iwmmxt_wmiawtt): Likewise.
4101 (iwmmxt_wmiawbbn): Likewise.
4102 (iwmmxt_wmiawbtn): Likewise.
4103 (iwmmxt_wmiawtbn): Likewise.
4104 (iwmmxt_wmiawttn): Likewise.
4105 (iwmmxt_wmerge): Likewise.
4106 (iwmmxt_tandc<mode>3): Likewise.
4107 (iwmmxt_torc<mode>3): Likewise.
4108 (iwmmxt_torvsc<mode>3): Likewise.
4109 (iwmmxt_textrc<mode>3): Likewise.
4110 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
4111 (wmmxt_pack): Likewise.
4112 (wmmxt_mult_c1): Likewise.
4113 (wmmxt_mult_c2): Likewise.
4114 (wmmxt_alu_c1): Likewise.
4115 (wmmxt_alu_c2): Likewise.
4116 (wmmxt_alu_c3): Likewise.
4117 (wmmxt_transfer_c1): Likewise.
4118 (wmmxt_transfer_c2): Likewise.
4119 (wmmxt_transfer_c3): Likewise.
4120 (marvell_f_iwmmxt_wstr): Likewise.
4121 (marvell_f_iwmmxt_wldr): Likewise.
4123 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
4125 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
4127 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4130 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4131 * lra-constraints.c (need_for_split_p): Check call used hard regs
4132 living through calls.
4134 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
4135 call used regs for call insn.
4137 2013-06-28 Jakub Jelinek <jakub@redhat.com>
4140 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
4141 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
4143 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
4145 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
4147 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4149 * lra-constraints.c (need_for_split_p): Check call used hard regs
4150 living through calls.
4152 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
4155 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
4156 to tie with any other modes. Eliminate Altivec vector mode tests,
4157 since these are a subset of ALTIVEC or VSX vector modes. Simplify
4158 code, to return 0 if testing MODE2 for a condition, if we've
4159 already tested MODE1 for the same condition.
4161 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4163 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
4166 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4168 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4169 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
4171 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4173 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
4175 (aarch64_symbolic_constant_p): Remove.
4176 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
4177 static. Fix line length and white space.
4178 (aarch64_symbolic_constant_p): Remove.
4179 * config/aarch64/predicates.md (aarch64_valid_symref):
4180 Use aarch64_classify_symbol_expression.
4182 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4184 * config/arm/constraints.md (Ts): New constraint.
4185 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
4187 (compare_scc): Use "Ts" constraint for operand 0.
4188 (ior_scc_scc): Likewise.
4189 (and_scc_scc): Likewise.
4190 (and_scc_scc_nodom): Likewise.
4191 (ior_scc_scc_cmp): Likewise for operand 7.
4192 (and_scc_scc_cmp): Likewise.
4193 * config/arm/thumb2.md (thumb2_movsi_insn):
4194 Add alternatives for 16-bit encodings.
4195 (thumb2_movhi_insn): Likewise.
4196 (thumb2_movsicc_insn): Likewise.
4197 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
4198 (thumb2_negscc): Use "Ts" constraint.
4199 Move mvn instruction outside cond_exec block.
4200 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
4201 for 16-bit encodings.
4203 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4205 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
4207 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
4208 (mulsi3subsi): Likewise.
4209 (mulsidi3adddi): Likewise.
4210 (mulsidi3_v6): Likewise.
4211 (umulsidi3_v6): Likewise.
4212 (umulsidi3adddi_v6): Likewise.
4213 (smulsi3_highpart_v6): Likewise.
4214 (umulsi3_highpart_v6): Likewise.
4215 (mulhisi3tb): Likewise.
4216 (mulhisi3bt): Likewise.
4217 (mulhisi3tt): Likewise.
4218 (maddhisi4): Likewise.
4219 (maddhisi4tb): Likewise.
4220 (maddhisi4tt): Likewise.
4221 (maddhidi4): Likewise.
4222 (maddhidi4tb): Likewise.
4223 (maddhidi4tt): Likewise.
4224 (zeroextractsi_compare0_scratch): Likewise.
4225 (insv_zero): Likewise.
4226 (insv_t2): Likewise.
4227 (anddi_notzesidi_di): Likewise.
4228 (anddi_notsesidi_di): Likewise.
4229 (andsi_notsi_si): Likewise.
4230 (iordi_zesidi_di): Likewise.
4231 (xordi_zesidi_di): Likewise.
4232 (andsi_iorsi3_notsi): Likewise.
4234 (smax_m1): Likewise.
4236 (not_shiftsi): Likewise.
4237 (unaligned_loadsi): Likewise.
4238 (unaligned_loadhis): Likewise.
4239 (unaligned_loadhiu): Likewise.
4240 (unaligned_storesi): Likewise.
4241 (unaligned_storehi): Likewise.
4242 (extv_reg): Likewise.
4243 (extzv_t2): Likewise.
4245 (udivsi3): Likewise.
4246 (arm_zero_extendhisi2addsi): Likewise.
4247 (arm_zero_extendqisi2addsi): Likewise.
4248 (compareqi_eq0): Likewise.
4249 (arm_extendhisi2_v6): Likewise.
4250 (arm_extendqisi2addsi): Likewise.
4251 (arm_movt): Likewise.
4252 (thumb2_ldrd): Likewise.
4253 (thumb2_ldrd_base): Likewise.
4254 (thumb2_ldrd_base_neg): Likewise.
4255 (thumb2_strd): Likewise.
4256 (thumb2_strd_base): Likewise.
4257 (thumb2_strd_base_neg): Likewise.
4258 (arm_negsi2): Add alternative for 16-bit encoding.
4259 (arm_one_cmplsi2): Likewise.
4261 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4263 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
4264 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
4265 (movdfcc): Likewise.
4266 * config/arm/vfp.md (*thumb2_movsf_vfp):
4267 Disable predication for arm_restrict_it.
4268 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
4269 (*thumb2_movdfcc_vfp): Likewise.
4270 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
4271 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
4272 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
4273 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
4274 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
4275 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
4276 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
4277 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
4278 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
4279 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
4280 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
4281 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
4282 Disable predication for arm_restrict_it.
4284 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
4286 * config/i386/bmiintrin.h (_bextr_u32): New.
4287 (_bextr_u64): Ditto.
4289 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
4291 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
4292 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
4293 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
4294 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
4295 * config/mips/n32-elf.h: ...this new file.
4297 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
4300 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
4301 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
4303 2013-06-27 Catherine Moore <clm@codesourcery.com>
4305 * config/mips/mips-tables.opt: Regenerate.
4306 * config/mips/mips-cpus.def: Add m14ke and m14kec.
4307 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
4308 * doc/invoke.texi: Add -m14kc.
4310 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4313 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
4314 constraints of operand 1 and 2.
4317 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
4318 to match RTL canonicalization. Swap predicates and
4319 constraints of operand 1 and 2.
4321 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
4323 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
4324 Process OP_INOUT regs for splitting too.
4326 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4328 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
4329 decl before the loop, initialize to NULL.
4330 (vectorizable_load): Initialize ptr_incr to NULL.
4332 2013-06-27 Martin Jambor <mjambor@suse.cz>
4335 * ipa-ref.h (ipa_maybe_record_reference): Declare.
4336 * ipa-ref.c (ipa_maybe_record_reference): New function.
4337 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
4338 * ipa-cp.c (create_specialized_node): Record potential references from
4340 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
4342 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4344 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
4345 parameter 'mode' of type 'enum machine_mode mode'; change to pass
4346 'mode' to force_reg.
4347 (aarch64_add_offset): Update calls to aarch64_force_temporary.
4348 (aarch64_expand_mov_immediate): Likewise.
4350 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4352 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
4353 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
4355 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4357 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
4358 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4359 (struct machine_function): Add tbegin_p.
4360 (s390_canonicalize_comparison): Fold CC mode compares to
4361 conditional jump if possible.
4362 (s390_emit_jump): Return the emitted jump.
4363 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4364 Handle CCRAWmode compares.
4365 (s390_option_override): Default to -mhtm if available.
4366 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4367 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4368 FPRs instead of df_regs_ever_live_p.
4369 (s390_optimize_nonescaping_tx): New function.
4370 (s390_init_frame_layout): Extend clobbered_regs array to cover
4372 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4373 (s390_expand_tbegin): New function.
4374 (enum s390_builtin): New enum definition.
4375 (code_for_builtin): New array definition.
4376 (s390_init_builtins): New function.
4377 (s390_expand_builtin): New function.
4378 (TARGET_INIT_BUILTINS): Define.
4379 (TARGET_EXPAND_BUILTIN): Define.
4380 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4381 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4382 (s390_alc_comparison): Likewise.
4383 * config/s390/s390-modes.def: Add CCRAWmode.
4384 * config/s390/s390.h (processor_flags): Add PF_TX.
4385 (TARGET_CPU_HTM): Define macro.
4386 (TARGET_HTM): Define macro.
4387 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4388 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4389 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4390 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4391 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4392 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4393 ("*cc_to_int"): Move up.
4394 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4395 constants other than 0.
4396 ("*ccraw_to_int"): New insn and splitter definition.
4397 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4398 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4399 ("tx_assist"): New expander.
4400 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4401 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4402 * config/s390/s390.opt: Add -mhtm option.
4403 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4404 * config/s390/htmxlintrin.h: New file.
4405 * config/s390/htmintrin.h: New file.
4406 * config/s390/s390intrin.h: New file.
4407 * doc/extend.texi: Document htm builtins.
4408 * config.gcc: Add the new header files to extra_headers.
4410 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4412 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4413 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4415 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4416 Pat Haugen <pthaugen@us.ibm.com>
4417 Peter Bergner <bergner@vnet.ibm.com>
4419 * config/rs6000/power8.md: New.
4420 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4421 setting for power8 entry.
4422 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4423 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4424 test for Power4/Power5 only.
4425 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4427 (force_new_group): Adjust comment.
4428 * config/rs6000/rs6000.md: Include power8.md.
4430 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4432 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4433 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4435 (tune_params): Update comment.
4436 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4437 (arm_max_conditional_execute): New function.
4438 (thumb2_final_prescan_insn): Use max_insn_skipped and
4439 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4441 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4443 PR tree-optimization/57705
4444 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4445 SSA_NAME step, provided that it is not defined inside the loop.
4446 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4447 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4449 2013-06-25 Martin Jambor <mjambor@suse.cz>
4452 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4453 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4455 (ipa_note_param_call): Initialize member_ptr flag.
4456 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4457 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4458 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4459 (ipa_read_indirect_edge_info): Likewise.
4461 2013-06-25 Richard Biener <rguenther@suse.de>
4464 * passes.c (init_optimization_passes): Move pass_fold_builtins
4465 and pass_dce earlier with -Og.
4467 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4469 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4470 <BIT_FIELD_REF>: Remove trailing TAB.
4471 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4474 2013-06-24 Martin Jambor <mjambor@suse.cz>
4476 PR tree-optimization/57358
4477 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4478 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4479 (ipa_analyze_params_uses): Generate pessimistic info when true.
4481 2013-06-24 Martin Jambor <mjambor@suse.cz>
4483 PR tree-optimization/57539
4484 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4485 global.inlined_to of the new node to it. All callers changed.
4486 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4487 inlining_into, pass it to cgraph_clone_node.
4488 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4489 ipa_free_edge_args_substructures.
4490 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4491 rdesc linked list. Do not assert rdesc edges have inlined caller.
4492 Assert we have found an rdesc in the rdesc list.
4494 2013-06-24 Richard Biener <rguenther@suse.de>
4496 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4497 (pointer_set_lookup): Declare.
4498 (class pointer_map): New template class implementing a
4499 generic pointer to T map.
4500 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4501 pointer_map<T>::contains, pointer_map<T>::insert,
4502 pointer_map<T>::traverse): New functions.
4503 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4504 (pointer_set_lookup): New function.
4505 (pointer_set_contains): Use pointer_set_lookup.
4506 (pointer_set_insert): Likewise.
4507 (insert_aux): Remove.
4508 (struct pointer_map_t): Embed a pointer_set_t.
4509 (pointer_map_create): Adjust.
4510 (pointer_map_destroy): Likewise.
4511 (pointer_map_contains): Likewise.
4512 (pointer_map_insert): Likewise.
4513 (pointer_map_traverse): Likewise.
4514 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4515 pointer_map<unsigned> instead of a pointer_map_t.
4516 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4517 (streamer_tree_cache_lookup): Likewise.
4518 (streamer_tree_cache_create): Likewise.
4519 (streamer_tree_cache_delete): Likewise.
4520 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4521 pointer_map<unsigned> instead of a pointer_map_t.
4522 (lto_init_tree_ref_encoder): Adjust.
4523 (lto_destroy_tree_ref_encoder): Likewise.
4524 * lto-section-out.c (lto_output_decl_index): Likewise.
4525 (lto_record_function_out_decl_state): Likewise.
4526 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4528 2013-06-24 Richard Biener <rguenther@suse.de>
4530 PR tree-optimization/57488
4531 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4533 2013-06-24 Alan Modra <amodra@gmail.com>
4535 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4536 (gen_easy_altivec_constant): Likewise.
4537 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4538 easy_vector_constant_msb): Likewise.
4540 2013-06-23 Jakub Jelinek <jakub@redhat.com>
4543 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4544 add missing return true.
4546 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4549 * config/sh/predicates.md (general_extend_operand): Invoke
4550 general_movsrc_operand for memory operands.
4551 (general_movsrc_operand): Allow reg+reg addressing, do not use
4552 general_operand for memory operands.
4554 2013-06-23 Sriraman Tallam <tmsriram@google.com>
4556 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4557 when current target options does not apply.
4558 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4559 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
4560 * config/i386/bmiintrin.h: Pass appropriate target
4561 attributes to header.
4562 * config/i386/mmintrin.h: Ditto.
4563 * config/i386/nmmintrin.h: Ditto.
4564 * config/i386/avx2intrin.h: Ditto.
4565 * config/i386/fxsrintrin.h: Ditto.
4566 * config/i386/tbmintrin.h: Ditto.
4567 * config/i386/xsaveintrin.h: Ditto.
4568 * config/i386/f16cintrin.h: Ditto.
4569 * config/i386/xtestintrin.h: Ditto.
4570 * config/i386/xsaveoptintrin.h: Ditto.
4571 * config/i386/bmi2intrin.h: Ditto.
4572 * config/i386/lzcntintrin.h: Ditto.
4573 * config/i386/smmintrin.h: Ditto.
4574 * config/i386/wmmintrin.h: Ditto.
4575 * config/i386/x86intrin.h: Remove all header include guards.
4576 * config/i386/prfchwintrin.h: Ditto.
4577 * config/i386/pmmintrin.h: Ditto.
4578 * config/i386/tmmintrin.h: Ditto.
4579 * config/i386/xmmintrin.h: Ditto.
4580 * config/i386/popcntintrin.h: Ditto.
4581 * config/i386/rdseedintrin.h: Ditto.
4582 * config/i386/ammintrin.h: Ditto.
4583 * config/i386/emmintrin.h: Ditto.
4584 * config/i386/immintrin.h: Remove all header include guards.
4585 * config/i386/fma4intrin.h: Ditto.
4586 * config/i386/lwpintrin.h: Ditto.
4587 * config/i386/xopintrin.h: Ditto.
4588 * config/i386/ia32intrin.h: Ditto.
4589 * config/i386/avxintrin.h: Ditto.
4590 * config/i386/rtmintrin.h: Ditto.
4591 * config/i386/fmaintrin.h: Ditto.
4592 * config/i386/mm3dnow.h: Ditto.
4594 2013-06-22 Sriraman Tallam <tmsriram@google.com>
4596 * common/config/i386/i386-common.c: Handle LZCNT.
4598 2013-06-22 Andi Kleen <ak@linux.intel.com>
4600 * doc/extend.texi: Use __atomic_store_n instead of
4601 __atomic_store in HLE example.
4603 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4605 * config/sh/sh.c: Remove <cstdlib> workaround.
4607 2013-06-21 Andi Kleen <ak@linux.intel.com>
4609 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
4611 2013-06-21 Andi Kleen <ak@linux.intel.com>
4613 * doc/extend.texi: Document that __atomic_clear and
4614 __atomic_test_and_set should only be used with bool.
4616 2013-06-20 Jan Hubicka <jh@suse.cz>
4618 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4620 * tree.c (decls_same_for_odr): New function.
4621 (same_for_edr): New function.
4622 (types_same_for_odr): New function.
4623 (get_binfo_at_offset): Use it.
4624 * tree.h (types_same_for_odr): Declare.
4626 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
4627 Jason Merrill <jason@redhat.com>
4629 * system.h: Include <cstdlib> as well as <stdlib.h>.
4631 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
4634 * config/i386/i386.c (construct_container): Report error if
4635 long double is used with disabled x87 float returns.
4637 2013-06-20 Jan Hubicka <jh@suse.cz>
4639 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4641 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4643 PR rtl-optimization/57425
4644 PR rtl-optimization/57569
4645 * alias.c (write_dependence_p): Remove parameters mem_mode and
4646 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4647 Changed all callers.
4648 (canon_anti_dependence): Get comments and semantics in sync.
4649 Add parameter mem_canonicalized. Changed all callers.
4650 * rtl.h (canon_anti_dependence): Update prototype.
4652 2013-06-20 Richard Biener <rguenther@suse.de>
4654 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4655 byte case, inline streamer_read_uchar and defer section
4658 2013-06-20 Richard Biener <rguenther@suse.de>
4660 PR tree-optimization/57584
4661 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4662 SSA names into the expanded expression that take part in
4663 abnormal coalescing.
4665 2013-06-19 Sharad Singhai <singhai@google.com>
4667 * gcov.c (print_usage): Handle new option.
4668 (process_args): Ditto.
4669 (get_gcov_intermediate_filename): New function.
4670 (output_intermediate_file): New function.
4671 (output_gcov_file): New function
4672 (generate_results): Handle new option.
4673 (release_function): Relase demangled name.
4674 (read_graph_file): Handle demangled name.
4675 (output_lines): Ditto.
4676 * doc/gcov.texi: Document gcov intermediate format.
4678 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4681 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4682 (lra_emit_add): Use the functions. Add comment about Y as an
4685 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
4688 * collect2.c (collect_atexit): New.
4689 (collect_exit): Delete.
4690 (main): Register collect_atexit with atexit.
4691 (collect_wait): Change collect_exit to exit.
4693 * collect2.h (collect_exit): Delete.
4694 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4696 2013-06-19 Wei Mi <wmi@google.com>
4698 PR rtl-optimization/57518
4699 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4700 if regno is used in paradoxical subreg.
4701 (update_equiv_regs): Check pdx_subregs[regno] before
4702 set a reg to be equivalent with a mem.
4704 2013-06-19 Matthias Klose <doko@ubuntu.com>
4707 * file-find.h (find_a_file): Add a mode parameter.
4708 * file-find.c (find_a_file): Likewise.
4709 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4710 with X_OK for the executables.
4711 * collect2.c (main): Call find_a_file with X_OK.
4713 2013-06-19 Steve Ellcey <sellcey@mips.com>
4716 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4717 Use NEXT_INSN instead of next_real_insn.
4719 2013-06-19 Jan Hubicka <jh@suse.cz>
4721 * cgraph.h (const_value_known_p): Replace by ...
4722 (ctor_for_folding): .. this one.
4723 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4724 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4725 * expr.c (expand_expr_real_1): Likewise.
4726 (string_constant): Likewise.
4727 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4728 * ipa.c (process_references): Likewise.
4729 (symtab_remove_unreachable_nodes): Likewise.
4730 * ipa-inline-analysis.c (param_change_prob): Likewise.
4731 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4732 (get_base_constructor): Likwise.
4733 * varpool.c (varpool_remove_node): Likewise.
4734 (varpool_remove_initializer): LIkewise.
4735 (dump_varpool_node): LIkwise.
4736 (const_value_known_p): Rewrite to ...
4737 (ctor_for_folding): ... this one.
4739 2013-06-19 Jakub Jelinek <jakub@redhat.com>
4742 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4743 PERSONALITY in $PATH derived prefixes.
4745 2013-06-19 Jeff Law <law@redhat.com>
4747 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4750 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4751 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4754 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4756 * config/arm/vfp.md: Move VFP instruction classification documentation
4758 * config/arm/arm.md: ... here. Update instruction classification
4761 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
4763 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4764 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4765 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4767 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4770 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4771 with NEXT_INSN. Use tablejump_p to check for jump table data
4774 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4777 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4778 that now in C++ the value is correct per the C++ standards.
4780 2013-06-19 Richard Biener <rguenther@suse.de>
4782 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4785 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4788 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4791 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4792 with next_active_insn.
4794 2013-06-18 Sriraman Tallam <tmsriram@google.com>
4796 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4797 functions are inlined during failures to flag an error.
4798 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4799 in early inline pass.
4801 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4803 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4806 2013-06-18 Julian Brown <julian@codesourcery.com>
4808 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4809 Permit virtual register pre-reload if !strict.
4810 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4812 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4814 * config/arm/neon.md (movmisalign<mode>): Use
4815 neon_perm_struct_or_reg_operand instead of
4816 neon_struct_or_register_operand.
4817 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4818 neon_permissive_struct_operand instead of neon_struct_operand.
4819 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4820 neon_vector_mem_operand.
4821 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4822 neon_vector_mem_operand.
4823 (neon_permissive_struct_operand): New.
4824 (neon_struct_or_register_operand): Rename to...
4825 (neon_perm_struct_or_reg_operand): This. Adjust call to
4826 neon_vector_mem_operand.
4828 2013-06-18 Richard Biener <rguenther@suse.de>
4830 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4831 * lto-streamer.h: Include pointer-set.h.
4832 (struct lto_decl_slot): Remove.
4833 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4834 Remove next_index entry.
4835 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4836 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4837 (lto_init_tree_ref_encoder): Adjust.
4838 (lto_destroy_tree_ref_encoder): Likewise.
4839 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4840 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4841 (lto_output_decl_index): Adjust.
4842 (lto_new_out_decl_state): Likewise.
4843 (lto_record_function_out_decl_state): Likewise.
4844 * lto-streamer-out.c (copy_function): Likewise.
4846 2013-06-18 Richard Biener <rguenther@suse.de>
4848 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4849 * cgraphunit.c: Include cfgloop.h.
4850 (init_lowered_empty_function): Initialize the loop tree.
4851 (assemble_thunk): Insert new BBs into loops.
4853 2013-06-18 Richard Biener <rguenther@suse.de>
4855 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4856 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4857 the map from cache entry to cache index optional.
4858 (streamer_tree_cache_replace_tree): Adjust accordingly.
4859 (streamer_tree_cache_append): Likewise.
4860 (streamer_tree_cache_delete): Likewise.
4861 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4862 streamer cache map from cache entry to cache index.
4863 * lto-streamer-out.c (create_output_block): Adjust.
4865 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4867 * config/arm/arm.md (attribute "insn"): Move multiplication and
4868 division attributes to...
4869 (attribute "type"): ... here. Remove mult.
4870 (attribute "mul32"): New attribute.
4871 (attribute "mul64"): Add umaal.
4872 (*arm_mulsi3): Update attributes.
4873 (*arm_mulsi3_v6): Likewise.
4874 (*thumb_mulsi3): Likewise.
4875 (*thumb_mulsi3_v6): Likewise.
4876 (*mulsi3_compare0): Likewise.
4877 (*mulsi3_compare0_v6): Likewise.
4878 (*mulsi_compare0_scratch): Likewise.
4879 (*mulsi_compare0_scratch_v6): Likewise.
4880 (*mulsi3addsi): Likewise.
4881 (*mulsi3addsi_v6): Likewise.
4882 (*mulsi3addsi_compare0): Likewise.
4883 (*mulsi3addsi_compare0_v6): Likewise.
4884 (*mulsi3addsi_compare0_scratch): Likewise.
4885 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4886 (*mulsi3subsi): Likewise.
4887 (*mulsidi3adddi): Likewise.
4888 (*mulsi3addsi_v6): Likewise.
4889 (*mulsidi3adddi_v6): Likewise.
4890 (*mulsidi3_nov6): Likewise.
4891 (*mulsidi3_v6): Likewise.
4892 (*umulsidi3_nov6): Likewise.
4893 (*umulsidi3_v6): Likewise.
4894 (*umulsidi3adddi): Likewise.
4895 (*umulsidi3adddi_v6): Likewise.
4896 (*smulsi3_highpart_nov6): Likewise.
4897 (*smulsi3_highpart_v6): Likewise.
4898 (*umulsi3_highpart_nov6): Likewise.
4899 (*umulsi3_highpart_v6): Likewise.
4900 (mulhisi3): Likewise.
4901 (*mulhisi3tb): Likewise.
4902 (*mulhisi3bt): Likewise.
4903 (*mulhisi3tt): Likewise.
4904 (maddhisi4): Likewise.
4905 (*maddhisi4tb): Likewise.
4906 (*maddhisi4tt): Likewise.
4907 (maddhidi4): Likewise.
4908 (*maddhidi4tb): Likewise.
4909 (*maddhidi4tt): Likewise.
4911 (udivsi3): Likewise.
4912 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4913 (thumb2_mulsi_short_compare0): Likewise.
4914 (thumb2_mulsi_short_compare0_scratch): Likewise.
4915 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4916 (1020mult2): Likewise.
4917 (1020mult3): Likewise.
4918 (1020mult4): Likewise.
4919 (1020mult5): Likewise.
4920 (1020mult6): Likewise.
4921 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4923 (cortex_a15_mult64): Likewise.
4924 (cortex_a15_sdiv): Likewise.
4925 (cortex_a15_udiv): Likewise.
4926 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4932 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4933 (pj4_ir_div): Likewise.
4934 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4935 (11_mult2): Likewise.
4936 (11_mult3): Likewise.
4937 (11_mult4): Likewise.
4938 (11_mult5): Likewise.
4939 (11_mult6): Likewise.
4940 (11_mult7): Likewise.
4941 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4942 (cortex_a8_mla): Likewise.
4943 (cortex_a8_mull): Likewise.
4944 (cortex_a8_smulwy): Likewise.
4945 (cortex_a8_smlald): Likewise.
4946 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4947 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4948 (cortex_r4_mul_3): Likewise.
4949 (cortex_r4_mla_4): Likewise.
4950 (cortex_r4_mla_3): Likewise.
4951 (cortex_r4_smlald): Likewise.
4952 (cortex_r4_mull): Likewise.
4953 (cortex_r4_sdiv): Likewise.
4954 (cortex_r4_udiv): Likewise.
4955 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4956 (cortex_a7_idiv): Likewise.
4957 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4958 (9_mult2): Likewise.
4959 (9_mult3): Likewise.
4960 (9_mult4): Likewise.
4961 (9_mult5): Likewise.
4962 (9_mult6): Likewise.
4963 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4964 (cortex_a53_sdiv): Likewise.
4965 (cortex_a53_udiv): Likewise.
4966 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
4967 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
4968 (mp626_mult2): Likewise.
4969 (mp626_mult3): Likewise.
4970 (mp626_mult4): Likewise.
4971 * config/arm/fa526.md (526_mult1): Update attribute change.
4972 (526_mult2): Likewise.
4973 * config/arm/arm-generic.md (mult): Update attribute change.
4974 (mult_ldsched_strongarm): Likewise.
4975 (mult_ldsched): Likewise.
4976 (multi_cycle): Likewise.
4977 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
4978 * config/arm/fa606te.md (606te_mult1): Update attribute change.
4979 (606te_mult2): Likewise.
4980 (606te_mult3): Likewise.
4981 (606te_mult4): Likewise.
4982 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
4983 (cortex_a9_mac16): Likewise.
4984 (cortex_a9_multiply): Likewise.
4985 (cortex_a9_mac): Likewise.
4986 (cortex_a9_multiply_long): Likewise.
4987 * config/arm/fa626te.md (626te_mult1): Update attribute change.
4988 (626te_mult2): Likewise.
4989 (626te_mult3): Likewise.
4990 (626te_mult4): Likewise.
4992 2013-06-18 Richard Biener <rguenther@suse.de>
4995 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
4997 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5000 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5001 with next_active_insn.
5003 2013-06-18 Alan Modra <amodra@gmail.com>
5005 * config/rs6000/rs6000.h (enum data_align): New.
5006 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5007 (DATA_ABI_ALIGNMENT): Define.
5008 (CONSTANT_ALIGNMENT): Correct comment.
5009 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5010 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5012 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5014 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5015 ATTRIBUTE_UNUSED marking.
5017 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5019 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5020 alternative and update.
5021 (aarch64_dup_lanedi): Delete.
5022 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5023 * config/aarch64/aarch64-simd-builtins.def: Update.
5025 2013-06-17 Richard Biener <rguenther@suse.de>
5027 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5028 (lto_input_scc): Declare.
5029 (lto_input_tree_1): Likewise.
5030 (struct lto_stats_d): Add num_tree_bodies_output and
5031 num_pickle_refs_output.
5032 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5033 (lto_read_tree_1): Split out from ...
5034 (lto_read_tree): ... this.
5035 (lto_input_scc): New function.
5036 (lto_input_tree_1): Split out from ...
5037 (lto_input_tree): ... this. Handle LTO_tree_scc.
5038 (lto_data_in_create): Create the streamer cache without hashes.
5039 * lto-streamer-out.c (create_output_block): Create the streamer
5040 cache with hashes when not doing WPA.
5041 (lto_write_tree_1): Split out from ...
5042 (lto_write_tree): ... this.
5043 (get_symbol_initial_value): New function.
5044 (lto_output_tree_1): Split out from ...
5045 (lto_output_tree): ... this. Write trees as series of SCCs
5046 using a DFS walk via DFS_write_tree.
5047 (struct sccs, struct scc_entry): New types.
5048 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5049 (DFS_write_tree_body): New function.
5050 (DFS_write_tree): Likewise.
5051 (hash_tree): Likewise.
5052 (scc_entry_compare): Likewise.
5053 (hash_scc): Likewise.
5054 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5055 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5056 TREE_CHAIN as regular reference.
5057 (streamer_read_integer_cst): Remove.
5058 (streamer_get_pickled_tree): Adjust.
5059 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5060 of DECL_EXTERNALs in BLOCK_VARS for now.
5061 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5063 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5064 Add hash value argument and record that if hashes are recorded
5066 (streamer_tree_cache_insert_1): Adjust.
5067 (streamer_tree_cache_insert): Likewise.
5068 (streamer_tree_cache_insert_at): Rename to ...
5069 (streamer_tree_cache_replace_tree): ... this and adjust.
5070 (streamer_tree_cache_append): Adjust.
5071 (record_common_node): Likewise.
5072 (streamer_tree_cache_create): Add argument whether to
5073 record hash values together with trees.
5074 (streamer_tree_cache_delete): Adjust.
5075 * tree-streamer.h (struct streamer_tree_cache_d): Add
5077 (streamer_read_integer_cst): Remove.
5078 (streamer_tree_cache_insert): Adjust.
5079 (streamer_tree_cache_append): Likewise.
5080 (streamer_tree_cache_insert_at): Rename to ...
5081 (streamer_tree_cache_replace_tree): ... this and adjust.
5082 (streamer_tree_cache_create): Add argument whether to record hashes.
5083 (streamer_tree_cache_get): Rename to ...
5084 (streamer_tree_cache_get_tree): ... this.
5085 (streamer_tree_cache_get_hash): New function.
5086 * tree.c (cache_integer_cst): New function.
5087 * tree.h (cache_integer_cst): Declare.
5088 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5089 * lto-symtab.c (lto_varpool_replace_node): Only release
5090 DECL_INITIAL of non-prevailing decls.
5091 * varpool.c (varpool_remove_initializer): Do not release
5092 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5094 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5096 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5097 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5098 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5099 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5100 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5101 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5102 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5103 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5104 instead of TARGET_64BIT.
5105 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5106 Require ISA_HAS_<D>DIV.
5108 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
5110 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
5111 (mips*-*-linux*): Move default with_llsc setting to where other
5113 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
5115 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
5116 Likewise. Remove default with_tune setting. Move default float
5117 setting to its own block. Handle with_llsc in the same block as above.
5119 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
5121 PR rtl-optimization/57425
5122 PR rtl-optimization/57569
5123 * alias.c (write_dependence_p): Add new parameters mem_mode,
5124 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
5125 Changed all callers.
5126 (canon_anti_dependence): New function.
5127 * cse.c (check_dependence): Use canon_anti_dependence.
5128 * cselib.c (cselib_invalidate_mem): Likewise.
5129 * rtl.h (canon_anti_dependence): Declare.
5131 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5133 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
5134 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
5135 ".set mips3" for 64-bit targets.
5137 2013-06-15 Dehao Chen <dehao@google.com>
5139 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
5140 * gimple-low.c (gimple_check_call_matching_types): Likewise.
5141 (gimple_check_call_args): Likewise.
5142 * value-prof.c (check_ic_target): Likewise.
5143 * ipa-inline.c (early_inliner): Likewise.
5144 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
5145 * cgraph.c (cgraph_create_edge_1): Likewise.
5146 (cgraph_make_edge_direct): Likewise.
5148 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
5151 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
5152 rs6000_output_move_128bit to handle emitting quad memory
5153 operations. Set attribute length to 8 bytes.
5155 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
5157 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
5159 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
5160 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
5161 (aarch64_<su>mlsl<mode>): Likewise.
5163 2013-06-14 Mike Stump <mikestump@comcast.net>
5165 * Makefile.in (TARGET_H): Add insn-codes.h.
5167 2013-06-14 Alan Modra <amodra@gmail.com>
5171 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
5172 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
5173 bitfield expansion when EXPAND_MEMORY.
5174 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
5176 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
5178 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
5179 test for clearing quad memory on 32-bit later.
5181 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5183 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
5184 (fold_negate_expr): Likewise.
5185 (fold_real_zero_addition_p): Handle vectors.
5186 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
5188 2013-06-14 Alan Modra <amodra@gmail.com>
5190 * varasm.c (force_const_mem): Revert 2013-06-07 change.
5192 2013-06-13 Jan Hubicka <jh@suse.cz>
5194 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
5195 Local comdats are not externally visible.
5196 * symtab.c (dump_symtab_base): Dump externally visible.
5197 (verify_symtab_base): Verify back links in the symtab hash.
5199 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5201 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
5202 CONVERT_EXPR as equal nodes.
5204 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5206 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
5208 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5210 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
5211 Generalize to complex and vector.
5212 * tree.c (build_all_ones_cst): New function.
5213 * tree.h (build_all_ones_cst): Declare it.
5215 2013-06-13 Alan Modra <amodra@gmail.com>
5217 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
5218 * config/rs6000/rs6000.md (signbittf2): New insn.
5219 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
5220 (abstf2_internal, cmptf_internal2): Likewise.
5221 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
5223 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5224 Pat Haugen <pthaugen@us.ibm.com>
5225 Peter Bergner <bergner@vnet.ibm.com>
5227 * config/rs6000/rs6000.c (emit_load_locked): Add support for
5228 power8 byte, half-word, and quad-word atomic instructions.
5229 (emit_store_conditional): Likewise.
5230 (rs6000_expand_atomic_compare_and_swap): Likewise.
5231 (rs6000_expand_atomic_op): Likewise.
5233 * config/rs6000/sync.md (larx): Add new modes for power8.
5235 (AINT): New mode iterator to include TImode as well as normal
5236 integer modes on power8.
5237 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
5238 that VSX registers are not considered. Use AINT mode iterator
5239 instead of INT1 to allow inclusion of quad word atomic operations
5241 (load_locked<mode>): Likewise.
5242 (store_conditional<mode>): Likewise.
5243 (atomic_compare_and_swap<mode>): Likewise.
5244 (atomic_exchange<mode>): Likewise.
5245 (atomic_nand<mode>): Likewise.
5246 (atomic_fetch_<fetchop_name><mode>): Likewise.
5247 (atomic_nand_fetch<mode>): Likewise.
5248 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
5250 (ATOMIC): On power8, add QImode, HImode modes.
5251 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
5252 modes that promote to SImode.
5253 (load_lockedti): Convert TImode arguments to PTImode, so that we
5254 get a guaranteed even/odd register pair.
5255 (load_lockedpti): Likewise.
5256 (store_conditionalti): Likewise.
5257 (store_conditionalpti): Likewise.
5259 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
5260 atomic load/store instructions.
5263 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
5265 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
5267 (insn_count): New attribute, with most cases extracted from...
5268 (length): ...here. Redefine most cases in terms of insn_count.
5269 (single_insn): Delete.
5270 (can_delay): Use insn_count to check for single instructions.
5271 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
5272 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
5273 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
5274 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
5275 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
5276 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
5277 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
5278 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
5279 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
5280 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
5281 rather than "length".
5282 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
5283 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
5284 Use "insn_count" rather than "length".
5285 * config/mips/mips-dsp.md
5286 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
5287 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
5290 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
5292 PR tree-optimization/57361
5293 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
5295 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
5297 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
5299 (aarch64_simd_combine<mode>): New instruction expansion.
5300 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
5302 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
5303 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
5305 2013-06-12 Jan Hubicka <jh@suse.cz>
5307 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
5308 decl has when in streaming stage.
5309 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
5310 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
5312 2013-06-12 Roland Stigge <stigge@antcom.de>
5315 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
5317 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5319 PR tree-optimization/57537
5320 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
5321 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
5323 2013-06-12 Richard Biener <rguenther@suse.de>
5325 * data-streamer.h (streamer_write_char_stream): CSE
5326 obs->current_pointer.
5327 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
5328 streamer_write_char_stream manually and optimize the resulting loop.
5329 (streamer_write_hwi_stream): Likewise.
5331 2013-06-12 Jan Hubicka <jh@suse.cz>
5333 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
5334 * cgraph.h (varpool_create_empty_node): Declare.
5335 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
5337 * symtab.c (symtab_unregister_node): Be lax about missin entries
5339 (symtab_get_node): Update comment.
5340 * varpool.c (varpool_create_empty_node): Break out from ...
5341 (varpool_node_for_decl): ... here.
5342 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
5344 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
5346 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
5347 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
5348 part. Use straight-line flow at the end.
5349 <COMPONENT_REF>: Remove superfluous else.
5350 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
5352 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5355 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
5356 target hook even for !TREE_PUBLIC decls. If no resolution info
5357 is available, return false for common and external decls.
5359 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5361 * config/rl78/constraints.md (U): New constraint.
5362 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5365 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5368 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5369 to allow returning address to AT_PLATFORM name.
5371 2013-06-11 Jan Hubicka <jh@suse.cz>
5373 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5374 * cgraph.h (symtab_node_base): Add weakref flag.
5375 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5376 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5377 (output_weakrefs): Use weakref flag.
5378 * fold-const.c (simple_operand_p): Handle WEAK.
5379 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5380 * ipa.c (varpool_externally_visible_p): Drop weakref.
5381 (function_and_variable_visibility): Update comment; fix weakref
5382 sanity checks; do not clear DECL_WEAK on them.
5383 * lto-cgraph.c (lto_output_node): update.
5384 (lto_output_varpool_node): Update.
5385 (input_overwrite_node): Update.
5386 (input_node): Update.
5387 (input_varpool_node): Update.
5388 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5389 (lto_symtab_merge_symbols): Add sanity check.
5390 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5391 * passes.c (rest_of_decl_compilation): Set static flag, too.
5392 * symtab.c (dump_symtab_base): Dump weakref.
5393 (verify_symtab_base): Sanity check weakrefs.
5394 (symtab_make_decl_local): Remove duplicated code.
5395 (symtab_alias_ultimate_target): Simplify.
5396 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5398 2013-06-11 DJ Delorie <dj@redhat.com>
5400 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5401 (rl78_unwind_word_mode): New.
5403 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5405 * final.c (debug_prefix_maps): Make static.
5407 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5409 * function.c (initial_trampoline): Remove stray copy.
5411 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5413 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5415 2013-06-11 Martin Jambor <mjambor@suse.cz>
5417 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5418 within bounds at the beginning of the function.
5420 2013-06-11 Alan Modra <amodra@gmail.com>
5422 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5424 (get_named_section): Don't NULL !DECL_P decl.
5426 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5428 * doc/invoke.texi (core-avx2): Document.
5430 (atom): Updated with MOVBE.
5432 2013-06-11 Richard Biener <rguenther@suse.de>
5434 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5436 2013-06-11 Anton Blanchard <anton@samba.org>
5438 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5439 correct shift value in little-endian mode.
5441 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5444 * varasm.c (get_variable_align): Move #endif to the right place.
5446 2013-06-10 Cary Coutant <ccoutant@google.com>
5448 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5449 for hash so that hash table traversal order is deterministic.
5451 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5452 Pat Haugen <pthaugen@us.ibm.com>
5453 Peter Bergner <bergner@vnet.ibm.com>
5455 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5456 of vectors in GPRS if they are direct moves or quad word load or
5459 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5461 (direct_move_p): Likewise.
5462 (quad_load_store_p): Likewise.
5464 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5465 classes into bins based on the physical register type.
5466 (reg_class_to_reg_type): Likewise.
5467 (IS_STD_REG_TYPE): Likewise.
5468 (IS_FP_VECT_REG_TYPE): Likewise.
5469 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5470 use direct move instructions.
5471 (reload_gpr_vsx): Likewise.
5472 (reload_vsx_gpr): Likewise.
5473 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5474 information that is a simplification of register classes. Also
5475 precalculate direct move reload helpers.
5476 (direct_move_p): New function to return true if the operation can
5477 be done as a direct move instruciton.
5478 (quad_load_store_p): New function to return true if the operation
5479 is a quad memory operation.
5480 (rs6000_legitimize_address): If quad memory, only allow register
5481 indirect for TImode addresses.
5482 (rs6000_legitimate_address_p): Likewise.
5483 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5484 (rs6000_reload_register_type): Likewise.
5485 (register_to_reg_type): Return register type.
5486 (rs6000_secondary_reload_simple_move): New helper function for
5487 secondary reload and secondary memory needed to identify anything
5488 that is a simple move, and does not need reloading.
5489 (rs6000_secondary_reload_direct_move): New helper function for
5490 secondary reload to identify cases that can be done with several
5491 instructions via the direct move instructions.
5492 (rs6000_secondary_reload_move): New helper function for secondary
5493 reload to identify moves between register types that can be done.
5494 (rs6000_secondary_reload): Add support for quad memory operations
5495 and for direct move.
5496 (rs6000_secondary_memory_needed): Likewise.
5497 (rs6000_debug_secondary_memory_needed): Change argument names.
5498 (rs6000_output_move_128bit): New function to return the move to
5499 use for 128-bit moves, including knowing about the various
5500 limitations of quad memory operations.
5502 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5503 memory operations. call rs6000_output_move_128bit for the actual
5504 instruciton(s) to generate.
5505 (vsx_movti_64bit): Likewise.
5507 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5508 (UNSPEC_P8V_MTVSRWZ): Likewise.
5509 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5510 (UNSPEC_P8V_MTVSRD): Likewise.
5511 (UNSPEC_P8V_XXPERMDI): Likewise.
5512 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5513 (UNSPEC_FUSION_GPR): Likewise.
5514 (FMOVE128_GPR): New iterator for direct move.
5515 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
5518 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5519 loads and direct move instructions.
5520 (zero_extendsidi2_lfiwzx): Likewise.
5521 (extendsidi2_lfiwax): Likewise.
5522 (extendsidi2_nocell): Likewise.
5523 (floatsi<mode>2_lfiwax): Likewise.
5525 (floatunssi<mode>2_lfiwzx): Likewise.
5527 (fix_trunc<mode>_stfiwx): Likewise.
5528 (fixuns_trunc<mode>_stfiwx): Likewise.
5529 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5530 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5531 (parity<mode>2_cmpb): Set length/type attr.
5532 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5533 for 'mr.' to fast_compare.
5534 (bpermd_<mode>): Change type attr to popcnt.
5535 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5536 (p8_mtvsrwz_1): Likewise.
5537 (p8_mtvsrwz_2): Likewise.
5538 (reload_fpr_from_gpr<mode>): Likewise.
5539 (p8_mtvsrd_1): Likewise.
5540 (p8_mtvsrd_2): Likewise.
5541 (p8_xxpermdi_<mode>): Likewise.
5542 (reload_vsx_from_gpr<mode>): Likewise.
5543 (reload_vsx_from_gprsf): Likewise.
5544 (p8_mfvsrd_3_<mode>): LIkewise.
5545 (reload_gpr_from_vsx<mode>): Likewise.
5546 (reload_gpr_from_vsxsf): Likewise.
5547 (p8_mfvsrd_4_disf): Likewise.
5548 (multi-word GPR splits): Do not split direct moves or quad memory
5551 2013-06-10 David Malcolm <dmalcolm@redhat.com>
5553 * tree-into-ssa.c (interesting_blocks): Make static.
5555 2013-06-10 Jakub Jelinek <jakub@redhat.com>
5558 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5559 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5560 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5561 (get_variable_align): New function.
5562 (get_variable_section, emit_bss, emit_common,
5563 assemble_variable_contents, place_block_symbol): Use
5564 get_variable_align instead of DECL_ALIGN.
5565 (assemble_noswitch_variable): Add align argument, use it
5566 instead of DECL_ALIGN.
5567 (assemble_variable): Adjust caller. Use get_variable_align
5568 instead of DECL_ALIGN.
5569 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5571 (DATA_ABI_ALIGNMENT): Define.
5572 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5573 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5574 opt is false, only return the psABI mandated alignment increase.
5575 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5576 (DATA_ABI_ALIGNMENT): ... this.
5577 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5578 (DATA_ABI_ALIGNMENT): ... this.
5579 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5580 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5581 (DATA_ABI_ALIGNMENT): ... this.
5582 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5583 * doc/tm.texi: Regenerated.
5585 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
5587 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5588 cmp_code to construct REG_EQUAL note.
5590 2013-06-09 Jakub Jelinek <jakub@redhat.com>
5593 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5594 that operands[2] doesn't overlap with operands[0].
5596 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
5597 Jan Hubicka <jh@suse.cz>
5599 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5600 hack to mark symbols as used.
5602 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5604 PR rtl-optimization/57559
5605 * lra-constraints.c (process_alt_operands): Don't discourage
5606 memory with known offset for offsetable memory constraint.
5607 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5609 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5611 * varasm.c (struct oc_local_state): Reorder fields.
5612 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5613 and adjust accordingly.
5614 (output_constructor): Reorder initialization code and adjust call to
5615 output_constructor_bitfield.
5617 2013-06-07 Jan Hubicka <jh@suse.cz>
5619 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5621 2013-06-07 David Malcolm <dmalcolm@redhat.com>
5623 * tree-object-size.c (unknown): Make const.
5625 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5627 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5628 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5629 for last alternative in the cpu_facility attribute.
5631 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5634 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5635 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5636 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5637 * config/arm/constraints.md (Dg): New constraint.
5638 * config/arm/neon.md (xordi3_neon): Remove.
5639 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5640 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5642 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5644 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5645 Clean up alternatives.
5647 2013-06-07 Alan Modra <amodra@gmail.com>
5649 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5652 2013-06-07 Alan Modra <amodra@gmail.com>
5654 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5656 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5658 * config/arm/constraints.md (Df): New constraint.
5659 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5660 Correct length attribute for last two alternatives.
5662 2013-06-07 Alan Modra <amodra@gmail.com>
5664 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5665 override user -mfp-in-toc.
5666 (offsettable_ok_by_alignment): Consider just the current access
5667 rather than the whole object, unless BLKmode. Handle
5668 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5669 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5670 for -mcmodel=medium.
5671 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5672 override user -mfp-in-toc or -msum-in-toc. Default to
5673 -mno-fp-in-toc for -mcmodel=medium.
5675 2013-06-06 DJ Delorie <dj@redhat.com>
5677 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5678 TARGET_VALID_POINTER_MODE.
5680 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5681 Pat Haugen <pthaugen@us.ibm.com>
5682 Peter Bergner <bergner@vnet.ibm.com>
5684 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5685 Document new power8 builtins.
5687 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5688 condition code register, to allow 128-bit logical operations to be
5689 done in the VSX or GPR registers.
5690 (nor<mode>3): Use the canonical form for nor.
5691 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5692 vclz*, and vpopcnt* vector instructions.
5693 (nand<mode>3): Likewise.
5694 (orc<mode>3): Likewise.
5695 (clz<mode>2): LIkewise.
5696 (popcount<mode>2): Likewise.
5698 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5699 that only the GPRs are recognized.
5701 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5702 support for new power8 builtins.
5704 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5706 (xscvdpspn): Likewise.
5712 (vpopcnt): Likewise.
5713 (vpopcntb): Likewise.
5714 (vpopcnth): Likewise.
5715 (vpopcntw): Likewise.
5716 (vpopcntd): Likewise.
5721 (eqv_v16qi3): Likewise.
5722 (eqv_v8hi3): Likewise.
5723 (eqv_v4si3): Likewise.
5724 (eqv_v2di3): Likewise.
5725 (eqv_v4sf3): Likewise.
5726 (eqv_v2df3): Likewise.
5728 (nand_v16qi3): Likewise.
5729 (nand_v8hi3): Likewise.
5730 (nand_v4si3): Likewise.
5731 (nand_v2di3): Likewise.
5732 (nand_v4sf3): Likewise.
5733 (nand_v2df3): Likewise.
5735 (orc_v16qi3): Likewise.
5736 (orc_v8hi3): Likewise.
5737 (orc_v4si3): Likewise.
5738 (orc_v2di3): Likewise.
5739 (orc_v4sf3): Likewise.
5740 (orc_v2df3): Likewise.
5742 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5743 allow power8 quad mode in 64-bit.
5744 (rs6000_builtin_vectorized_function): Add support to vectorize
5745 ISA 2.07 count leading zeros, population count builtins.
5746 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5747 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5748 (builtin_function_type): Add vgbbd builtin function which takes an
5750 (altivec_expand_vec_perm_const): Add support for new power8 merge
5753 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5754 that does not include TImdoe for use with 32-bit.
5755 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5757 (UNSPEC_VSX_CVDPSPN): Likewise.
5758 (vsx_xscvdpspn): Likewise.
5759 (vsx_xscvspdpn): Likewise.
5760 (vsx_xscvdpspn_scalar): Likewise.
5761 (vsx_xscvspdpn_directmove): Likewise.
5762 (vsx_and<mode>3): Split logical operations into 32-bit and
5763 64-bit. Add support to do logical operations on TImode as well as
5764 VSX vector types. Allow logical operations to be done in either
5765 VSX registers or in general purpose registers in 64-bit mode. Add
5766 splitters if GPRs were used. For AND, add clobber of CCmode to
5767 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5769 (vsx_and<mode>3_32bit): Likewise.
5770 (vsx_and<mode>3_64bit): Likewise.
5771 (vsx_ior<mode>3): Likewise.
5772 (vsx_ior<mode>3_32bit): Likewise.
5773 (vsx_ior<mode>3_64bit): Likewise.
5774 (vsx_xor<mode>3): Likewise.
5775 (vsx_xor<mode>3_32bit): Likewise.
5776 (vsx_xor<mode>3_64bit): Likewise.
5777 (vsx_one_cmpl<mode>2): Likewise.
5778 (vsx_one_cmpl<mode>2_32bit): Likewise.
5779 (vsx_one_cmpl<mode>2_64bit): Likewise.
5780 (vsx_nor<mode>3): Likewise.
5781 (vsx_nor<mode>3_32bit): Likewise.
5782 (vsx_nor<mode>3_64bit): Likewise.
5783 (vsx_andc<mode>3): Likewise.
5784 (vsx_andc<mode>3_32bit): Likewise.
5785 (vsx_andc<mode>3_64bit): Likewise.
5786 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5787 and xxlorc instructions.
5788 (vsx_eqv<mode>3_64bit): Likewise.
5789 (vsx_nand<mode>3_32bit): Likewise.
5790 (vsx_nand<mode>3_64bit): Likewise.
5791 (vsx_orc<mode>3_32bit): Likewise.
5792 (vsx_orc<mode>3_64bit): Likewise.
5794 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5796 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5798 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5799 (p8_vmrgow): Likewise.
5800 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5801 GPRs to be split under VSX.
5802 (p8v_clz<mode>2): Add power8 count leading zero support.
5803 (p8v_popcount<mode>2): Add power8 population count support.
5804 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5807 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5810 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5812 (vec_nand): Likewise.
5813 (vec_vclz): Likewise.
5814 (vec_vclzb): Likewise.
5815 (vec_vclzd): Likewise.
5816 (vec_vclzh): Likewise.
5817 (vec_vclzw): Likewise.
5818 (vec_vgbbd): Likewise.
5819 (vec_vmrgew): Likewise.
5820 (vec_vmrgow): Likewise.
5821 (vec_vpopcnt): Likewise.
5822 (vec_vpopcntb): Likewise.
5823 (vec_vpopcntd): Likewise.
5824 (vec_vpopcnth): Likewise.
5825 (vec_vpopcntw): Likewise.
5827 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5829 PR rtl-optimization/57468
5830 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5833 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5835 PR rtl-optimization/57459
5836 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5837 type when setting live regs.
5839 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5841 * config/s390/s390.opt (mlra): New option.
5842 * config/s390/s390.c (s390_decompose_address): Check displacement
5843 for all registers for LRA.
5844 (s390_secondary_reload): Don't used secondary reloads for LRA.
5845 (s390_lra_p): New function.
5846 (TARGET_LRA_P): Define.
5847 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5848 of attribute cpu_facility to zarch for the last alternative.
5849 (*cmpmem_short): Ditto.
5851 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5853 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5854 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5855 (arm_expand_prologue): Likewise.
5857 2013-06-06 Teresa Johnson <tejohnson@google.com>
5860 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5861 as this is now done by redirect_edge_and_branch_force.
5862 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5863 barriers, and fix interaction with splitting.
5864 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5865 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5866 reflect changes made in the routine.
5867 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5868 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5869 since this is called in cfglayout mode, and replace partition fixup
5870 with assert as that is now done by force_nonfallthru_and_redirect.
5871 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5872 already be marked with region crossing note.
5873 (insert_section_boundary_note): Make non-static, gate on flag
5874 has_bb_partition, rewrite to also check for multiple partitions.
5875 (rest_of_handle_reorder_blocks): Remove call to
5876 insert_section_boundary_note, now done later during free_cfg.
5877 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5878 * bb-reorder.h (insert_section_boundary_note): Declare.
5879 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5880 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5881 invoke insert_section_boundary_note.
5882 (try_redirect_by_replacing_jump): Remove unnecessary
5883 check for region crossing note.
5884 (fixup_partition_crossing): New function.
5885 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5886 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5887 in non-cfglayout mode.
5888 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5889 remove old code that tried to do this. Emit barrier correctly
5890 when we are in cfglayout mode.
5891 (last_bb_in_partition): New function.
5892 (rtl_split_edge): Correctly fixup partition boundaries.
5893 (commit_one_edge_insertion): Remove old code that tried to
5894 fixup region crossing edge since this is now handled in
5895 split_block, and set up insertion point correctly since
5896 block may now end in a jump.
5897 (verify_hot_cold_block_grouping): Guard against checking when not in
5898 linearized RTL mode.
5899 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5901 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5902 rtl_verify_flow_info, so not called in cfglayout mode.
5903 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5904 (fixup_reorder_chain): Remove old code that attempted to fixup region
5905 crossing note as this is now handled in force_nonfallthru_and_redirect.
5906 (duplicate_insn_chain): Don't duplicate switch section notes.
5907 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5909 * basic-block.h (emit_barrier_after_bb): Declare.
5911 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5913 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5914 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5915 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5917 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5919 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5921 * config/arm/ldmstm.md: Regenerate.
5923 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5925 * config/arm/sync.md (atomic_loaddi_1):
5926 Disable predication for arm_restrict_it.
5927 (arm_load_exclusive<mode>): Likewise.
5928 (arm_load_exclusivesi): Likewise.
5929 (arm_load_exclusivedi): Likewise.
5930 (arm_load_acquire_exclusive<mode>): Likewise.
5931 (arm_load_acquire_exclusivesi): Likewise.
5932 (arm_load_acquire_exclusivedi): Likewise.
5933 (arm_store_exclusive<mode>): Likewise.
5934 (arm_store_exclusive<mode>): Likewise.
5935 (arm_store_release_exclusivedi): Likewise.
5936 (arm_store_release_exclusive<mode>): Likewise.
5938 2013-06-06 Richard Biener <rguenther@suse.de>
5940 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5942 (lto_tag_is_tree_code_p): Adjust.
5943 (lto_tag_is_gimple_code_p): Likewise.
5944 (lto_gimple_code_to_tag): Likewise.
5945 (lto_tag_to_gimple_code): Likewise.
5946 (lto_tree_code_to_tag): Likewise.
5947 (lto_tag_to_tree_code): Likewise.
5948 * data-streamer.h (streamer_write_hwi_in_range): Use
5949 uhwi streaming to stream the normalized range.
5950 (streamer_read_hwi_in_range): Likewise.
5952 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5954 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5955 (predicable_short_it): Likewise.
5956 (predicated): Likewise.
5957 (enabled): Handle above.
5958 (define_cond_exec): Set predicated attribute to yes.
5960 2013-06-05 Mike Stump <mikestump@comcast.net>
5962 * gdbinit.in (__FUNCTION__): Add.
5964 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
5966 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
5967 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
5969 2013-06-05 Jan Hubicka <jh@suse.cz>
5971 * varasm.c (mark_decl_referenced): Revert the removal until targets
5974 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
5976 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
5977 instead of mark_decl_referenced.
5979 2013-06-05 Jan Hubicka <jh@suse.cz>
5981 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
5982 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
5983 and symtab_used_from_object_file_p.
5984 (cgraph_make_node_local_1): Clear forced_by_abi.
5985 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
5986 * cgraph.h (symtab_node_base): Add forced_by_abi.
5987 (decide_is_variable_needed): Remove.
5988 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
5989 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
5990 (decide_is_symbol_needed): ... this one; handle symbols in general;
5991 always analyze virtuals; honnor forced_by_abi.
5992 (cgraph_finalize_function): Update.
5993 (varpool_finalize_decl): Update.
5994 (symbol_defined_and_needed): Remove.
5995 (analyze_functions): Update.
5996 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
5997 output_refs, input_overwrite_node): Handle forced_by_abi.
5998 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
5999 (address_taken_from_non_vtable_p): ... this one.
6000 (comdat_can_be_unshared_p_1): New function.
6001 (cgraph_comdat_can_be_unshared_p): Rename to ...
6002 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6003 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6004 (function_and_variable_visibility): Clear forced_by_abi as needed.
6005 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6006 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6007 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6008 * varpool.c (decide_is_variable_needed): Remove.
6010 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6012 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6013 (arm_option_override): Override arm_restrict_it where appropriate.
6014 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6015 * config/arm/arm.opt (mrestrict-it): New command-line option.
6016 * doc/invoke.texi: Document -mrestrict-it.
6018 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6020 * tsan.c (tsan_atomic_table): Make const.
6022 2013-06-05 Richard Biener <rguenther@suse.de>
6024 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6025 index associated with the tree we are supposed to replace.
6026 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6027 TREE_ASM_WRITTEN as zero for everything but SSA names.
6029 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6031 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6033 2013-06-04 Jan Hubicka <jh@suse.cz>
6035 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6036 (input_node, input_varpool_node): Handle correctly external same
6038 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6039 nodes at ltrans stage.
6041 2013-06-04 Jan Hubicka <jh@suse.cz>
6043 * ipa-inline.c (update_caller_keys): Fix availability test.
6044 (update_callee_keys): Likewise.
6045 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6046 to follow ELF standard.
6048 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6050 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6051 (mips64r5900el-*-elf*): New configurations.
6052 * config/mips/mips-cpus.def (r5900): New processor.
6053 * config/mips/mips-tables.opt: Regenerate.
6054 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6055 (mips_issue_rate): Handle PROCESSOR_R5900.
6056 (mips_reorg_process_insns): Force reorder mode for the R5900.
6057 * config/mips/mips.h (TARGET_MIPS5900): Define.
6058 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6060 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6062 * config/mips/mips.md (processor): Add r5900.
6063 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6065 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6067 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6068 into function to generate MOVI instruction.
6069 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6070 (aarch64_preferred_simd_mode): Turn into wrapper.
6071 (aarch64_output_scalar_simd_mov_immediate): New function.
6072 * config/aarch64/aarch64-protos.h: Add prototype for above.
6074 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6076 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6077 element_char member.
6078 (sizetochar): Return signed char.
6079 (aarch64_simd_valid_immediate): Remove elchar and other
6080 unnecessary variables.
6081 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6082 Calculate element_char as required.
6083 * config/aarch64/aarch64-protos.h: Update and move prototype
6084 for aarch64_output_simd_mov_immediate.
6085 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6088 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6090 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6091 information completed by aarch64_simd_valid_immediate.
6092 (aarch64_legitimate_constant_p): Update arguments.
6093 (aarch64_simd_valid_immediate): Work with struct rather than many
6095 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6096 (aarch64_simd_make_constant): Update arguments.
6097 (aarch64_output_simd_mov_immediate): Work with struct rather than
6098 many pointers. Output immediate directly rather than as operand.
6099 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6101 * config/aarch64/constraints.md (Dn): Update arguments.
6103 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6105 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6107 (aarch64_simd_immediate_valid_for_move): Remove.
6108 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
6109 (aarch64_simd_make_constant): Update call.
6110 (aarch64_output_simd_mov_immediate): Update call.
6111 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6113 * config/aarch64/constraints.md (Dn): Update call.
6115 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6117 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
6118 return type to bool for prototype.
6119 (aarch64_legitimate_constant_p): Check for true instead of not -1.
6120 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
6121 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
6123 2013-06-04 Catherine Moore <clm@codesourcery.com>
6125 * config/mips/mips.opt (meva): New.
6126 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
6127 (ASM_SPEC): Handle -meva.
6128 * doc/invoke.texi (meva): Document.
6130 2013-06-04 Alan Modra <amodra@gmail.com>
6132 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
6135 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6137 * rtl.def: Add extra fourth optional field to define_cond_exec.
6138 * gensupport.c (process_one_cond_exec): Process attributes from
6140 * doc/md.texi: Document fourth field in define_cond_exec.
6142 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
6144 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
6145 out the processing order as in store_bit_field_1.
6147 2013-06-04 Jan Hubicka <jh@suse.cz>
6150 * cgraphunit.c (cgraph_process_same_body_aliases): Create
6151 non-VAR_DECL node if it does not exist yet.
6153 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
6155 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
6156 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
6157 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
6158 target_cpu_default setting.
6160 2013-06-03 Teresa Johnson <tejohnson@google.com>
6162 * dumpfile.c (opt_info_switch_p): Change -fopt-info
6163 default to -fopt-info=optimized instead of all.
6164 * doc/invoke.texi: Ditto.
6165 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
6166 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
6167 (execute_vect_slp): Emit BB vectorization success under
6168 MSG_OPTIMIZED_LOCATIONS.
6169 * tree-vect-slp.c (vect_slp_transform_bb): Change
6170 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
6171 * tree-vect-loop.c (vect_transform_loop): Ditto.
6173 2013-06-03 Jason Merrill <jason@redhat.com>
6176 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
6177 Use TARGET_EXPR for C++.
6179 2013-06-03 Jakub Jelinek <jakub@redhat.com>
6181 PR rtl-optimization/57268
6182 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
6183 if DEBUG_INSN_P (insn).
6186 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6188 PR rtl-optimization/57268
6189 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6190 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6192 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6194 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
6195 (ix86_avoid_lea_for_addr): Likewise.
6196 (exact_dependency_1): Likewise.
6197 (ix86_adjust_cost): Likewise.
6198 (swap_top_of_ready_list): Fix formatting and !reload_completed check
6200 (do_reorder_for_imul): Fix typo, formatting and
6201 !reload_completed check removed.
6202 (ix86_sched_reorder): Fix typo and formatting.
6203 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
6206 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
6208 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
6210 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
6212 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
6213 <STRING_CST>: Likewise.
6214 <VECTOR_CST>: Likewise.
6216 2013-06-01 Janus Weil <janus@gcc.gnu.org>
6217 Mikael Morin <mikael@gcc.gnu.org>
6219 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
6220 * config.in: Regenerated.
6221 * configure: Regenerated.
6223 2013-06-01 Jan Hubicka <jh@suse.cz>
6226 * cgraphunit.c (compile): When weakref is not supported,
6227 set up transparent aliases before final output pass.
6228 * varasm.c (assemble_alias): Do not try to do it here.
6230 2013-06-01 Jan Hubicka <jh@suse.cz>
6233 * passes.c (for_per_function): Skip unanalyzed functions.
6235 2013-06-01 Jan Hubicka <jh@suse.cz>
6237 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
6238 (lto_symtab_merge_symbols_1): ... this one.
6239 (lto_symtab_merge_cgraph_nodes): Rename to ...
6240 (lto_symtab_merge_symbols): ... this one; simplify.
6241 * cgraph.c (same_body_aliases_done): Rename to ...
6242 (cpp_implicit_aliases_done): ... this one.
6243 (cgraph_create_function_alias): Update.
6244 (cgraph_same_body_alias): Update.
6245 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
6246 (verify_edge_corresponds_to_fndecl): Simplify.
6247 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
6248 (cgraph_node): Remove same_body_alias.
6249 (varpool_node): Remove alias_of and extra_name_alias.
6250 (same_body_aliases_done): Rename to ..
6251 (cpp_implicit_aliases_done): ... this one.
6252 (symtab_alias_ultimate_target): Add default parameter.
6253 (symtab_resolve_alias): New function.
6254 (fixup_same_cpp_alias_visibility): Declare.
6255 (cgraph_function_node): Add default parameter.
6256 (cgraph_node_asm_name): Likewise.
6257 (cgraph_function_or_thunk_node): Add default parameter; do
6258 not ICE when it is NULL.
6259 (varpool_variable_node): Likewise.
6260 * tree-emutls.c (create_emultls_var): Update.
6261 (ipa_lower_emutls): Update.
6262 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
6263 (cgraph_reset_node): Reset alias info.
6264 (cgraph_finalize_function): Update.
6265 (fixup_same_cpp_alias_visibility): Move to symtab.c.
6266 (analyze_function): Simplify.
6267 (cgraph_process_same_body_aliases): Simplify.
6268 (analyze_functions): Fixup same body aliases.
6269 (handle_alias_pairs): Simplify.
6270 (assemble_thunk): Update.
6271 (assemble_thunks_and_aliases): Update.
6272 (output_weakrefs): Rewrite.
6273 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
6274 (lto_output_varpool_node): Likewise.
6275 (compute_ltrans_boundary): Remve assert.
6276 (get_alias_symbol): New functoin.
6277 (input_node): Rewrite alias handling.
6278 (input_varpool_node): Likewise.
6279 * ipa-pure-const.c (propagate_pure_const): Fix formating.
6280 * ipa.c (process_references): Handle weakrefs correctly.
6281 (symtab_remove_unreachable_nodes): Likewise.
6282 * trans-mem.c (get_cg_data): Update.
6283 (ipa_tm_create_version_alias): Update.
6284 (ipa_tm_execute): Update.
6285 * symtab.c (dump_symtab_base): Dump aliases.
6286 (verify_symtab_base): Verify aliases.
6287 (symtab_node_availability): New function.
6288 (symtab_alias_ultimate_target): Simplify.
6289 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
6290 handle all the fixup cases.
6291 (symtab_resolve_alias): New function.
6292 * passes.c (ipa_write_summaries): Handle weakrefs.
6293 * varpool.c (varpool_analyze_node): Simplify.
6294 (assemble_aliases): Update.
6295 (varpool_create_variable_alias): Simplify.
6296 (varpool_extra_name_alias): Simplify.
6297 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
6298 (lto_symtab_merge_symbols): ... this one.
6300 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
6303 PR rtl-optimization/57268
6304 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6305 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6307 2013-06-01 Tobias Burnus <burnus@net-b.de>
6310 2013-05-31 Tobias Burnus <burnus@net-b.de>
6313 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
6316 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6318 PR rtl-optimization/57268
6319 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6320 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6322 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6324 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
6325 unordered comparison operators when -fno-trapping-math is in effect
6327 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
6328 and implement unordered comparison operators properly on the e500.
6330 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6332 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
6333 for constant scalar integers.
6334 (simplify_relational_operation_1): Likewise.
6336 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
6338 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
6339 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
6342 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6343 Igor Zamyatin <igor.zamyatin@intel.com>
6345 Silvermont (SLM) architecture performance tuning.
6346 * config/i386/i386.h (enum ix86_tune_indices): Add
6347 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
6348 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
6350 * config/i386/i386.c (initial_ix86_tune_features)
6351 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
6352 (ix86_lea_outperforms): Handle Silvermont tuning.
6353 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
6355 (ix86_use_lea_for_mov): Likewise.
6356 (ix86_avoid_lea_for_addr): Likewise.
6357 (ix86_lea_for_add_ok): Likewise.
6358 (exact_dependency_1): New function.
6359 (exact_store_load_dependency): Likewise.
6360 (ix86_adjust_cost): Handle Silvermont tuning.
6361 (do_reoder_for_imul): Likewise.
6362 (swap_top_of_ready_list): New function.
6363 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6365 * config/i386/i386.md (peepholes that split memory operand in fp
6368 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6370 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6371 Remove un-necessary braces.
6373 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6375 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6376 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6378 2013-05-31 Tobias Burnus <burnus@net-b.de>
6381 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6383 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6386 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6387 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6388 * config/arm/neon.md (iordi3_neon): Remove.
6389 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6390 * config/arm/predicates.md (imm_for_neon_logic_operand):
6391 Move to earlier in the file.
6392 (neon_logic_op2): Likewise.
6393 (arm_iordi_operand_neon): New predicate.
6395 2013-05-31 Richard Biener <rguenther@suse.de>
6397 PR tree-optimization/57478
6398 PR tree-optimization/57453
6399 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6402 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6404 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6405 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6407 2013-05-30 Tobias Burnus <burnus@net-b.de>
6408 Thomas Koenig <tkoenig@gcc.gnu.org>
6411 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6412 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6414 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6416 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6418 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6420 * target.def (register_usage_leveling_p): New hook.
6421 * targhooks.c (default_register_usage_leveling_p): New.
6422 * targhooks.h (default_register_usage_leveling_p): New prototype.
6423 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6424 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6425 * doc/tm.texi: Update.
6426 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6428 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6430 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6431 (*insv_reg<mode>): New define_insn.
6433 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6435 PR rtl-optimization/57439
6436 * postreload.c (move2add_valid_value_p): Check that we have
6437 a zero subreg_regno_offset when accessing the register in
6440 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6441 Igor Zamyatin <igor.zamyatin@intel.com>
6443 Silvermont (SLM) architecture pipeline model, tuning and
6445 * config.gcc: Add slm config options and target.
6447 * config/i386/slm.md: New.
6449 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6451 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6453 (ix86_target_macros_internal): Likewise.
6455 * gcc/config/i386/i386.c (slm_cost): New cost.
6456 (m_SLM): New macro flag.
6457 (initial_ix86_tune_features): Set m_SLM.
6458 (x86_accumulate_outgoing_args): Likewise.
6459 (x86_arch_always_fancy_math_387): Likewise.
6460 (processor_target_table): Add slm cost.
6461 (cpu_names): Add slm cpu name.
6462 (x86_option_override_internal): Set SLM ISA.
6463 (ix86_issue_rate): New case PROCESSOR_SLM.
6464 (ia32_multipass_dfa_lookahead): Likewise.
6465 (fold_builtin_cpu): Add slm.
6467 * config/i386/i386.h (TARGET_SLM): New target macro.
6468 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6469 (processor_type): Add PROCESSOR_SLM.
6471 * config/i386/i386.md (cpu): Add new value "slm".
6472 (slm.md): Include slm.md.
6474 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6475 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6477 * config/arm/arm-protos.h: Add and update function protos.
6478 * config/arm/arm.c (use_simple_return_p): New added.
6479 (thumb2_expand_return): Check simple_return flag.
6480 * config/arm/arm.md: Add simple_return and conditional simple_return.
6481 * config/arm/iterators.md: Add iterator for return and simple_return.
6483 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6485 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6486 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6487 (arm_emit_vfp_multi_reg_pop): Likewise.
6488 (thumb2_emit_ldrd_pop): Likewise.
6489 (arm_expand_epilogue): Add misc REG_CFA notes.
6490 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6492 2013-05-29 Lawrence Crowl <crowl@google.com>
6494 * config/arm/t-arm: Update for below.
6496 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6497 Change type to hash_table. Update dependent calls and types.
6499 * config/i386/t-cygming: Update for below.
6501 * config/i386/t-interix: Update for below.
6503 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6504 Change type to hash_table. Update dependent calls and types.
6505 (i386_find_on_wrapper_list::wrappers): Likewise.
6507 * config/ia64/t-ia64: Update for below.
6509 * config/ia64/ia64.c (bundle_state_table):
6510 Change type to hash_table. Update dependent calls and types.
6512 * config/mips/mips.c (mips_reorg_process_insns::htab):
6513 Change type to hash_table. Update dependent calls and types.
6515 * config/sol2.c (solaris_comdat_htab):
6516 Change type to hash_table. Update dependent calls and types.
6518 * config/t-sol2: Update for above.
6520 2013-05-29 Teresa Johnson <tejohnson@google.com>
6522 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6523 functions are not yet marked as defined.
6525 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6526 Pat Haugen <pthaugen@us.ibm.com>
6527 Peter Bergner <bergner@vnet.ibm.com>
6529 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6533 (vrotl<mode>3): Likewise.
6534 (vashl<mode>3): Likewise.
6535 (vlshr<mode>3): Likewise.
6536 (vashr<mode>3): Likewise.
6538 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6539 support for power8 V2DI builtins.
6541 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6542 power8 V2DI builtins.
6543 (vupkhsw): Likewise.
6544 (vupklsw): Likewise.
6545 (vaddudm): Likewise.
6550 (vpkudum): Likewise.
6551 (vpksdss): Likewise.
6552 (vpkudus): Likewise.
6553 (vpksdus): Likewise.
6558 (vsubudm): Likewise.
6559 (vcmpequd): Likewise.
6560 (vcmpgtsd): Likewise.
6561 (vcmpgtud): Likewise.
6562 (vcmpequd_p): Likewise.
6563 (vcmpgtsd_p): Likewise.
6564 (vcmpgtud_p): Likewise.
6565 (vupkhsw): Likewise.
6566 (vupklsw): Likewise.
6567 (vaddudm): Likewise.
6572 (vpksdss): Likewise.
6573 (vpksdus): Likewise.
6574 (vpkudum): Likewise.
6575 (vpkudus): Likewise.
6580 (vsubudm): Likewise.
6582 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6583 support for power8 V2DI instructions.
6585 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6586 power8 V2DI instructions. Combine pack and unpack insns to use an
6587 iterator for each mode. Check whether a particular mode supports
6588 Altivec instructions instead of just checking TARGET_ALTIVEC.
6589 (UNSPEC_VPKUWUM): Likewise.
6590 (UNSPEC_VPKSHSS): Likewise.
6591 (UNSPEC_VPKSWSS): Likewise.
6592 (UNSPEC_VPKUHUS): Likewise.
6593 (UNSPEC_VPKSHUS): Likewise.
6594 (UNSPEC_VPKUWUS): Likewise.
6595 (UNSPEC_VPKSWUS): Likewise.
6596 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6597 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6598 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6599 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6600 (UNSPEC_VUPKHSB): Likewise.
6601 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6602 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6603 (UNSPEC_VUPKHSH): Likewise.
6604 (UNSPEC_VUPKLSB): Likewise.
6605 (UNSPEC_VUPKLSH): Likewise.
6607 (VI_char): Likewise.
6608 (VI_scalar): Likewise.
6609 (VI_unit): Likewise.
6611 (VP_small): Likewise.
6612 (VP_small_lc): Likewise.
6613 (VU_char): Likewise.
6614 (add<mode>3): Likewise.
6615 (altivec_vaddcuw): Likewise.
6616 (altivec_vaddu<VI_char>s): Likewise.
6617 (altivec_vadds<VI_char>s): Likewise.
6618 (sub<mode>3): Likewise.
6619 (altivec_vsubcuw): Likewise.
6620 (altivec_vsubu<VI_char>s): Likewise.
6621 (altivec_vsubs<VI_char>s): Likewise.
6622 (altivec_vavgs<VI_char>): Likewise.
6623 (altivec_vcmpbfp): Likewise.
6624 (altivec_eq<mode>): Likewise.
6625 (altivec_gt<mode>): Likewise.
6626 (altivec_gtu<mode>): Likewise.
6627 (umax<mode>3): Likewise.
6628 (smax<mode>3): Likewise.
6629 (umin<mode>3): Likewise.
6630 (smin<mode>3): Likewise.
6631 (altivec_vpkuhum): Likewise.
6632 (altivec_vpkuwum): Likewise.
6633 (altivec_vpkshss): Likewise.
6634 (altivec_vpkswss): Likewise.
6635 (altivec_vpkuhus): Likewise.
6636 (altivec_vpkshus): Likewise.
6637 (altivec_vpkuwus): Likewise.
6638 (altivec_vpkswus): Likewise.
6639 (altivec_vpks<VI_char>ss): Likewise.
6640 (altivec_vpks<VI_char>us): Likewise.
6641 (altivec_vpku<VI_char>us): Likewise.
6642 (altivec_vpku<VI_char>um): Likewise.
6643 (altivec_vrl<VI_char>): Likewise.
6644 (altivec_vsl<VI_char>): Likewise.
6645 (altivec_vsr<VI_char>): Likewise.
6646 (altivec_vsra<VI_char>): Likewise.
6647 (altivec_vsldoi_<mode>): Likewise.
6648 (altivec_vupkhsb): Likewise.
6649 (altivec_vupkhs<VU_char>): Likewise.
6650 (altivec_vupkls<VU_char>): Likewise.
6651 (altivec_vupkhsh): Likewise.
6652 (altivec_vupklsb): Likewise.
6653 (altivec_vupklsh): Likewise.
6654 (altivec_vcmpequ<VI_char>_p): Likewise.
6655 (altivec_vcmpgts<VI_char>_p): Likewise.
6656 (altivec_vcmpgtu<VI_char>_p): Likewise.
6657 (abs<mode>2): Likewise.
6658 (vec_unpacks_hi_v16qi): Likewise.
6659 (vec_unpacks_hi_v8hi): Likewise.
6660 (vec_unpacks_lo_v16qi): Likewise.
6661 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6662 (vec_unpacks_lo_v8hi): Likewise.
6663 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6664 (vec_pack_trunc_v8h): Likewise.
6665 (vec_pack_trunc_v4si): Likewise.
6666 (vec_pack_trunc_<mode>): Likewise.
6668 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6670 (vec_vmaxsd): Likewise.
6671 (vec_vmaxud): Likewise.
6672 (vec_vminsd): Likewise.
6673 (vec_vminud): Likewise.
6674 (vec_vpksdss): Likewise.
6675 (vec_vpksdus): Likewise.
6676 (vec_vpkudum): Likewise.
6677 (vec_vpkudus): Likewise.
6678 (vec_vrld): Likewise.
6679 (vec_vsld): Likewise.
6680 (vec_vsrad): Likewise.
6681 (vec_vsrd): Likewise.
6682 (vec_vsubudm): Likewise.
6683 (vec_vupkhsw): Likewise.
6684 (vec_vupklsw): Likewise.
6686 2013-05-29 Jan Hubicka <jh@suse.cz>
6688 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6689 flags; reorder rest of fields in more consistent way.
6690 (varpool_node): Remove analyzed, finalized and alias.
6691 (cgraph_ndoe): Likewise.
6692 (symtab_alias_ultimate_target): New function.
6693 (cgraph_function_node): Move offline.
6694 (cgraph_reset_node): Declare.
6695 (cgraph_comdat_can_be_unshared_p): Remove.
6696 (varpool_remove_initializer): Declare.
6697 (varpool_first_defined_variable, varpool_next_defined_variable
6698 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6699 (cgraph_function_with_gimple_body_p): Update.
6700 (varpool_all_refs_explicit_p): Update.
6701 (symtab_alias_target): New function.
6702 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6703 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
6704 (cgraph_function_or_thunk_node): Simplify using
6705 symtab_alias_ultimate_target.
6706 (varpool_variable_node): Likewise.
6707 * cgraph.c (cgraph_create_function_alias): Update.
6708 (cgraph_add_thunk): Update.
6709 (cgraph_remove_node): Update.
6710 (dump_cgraph_node): Do not dump removed flags.
6711 (cgraph_function_body_availability): Update.
6712 (cgraph_propagate_frequency): Update.
6713 (verify_cgraph_node): Check sanity of local flag.
6714 (cgraph_function_node): Move here from cgraph.h; revamp for
6715 cgraph_function_or_thunk_node.
6716 * lto-symtab.c (lto_varpool_replace_node): Update.
6717 (lto_symtab_resolve_can_prevail_p): Update.
6718 (lto_symtab_merge_cgraph_nodes): Update.
6719 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6720 propagate_constants_accross_call, devirtualization_time_bonus,
6721 ipcp_propagate_stage): Update.
6722 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
6723 * ipa-inline-transform.c (clone_inlined_nodes,
6724 preserve_function_body_p): Update.
6725 * ipa-reference.c (propagate): Update.
6726 (write_node_summary_p): Update.
6727 * toplev.c (wrapup_global_declaration_2): Update.
6728 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6729 (analyze_function) ... this one.
6730 (cgraph_process_new_functions): Update.
6731 (cgraph_reset_node): Export.
6732 (cgraph_finalize_function): Update.
6733 (cgraph_add_new_function): Update.
6734 (process_function_and_variable_attributes): Update.
6735 (varpool_finalize_decl): Update.
6736 (symbol_finalized): Remove.
6737 (symbol_finalized_and_needed): Rename to ...
6738 (symbol_defined_and_needed): ... update.
6739 (cgraph_analyze_functions): Update.
6740 (handle_alias_pairs): Update.
6741 (mark_functions_to_output): Update.
6742 (assemble_thunk): Update.
6743 (output_in_order): Update.
6744 (output_weakrefs): Update.
6745 (finalize_compilation_unit): Update.
6746 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6747 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6748 input_node, input_varpool_node): Update.
6749 * dbxout.c (dbxout_expand_expr): Update.
6750 * cgraphclones.c (cgraph_clone_node): Update.
6751 (cgraph_copy_node_for_versioning): Update.
6752 (cgraph_materialize_clone): Update.
6753 (cgraph_materialize_all_clones): Update.
6754 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6755 propagate_pure_const, propagate_nothrow): Update.
6756 * lto-streamer-out.c (lto_output, write_symbol): Update.
6757 * ipa-utils.c (ipa_reverse_postorder): Update.
6758 * ipa-inline.c (can_inline_edge_p): Update.
6759 (update_caller_keys, ipa_inline): Update.
6760 * dwarf2out.c (reference_to_unused,
6761 premark_types_used_by_global_vars_helper): Update.
6762 * tree-eh.c (tree_could_trap_p): Update.
6763 * ipa-split.c (consider_split, execute_split_functions): Update.
6764 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
6765 has_addr_references_p): Update; move ahead in file for better
6767 (process_references): Simplify.
6768 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6770 (cgraph_comdat_can_be_unshared_p): Make static.
6771 (cgraph_externally_visible_p): Update.
6772 (varpool_externally_visible_p): Update.
6773 (function_and_variable_visibility): Update.
6774 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6775 ipa_tm_mark_force_output_node): Update.
6776 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6777 estimate_edge_devirt_benefit, inline_generate_summary,
6778 inline_write_summary): Update.
6779 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6780 * ipa-prop.c (ipa_compute_jump_functions): Update.
6781 (ipa_print_node_params, ipa_prop_read_section,
6782 ipa_update_after_lto_read, read_replacements_section): Update.
6783 * varasm.c (mark_decl_referenced): Update.
6784 (assemble_alias, dump_tm_clone_pairs): Update.
6785 * tree-inline.c (copy_bb): Update.
6786 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6788 * symtab.c (dump_symtab_base): Print new flags.
6789 (verify_symtab_base): Verify new flags.
6790 (symtab_alias_ultimate_target): New function.
6791 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6792 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6794 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6796 * i386.c (ix86_get_function_versions_dispatcher,
6797 ix86_generate_version_dispatcher_body): Update.
6798 (fold_builtin_cpu): Use varpool_add_new_variable.
6799 * varpool.c (varpool_remove_initializer): Break out from ...
6800 (varpool_remove_node): ... this one.
6801 (dump_varpool_node, varpool_node_for_asm,
6802 cgraph_variable_initializer_availability, varpool_analyze_node,
6803 varpool_assemble_decl, varpool_remove_unreferenced_decls,
6804 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6807 2013-05-29 Jan Hubicka <jh@suse.cz>
6809 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6811 2013-05-29 Easwaran Raman <eraman@google.com>
6813 PR tree-optimization/57442
6814 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6815 when control exits the main loop.
6817 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6819 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6821 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
6822 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6823 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6826 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6828 PR tree-optimization/57441
6829 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6830 Don't limit size of incr_vec to number of candidates.
6832 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
6834 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6835 and mips16 directories.
6836 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
6837 (MULTILIB_DIRNAMES): Ditto.
6838 (MULTILIB_EXCEPTIONS): Add new exceptions.
6839 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6840 (MULTILIB_DIRNAMES): Ditto.
6841 (MULTILIB_EXCEPTIONS): Add new exceptions.
6843 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6844 Marcus Shawcroft <marcus.shawcroft@arm.com>
6846 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6847 SYMBOL_TINY_ABSOLUTE.
6848 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6849 SYMBOL_TINY_ABSOLUTE.
6850 (aarch64_expand_mov_immediate): Likewise.
6851 (aarch64_classify_symbol): Likewise.
6852 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6853 Permit SYMBOL_TINY_ABSOLUTE.
6854 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6856 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6857 Marcus Shawcroft <marcus.shawcroft@arm.com>
6859 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6860 Refactor if/switch. Replace gcc_assert with if.
6862 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6864 * config/i386/i386.c (initial_ix86_tune_features): Enable
6865 FP Reassociation for AMD bdver1 and bdver2.
6867 2013-05-29 Martin Jambor <mjambor@suse.cz>
6869 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6870 and IMAGPART_EXPR do not occur within other handled_components.
6872 2013-05-29 Richard Biener <rguenther@suse.de>
6874 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6875 access on whether the use is in the BB we currently try to
6877 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6878 vectorize to vect_bb_slp_scalar_cost.
6880 2013-05-29 Richard Biener <rguenther@suse.de>
6882 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6883 computing scalar cost offsetted by stmts that are kept live
6885 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6886 for computation of scalar cost.
6888 2013-05-28 Steve Ellcey <sellcey@mips.com>
6890 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6892 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6894 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6895 array notation built-in reduction functions.
6896 * doc/passes.texi (Passes): Added documentation about changes done
6898 * doc/invoke.texi (C Dialect Options): Added documentation about
6899 the -fcilkplus flag.
6900 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6901 (BUILTINS_DEF): Depend on cilkplus.def.
6902 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6903 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6904 * cilkplus.def: New file.
6906 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6908 PR rtl-optimization/57439
6909 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6911 2013-05-28 Easwaran Raman <eraman@google.com>
6913 PR tree-optimization/57337
6914 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6915 (find_insert_point): Correctly identify the insertion point
6916 when two statements with the same UID is compared.
6918 2013-05-28 Richard Biener <rguenther@suse.de>
6920 PR tree-optimization/56787
6921 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6922 from the list of data references.
6923 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6925 (vect_analyze_loop_operations): Likewise.
6926 (vect_transform_loop): Remove clobbers.
6928 2013-05-28 Martin Jambor <mjambor@suse.cz>
6930 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6931 and REALPART_EXPRs have scalar type.
6933 2013-05-28 Richard Biener <rguenther@suse.de>
6935 PR tree-optimization/57411
6936 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6938 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6939 virtual operand propagation.
6941 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6943 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6944 destination register for bmasksi_vis.
6945 (vector_init_bshuffle): Likewise.
6946 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6948 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6950 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6951 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6952 mode if the instruction isn't available in the original mode.
6953 * config/sparc/sparc.opt (mfix-ut699): New option.
6954 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6955 (divdf3): Turn into expander.
6956 (divdf3_nofix): New insn.
6957 (divdf3_fix): Likewise.
6958 (divsf3): Disable if -mfix-ut699.
6959 (sqrtdf2): Turn into expander.
6960 (sqrtdf2_nofix): New insn.
6961 (sqrtdf2_fix): Likewise.
6962 (sqrtsf2): Disable if -mfix-ut699.
6964 2013-05-27 Richard Biener <rguenther@suse.de>
6967 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
6968 block for the new loop.
6970 2013-05-27 Richard Biener <rguenther@suse.de>
6972 PR tree-optimization/57343
6973 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
6974 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
6975 (number_of_iterations_cond): Do not build the folded tree.
6977 2013-05-27 Richard Biener <rguenther@suse.de>
6981 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
6982 OEP_CONSTANT_ADDRESS_OF retained.
6984 PR tree-optimization/57417
6985 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
6987 (set_ssa_val_to): Compare addresses using
6988 get_addr_base_and_unit_offset.
6990 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
6992 PR rtl-optimization/56833
6993 * postreload.c (move2add_record_mode): New function.
6994 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
6995 (move2add_use_add2_insn): Use move2add_record_sym_value.
6996 (move2add_use_add3_insn): Likewise.
6997 (reload_cse_move2add): Use move2add_valid_value_p and
6998 move2add_record_mode. Invalidate call-clobbered and REG_INC
6999 affected regs by setting reg_mode to VOIDmode.
7000 (move2add_note_store): Don't pretend the inside of a SUBREG is
7001 the actual destination. Invalidate single/leading registers by
7002 setting reg_mode to VOIDmode.
7003 Use move2add_record_sym_value, move2add_valid_value_p and
7004 move2add_record_mode.
7006 2013-05-27 Richard Biener <rguenther@suse.de>
7008 PR tree-optimization/57396
7009 * tree-affine.c (double_int_constant_multiple_p): Properly
7010 return false for val == 0 and div != 0.
7012 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7014 * config/mips/mips.h: Use #elif in preprocessor conditions.
7016 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7019 * config/mips/constraints.md (kl): New constraint.
7020 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7021 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7022 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7024 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7026 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7029 * config/mips/mips.c (mips_can_inline_p): New function.
7030 (TARGET_CAN_INLINE_P): Define.
7032 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7034 * sched-int.h (ds_t, dw_t): Make unsigned int.
7035 Fix documentation that describes how all the ds_t bits are used.
7036 Reserve the last bit for delayed-branch scheduling.
7037 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7038 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7039 (gen_dep_weak_1): Remove prototype.
7040 * sched-deps.c (get_dep_weak_1): Make static.
7041 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7042 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7043 * doc/tm.texi: Regenerate.
7044 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7046 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7049 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7051 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7052 Sandra Loosemore <sandra@codesourcery.com>
7054 * config.gcc (powerpc-*): Allow native for with-cpu.
7056 2013-05-24 Jeff Law <law@redhat.com>
7058 PR tree-optimization/57124
7059 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7060 conversion feeding a condition if the range has an overflow
7061 if -fstrict-overflow. Add warnings for when we do make the
7064 2013-05-24 Dehao Chen <dehao@google.com>
7066 * tree-cfg.c (locus_discrim_map): Fix the typo.
7067 (locus_discrim_hasher): Likewise.
7068 (locus_discrim_hasher::hash): Likewise.
7069 (locus_discrim_hasher::equal): Likewise.
7071 2013-05-24 Martin Jambor <mjambor@suse.cz>
7073 PR tree-optimization/57294
7074 * cgraph.h (ipa_record_stmt_references): Declare.
7075 * cgraphbuild.c (ipa_record_stmt_references): New function.
7076 (build_cgraph_edges): Use ipa_record_stmt_references.
7077 (rebuild_cgraph_edges): Likewise.
7078 (cgraph_rebuild_references): Likewise.
7079 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7080 associated with the old statement and build references from the
7081 newly built statements.
7082 * ipa-ref.c (ipa_remove_stmt_references): New function.
7083 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7085 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7087 * lra-constraints.c (emit_spill_move): Use smaller mode for
7089 (check_and_process_move): Consider mem-reg moves for secondary
7091 (curr_insn_transform): Don't lose insns emitted before for
7092 secondary memory moves.
7093 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7094 reg set up in the current insn.
7096 2013-05-24 Dehao Chen <dehao@google.com>
7098 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7100 (locus_descrim_hasher::equal): Likewise.
7101 (build_gimple_cfg): New discriminator assignment algorithm.
7102 (make_edges): Likewise.
7103 (next_discriminator_for_locus): Likewise.
7104 (same_line_p): Likewise.
7105 (assign_discriminators): Likewise.
7106 (make_cond_expr_edges): Likewise.
7107 (make_gimple_switch_edges): Likewise.
7108 (make_goto_expr_edges): Likewise.
7109 (make_gimple_asm_edges): Likewise.
7111 2013-05-24 Ian Bolton <ian.bolton@arm.com>
7113 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
7114 X format specifier to only display bottom 16 bits.
7115 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
7116 immediate to match for operand 2, since it will be masked.
7118 2013-05-24 Richard Biener <rguenther@suse.de>
7120 PR tree-optimization/57287
7121 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
7122 all SSA names that occur in abnormal PHIs.
7124 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
7127 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
7128 that index is not negative.
7130 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
7132 PR rtl-optimization/55177
7133 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
7134 (simplify_byte_swapping_operation): New.
7135 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
7136 (simplify_relational_operation_1): Deal with BSWAP.
7138 2013-05-23 Richard Henderson <rth@redhat.com>
7141 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
7142 (ix86_reorg): Call it.
7144 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
7147 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
7148 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
7149 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
7151 2013-05-23 Christian Bruel <christian.bruel@st.com>
7154 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
7156 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7157 Marcus Shawcroft <marcus.shawcroft@arm.com>
7159 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
7160 * config/aarch64/constraints.md (Usa): Remove.
7161 * doc/md.texi (AArch64 Usa): Remove.
7163 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7164 Marcus Shawcroft <marcus.shawcroft@arm.com>
7166 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
7167 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
7168 * config/aarch64/predicates.md (aarch64_const_address): Remove.
7169 (aarch64_mov_operand): Use aarch64_mov_operand_p.
7171 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
7173 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
7174 instruction (AdvSIMD).
7175 * config/aarch64/aarch64-builtins.c
7176 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
7177 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
7179 2013-05-23 Martin Jambor <mjambor@suse.cz>
7182 * tree.h (contains_bitfld_component_ref_p): Declare.
7183 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
7184 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
7186 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
7187 not access a bit-field. Assert all final offsets are byte-aligned.
7189 2013-05-23 Richard Biener <rguenther@suse.de>
7191 PR tree-optimization/57380
7192 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
7193 least one invariant or re-used load.
7194 * passes.c (init_optimization_passes): Move pass_phiprop before
7197 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
7199 * config/aarch64/aarch64-simd.md
7200 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
7202 2013-05-23 Richard Biener <rguenther@suse.de>
7205 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7206 OEP_CONSTANT_ADDRESS_OF retained.
7208 2013-05-23 Jakub Jelinek <jakub@redhat.com>
7211 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
7212 don't lower unit. Handle unit not being always BITS_PER_WORD.
7214 2013-05-23 Richard Biener <rguenther@suse.de>
7216 PR rtl-optimization/57341
7217 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
7218 instead of true_dependence.
7220 2013-05-22 David Malcolm <dmalcolm@redhat.com>
7222 * bb-reorder.c (branch_threshold): Make const.
7223 (exec_threshold): Ditto.
7225 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7226 Pat Haugen <pthaugen@us.ibm.com>
7227 Peter Bergner <bergner@vnet.ibm.com>
7229 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
7230 documentation for the power8 crypto builtins.
7232 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
7234 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
7235 macros for defining power8 builtin functions.
7236 (BU_P8V_AV_2): Likewise.
7237 (BU_P8V_AV_P): Likewise.
7238 (BU_P8V_VSX_1): Likewise.
7239 (BU_P8V_OVERLOAD_1): Likewise.
7240 (BU_P8V_OVERLOAD_2): Likewise.
7241 (BU_CRYPTO_1): Likewise.
7242 (BU_CRYPTO_2): Likewise.
7243 (BU_CRYPTO_3): Likewise.
7244 (BU_CRYPTO_OVERLOAD_1): Likewise.
7245 (BU_CRYPTO_OVERLOAD_2): Likewise.
7246 (XSCVSPDP): Fix typo, point to the correct instruction.
7247 (VCIPHER): Add power8 crypto builtins.
7248 (VCIPHERLAST): Likewise.
7249 (VNCIPHER): Likewise.
7250 (VNCIPHERLAST): Likewise.
7251 (VPMSUMB): Likewise.
7252 (VPMSUMH): Likewise.
7253 (VPMSUMW): Likewise.
7254 (VPERMXOR_V2DI): Likewise.
7255 (VPERMXOR_V4SI: Likewise.
7256 (VPERMXOR_V8HI: Likewise.
7257 (VPERMXOR_V16QI: Likewise.
7258 (VSHASIGMAW): Likewise.
7259 (VSHASIGMAD): Likewise.
7261 (VPERMXOR): Likewise.
7262 (VSHASIGMA): Likewise.
7264 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7265 __CRYPTO__ if the crypto instructions are available.
7266 (altivec_overloaded_builtins): Add support for overloaded power8
7269 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
7270 support for power8 crypto builtins.
7271 (builtin_function_type): Likewise.
7272 (altivec_init_builtins): Add support for builtins that take vector
7273 long long (V2DI) arguments.
7275 * config/rs6000/crypto.md: New file, define power8 crypto
7278 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7279 Pat Haugen <pthaugen@us.ibm.com>
7280 Peter Bergner <bergner@vnet.ibm.com>
7282 * doc/invoke.texi (Option Summary): Add power8 options.
7283 (RS/6000 and PowerPC Options): Likewise.
7285 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
7286 constraints.md instead of rs6000.h. Reorder w* constraints. Add
7287 wm, wn, wr documentation.
7289 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
7290 registers if direct move instructions are enabled.
7291 (wn): New constraint for no registers.
7292 (wq): New constraint for quad word even GPR registers.
7293 (wr): New constraint if 64-bit instructions are enabled.
7294 (wv): New constraint if power8 vector instructions are enabled.
7295 (wQ): New constraint for quad word memory locations.
7297 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
7298 constraint for 0..15 for crypto instructions.
7299 (gpc_reg_operand): If VSX allow registers in VSX registers as well
7300 as GPR and floating point registers.
7301 (int_reg_operand): New predicate to match only GPR registers.
7302 (base_reg_operand): New predicate to match base registers.
7303 (quad_int_reg_operand): New predicate to match even GPR registers
7304 for quad memory operations.
7305 (vsx_reg_or_cint_operand): New predicate to allow vector logical
7306 operations in both GPR and VSX registers.
7307 (quad_memory_operand): New predicate for quad memory operations.
7308 (reg_or_indexed_operand): New predicate for direct move support.
7310 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
7311 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
7312 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
7313 (POWERPC_MASKS): Add power8 options.
7314 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
7317 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7318 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
7320 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
7321 (-mpower8-fusion): New power8 options.
7322 (-mpower8-fusion-sign): Likewise.
7323 (-mpower8-vector): Likewise.
7324 (-mcrypto): Likewise.
7325 (-mdirect-move): Likewise.
7326 (-mquad-memory): Likewise.
7328 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
7330 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
7332 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
7333 (rs6000_debug_vector_unit): Add p8_vector.
7334 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
7335 definitions. Also print fusion state.
7336 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
7337 (rs6000_builtin_mask_calculate): Add power8 builtin support.
7338 (rs6000_option_override_internal): Add support for power8.
7339 (rs6000_common_init_builtins): Add debugging for skipped builtins
7341 (rs6000_adjust_cost): Add power8 support.
7342 (rs6000_issue_rate): Likewise.
7343 (insn_must_be_first_in_group): Likewise.
7344 (insn_must_be_last_in_group): Likewise.
7345 (force_new_group): Likewise.
7346 (rs6000_register_move_cost): Likewise.
7347 (rs6000_opt_masks): Likewise.
7349 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
7350 power8 capable assembler, default to power7 options.
7351 (TARGET_DIRECT_MOVE): Likewise.
7352 (TARGET_CRYPTO): Likewise.
7353 (TARGET_P8_VECTOR): Likewise.
7354 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
7355 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
7356 (VECTOR_MEM_P8_VECTOR_P): Likewise.
7357 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
7358 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
7359 (TARGET_XSCVDPSPN): Likewise.
7360 (TARGET_XSCVSPDPN): Likewsie.
7361 (TARGET_SYNC_HI_QI): Likewise.
7362 (TARGET_SYNC_TI): Likewise.
7363 (MASK_CRYPTO): Likewise.
7364 (MASK_DIRECT_MOVE): Likewise.
7365 (MASK_P8_FUSION): Likewise.
7366 (MASK_P8_VECTOR): Likewise.
7367 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7368 temporary used by some of the direct move instructions to get two FP
7369 temporary registers does not force creation of a stack frame.
7370 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7371 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7372 that any VSX registers are tieable, even if they are also an
7373 Altivec vector mode.
7374 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7375 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7376 (RS6000_BTM_CRYPTO): Likewise.
7377 (RS6000_BTM_COMMON): Likewise.
7379 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7380 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7381 (enum rs6000_vector): Add power8 vector support.
7383 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7387 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7388 (any_sibcall_could_use_r3): this and handle indirect calls.
7389 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7391 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7393 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7395 2013-05-22 Richard Biener <rguenther@suse.de>
7398 * profile.c (branch_prob): Do not split blocks that are
7399 abnormally receiving from ECF_RETURNS_TWICE functions.
7401 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7403 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7404 address mode. Move pointer mode initialization to the same place.
7406 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7408 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7409 while it has any effect.
7411 2013-05-21 Easwaran Raman <eraman@google.com>
7413 PR tree-optimization/57322
7414 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7415 UID of the statement added to the BB to be 1.
7417 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7419 PR tree-optimization/57331
7420 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7421 of conversion from pointer type to integral type with integer.
7423 2013-05-21 Martin Jambor <mjambor@suse.cz>
7426 * ipa-prop.c (ipa_read_node_info): Process param_used and
7427 controlled_uses in the same order as when writing.
7429 2013-05-21 Magnus Granberg <baldrick@free.fr>
7432 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7434 2013-05-21 Richard Biener <rguenther@suse.de>
7436 PR tree-optimization/57318
7437 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7438 estimate stmts with side-effects as likely eliminated.
7440 2013-05-21 Richard Biener <rguenther@suse.de>
7442 PR tree-optimization/57330
7443 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7444 preserve the call stmts fntype.
7446 2013-05-21 Richard Biener <rguenther@suse.de>
7448 PR tree-optimization/57303
7449 * tree-ssa-sink.c (statement_sink_location): Improve killing
7450 stmt detection and properly handle self-assignments.
7452 2013-05-21 Christian Bruel <christian.bruel@st.com>
7454 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7455 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7456 registers. Set register size out of the PARALLEL loop.
7458 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7461 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7462 (*fmasf4, *fmasf4_media): New insns.
7464 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7466 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7467 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7468 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7469 (mips_idiv_insns): Update the comments to say that the returned
7470 instruction counts are in units of BASE_INSN_LENGTH.
7471 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7472 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7473 using 2 rather than 4 as the length of indirect MIPS16 and
7474 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7475 length of a NOP. Don't divide MIPS16 lengths by 2.
7476 (mips16_split_long_branches): Assume a branch is long if the
7477 length is greater than 4 rather than 8.
7478 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7479 rather than multiplying them by 2. Multiply instruction counts
7480 by BASE_INSN_LENGTH rather than 4.
7481 (*jump_mips16, tls_get_tp_mips16_<mode>)
7482 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7484 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7486 * config/mips/mips.md (extended_mips16): Remove branch case.
7487 (length): Remove duplicated extended_mips16 test.
7489 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7491 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7493 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7495 * recog.h (Recog_data): Rename to...
7496 (recog_data_d): ...this.
7497 (recog_data): Update accordingly.
7498 * recog.c (recog_data): Likewise.
7499 * reload.c (save_recog_data): Likewise.
7500 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7501 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7503 2013-05-17 Julian Brown <julian@codesourcery.com>
7505 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7506 found in a REG_EQUAL note, invalidate it.
7508 2013-05-17 Easwaran Raman <eraman@google.com>
7510 * tree-ssa-reassoc.c (find_insert_point): New function.
7511 (insert_stmt_after): Likewise.
7512 (get_def_stmt): Likewise.
7513 (ensure_ops_are_available): Likewise.
7514 (not_dominated_by): Likewise.
7515 (rewrite_expr_tree): Do not move statements beyond what is
7516 necessary. Remove call to swap_ops_for_binary_stmt...
7517 (reassociate_bb): ... and move it here.
7518 (build_and_add_sum): Assign UIDs for new statements.
7519 (linearize_expr): Likewise.
7520 (do_reassoc): Renumber gimple statement UIDs.
7522 2013-05-17 Jan Hubicka <jh@suse.cz>
7524 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7526 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
7527 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7529 (output_weakrefs): Update.
7531 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7532 Martin Jambor <mjambor@suse.cz>
7535 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7536 value that corresponds to the given aggval is found in values vector.
7538 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
7540 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7541 sse, sse2, sse3, ssse3 and sse4a flags to options.
7543 2013-05-17 David Malcolm <dmalcolm@redhat.com>
7545 * gengtype-state.c: (s_expr_writer): New class, to handle
7546 prettifying of output layout of s-expressions.
7547 (state_writer): New class, to write out gtype.state.
7548 (state_written_type_count): Move this variable into member data of
7550 (s_expr_writer::s_expr_writer): New code: constructor for new class
7551 (state_writer::state_writer(): ditto
7552 (s_expr_writer::write_new_line): New function
7553 (s_expr_writer::write_any_indent): ditto
7554 (s_expr_writer::begin_s_expr): ditto
7555 (s_expr_writer::end_s_expr): ditto
7556 (write_state_fileloc): convert to method of state_writer...
7557 (state_writer:: write_state_fileloc): ...and use methods of
7558 s_expr_writer to write indentation into the gtype.state output file
7559 to visually represent the hierarchical structure of the list
7561 (write_state_fields): ditto, renaming to...
7562 (state_writer::write_state_fields)
7563 (write_state_a_string): ditto, renaming to...
7564 (state_writer::write_state_a_string)
7565 (write_state_string_option): ditto, renaming to...
7566 (state_writer::write_state_string_option)
7567 (write_state_type_option): ditto, renaming to...
7568 (state_writer::write_state_type_option)
7569 (write_state_nested_option): ditto, renaming to...
7570 (state_writer::write_state_nested_option)
7571 (write_state_option): ditto, renaming to...
7572 (state_writer::write_state_option)
7573 (write_state_options): ditto, renaming to...
7574 (state_writer::write_state_options)
7575 (write_state_lang_bitmap): ditto, renaming to...
7576 (state_writer::write_state_lang_bitmap)
7577 (write_state_version): ditto, renaming to...
7578 (state_writer::write_state_version)
7579 (write_state_scalar_type): ditto, renaming to...
7580 (state_writer::write_state_scalar_type)
7581 (write_state_string_type): ditto, renaming to...
7582 (state_writer::write_state_string_type)
7583 (write_state_undefined_type): ditto, renaming to...
7584 (state_writer::write_state_undefined_type)
7585 (write_state_struct_union_type): ditto, renaming to...
7586 (state_writer::write_state_struct_union_type)
7587 (write_state_struct_type): ditto, renaming to...
7588 (state_writer::write_state_struct_type)
7589 (write_state_user_struct_type): ditto, renaming to...
7590 (state_writer::write_state_user_struct_type)
7591 (write_state_lang_struct_type): ditto, renaming to...
7592 (state_writer::write_state_lang_struct_type)
7593 (write_state_param_struct_type): ditto, renaming to...
7594 (state_writer::write_state_param_struct_type)
7595 (write_state_pointer_type): ditto, renaming to...
7596 (state_writer::write_state_pointer_type)
7597 (write_state_array_type): ditto, renaming to...
7598 (state_writer::write_state_array_type)
7599 (write_state_gc_used): ditto, renaming to...
7600 (state_writer::write_state_gc_used)
7601 (write_state_common_type_content): ditto, renaming to...
7602 (state_writer::write_state_common_type_content)
7603 (write_state_type): ditto, renaming to...
7604 (state_writer::write_state_type)
7605 (write_state_pair_list): ditto, renaming to...
7606 (state_writer::write_state_pair_list)
7607 (write_state_pair): ditto, renaming to...
7608 (state_writer::write_state_pair)
7609 (write_state_typedefs): ditto, renaming to...
7610 (state_writer::write_state_typedefs)
7611 (write_state_structures): ditto, renaming to...
7612 (state_writer::write_state_structures)
7613 (write_state_param_structs): ditto, renaming to...
7614 (state_writer::write_state_param_structs)
7615 (write_state_variables): ditto, renaming to...
7616 (state_writer::write_state_variables)
7617 (write_state_srcdir): ditto, renaming to...
7618 (state_writer::write_state_srcdir)
7619 (write_state_files_list): ditto, renaming to...
7620 (state_writer::write_state_files_list)
7621 (write_state_languages): ditto, renaming to...
7622 (state_writer::write_state_languages)
7623 (write_state): create a state_writer instance and use it when
7624 writing out the state file
7626 2013-05-17 Mike Stump <mikestump@comcast.net>
7628 PR rtl-optimization/57304
7629 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7630 accessing DF_REF_REAL_LOC.
7632 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7634 PR rtl-optimization/57281
7635 PR rtl-optimization/57300
7636 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7637 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7638 what the other splitter did if the registers are dead.
7640 2013-05-17 Richard Biener <rguenther@suse.de>
7642 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7645 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7647 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7650 2013-05-17 Marek Polacek <polacek@redhat.com>
7652 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7653 length when doing non-zero store of storing '\0' to '\0'.
7655 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7657 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7658 vect_external_def oprnd1 with loop_vinfo, try to emit
7659 optional cast, negation and and stmts on the loop preheader
7660 edge instead of into the pattern def seq.
7662 PR tree-optimization/57051
7663 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7664 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7666 2013-05-16 Nick Clifton <nickc@redhat.com>
7668 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7669 (rl78_is_naked_func): New function.
7670 (rl78_expand_prologue): Skip prologue generation for naked functions.
7671 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
7672 * doc/extend.texi (naked): Add RL78 to the list of processors
7673 that supports this attribute.
7675 2013-05-16 Jeff Law <law@redhat.com>
7677 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7679 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7681 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7682 cache parameters using detect_caches_amd also for CYRIX,
7683 NSC and TM2 signatures.
7685 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7686 Dzianis Kahanovich <mahatma@eu.by>
7690 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7691 VIA/Centaur processors and determine their cache parameters
7692 using detect_caches_amd.
7694 2013-05-16 Teresa Johnson <tejohnson@google.com>
7696 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7697 (rtl_verify_edges): New function.
7698 (rtl_verify_bb_insns): Ditto.
7699 (rtl_verify_bb_pointers): Ditto.
7700 (rtl_verify_bb_insn_chain): Ditto.
7701 (rtl_verify_fallthru): Ditto.
7702 (rtl_verify_bb_layout): Ditto.
7703 (rtl_verify_flow_info_1): Outline checks into new functions.
7704 (rtl_verify_flow_info): Ditto.
7706 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
7708 * cfghooks.c (copy_bbs): Add update_dominance argument.
7709 * cfghooks.h (copy_bbs): Update prototype.
7710 * tree-cfg.c (gimple_duplicate_sese_region):
7711 Add update_dominance argument.
7712 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7713 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7714 gimple_duplicate_sese_region call.
7715 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7716 Update copy_bbs call.
7717 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7718 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7720 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7722 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7723 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7724 vect_recog_rotate_pattern.
7725 (vect_recog_rotate_pattern): New function.
7727 2013-05-16 Jason Merrill <jason@redhat.com>
7729 * Makefile.in (LLINKER): New variable.
7730 (mostlyclean): Remove link mutex.
7731 * configure.ac: Handle --enable-link-mutex.
7732 * lock-and-run.sh: New script.
7734 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7737 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7740 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7742 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7744 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7746 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7747 * config/arm/arm.c (next_consecutive_mem): New function.
7748 (gen_movmem_ldrd_strd): Likewise.
7749 * config/arm/arm.md (movmemqi): Update condition and code.
7750 (unaligned_loaddi, unaligned_storedi): New patterns.
7752 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7754 * config.gcc: Obsolete *-*-solaris2.9*.
7755 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7757 2013-05-16 Richard Biener <rguenther@suse.de>
7759 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7760 earlier, after GRAPHITE transforms and IV canonicalization.
7762 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7764 * omp-low.c (extract_omp_for_data): For collapsed loops,
7765 if at least one of the loops is known at compile time to
7766 iterate zero times, set count to 0.
7767 (expand_omp_regimplify_p): New function.
7768 (expand_omp_for_generic): For collapsed loops, if at least
7769 one of the loops isn't known to iterate at least once,
7770 add runtime check with setting count to 0.
7771 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7772 For unsigned types if it isn't known at compile time that
7773 the loop will iterate at least once, add runtime check to bypass
7774 the whole loop if initial condition isn't true.
7776 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7778 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7780 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
7783 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7784 transformations to avoid an infinite loop.
7786 2013-05-16 Marek Polacek <polacek@redhat.com>
7788 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7790 2013-05-15 Leif Ekblad <leif@rdos.net>
7792 * config/i386/i386.c (ix86_decompose_address): Use
7793 DEFAULT_TLS_SEG_REG to access TLS segment register.
7794 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7795 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7796 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7798 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7801 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7802 sibling calls to functions that would normally be lazily bound,
7803 unless $gp is call-clobbered.
7805 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
7807 * config/i386/i386.c (ix86_option_override_internal): Update
7808 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
7809 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7810 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
7811 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7813 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7815 2013-05-15 Andreas Schwab <schwab@suse.de>
7817 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7818 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7821 2013-05-15 Teresa Johnson <tejohnson@google.com>
7823 * loop-unroll.c (report_unroll_peel): Check decision before
7824 emitting unroll/peel message.
7826 2013-05-15 Teresa Johnson <tejohnson@google.com>
7828 * function.h (has_bb_partition): New rtl_data flag.
7829 (bb_reorder_complete): Ditto.
7830 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7831 instead of flag_reorder_blocks_and_partition.
7832 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7833 with some enhancements.
7834 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7835 * bb-reorder.c (connect_traces): Check for has_bb_partition
7836 instead of flag_reorder_blocks_and_partition.
7837 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7838 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7839 verify_hot_cold_block_grouping.
7840 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7842 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7845 * config/arm/predicates.md (call_insn_operand): New predicate.
7846 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7847 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7848 if insn is not a tail call.
7849 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7851 * config/arm/arm.h (enum reg_class): New caller save register class.
7852 (REG_CLASS_NAMES): Likewise.
7853 (REG_CLASS_CONTENTS): Likewise.
7854 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7857 2013-05-15 Richard Biener <rguenther@suse.de>
7859 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7860 of MSG_OPTIMIZED_LOCATIONS.
7861 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7862 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7864 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7865 of MSG_OPTIMIZED_LOCATIONS.
7866 (execute_vect_slp): Likewise.
7867 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7868 (vect_create_cond_for_alias_checks): Likewise.
7869 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7870 (vect_recog_widen_mult_pattern): Likewise.
7871 (vect_recog_widen_sum_pattern): Likewise.
7872 (vect_recog_over_widening_pattern): Likewise.
7873 (vect_recog_widen_shift_pattern): Likewise.
7874 (vect_recog_vector_vector_shift_pattern): Likewise.
7875 (vect_recog_divmod_pattern): Likewise.
7876 (vect_recog_mixed_size_cond_pattern): Likewise.
7877 (vect_recog_bool_pattern): Likewise.
7878 (vect_pattern_recog_1): Likewise.
7880 2013-05-15 Martin Jambor <mjambor@suse.cz>
7882 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7883 non-functions to builtin_unreachable.
7884 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7885 correct when new direct edges were discovered.
7887 2013-05-15 Martin Jambor <mjambor@suse.cz>
7889 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7890 header, print symbol order instead of node uid, print more information
7891 about indirect edge targets.
7892 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7894 (ipa_make_edge_direct_to_target): Likewise.
7895 (remove_described_reference): Likewise.
7896 (propagate_controlled_uses): Likewise.
7897 (ipa_print_node_params): Also print symbol order.
7898 (ipcp_transform_function): Print symbol order instead of node uids.
7899 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7900 (cgraph_get_create_real_symbol_node): Likewise.
7901 * ipa-cp.c (print_lattice): Likewise.
7902 (print_all_lattices): Likewise.
7903 (determine_versionability): Likewise.
7904 (initialize_node_lattices): Likewise.
7905 (estimate_local_effects): Likewise.
7906 (update_profiling_info): Likewise.
7907 (create_specialized_node): Likewise.
7908 (perhaps_add_new_callers): Likewise.
7909 (decide_about_value): Likewise.
7910 (decide_whether_version_node): Likewise.
7911 (identify_dead_nodes): Likewise.
7912 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7913 (dump_inline_summary): Likewise.
7914 (estimate_node_size_and_time): Likewise.
7915 (inline_analyze_function): Likewise.
7916 * ipa-inline.c (report_inline_failed_reason): Likewise.
7917 (want_early_inline_function_p): Likewise.
7918 (edge_badness): Likewise.
7919 (update_edge_key): Likewise.
7920 (inline_small_functions): Likewise. Add dumping of order to two other
7922 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7923 instead of node uids.
7924 (propagate_pure_const): Likewise.
7925 (propagate_pure_const): Likewise.
7926 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7927 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7928 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7930 * tree-pretty-print.c (dump_function_header): Likewise.
7931 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7932 Print symbol order instead of node uids.
7934 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7936 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7937 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7939 2013-05-15 Richard Biener <rguenther@suse.de>
7941 PR tree-optimization/57275
7942 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7943 return value for fail to do runtime alias checks for gather loads.
7945 2013-05-15 Jan Hubicka <jh@suse.cz>
7949 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7950 weakrefs are not external.
7951 (lto_symtab_merge_decls): Fix thinko when dealing with
7952 non-lto_symtab decls.
7953 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7954 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7955 * varpool.c (dump_varpool_node): Dump more flags.
7957 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7959 * config/i386/i386.c (processor_alias_table): Add instruction
7960 FSGSBASE for AMD bdver3 architecture.
7962 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7964 * tree.c (warn_deprecated_use): Print file:line using locus color.
7965 * diagnostic.c (diagnostic_report_current_module): Print file:line
7966 and file:line:column using locus color.
7968 2013-05-14 Mike Stump <mikestump@comcast.net>
7970 * gdbinit.in: Add __null.
7972 2013-05-14 Mike Stump <mikestump@comcast.net>
7974 * recog.h: Rename struct recog_data to Recog_data.
7975 * recog.c: Likewise.
7976 * reload.c (can_reload_into): Likewise.
7977 * config/picochip/picochip.c: Likewise.
7979 2013-05-14 Mike Stump <mikestump@comcast.net>
7981 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
7983 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
7985 * resource.h (struct resources): Remove unch_memory member.
7986 (CLEAR_RESOURCE): Don't clear unch_memory.
7987 * resource.c (mark_referenced_resources): Don't set it.
7988 (mark_set_resources): Likewise.
7989 (mark_target_live_regs): Don't clear it.
7990 (init_resource_info): Likewise.
7991 * reorg.c (resource_conflicts_p): Don't compare it.
7992 (redundant_insn): Don't set it.
7994 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
7996 * emit-rtl.c (next_label): Remove unused function.
7997 (skip_consecutive_labels, link_cc0_insns): Move to ...
7998 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
7999 only place where these functions are used, and make them static.
8001 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8003 * fold-const.c (fold_negate_expr): Handle vectors.
8004 (fold_truth_not_expr): Make it static.
8005 (fold_invert_truthvalue): New static function.
8006 (invert_truthvalue_loc): Handle vectors. Do not call
8007 fold_truth_not_expr directly.
8008 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8009 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8010 (fold_comparison): Handle vector constants.
8011 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8012 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8013 * tree.h (fold_truth_not_expr): Remove declaration.
8015 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8017 * config/aarch64/aarch64-simd.md
8018 (aarch64_vcond_internal<mode>): Rename to...
8019 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8020 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8021 float modes. Clarify all iterator modes.
8022 (vcond<mode><mode>): Use new name for vcond expanders.
8023 (vcond<v_cmp_result><mode>): Likewise.
8024 (vcondu<mode><mode>: Likewise.
8025 * config/aarch64/iterators.md (VDQF_COND): New.
8027 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8030 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8031 variable for the shift amount. Check that we shift by non-negative
8034 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8037 * config/arm/arm.h (EPILOGUE_USES): Only return true
8038 for LR_REGNUM after epilogue_completed.
8040 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8042 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8045 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8048 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8049 and Solaris 11+/x86 with gld.
8050 * configure: Regenerate.
8052 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8054 * expmed.c (expand_shift_1): Canonicalize rotates by
8055 constant bitsize / 2 to bitsize - 1.
8056 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8057 case ROTATERT>: Likewise.
8060 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8062 * config/i386/i386.md (rotateinv): New code attr.
8063 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8064 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8065 roll $31, %eax, etc.
8067 2013-05-14 Richard Biener <rguenther@suse.de>
8070 * tree-eh.c (sink_clobbers): Give up for successors with
8071 multiple predecessors and no virtual uses.
8073 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8075 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8076 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8078 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8081 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8082 the case when both op0 and op1 have VOIDmode.
8084 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8086 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8087 in multiply-accumulate mode.
8089 2013-05-13 Guozhi Wei <carrot@google.com>
8091 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8093 2013-05-13 Kai Tietz <ktietz@redhat.com>
8096 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8097 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8098 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8099 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8100 (ix86_expand_prologue): Likewise.
8101 (ix86_expand_split_stack_prologue): Likewise.
8102 (legitimate_pic_address_disp_p): Likewise.
8103 (legitimize_pic_address): Likewise.
8104 (legitimize_tls_address): Likewise.
8105 (legitimize_pe_coff_symbol): Likewise.
8106 (output_pic_addr_const): Likewise.
8107 (construct_plt_address): Likewise.
8108 (ix86_expand_call): Likewise.
8109 (x86_output_mi_thunk): Likewise.
8110 (x86_function_profiler): Likewise.
8112 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
8114 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
8115 similar switch cases.
8116 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
8117 (aarch64_simd_mov_to_<mode>low): Delete.
8118 (aarch64_simd_mov_to_<mode>high): Delete.
8119 (move_lo_quad_<mode>): Add w<-r alternative.
8120 (aarch64_simd_move_hi_quad_<mode>): Likewise.
8121 (aarch64_simd_mov_from_*): Update type attribute.
8122 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
8125 2013-05-13 Jan Hubicka <jh@suse.cz>
8127 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
8128 * config/i386/i386.c (ix86_compute_frame_layout,
8129 ix86_expand_epilogue, emit_i387_cw_initialization,
8130 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
8131 ix86_local_alignment): Fix use of size/speed predicates.
8133 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8135 PR tree-optimization/45216
8136 PR tree-optimization/57157
8137 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
8138 the (-Y) & (B - 1) variant if OP is |.
8139 * expmed.c (expand_shift_1): For rotations by const0_rtx just
8140 return shifted. Use (-op1) & (prec - 1) as other_amount
8141 instead of prec - op1.
8143 2013-05-13 Martin Jambor <mjambor@suse.cz>
8146 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
8147 (ipa_constant_data): New type.
8148 (ipa_jump_func): Use ipa_constant_data to hold information about
8149 constant jump functions.
8150 (ipa_get_jf_constant): Adjust to jump function type changes.
8151 (ipa_get_jf_constant_rdesc): New function.
8152 (ipa_param_descriptor): New field controlled_uses.
8153 (ipa_get_controlled_uses): New function.
8154 (ipa_set_controlled_uses): Likewise.
8155 * ipa-ref.h (ipa_find_reference): Declare.
8156 * ipa-prop.c (ipa_cst_ref_desc): New type.
8157 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
8159 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
8160 New parameter cs. Adjust all callers.
8161 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
8162 (remove_described_reference): New function.
8163 (jfunc_rdesc_usable): Likewise.
8164 (try_make_edge_direct_simple_call): Decrement controlled use count,
8165 attempt to remove reference if it hits zero.
8166 (combine_controlled_uses_counters): New function.
8167 (propagate_controlled_uses): Likewise.
8168 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
8169 (ipa_edge_duplication_hook): Duplicate reference descriptions.
8170 (ipa_print_node_params): Print described use counter.
8171 (ipa_write_jump_function): Adjust to jump function type changes.
8172 (ipa_read_jump_function): New parameter CS, pass it to
8173 ipa_set_jf_constant. Adjust caller.
8174 (ipa_write_node_info): Stream controlled use count
8175 (ipa_read_node_info): Likewise.
8176 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
8178 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
8179 count. Remove cloning-added reference if it reaches zero.
8180 * ipa-ref.c (ipa_find_reference): New function.
8182 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8184 * config/i386/i386.c (processor_target_table): Modified default
8185 alignment values for AMD BD and BT architectures.
8187 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
8189 * tree-vect-generic.c (uniform_vector_p): Move ...
8190 * tree.c (uniform_vector_p): ... here.
8191 * tree.h (uniform_vector_p): Declare it.
8192 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
8195 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8197 PR tree-optimization/57230
8198 * tree-ssa-strlen.c (handle_char_store): Record length for
8199 array store from STRING_CST.
8201 PR tree-optimization/57230
8202 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
8205 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
8207 * config/epiphany/epiphany.c (epiphany_init): Check size of
8208 NUM_MODES_FOR_MODE_SWITCHING.
8209 (epiphany_expand_prologue):
8210 Remove CONFIG_REGNUM initial value handling code.
8211 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
8212 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
8213 (emit_set_fp_mode, epiphany_mode_after): Likewise.
8214 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
8215 Don't return 1 for FP_MODE_NONE.
8216 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
8217 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
8218 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
8219 * config/epiphany/epiphany.md (save_config): New pattern.
8221 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
8223 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
8225 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
8227 * config/i386/i386.md (memory): Handle sseishft1.
8228 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
8229 (*vec_extractv2di_1): Ditto.
8231 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
8233 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
8236 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
8238 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
8239 Add mthumb/march=armv7-a multilib.
8240 Add mthumb/march=armv7-r multilib.
8241 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
8243 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
8245 * config/v850/t-rtems: Add more multilibs.
8247 2013-05-10 Richard Biener <rguenther@suse.de>
8249 PR tree-optimization/57214
8250 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
8251 not propagate from SSA names that occur in abnormal PHI nodes.
8253 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
8255 * stor-layout.c (element_precision): New function.
8256 * machmode.h (element_precision): Declare it.
8257 * tree.c (build_minus_one_cst): New function.
8258 (element_precision): Likewise.
8259 * tree.h (build_minus_one_cst): Declare new function.
8260 (element_precision): Likewise.
8261 * fold-const.c (operand_equal_p): Use element_precision.
8262 (fold_binary_loc): Handle vector types.
8263 * convert.c (convert_to_integer): Use element_precision.
8264 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
8267 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
8269 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
8270 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
8271 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
8272 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8273 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
8274 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
8275 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
8276 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8277 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
8278 (Uuw8): New constraints.
8279 (Usb4): Move into alphabetical order.
8280 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
8281 (sd8_operand, ub8_operand, uw8_operand): New predicates.
8282 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
8283 previously unnamed patterns.
8284 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
8285 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
8286 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
8287 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
8288 of set_attr_alternative/if_then_else. Use extended_mips16 instead
8289 of specific lengths.
8291 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8293 * config/i386/i386.md (rotateinv): New code attr.
8294 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8295 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8296 roll $31, %eax, etc.
8298 PR tree-optimization/45216
8299 PR tree-optimization/57157
8300 * tree-ssa-forwprop.c (simplify_rotate): New function.
8301 (ssa_forward_propagate_and_combine): Call it.
8303 2013-05-10 Richard Biener <rguenther@suse.de>
8305 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
8306 disable peeling when we version for aliasing.
8307 (vector_alignment_reachable_p): Honor explicit user alignment.
8308 (vect_supportable_dr_alignment): Likewise.
8309 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
8310 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
8311 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
8312 then peeling to arrange for the cost-model check to come first.
8314 2013-05-10 Alan Modra <amodra@gmail.com>
8316 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
8317 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
8318 * configure: Regenerate.
8320 2013-05-10 Alan Modra <amodra@gmail.com>
8323 * varasm.c (default_elf_select_section): Move !DECL_P check..
8324 (get_named_section): ..to here before calling get_section_name.
8326 (default_section_type_flags): Add DECL_P check.
8327 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
8328 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
8330 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
8332 * config/epiphany/epiphany.c (epiphany_expand_prologue):
8333 When using gen_stack_adjust_str with a register offset, add a
8334 REG_FRAME_RELATED_EXPR note.
8336 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
8338 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
8339 (*vec_extractv4si_zext_mem): Ditto.
8340 (*vec_extractv2di): Add 0->x and x->x alternatives.
8341 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
8342 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
8344 2013-05-09 Jason Merrill <jason@redhat.com>
8346 N3639 C++1y VLA support
8347 * gimplify.c (gimplify_vla_decl): Don't touch an existing
8350 * tree.c (build_constructor_va): New.
8351 * tree.h: Declare it.
8353 2013-05-09 Martin Jambor <mjambor@suse.cz>
8356 * gimple-fold.c (canonicalize_constructor_val): Call
8357 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
8359 2013-05-09 Jan Hubicka <jh@suse.cz>
8360 Richard Biener <rguenther@suse.de>
8363 * symtab.c (symtab_make_decl_local): Do not add private names.
8365 2013-05-09 Jan Hubicka <jh@suse.cz>
8368 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8369 (unlink_from_assembler_name_hash): Likewise.
8370 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8371 symtab_unregister_node, symtab_initialize_asm_name_hash,
8372 change_decl_assembler_name): Update.
8374 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8376 * config/aarch64/aarch64.md: New movtf split.
8377 (*movtf_aarch64): Update.
8378 (aarch64_movdi_tilow): Handle TF modes and rename to
8379 aarch64_movdi_<mode>low.
8380 (aarch64_movdi_tihigh): Handle TF modes and rename to
8381 aarch64_movdi_<mode>high
8382 (aarch64_movtihigh_di): Handle TF modes and rename to
8383 aarch64_mov<mode>high_di
8384 (aarch64_movtilow_di): Handle TF modes and rename to
8385 aarch64_mov<mode>low_di
8386 (aarch64_movtilow_tilow): Remove spurious whitespace.
8387 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8389 (aarch64_print_operand): Update.
8391 2013-05-09 Alan Modra <amodra@gmail.com>
8393 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8395 * configure: Regenerate.
8397 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8399 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8400 splitter preparation statements.
8401 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8402 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8405 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8407 * gimple-ssa-strength-reduction.c (count_candidates): Change
8408 return value to int.
8409 (analyze_candidates_and_replace): Change type of length to int.
8411 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8413 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8414 (*vec_extract<mode>): Use VI12_128 mode iterator.
8415 (*vec_extract<mode>_mem): Ditto.
8416 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8419 2013-05-08 Diego Novillo <dnovillo@google.com>
8424 2012-08-17 Diego Novillo <dnovillo@google.com>
8427 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8428 * config.in: Regenerate.
8429 * configure: Regenerate.
8430 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8432 2013-05-08 Jan Hubicka <jh@suse.cz>
8435 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8436 * cgraph.h (symtab_node_base): Add unique_name.
8437 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8438 input_overwrite_node, input_varpool_node): Stream unique_name.
8439 * cgraphclones.c (cgraph_create_virtual_clone,
8440 cgraph_function_versioning): Set unique_name.
8441 * ipa.c (function_and_variable_visibility): Set unique_name.
8443 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8445 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8446 (alloc_cand_and_find_basis): Restrict conditional candidate
8447 processing to CAND_MULTs.
8449 2013-05-08 Jan Hubicka <jh@suse.cz>
8452 lto-symtab.c (lto_symtab_symbol_p): New function.
8453 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8454 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8455 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8456 Skip static symbols.
8458 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8460 PR tree-optimization/57200
8461 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8462 Only call inform if the preceding warning_at returns true.
8464 2013-05-07 Han Shen <shenhan@google.com>
8466 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8467 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8468 * common.opt (fstack-protector-strong): New option.
8469 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8470 * doc/invoke.texi (Optimization Options): Document
8471 "-fstack-protector-strong".
8472 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8474 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8476 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8478 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8480 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8481 it is not lhs of the insn.
8482 (setup_reg_equiv): Remove insn having equiv memory which it is not
8484 * lra-constraints.c (process_address): Try to improve generation
8485 code for address base + disp.
8486 (lra_constraints): Make correct the code for checking insn setting
8487 up backward equivalence. Remove insn only if it is in the init
8489 * lra-eliminations.c (update_reg_eliminate): Change return value.
8490 (lra_eliminate): Use the result.
8492 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8494 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8495 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8496 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8497 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8498 register target operands.
8499 (*vec_extractv8hi_sse2): New pattern.
8500 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8501 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8502 (*vec_extract<mode>_mem): New insn and split pattern.
8504 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8506 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8507 (TARGET_ASAN_SHADOW_OFFSET): Define.
8508 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8509 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8511 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8513 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8514 (incr_vec_index): Return -1 if increment not found.
8515 (create_add_on_incoming_edge): Assert if increment not found.
8516 (record_increment): Limit number of increments recorded.
8517 (all_phi_incrs_profitable): Return false if an increment not found.
8518 (replace_profitable_candidates): Don't process increments that were
8520 (analyze_candidates_and_replace): Limit size of incr_vec.
8522 2013-05-07 Richard Biener <rguenther@suse.de>
8524 * calls.c (special_function_p): setjmp-like functions are leaf.
8525 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8526 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8528 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8530 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8531 (aarch64_simd_mov<mode>): New expander.
8532 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8533 (aarch64_simd_mov_to_<mode>high): Likewise.
8534 (aarch64_simd_mov_from_<mode>low): Likewise.
8535 (aarch64_simd_mov_from_<mode>high): Likewise.
8536 (aarch64_dup_lane<mode>): Update.
8537 (aarch64_dup_lanedi): New instruction pattern.
8538 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8539 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8541 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8543 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8544 (replace_mult_candidate): Remove unnecessary argument; remove
8545 unnecessary parameter from call to introduce_cast_before_cand.
8546 (replace_unconditional_candidate): Remove unnecessary parameter
8547 from call to replace_mult_candidate.
8548 (replace_conditional_candidate): Likewise.
8549 (insert_initializers): Use make_temp_ssa_name.
8550 (introduce_cast_before_cand): Remove unnecessary argument; use
8552 (replace_one_candidate): Remove unnecessary argument; remove
8553 unnecessary parameter from calls to introduce_cast_before_cand.
8554 (replace_profitable_candidates): Remove unnecessary parameters
8555 from calls to replace_one_candidate.
8557 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8559 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8560 phi def as possibly hiding a basis for a CAND_ADD whose operands
8561 have been commuted in the analysis.
8562 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8564 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8566 * config/aarch64/aarch64.md
8567 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8568 shift value between 0-4.
8570 2013-05-07 Richard Biener <rguenther@suse.de>
8572 * double-int.h (rshift): New overload.
8573 * double-int.c (rshift): New function.
8574 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8575 (create_reference_ops_from_ref): Remove.
8576 (vn_reference_insert): Use shared ops for constructing the
8577 reference and copy it.
8579 2013-05-07 Richard Biener <rguenther@suse.de>
8582 * tree-eh.c (sink_clobbers): Properly propagate
8583 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8585 2013-05-07 Jakub Jelinek <jakub@redhat.com>
8587 PR tree-optimization/57149
8588 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8589 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8590 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8591 uninit_undefined_value_p instead of ssa_undefined_value_p.
8594 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8595 for modifier == EXPAND_INITIALIZER.
8597 2013-05-07 Anton Blanchard <anton@samba.org>
8599 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8600 for powerpc64 little endian.
8601 * configure: Regenerate.
8603 2013-05-06 Graham Stott <grahams@btinternet.com>
8605 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8606 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8607 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8608 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8610 2013-05-06 Graham Stott <grahams@btinternet.com>
8612 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8613 codes which allow non-lvalues.
8615 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
8617 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8618 components are all 1s.
8619 (integer_minus_onep): New function.
8620 * tree.h (integer_minus_onep): Declare it.
8621 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8622 integer_minus_onep instead of integer_all_onesp.
8624 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8627 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8628 variations of these patterns.
8630 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8632 * config/i386/i386.md (isa): Add x64_sse4 member.
8633 (enabled): Handle x64_sse4.
8634 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8635 instruction for 64bit SSE4_1 targets. Update insn attributes.
8636 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8637 instruction for SSE4_1 targets. Update insn attributes.
8638 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8639 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8640 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8642 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8643 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8644 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8646 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8649 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8651 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8653 * final.c (do_assembler_dialects): Don't handle curly braces and
8654 vertical bar escaped by % as dialect delimiters.
8655 (output_asm_insn): Print curly braces and vertical bar if escaped
8656 by % and ASSEMBLER_DIALECT defined.
8657 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8658 * doc/tm.texi: Regenerated.
8660 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8662 * config/mips/mips.c: Include tree-pass.h.
8663 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8664 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8665 (pass_mips_machine_reorg2): New machine specific pass.
8666 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8667 (mips_option_override): Register the new pass.
8668 * rtl.h (cleanup_barriers): Remove prototype.
8669 (dbr_schedule): Likewise.
8670 * jump.c (cleanup_barriers): Make static.
8671 * reorg.c (dbr_schedule): Likewise.
8673 2013-05-06 Richard Biener <rguenther@suse.de>
8675 PR tree-optimization/57185
8676 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8679 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8682 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8683 between "first_arg" and "insn", not "last" and "insn".
8685 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8687 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8688 (find_candidates_in_block): Re-enable slsr_process_phi.
8689 (create_phi_basis): Fix double counting of candidate adjustment.
8691 2013-05-06 Richard Biener <rguenther@suse.de>
8694 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8695 the edge is also fallthru, preserve it and just clear the
8697 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8698 also complex, preserve that and just clear the fallthru flag.
8699 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8700 update virtual operands.
8702 2013-05-06 Alan Modra <amodra@gmail.com>
8704 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8705 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8706 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8707 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8708 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8709 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8711 2013-05-06 Alan Modra <amodra@gmail.com>
8713 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8714 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8715 (DEFAULT_ASM_ENDIAN): Define.
8716 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8717 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8718 Update -K PIC clause from sysv4.h.
8719 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8720 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8722 2013-05-06 Alan Modra <amodra@gmail.com>
8724 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8725 twice for little-endian.
8726 (ashrdi3_no_power, ashrdi3): Support little-endian.
8728 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8731 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8732 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8734 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8735 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8736 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8739 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8740 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8742 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8743 * config/pa/pa.opt: Make mbig-switch a no-op.
8744 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8745 (CASE_VECTOR_MODE): Always return SImode.
8746 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8747 for the !TARGET_BIG_SWITCH case.
8748 * config/pa/pa-linux.h: Likewise.
8749 * config/pa/pa-openbsd.h: Likewise.
8750 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8751 * config/pa/pa.md (short_jump): Remove define_insn.
8752 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8753 (casesi0): Remove define_insn.
8754 (type): Remove btable_branch.
8755 (pa_combine_type): Likewise.
8756 (in_nullified_branch_delay): Likewise.
8757 (in_call_delay): Likewise.
8758 (define_delay): Likewise.
8759 (define_insn_reservation "Z3"): Likewise.
8760 (define_insn_reservation "Z4"): Likewise.
8761 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8762 (pa_adjust_insn_length): Remove adjustment for btable branches.
8763 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8766 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
8768 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8769 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8770 Add m->r,x alternatives.
8771 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8772 splitters using SWI48x mode iterator.
8773 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8774 TARGET_64BIT. Add m->x alternative.
8775 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8776 Add o->x alternative. Enable for TARGET_SSE.
8777 (sse_storeq): Remove expander.
8778 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8779 with memory input operand.
8780 (*vec_extractv2di_1 splitter): New.
8781 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8782 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8784 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8786 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8787 (INT_LOWPART): Delete.
8788 (extract_MB): Adjust.
8789 (extract_ME): Adjust.
8790 (print_operand): Adjust.
8792 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8794 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8795 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8796 (reg_or_logical_cint_operand, easy_fp_constant,
8797 logical_const_operand): Delete "CONST_DOUBLE" case.
8798 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8799 "HOST_BITS_PER_WIDE_INT == 64" test.
8800 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
8801 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8802 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8803 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8805 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8806 CONST_DOUBLE DImode/VOIDmode case.
8807 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8808 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8809 CONST_DOUBLE VOIDmode case.
8810 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8811 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8812 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8813 Delete CONST_DOUBLE case.
8814 (splitters for mov FMOVE64 const_double): Delete
8815 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8816 "HOST_BITS_PER_WIDE_INT >= 64" test.
8817 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8819 (mov DI const_double): Delete.
8821 2013-05-04 Jakub Jelinek <jakub@redhat.com>
8823 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8824 on op shows all bits zero in mode of a lowpart subreg, return zero.
8826 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8829 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8830 to save TFmode registers and DImode to save TImode registers for
8831 caller save operations.
8832 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8833 mark being partially clobbered since they only use the first
8836 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8837 and TDmode only use the upper 64-bits of each VSX register.
8839 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8841 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8842 (find_candidates_in_block): Disable slsr_process_phi.
8844 2013-05-03 Guozhi Wei <carrot@google.com>
8846 * coverage.c (coverage_obj_init): Move the construction of gcov
8848 (build_init_ctor): ... here.
8850 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8852 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8853 (slsr_cand_d): Redefine def_phi.
8854 (stride_status, phi_adjust_status, count_phis_status): New enums.
8855 (find_phi_def): New.
8856 (find_basis_for_base_expr): New.
8857 (find_basis_for_candidate): Handle hidden bases.
8858 (alloc_cand_and_find_basis): Handle phi candidates.
8859 (slsr_process_phi): New.
8860 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8861 (create_mul_imm_cand): Likewise.
8862 (create_add_ssa_cand): Exclude phi base candidates.
8863 (create_add_imm_cand): Likewise.
8864 (slsr_process_cast): Likewise.
8865 (slsr_process_copy): Likewise.
8866 (find_candidates_in_block): Handle phi candidates.
8867 (dump_candidate): Likewise.
8868 (unconditional_cands): Delete.
8869 (unconditional_cands_with_known_stride_p): Delete.
8870 (phi_dependent_cand_p): New.
8871 (cand_increment): Handle phi-dependent candidates.
8872 (replace_dependent): Delete.
8873 (replace_mult_candidate): New.
8874 (replace_unconditional_candidate): New.
8875 (incr_vec_index): Move to avoid forward reference.
8876 (create_add_on_incoming_edge): New.
8877 (create_phi_basis): New.
8878 (replace_dependents): Delete.
8879 (replace_conditional_candidate): New.
8880 (phi_add_costs): New.
8881 (replace_uncond_cands_and_profitable_phis): New.
8882 (record_increment): Handle phi adjustments.
8883 (record_phi_increments): New.
8884 (record_increments): Handle phi adjustments.
8885 (phi_incr_cost): New.
8886 (lowest_cost_path): Handle phis.
8887 (total_savings): Likewise.
8888 (analyze_increments): Likewise.
8889 (ncd_with_phi): New.
8890 (ncd_of_cand_and_phis): New.
8891 (nearest_common_dominator_for_cands): Handle phi increments.
8892 (all_phi_incrs_profitable): New.
8893 (replace_profitable_candidates): Handle phi-dependent candidates.
8894 (analyze_candidates_and_replace): Likewise.
8896 2013-05-03 Teresa Johnson <tejohnson@google.com>
8899 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8900 do not exceed REG_BR_PROB_BASE.
8902 2013-05-03 Jeff Law <law@redhat.com>
8904 PR tree-optimization/57144
8905 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8906 operand of the condition will bit into the new type when eliminating
8907 a cast feeding a condition.
8909 2013-05-03 Jakub Jelinek <jakub@redhat.com>
8911 PR rtl-optimization/57130
8912 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8913 of COMPARE as in_code to the recursive call if needed.
8915 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
8917 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8918 (enabled): Handle new members.
8919 * config/i386/sse.md (*vec_concatv2si): Merge from
8920 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8921 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8923 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8925 PR tree-optimization/57027
8926 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8927 for fnms opportunity, check we got the prerequisite kind
8928 of tree / gimple before using accessor functions.
8930 2013-05-03 Richard Biener <rguenther@suse.de>
8932 * double-int.h (lshift): New overload without precision
8934 (operator *=, operator +=, operator -=): Move ...
8935 * double-int.c (operator *=, operator +=, operator -=): ... here
8936 and implement more efficiently.
8937 (mul_double_with_sign): Remove.
8938 (lshift_double): Adjust to take unsinged shift argument, push
8939 dispatching code to callers.
8940 (mul_double_wide_with_sign): Add early out for callers that
8941 are not interested in high parts or overflow.
8942 (lshift): New function.
8943 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8945 (lrotate, rrotate): Use logical shifts.
8946 * expr.c (get_inner_reference): Use lshift.
8947 * fixed-value.c (do_fixed_divide): Likewise.
8948 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8949 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8950 (indirect_refs_may_alias_p): Likewise.
8951 (stmt_kills_ref_p_1): Likewise.
8953 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8955 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8957 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8959 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8960 scalar form of FABD instruction.
8962 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8964 * lra-constraints.c (process_alt_operands): Add checking alt
8965 number to choose the best alternative.
8967 2013-05-02 Richard Biener <rguenther@suse.de>
8969 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
8970 bitmap and its handling.
8971 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
8973 2013-05-02 Richard Biener <rguenther@suse.de>
8976 * tree-inline.c (copy_loops): Properly handle removed loops.
8977 (copy_cfg_body): Mark destination loops for fixup if source
8980 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
8983 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
8984 generating simple_return for naked functions.
8986 2013-05-02 Martin Jambor <mjambor@suse.cz>
8989 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
8990 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
8992 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
8993 ipa_agg_replacement_value structures.
8994 (known_aggs_to_agg_replacement_list): Likewise.
8995 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
8996 (read_agg_replacement_chain): Likewise.
8997 (ipcp_transform_function): Also check that by_ref flags match.
8999 2013-05-02 Richard Biener <rguenther@suse.de>
9001 * graphds.h (struct graph): Add obstack member.
9002 * graphds.c (new_graph): Initialize obstack and allocate
9004 (add_edge): Allocate edge from the obstack.
9005 (free_graph): Free the obstack instead of all edges and vertices.
9007 2013-05-02 Teresa Johnson <tejohnson@google.com>
9009 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9011 * cfg.c (update_bb_profile_for_threading): Ditto.
9012 * tree-inline.c (copy_bb): Ditto.
9013 (copy_edges_for_bb): Ditto.
9014 (initialize_cfun): Ditto.
9015 (copy_cfg_body): Ditto.
9016 (expand_call_inline): Ditto.
9017 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9018 (estimate_node_size_and_time): Ditto.
9019 (inline_merge_summary): Ditto.
9020 * cgraphclones.c (cgraph_clone_edge): Ditto.
9021 (cgraph_clone_node): Ditto.
9022 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9023 (compute_trg_info): Ditto.
9025 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9027 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9028 S reg when fp attribute set.
9029 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9031 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9033 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9035 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9036 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9037 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9039 2013-05-02 Richard Biener <rguenther@suse.de>
9041 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9042 (struct instantiate_cache_entry): New type.
9043 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9044 (struct instantiate_cache_type): New type.
9045 (set_instantiated_value, get_instantiated_value): Remove.
9046 (get_instantiated_value_entry): New function.
9047 (instantiate_scev_name): Use the new cache and adjust.
9048 (instantiate_scev_poly): Adjust.
9049 (instantiate_scev_binary): Likewise.
9050 (instantiate_array_ref): Likewise.
9051 (instantiate_scev_convert): Likewise.
9052 (instantiate_scev_not): Likewise.
9053 (instantiate_scev_3): Likewise.
9054 (instantiate_scev_2): Likewise.
9055 (instantiate_scev_r): Likewise.
9056 (instantiate_scev): Likewise.
9057 (resolve_mixers): Likewise.
9059 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9062 * lra-constraints.c (best_small_class_operands_num): Remove.
9063 (process_alt_operands): Remove small_class_operands_num. Take
9064 small classes operands into losers and only if the operand is not
9065 matched. Modify debugging output.
9066 (curr_insn_transform): Remove best_small_class_operands_num.
9069 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9071 * config/aarch64/aarch64-builtins.c
9072 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9073 * config/aarch64/aarch64-simd-builtins.def
9074 (reduc_splus_): Add new modes.
9075 (reduc_uplus_): New.
9076 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9077 (reduc_uplus_v4sf): Likewise.
9078 (reduc_splus_v4sf): Likewise.
9079 (aarch64_addv<mode>): Likewise.
9080 (reduc_uplus_<mode>): Likewise.
9081 (reduc_splus_<mode>): Likewise.
9082 (aarch64_addvv2di): Likewise.
9083 (reduc_uplus_v2di): Likewise.
9084 (reduc_splus_v2di): Likewise.
9085 (aarch64_addvv2si): Likewise.
9086 (reduc_uplus_v2si): Likewise.
9087 (reduc_splus_v2si): Likewise.
9088 (reduc_<sur>plus_<mode>): New.
9089 (reduc_<sur>plus_v2di): Likewise.
9090 (reduc_<sur>plus_v2si): Likewise.
9091 (reduc_<sur>plus_v4sf): Likewise.
9092 (aarch64_addpv4sf): Likewise.
9093 * config/aarch64/arm_neon.h
9094 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9095 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9096 add UNSPEC_SADDV, UNSPEC_UADDV.
9098 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9100 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9102 * config/aarch64/arm_neon.h
9103 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9105 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9107 * config/aarch64/aarch64-builtins
9108 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
9110 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9112 * config/aarch64/aarch64-simd-builtins.def
9114 (reduc_smin_): Likewise.
9115 (reduc_umax_): Likewise.
9116 (reduc_umin_): Likewise.
9117 (reduc_smax_nan_): Likewise.
9118 (reduc_smin_nan_): Likewise.
9121 (smax): Update for V2SF, V4SF and V2DF modes.
9124 (smin_nan): Likewise.
9125 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
9126 (<su><maxmin><mode>3): ...This, refactor.
9127 (s<maxmin><mode>3): New.
9128 (<maxmin_uns><mode>3): Likewise.
9129 (reduc_<maxmin_uns>_<mode>): Refactor.
9130 (reduc_<maxmin_uns>_v4sf): Likewise.
9131 (reduc_<maxmin_uns>_v2si): Likewise.
9132 (aarch64_<fmaxmin><mode>: Remove.
9133 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
9135 (vmin<q>_f<32,64>): Likewise.
9136 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9138 (su): Add mappings for smax, smin, umax, umin.
9140 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9141 (FMAXMIN): Rename as...
9142 (FMAXMIN_UNS): ...This.
9144 (fmaxminv): Likewise.
9145 (fmaxmin): Likewise.
9147 (maxmin_uns_op): Likewise.
9149 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9151 * config/aarch64/arm_neon.h
9152 (vac<ge, gt><sd>_f<32, 64>): Rename to...
9153 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
9154 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
9156 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9158 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
9159 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
9161 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9163 * config/aarch64/aarch64-simd.md
9164 (vcond<mode>_internal): Handle special cases for constant masks.
9165 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
9166 (vcondu<mode><mode>): Likewise.
9167 (vcond<v_cmp_result><mode>): New.
9169 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9171 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
9172 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
9173 * config/aarch64/aarch64-simd-builtins.def
9174 (cmeq): Update to BUILTIN_VALLDI.
9179 * config/aarch64/arm_neon.h
9180 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
9181 to builtins or C as appropriate.
9183 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9185 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
9187 (cmhi): Rename to...
9189 * config/aarch64/aarch64-simd.md
9190 (simd_mode): Add SF.
9191 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
9192 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
9193 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
9194 (cstore<mode>_neg): ...This.
9195 * config/aarch64/iterators.md
9197 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
9199 (UCOMPARISONS): Likewise.
9200 (optab): Add missing comparisons.
9208 (V_cmp_result): Add DF, SF modes.
9209 (v_cmp_result): Likewise.
9212 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
9214 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
9216 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
9217 define_insn to define_insn_and_split.
9218 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
9219 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
9220 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
9221 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
9222 (thumb2_negscc): Likewise.
9224 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9226 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
9228 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9230 * config/arm/thumb2.md: Remove trailing whitespaces.
9232 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9234 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
9235 Use gen_int_mode rather than GEN_INT.
9237 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
9239 * value-prof.c (stream_in_histogram_value): Remove the strayed
9242 2013-04-30 Richard Biener <rguenther@suse.de>
9245 * cfghooks.c (split_edge): Properly check for the loop latch edge.
9247 2013-04-30 Richard Biener <rguenther@suse.de>
9250 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
9252 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
9254 PR rtl-optimization/56957
9255 PR rtl-optimization/57105
9256 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
9257 variable. Use just INSN_UID for determining whether an insn
9258 should be only disconnected from the insn stream.
9259 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
9261 2013-04-30 Jakub Jelinek <jakub@redhat.com>
9263 PR tree-optimization/57104
9264 * tsan.c (instrument_expr): Don't instrument accesses to
9265 DECL_HARD_REGISTER VAR_DECLs.
9267 2013-04-30 Richard Biener <rguenther@suse.de>
9269 * function.h (loops_for_fn): New inline function.
9270 (set_loops_for_fn): Likewise.
9271 * cfgloop.h (place_new_loop): Add struct function parameter.
9272 (get_loop): Likewise.
9273 (get_loops): Likewise.
9274 (number_of_loops): Likewise.
9276 (fel_init): Likewise.
9277 * cfg.c (get_loop_copy): Adjust.
9278 * cfgloop.c (flow_loops_dump): Likewise.
9279 (record_loop_exits): Likewise.
9280 (verify_loop_structure): Likewise.
9281 * cfgloopanal.c (mark_irreducible_loops): Likewise.
9282 (estimate_reg_pressure_cost): Likewise.
9283 (mark_loop_exit_edges): Likewise.
9284 * cfgloopmanip.c (place_new_loop): Likewise.
9285 (add_loop): Likewise.
9286 (duplicate_loop): Likewise.
9287 * graph.c (draw_cfg_nodes): Likewise.
9288 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
9289 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
9290 (extract_affine_chrec): Likewise.
9291 (build_scop_iteration_domain): Likewise.
9292 * graphite.c (graphite_initialize): Likewise.
9293 * ira-build.c (create_loop_tree_nodes): Likewise.
9294 (more_one_region_p): Likewise.
9295 (rebuild_regno_allocno_maps): Likewise.
9296 (mark_loops_for_removal): Likewise.
9297 (mark_all_loops_for_removal): Likewise.
9298 (remove_unnecessary_regions): Likewise.
9299 (ira_build): Likewise.
9300 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
9301 * loop-init.c (fix_loop_structure): Likewise.
9302 (gate_rtl_move_loop_invariants): Likewise.
9303 (gate_rtl_unswitch): Likewise.
9304 (gate_rtl_unroll_and_peel_loops): Likewise.
9305 (rtl_doloop): Likewise.
9306 * lto-streamer-in.c (input_cfg): Likewise.
9307 * lto-streamer-out.c (output_cfg): Likewise.
9308 * modulo-sched.c (sms_schedule): Likewise.
9309 * predict.c (tree_estimate_probability): Likewise.
9310 (tree_estimate_probability_driver): Likewise.
9311 (estimate_loops): Likewise.
9312 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
9313 (move_sese_region_to_fn): Likewise.
9314 (debug_loop_num): Likewise.
9315 * tree-chrec.c (chrec_evaluate): Likewise.
9316 (hide_evolution_in_other_loops_than_loop): Likewise.
9317 (chrec_component_in_loop_num): Likewise.
9318 (reset_evolution_in_loop): Likewise.
9319 (evolution_function_is_invariant_rec_p): Likewise.
9320 * tree-if-conv.c (main_tree_if_conversion): Likewise.
9321 * tree-inline.c (copy_loops): Likewise.
9322 (copy_cfg_body): Likewise.
9323 (tree_function_versioning): Likewise.
9324 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
9325 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
9327 (add_to_evolution_1): Likewise.
9328 (scev_const_prop): Likewise.
9329 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
9330 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
9331 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
9332 (tree_ssa_lim_initialize): Likewise.
9333 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
9334 (verify_loop_closed_ssa): Likewise.
9335 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
9336 (tree_ssa_loop_im): Likewise.
9337 (tree_ssa_loop_unswitch): Likewise.
9338 (tree_vectorize): Likewise.
9339 (check_data_deps): Likewise.
9340 (tree_ssa_loop_ivcanon): Likewise.
9341 (tree_ssa_loop_bounds): Likewise.
9342 (tree_complete_unroll): Likewise.
9343 (tree_complete_unroll_inner): Likewise.
9344 (tree_parallelize_loops): Likewise.
9345 (tree_ssa_loop_prefetch): Likewise.
9346 (tree_ssa_loop_ivopts): Likewise.
9347 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
9348 * tree-vectorizer.c (vectorize_loops): Likewise.
9350 2013-04-29 Mike Frysinger <vapier@gentoo.org>
9352 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
9353 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
9354 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
9355 with EABI_LINK_SPEC.
9357 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9360 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9363 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9366 * lra-constraints.c (process_alt_operands): Discourage a bit more
9367 using memory for pseudos. Print cost dump for alternatives.
9368 Modify cost values for conflicts with early clobbers.
9369 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9371 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9374 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9376 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9378 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9380 (movdi_aarch64): Support LDR/STR from/to D register.
9382 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9384 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9385 or higher by default.
9387 2013-04-29 Richard Biener <rguenther@suse.de>
9390 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9391 even if not adding abnormal edges for calls that can make
9394 2013-04-29 Richard Biener <rguenther@suse.de>
9397 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9399 (move_stmt_r): Remove redundant checking.
9401 2013-04-29 Teresa Johnson <tejohnson@google.com>
9404 * basic-block.h (apply_scale): New function.
9405 (apply_probability): Use apply_scale.
9406 * gimple-streamer-in.c (input_bb): Ditto.
9407 * lto-streamer-in.c (input_cfg): Ditto.
9408 * lto-cgraph.c (merge_profile_summaries): Ditto.
9409 * tree-optimize.c (execute_fixup_cfg): Ditto.
9410 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9411 (copy_edges_for_bb): Ditto.
9412 (copy_cfg_body): Ditto.
9414 2013-04-29 Tom de Vries <tom@codesourcery.com>
9416 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9417 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9418 (tail_merge_optimize): Handle current_loops == NULL.
9420 2013-04-26 Jeff Law <law@redhat.com>
9422 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9423 (simplify_cond_using_ranges): Generalize code to simplify
9424 COND_EXPRs where one argument is a constant and the other
9425 is an SSA_NAME created by an integral type conversion.
9427 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9429 * config/arm/arm.md (store_minmaxsi): Use only when
9430 optimize_insn_for_size_p.
9432 2013-04-29 Christian Bruel <christian.bruel@st.com>
9435 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9437 2013-04-29 Richard Biener <rguenther@suse.de>
9440 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9441 loop tree make sure to schedule a fixup for the child as well.
9442 (expand_omp_for_generic): Properly add loops.
9443 (expand_omp_for_static_nochunk): Likewise.
9444 (expand_omp_for_static_chunk): Likewise.
9445 (expand_omp_for): For the degenerate case fixup loops.
9446 (expand_omp_sections): Fix default bb placement in loops.
9447 (expand_omp_atomic_pipeline): Properly add loops.
9449 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9451 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9453 2013-04-29 Tom de Vries <tom@codesourcery.com>
9455 * tree-ssa-tail-merge.c: Update header comment.
9457 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9459 * config/aarch64/arm_neon.h
9460 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9461 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9462 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9463 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9464 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9465 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9466 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9467 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9469 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9471 * config/aarch64/aarch64-simd.md
9472 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9473 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9474 fix_trunc, fixuns_trunc.
9475 (ftrunc<VDQF:mode>2): New.
9476 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9477 (fix_trunc_optab): New.
9479 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9481 * config/aarch64/aarch64-builtins.c
9482 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9483 iceilf, lround, iroundf.
9485 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9488 * config/i386/i386.h (enum ix86_tune_indices)
9489 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9490 New, split from X86_TUNE_INTER_UNIT_MOVES.
9491 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9492 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9493 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9494 (TARGET_INTER_UNIT_MOVES): Remove.
9495 * config/i386/i386.c (initial_ix86_tune_features): Update.
9496 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9497 (ix86_expand_convert_uns_didf_sse): Use
9498 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9499 (ix86_expand_vector_init_one_nonzero): Ditto.
9500 (ix86_expand_vector_init_interleave): Ditto.
9501 (inline_secondary_memory_needed): Return true for moves from SSE class
9502 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9503 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9504 * config/i386/constraints.md (Yi, Ym): Depend on
9505 TARGET_INTER_UNIT_MOVES_TO_VEC.
9506 (Yj, Yn): New constraints.
9507 * config/i386/i386.md (*movdi_internal): Change constraints of
9508 operand 1 from Yi to Yj and from Ym to Yn.
9509 (*movsi_internal): Ditto.
9510 (*movdf_internal): Ditto.
9511 (*movsf_internal): Ditto.
9512 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9513 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9514 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9515 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9516 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9517 * config/i386/sse.md (movdi_to_sse): Ditto.
9518 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9519 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9520 TARGET_INTER_UNIT_MOVES.
9521 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9522 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9523 instead of TARGET_INTER_UNIT_MOVES.
9524 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9525 operand 1 from Yi to Yj and from Ym to Yn.
9527 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9529 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9530 (float_truncate_hi_): Likewise.
9531 (float_extend_lo_): Likewise.
9532 (float_truncate_lo_): Likewise.
9533 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9534 (aarch64_float_extend_lo_v2df): Likewise.
9535 (vec_unpacks_hi_v4sf): Likewise.
9536 (aarch64_float_truncate_lo_v2sf): Likewise.
9537 (aarch64_float_truncate_hi_v4sf): Likewise.
9538 (vec_pack_trunc_v2df): Likewise.
9539 (vec_pack_trunc_df): Likewise.
9541 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9543 * config/aarch64/aarch64-builtins.c
9544 (aarch64_fold_builtin): Fold float conversions.
9545 * config/aarch64/aarch64-simd-builtins.def
9546 (floatv2si, floatv4si, floatv2di): New.
9547 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9548 * config/aarch64/aarch64-simd.md
9549 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9550 * config/aarch64/iterators.md (FLOATUORS): New.
9551 (optab): Add float, floatuns.
9552 (su_optab): Likewise.
9554 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9556 * config/aarch64/aarch64-builtins.c
9557 (aarch64_builtin_vectorized_function): Use new names for
9559 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9560 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9561 (fcvtzu): Split as...
9562 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9563 (fcvtas): Split as...
9564 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9565 (fcvtau): Split as...
9566 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9567 (fcvtps): Split as...
9568 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9569 (fcvtpu): Split as...
9570 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9571 (fcvtms): Split as...
9572 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9573 (fcvtmu): Split as...
9574 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9575 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9576 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9577 (lfrintnusf, lfrintnudf): Likewise.
9578 * config/aarch64/aarch64-simd.md
9579 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9581 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9582 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9583 (fcvt_pattern): Likewise.
9585 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9587 * config/aarch64/aarch64-simd.md
9588 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9589 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9591 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9593 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9594 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9595 (vrnd<a,m,n,p>_f32): Implement using builtins.
9596 (vrnd<i,x><q>_f<32, 64>): New.
9598 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9600 * config/aarch64/aarch64-builtins.c
9601 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9602 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9603 (frintz): Rename to...
9605 (frintp): Rename to...
9607 (frintm): Rename to...
9609 (frinti): Rename to...
9610 (nearbyint): ...this.
9611 (frintx): Rename to...
9613 (frinta): Rename to...
9615 * config/aarch64/aarch64-simd.md
9616 (aarch64_frint<frint_suffix><mode>): Delete.
9617 (<frint_pattern><mode>2): Convert to insn.
9618 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9619 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9620 (frint_pattern): Likewise.
9621 (frint_suffix): Likewise.
9623 2013-04-29 Richard Biener <rguenther@suse.de>
9625 PR tree-optimization/57081
9626 * loop-init.c: Include tree-flow.h.
9627 (loop_optimizer_finalize): Free number of iteration estimates.
9628 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9630 2013-04-29 Jakub Jelinek <jakub@redhat.com>
9632 PR tree-optimization/57083
9633 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9634 non-singleton shift count range, zero extend low_bound for uns case.
9636 * config/i386/predicates.md (general_vector_operand): New predicate.
9637 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9638 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9639 if they aren't nonimmediate operands. If their original values
9640 satisfy const_vector_equal_evenodd_p, don't shift them.
9641 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9642 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9643 if not nonimmediate_operand.
9644 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9645 instead of register_operand.
9646 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9648 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
9650 * stor-layout.c (finalize_size_functions): Allocate a structure and
9651 reset cfun before dumping the functions.
9653 2013-04-27 Jakub Jelinek <jakub@redhat.com>
9655 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9658 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9659 use xop_pmacsdqh if uns_p.
9660 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9661 the immediate rotate count.
9663 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9665 * rtl.h (struct rtx_def): Add comment for field jump.
9666 (LRA_SUBREG_P): New macro.
9667 * recog.c (register_operand): Check LRA_SUBREG_P.
9668 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9670 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9671 (lra_final_code_change): Skip subreg change for operators.
9672 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9673 if there are no operand changes.
9674 * lra-constraints.c (curr_insn_set): New.
9675 (match_reload): Set LRA_SUBREG_P.
9676 (emit_spill_move): Ditto.
9677 (check_and_process_move): Use curr_insn_set. Process only single
9678 set insns. Don't initialize sec_mem_p and change_p.
9679 (simplify_operand_subreg): Use LRA_SUBREG_P.
9680 (reg_in_class_p): New function.
9681 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9682 of #ifdef. Add code to remove cycling.
9683 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9684 non-null disp. Reload inner instead of disp when base and index
9685 are null. Try to put lo_sum into register.
9686 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9687 (check_and_process_move): Move code for move cost check to
9688 simple_move_p. Remove equiv_substitution.
9689 (simple_move_p): New function.
9690 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9691 curr_insn_set. Call check_and_process_move only for single set
9692 insns. Use the new function. Move call of check_and_process_move
9693 after operand equiv substitution and address process.
9695 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9698 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9699 with nonlocal goto receivers or returns twice calls, ignore
9700 unininitialized values from abnormal edges to nl goto receiver
9701 or returns twice call.
9703 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9705 PR tree-optimization/57051
9706 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9707 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9710 2013-04-26 Richard Biener <rguenther@suse.de>
9712 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9713 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9714 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9715 (fixup_loop_arrays_after_move): New function.
9716 (move_sese_region_to_fn): Properly outline the loop tree parts
9719 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
9721 * config/i386/i386.md (type, unit): Fix long lines.
9723 2013-04-26 Richard Biener <rguenther@suse.de>
9725 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9726 (lto-streamer-out.o): Likewise.
9727 * cfgloop.c (init_loops_structure): Export, add struct function
9728 argument and adjust.
9729 (flow_loops_find): Adjust.
9730 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9731 (init_loops_structure): Declare.
9732 * lto-streamer-in.c: Include cfgloop.h.
9733 (input_cfg): Input the loop tree.
9734 * lto-streamer-out.c: Include cfgloop.h.
9735 (output_cfg): Output the loop tree.
9736 (output_struct_function_base): Do not drop PROP_loops.
9738 2013-03-26 Richard Biener <rguenther@suse.de>
9740 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9741 (pass_build_cfg): Provide PROP_loops.
9742 (move_sese_region_to_fn): Remove loops that are outlined into fn
9744 * tree-inline.c: Include cfgloop.h.
9745 (initialize_cfun): Do not drop PROP_loops.
9746 (copy_loops): New function.
9747 (copy_cfg_body): Copy loop structure.
9748 (tree_function_versioning): Initialize destination loop tree.
9749 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9750 (pass_parallelize_loops): Do IL verification.
9751 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9752 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9753 the CFG make sure we fixup loops as well.
9754 * tree-ssa-tail-merge.c: Include cfgloop.h.
9755 (replace_block_by): When merging loop latches mark loops for fixup.
9756 * lto-streamer-out.c (output_struct_function_base): Drop
9758 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9759 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9760 * ipa-split.c: Include cfgloop.h.
9761 (split_function): Add the new return block to the loop tree root.
9762 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9763 whether we have removed the forwarder block.
9764 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9765 * cfgloop.h (place_new_loop): Declare.
9766 * cfgloopmanip.c (place_new_loop): Export.
9767 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9768 (tree-switch-conversion.o): Likewise.
9769 (tree-complex.o): Likewise.
9770 (tree-inline.o): Likewise.
9771 (tree-ssa-tailmerge.o): Likewise.
9772 (ipa-split.o): Likewise.
9773 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9774 (tree-ssa-copy.o): Likewise.
9775 * tree-switch-conversion.c: Include cfgloop.h
9776 (process_switch): If we emit a bit-test cascade, schedule loops
9778 * tree-complex.c: Include cfgloop.h.
9779 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9780 * asan.c: Include cfgloop.h.
9781 (create_cond_insert_point): Properly add new basic-blocks to
9782 loops, schedule loop fixup.
9783 * cfgloop.c (verify_loop_structure): Check that looks are not
9785 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9787 (expand_omp_for_generic): Likewise.
9788 (expand_omp_sections): Likewise.
9789 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9790 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9791 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9792 is initialized, not when loops are present.
9793 * tree-parloops.c (parallelize_loops): Remove checking here.
9794 * passes.c (init_optimization_passes): Schedule a copy-propagation
9795 pass before complete unrolling of inner loops.
9797 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9799 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9800 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9801 (colorize_init): Add argument to _WIN32 version.
9802 * toplev.c: Include diagnostic-color.h.
9803 (process_options): Default to -fdiagnostics-color=auto if
9804 GCC_COLORS env var is in the environment.
9805 * common.opt (fdiagnostics-color=): Add Var and Init.
9806 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9807 env var is in the environment, the default is auto rather than never.
9809 * diagnostic.h (file_name_as_prefix): Add context argument.
9810 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9811 the string as locus.
9812 * langhooks.c (lhd_print_error_function): Adjust caller.
9814 2013-04-25 Lawrence Crowl <crowl@google.com>
9816 * var-tracking.c (shared_hash_def::htab):
9817 Change type to hash_table. Update dependent calls and types.
9819 2013-04-25 Lawrence Crowl <crowl@google.com>
9821 * Makefile.in: Update as needed below.
9823 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9824 Move declaration to after the type's method definitons.
9826 * attribs.c (htab_t scoped_attributes::attribute_hash):
9827 Change type to hash_table. Update dependent calls and types.
9829 * bitmap.c (htab_t bitmap_desc_hash):
9830 Change type to hash_table. Update dependent calls and types.
9832 * cselib.c (htab_t cselib_hash_table):
9833 Change type to hash_table. Update dependent calls and types.
9835 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9836 (hash_string_slot_node): Move implementation into lto-streamer.h
9837 struct string_slot_hasher.
9838 (eq_string_slot_node): Likewise.
9840 * data-streamer-out.c: Update output_block::string_hash_table
9841 dependent calls and types.
9843 * dwarf2cfi.c (htab_t trace_index):
9844 Change type to hash_table. Update dependent calls and types.
9846 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9847 Change type to hash_table. Update dependent calls and types.
9848 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9849 (htab_t optimize_external_refs::map): Likewise.
9850 (htab_t output_comp_unit::extern_map): Likewise.
9851 (htab_t output_comdat_type_unit::extern_map): Likewise.
9852 (htab_t output_macinfo::macinfo_htab): Likewise.
9853 (htab_t optimize_location_lists::htab): Likewise.
9854 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9856 * except.c (htab_t ehspec_hash_type):
9857 Change type to hash_table. Update dependent calls and types.
9858 (assign_filter_values::ttypes): Likewise.
9859 (assign_filter_values::ehspec): Likewise.
9860 (sjlj_assign_call_site_values::ar_hash): Likewise.
9861 (convert_to_eh_region_ranges::ar_hash): Likewise.
9863 * gcse.c (htab_t pre_ldst_table):
9864 Change type to hash_table. Update dependent calls and types.
9866 * ggc-common.c (htab_t saving_htab):
9867 Change type to hash_table. Update dependent calls and types.
9868 (htab_t loc_hash): Likewise.
9869 (htab_t ptr_hash): Likewise.
9870 (call_count): Rename ggc_call_count.
9871 (call_alloc): Rename ggc_call_alloc.
9872 (loc_descriptor): Rename make_loc_descriptor.
9873 (add_statistics): Rename ggc_add_statistics.
9875 * ggc-common.c (saving_htab):
9876 Change type to hash_table. Update dependent calls and types.
9878 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9879 (push_gimplify_context): Likewise.
9880 (pop_gimplify_context): Likewise.
9881 (struct gimple_temp_hash_elt): Added.
9882 (struct gimplify_hasher): Likewise.
9883 (struct gimplify_ctx.temp_htab):
9884 Change type to hash_table. Update dependent calls and types.
9886 * gimple-fold.c: Include gimplify-ctx.h.
9888 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9889 Change type to hash_table. Update dependent calls and types.
9890 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9891 avoid potential global name collision.
9893 * gimplify.c: Include gimplify-ctx.h.
9894 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9895 (htab_t gimplify_ctx::temp_htab):
9896 Update dependent calls and types for new type hash_table.
9897 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9898 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9900 * gimplify-ctx.h: New.
9901 (struct gimple_temp_hash_elt): Move from gimplify.c.
9902 (class gimplify_hasher): New.
9903 (struct gimplify_ctx): Move from gimple.h.
9904 (htab_t gimplify_ctx::temp_htab):
9905 Change type to hash_table. Update dependent calls and types.
9907 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9908 (htab_t ivs_params::newivs_index):
9909 Change type to hash_table. Update dependent calls and types.
9910 (htab_t ivs_params::params_index): Likewise.
9911 (htab_t print_generated_program::params_index): Likewise.
9912 (htab_t gloog::newivs_index): Likewise.
9913 (htab_t gloog::params_index): Likewise.
9915 * graphite.c: Include graphite-htab.h.
9916 4htab_t graphite_transform_loops::bb_pbb_mapping):
9917 Change type to hash_table. Update dependent calls and types.
9919 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9920 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9921 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9923 * graphite-dependences.c: Include graphite-htab.h.
9924 (loop_is_parallel_p): Change hash table type of parameter.
9926 * graphite-htab.h: New.
9927 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9928 (extern find_pbb_via_hash): Move from graphite-poly.h.
9929 (extern loop_is_parallel_p): Move from graphite-poly.h.
9930 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9932 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9933 (extern loop_is_parallel_p): Move to graphite-htab.h.
9934 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9936 * haifa-sched.c (htab_t delay_htab):
9937 Change type to hash_table. Update dependent calls and types.
9938 (htab_t delay_htab_i2): Likewise.
9940 * ira-color.c (htab_t allocno_hard_regs_htab):
9941 Change type to hash_table. Update dependent calls and types.
9943 * ira-costs.c (htab_t cost_classes_htab):
9944 Change type to hash_table. Update dependent calls and types.
9946 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9947 Change type to hash_table. Update dependent calls and types.
9949 * loop-iv.c (htab_t bivs):
9950 Change type to hash_table. Update dependent calls and types.
9952 * loop-unroll.c (htab_t opt_info::insns_to_split):
9953 Change type to hash_table. Update dependent calls and types.
9954 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9956 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9957 (struct string_slot_hasher): New.
9958 (htab_t output_block::string_hash_table):
9959 Change type to hash_table. Update dependent calls and types.
9961 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9962 (htab_t file_name_hash_table):
9963 Change type to hash_table. Update dependent calls and types.
9965 * lto-streamer-out.c: Update output_block::string_hash_table dependent
9968 * lto-streamer.c (htab_t tree_htab):
9969 Change type to hash_table. Update dependent calls and types.
9971 * omp-low.c: Include gimplify-ctx.h.
9973 * passes.c (htab_t name_to_pass_map):
9974 Change type to hash_table. Update dependent calls and types.
9975 (pass_traverse): Rename to passes_pass_traverse.
9977 * plugin.c (htab_t event_tab):
9978 Change type to hash_table. Update dependent calls and types.
9980 * postreload-gcse.c (htab_t expr_table):
9981 Change type to hash_table. Update dependent calls and types.
9982 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
9984 * sese.c (debug_rename_map_1): Make extern.
9985 (htab_t copy_bb_and_scalar_dependences::rename_map):
9986 Change type to hash_table. Update dependent calls and types.
9988 * sese.h (extern debug_rename_map): Move to .c file.
9990 * store-motion.c (htab_t store_motion_mems_table):
9991 Change type to hash_table. Update dependent calls and types.
9993 * trans-mem.c (htab_t tm_new_mem_hash):
9994 Change type to hash_table. Update dependent calls and types.
9996 * tree-browser.c (htab_t TB_up_ht):
9997 Change type to hash_table. Update dependent calls and types.
9999 * tree-cfg.c (htab_t discriminator_per_locus):
10000 Change type to hash_table. Update dependent calls and types.
10002 * tree-complex.c: Include tree-hasher.h
10003 (htab_t complex_variable_components):
10004 Change type to hash_table. Update dependent calls and types.
10006 * tree-eh.c (htab_t finally_tree):
10007 Change type to hash_table. Update dependent calls and types.
10009 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10010 struct int_tree_hasher.
10011 (extern int_tree_map_eq): Likewise.
10012 (uid_decl_map_hash): Removed.
10013 (extern decl_tree_map_eq): Likewise.
10015 * tree-hasher.h: New.
10016 (struct int_tree_hasher): New.
10017 (typedef int_tree_htab_type): New.
10019 * tree-inline.c: Include gimplify-ctx.h.
10021 * tree-mudflap.c: Include gimplify-ctx.h.
10023 * tree-parloops.c: Include tree-hasher.h.
10024 (htab_t eliminate_local_variables_stmt::decl_address):
10025 Change type to hash_table. Update dependent calls and types.
10026 (htab_t separate_decls_in_region::decl_copies): Likewise.
10028 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10029 Change type to hash_table. Update dependent calls and types.
10031 * tree-sra.c (candidates):
10032 Change type to hash_table. Update dependent calls and types.
10034 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10036 (int_tree_map_hash): Likewise.
10038 * tree-ssa-dom.c (htab_t avail_exprs):
10039 Change type to hash_table. Update dependent calls and types.
10041 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10042 Change type to hash_table. Update dependent calls and types.
10044 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10045 Change type to hash_table. Update dependent calls and types.
10047 * tree-ssa-phiopt.c (seen_ssa_names):
10048 Change type to hash_table. Update dependent calls and types.
10050 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10051 Change type to hash_table. Update dependent calls and types.
10053 * tree-ssa-uncprop.c (equiv):
10054 Change type to hash_table. Update dependent calls and types.
10056 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10058 PR rtl-optimization/57003
10059 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10060 call note_stores with kill_clobbered_value callback again after
10061 killing regs_invalidated_by_call.
10063 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10065 * config/aarch64/aarch64-simd.md
10066 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10067 (aarch64_simd_bsl<mode>): Likewise.
10068 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10070 2013-04-25 Marek Polacek <polacek@redhat.com>
10072 PR tree-optimization/57066
10073 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10075 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10077 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10079 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10081 * config/aarch64/aarch64-builtins.c
10082 (aarch64_fold_builtin): New.
10083 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10084 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10085 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10086 * config/aarch64/arm_neon.h
10087 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10089 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10090 Tejas Belagod <tejas.belagod@arm.com>
10092 * config/aarch64/aarch64-builtins.c
10093 (aarch64_gimple_fold_builtin): New.
10094 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10095 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10096 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10097 (addvv4sf): Update.
10098 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10100 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10102 * config/aarch64/aarch64.md
10103 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10105 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10107 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10108 (*ngcsi_uxtw): New pattern.
10110 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10111 Julian Brown <julian@codesourcery.com>
10113 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
10114 (TB_DREG): Add T_V4HF.
10115 (v4hf_UP): New macro.
10116 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10117 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10118 Handle initialisation of V4HF. Adjust initialisation of reinterpret
10120 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10121 (arm_vector_mode_supported_p): Handle V4HF.
10122 (arm_mangle_map): Handle V4HFmode.
10123 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
10124 * config/arm/arm_neon_builtins.def: Add entries for
10125 vcvtv4hfv4sf, vcvtv4sfv4hf.
10126 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
10127 (neon_vcvtv4hfv4sf): Likewise.
10128 * config/arm/neon-gen.ml: Handle half-precision floating point
10130 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
10131 * config/arm/arm_neon.h: Regenerate.
10132 * config/arm/neon.ml (type elts): Add F16.
10133 (type vectype): Add T_float16x4, T_floatHF.
10134 (type vecmode): Add V4HF.
10135 (type features): Add Requires_FP_bit feature.
10136 (elt_width): Handle F16.
10137 (elt_class): Likewise.
10138 (elt_of_class_width): Likewise.
10139 (mode_of_elt): Refactor.
10140 (type_for_elt): Handle F16, fix error messages.
10141 (vectype_size): Handle T_float16x4.
10142 (vcvt_sh): New function.
10143 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
10144 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
10145 (string_of_mode): Handle V4HF.
10146 * doc/arm-neon-intrinsics.texi: Regenerate.
10148 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10150 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
10151 format specifier in 'X' case.
10153 2013-04-25 Alan Modra <amodra@gmail.com>
10156 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
10157 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
10158 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
10159 Repeat for many other rotate/shift and mask patterns using subregs.
10160 Name lshiftrt insns.
10161 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
10162 on WORDS_BIG_ENDIAN.
10164 2013-04-25 Alan Modra <amodra@gmail.com>
10166 * config.gcc: Support little-endian powerpc-linux targets.
10167 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
10168 (LINK_OS_LINUX_SPEC): Define.
10169 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
10170 Preserve MASK_LITTLE_ENDIAN.
10171 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
10172 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
10173 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
10174 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
10175 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
10176 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
10177 Correct fp word order for little-endian. Don't shift toc entries
10178 smaller than a word for little-endian.
10179 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
10180 (bswapdi2 splits): Correct low-part subreg for little-endian.
10181 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
10182 low/high where such is correct only for be.
10183 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
10184 little-endian for -mcall-aixdesc.
10186 2013-04-25 Alan Modra <amodra@gmail.com>
10188 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10189 replace_equiv_address_nv.
10191 2013-04-25 Alan Modra <amodra@gmail.com>
10193 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
10195 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10198 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10199 * rtl.h (struct rtx_def): ...
10201 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10203 PR rtl-optimizations/57046
10204 * lra-constraints (split_reg): Set up lra_risky_transformations_p
10205 for multi-reg splits.
10207 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
10209 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
10211 2013-04-24 Sterling Augustine <saugustine@google.com>
10213 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
10214 (comp_dir_string, debug_str_dwo_section): New.
10215 (DEBUG_STR_DWO_SECTION): Rename to ...
10216 (DEBUG_DWO_STR_SECTION): ... this.
10217 (DEBUG_NORM_STR_SECTION): Delete.
10218 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
10219 (DEBUG_STR_DWO_SECTION_FLAGS): New.
10220 (find_AT_string): Move most logic to ...
10221 (find_AT_string_in_table): ... here. New.
10222 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
10223 add_skeleton_AT_string. Delete logic.
10224 (output_skeleton_debug_sections): Remove call to
10225 add_top_level_skeleton_die_attrs.
10226 (add_comp_dir_attribute): Move logic to comp_dir_string.
10227 (dwarf2out_init): Initialize debug_str_dwo_section.
10228 (output_indirect_string): Call find_string_form.
10229 (output_indirect_strings): Rewrite.
10230 (prune_unused_types): Empty skeleton_debug_str_hash.
10231 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
10232 (dwarf2out_finish): Call output_indirect_strings.
10234 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10236 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
10238 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10240 * rtl.h (struct rtx_def): Add comment for field jump.
10241 (LRA_SUBREG_P): New macro.
10242 * recog.c (register_operand): Check LRA_SUBREG_P.
10243 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10245 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10246 (lra_final_code_change): Skip subreg change for operators.
10247 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10248 if there are no operand changes.
10249 * lra-constraints.c (curr_insn_set): New.
10250 (match_reload): Set LRA_SUBREG_P.
10251 (emit_spill_move): Ditto.
10252 (check_and_process_move): Use curr_insn_set. Process only single
10253 set insns. Don't initialize sec_mem_p and change_p.
10254 (simplify_operand_subreg): Use LRA_SUBREG_P.
10255 (reg_in_class_p): New function.
10256 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10257 of #ifdef. Add code to remove cycling.
10258 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10259 non-null disp. Reload inner instead of disp when base and index
10260 are null. Try to put lo_sum into register.
10261 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10262 (check_and_process_move): Move code for move cost check to
10263 simple_move_p. Remove equiv_substitution.
10264 (simple_move_p): New function.
10265 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10266 curr_insn_set. Call check_and_process_move only for single set
10267 insns. Use the new function. Move call of check_and_process_move
10268 after operand equiv substitution and address process.
10270 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
10272 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
10273 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
10274 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
10276 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10278 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
10280 2013-04-24 Marek Polacek <polacek@redhat.com>
10282 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
10283 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
10284 (select_loops_exit_conditions): Likewise.
10285 (number_of_iterations_for_all_loops): Likewise.
10286 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
10287 (scev_analysis): Likewise.
10289 2013-04-02 Catherine Moore <clm@codesourcery.com>
10290 Chao-ying Fu <fu@mips.com>
10292 * config/mips/micromips.md (jraddiusp): New pattern.
10293 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
10294 instruction if possible.
10296 2013-04-24 Alan Modra <amodra@gmail.com>
10298 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
10300 2013-04-24 Julian Brown <julian@codesourcery.com>
10301 Chung-Lin Tang <cltang@codesourcery.com>
10303 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
10304 dependency behavior in enumeration type DIE generation. Add TODO note
10305 to comments about future DW_FORM_sdata/udata re-work of related code.
10307 2013-04-23 Lawrence Crowl <crowl@google.com>
10309 * Makefile.in: Update as needed below.
10311 * hash-table.h (class hash_table):
10312 Correct many methods with parameter types compare_type to the correct
10313 value_type. (Correct code was unlikely to notice the change.)
10314 (hash_table::elements_with_deleted) New.
10315 (class hashtable::iterator): New.
10316 (hashtable::begin()): New.
10317 (hashtable::end()): New.
10318 (FOR_EACH_HASH_TABLE_ELEMENT): New.
10320 * statistics.c (statistics_hashes):
10321 Change type to hash_table. Update dependent calls and types.
10323 * tree-into-ssa.c (var_infos):
10324 Change type to hash_table. Update dependent calls and types.
10326 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
10327 Change type to hash_table. Update dependent calls and types.
10329 * tree-ssa-loop-im.c (struct mem_ref.refs):
10330 Change type to hash_table. Update dependent calls and types.
10332 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
10333 Change type to hash_table. Update dependent calls and types.
10335 * tree-ssa-sccvn.c (vn_tables_s::nary):
10336 Change type to hash_table. Update dependent calls and types.
10337 (vn_tables_s::phis): Likewise.
10338 (vn_tables_s::references): Likewise.
10340 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
10341 (vn_reference_eq): Update parameter and return types.
10343 * tree-ssa-structalias.c (pointer_equiv_class_table):
10344 Change type to hash_table. Update dependent calls and types.
10345 (location_equiv_class_table): Likewise.
10347 * tree-vect-data-refs.c: Consequential changes for making
10348 peeling a hash_table.
10350 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
10351 (destroy_loop_vec_info): Dependent hash_table update.
10353 * tree-vectorizer.h (peeling_htab):
10354 Change type to hash_table. Update dependent calls and types.
10356 2013-04-23 Shiva Chen <shiva0217@gmail.com>
10358 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
10359 to check the register content is equal or not.
10360 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10361 to assign register content record.
10362 * lra-eliminations.c (update_reg_eliminate): Use
10363 lra_update_reg_val_offset to update register content offset.
10364 * lra-int.h (struct lra_reg): Add offset member.
10365 (lra_reg_val_equal_p): New static inline function.
10366 (lra_update_reg_val_offset): New static inline function.
10367 (lra_assign_reg_val): New static inline function.
10368 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10369 to assign register content record.
10370 (initialize_lra_reg_info_element): Initial offset to zero.
10372 2013-04-23 Catherine Moore <clm@codesourcery.com>
10374 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10375 operands. Record compression.
10377 2013-04-23 Xinliang David Li <davidxl@google.com>
10379 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10381 2013-04-23 Richard Biener <rguenther@suse.de>
10383 PR middle-end/57036
10384 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10385 parameter, only add abnormal goto edges from the copied body
10386 if the call could perform abnormal gotos.
10387 (copy_cfg_body): Adjust.
10389 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10391 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10393 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10395 * coretypes.h (gimple_stmt_iterator): Add struct to make
10398 2013-04-23 Richard Biener <rguenther@suse.de>
10400 PR tree-optimization/57026
10401 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10402 from SSA names occuring in abnormal PHI nodes.
10404 2013-04-22 Andi Kleen <ak@linux.intel.com>
10406 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10408 2013-04-22 Andi Kleen <ak@linux.intel.com>
10410 * lto/lto.c (print_lto_report_1): Declare early.
10411 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10413 2013-04-22 Andi Kleen <ak@linux.intel.com>
10415 * common.opt (-flto-report-wpa): Add.
10416 * doc/invoke.texi (-flto-report-wpa): Add.
10417 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10420 2013-04-22 Xinliang David Li <davidxl@google.com>
10422 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10423 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10424 * Makefile.in: New dependency
10426 David Daney <ddaney.cavm@gmail.com>
10428 * configure.ac (gcc_cv_as_micromips_support): Use the
10429 --fatal-warnings option.
10430 * configure: Regenerate.
10432 2013-04-22 Marek Polacek <polacek@redhat.com>
10435 * tsan.c (instrument_expr): Don't instrument expression
10436 in case its size is zero.
10438 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10442 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10444 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10446 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10448 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10449 (gimple_stmt_iterator): New typedef.
10450 * gimple.h (gimple_stmt_iterator): Rename to...
10451 (gimple_stmt_iterator_d): ... This.
10452 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10453 trees be valid for GIMPLE and GENERIC.
10454 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10455 * gimple-fold.c (gimple_fold_call): Call target hook
10456 gimple_fold_builtin.
10457 * hooks.c (hook_bool_gsiptr_false): New.
10458 * hooks.h (hook_bool_gsiptr_false): New.
10459 * target.def (fold_stmt): New.
10460 * doc/tm.texi: Regenerate.
10462 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10465 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10466 a set sp if no stack realignment.
10468 2013-04-22 Nick Clifton <nickc@redhat.com>
10470 * config.gcc (tilegx-linux): Extend extra_objs rather than
10472 (tilepro-linux): Likewise.
10474 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10476 * config/aarch64/aarch64-builtins.c
10478 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10479 (VAR<1-12>): Add MAP parameter.
10480 (BUILTIN_*): Likewise.
10481 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10482 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10483 (aarch64_ushl_n<mode>): Likewise.
10484 (aarch64_sshr_n<mode>): Likewise.
10485 (aarch64_ushr_n<mode>): Likewise.
10486 (aarch64_<maxmin><mode>): Likewise.
10487 (aarch64_sqrt<mode>): Likewise.
10488 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10489 (vshr<q>_n_*): Likewise.
10491 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10493 * config/aarch64/aarch64-builtins.c
10494 (aarch64_simd_builtin_type_mode): Handle SF types.
10496 (BUILTIN_GPF): Define.
10497 (aarch64_init_simd_builtins): Handle SF types.
10498 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10499 (frecps): Likewise.
10500 (frecpx): Likewise.
10501 * config/aarch64/aarch64-simd.md
10502 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10503 (aarch64_frecpe<mode>): New.
10504 (aarch64_frecps<mode>): Likewise.
10505 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10506 (v8type): Add frecp<esx>.
10507 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10508 (aarch64_frecps<mode>): Likewise.
10509 * config/aarch64/iterators.md (FRECP): New.
10510 (frecp_suffix): Likewise.
10511 * config/aarch64/arm_neon.h
10512 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10514 2013-04-22 Christian Bruel <christian.bruel@st.com>
10517 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10518 (REG_CLASS_NAMES): Idem.
10519 (REG_CLASS_CONTENTS): Idem.
10520 (REGCLASS_HAS_FP_REG): Idem.
10521 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10522 (sh_conditional_register_usage): Idem.
10524 2013-04-21 Jeff Law <law@redhat.com>
10526 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10527 (ssa_forward_propagate_and_combine): Use it.
10529 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10531 * lra.c: Update the flow chart diagram.
10533 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10535 PR rtl-optimization/56847
10536 * lra-constraints.c (process_alt_operands): Discourage alternative
10537 with non-matche doffsettable memory constraint fro memory with
10540 2013-04-19 Richard Biener <rguenther@suse.de>
10542 PR tree-optimization/56982
10543 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10545 * gimplify.c (gimplify_call_expr): Notice special calls.
10546 (gimplify_modify_expr): Likewise.
10547 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10548 abnormal control flow receivers.
10549 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10550 in the same way as cfun->has_nonlocal_labels.
10551 (gimple_purge_dead_abnormal_call_edges): Likewise.
10552 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10553 receivers start a basic-block.
10555 2013-04-19 Richard Biener <rguenther@suse.de>
10557 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10559 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10560 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10561 (SLP_TREE_LOAD_PERMUTATION): Add.
10562 (vect_transform_slp_perm_load): Adjust prototype.
10563 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10564 (vect_free_slp_instance): Likewise.
10565 (vect_create_new_slp_node): Likewise.
10566 (vect_supported_slp_permutation_p): Remove.
10567 (vect_slp_rearrange_stmts): Adjust.
10568 (vect_supported_load_permutation_p): Likewise. Inline
10569 vect_supported_slp_permutation_p here.
10570 (vect_analyze_slp_instance): Compute load permutations per
10571 slp node instead of per instance.
10572 (vect_get_slp_defs): Adjust.
10573 (vect_transform_slp_perm_load): Likewise.
10574 (vect_schedule_slp_instance): Remove redundant code.
10575 (vect_schedule_slp): Remove hack for PR56270, add it ...
10576 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10577 CSE loads for SLP. Adjust.
10579 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10581 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10582 spelling in two comments.
10584 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10587 * config/arm/arm.c (load_multiple_sequence): Require SP
10588 as base register for loads if SP is in the register list.
10590 2013-04-19 Martin Jambor <mjambor@suse.cz>
10592 PR tree-optimization/56718
10593 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10594 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10595 and made public. Adjusted all callers.
10596 (ipa_intraprocedural_devirtualization): New function.
10597 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10598 (ipa_intraprocedural_devirtualization): Likewise.
10599 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10601 2013-04-19 Richard Biener <rguenther@suse.de>
10603 PR tree-optimization/57000
10604 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10606 2013-04-19 Terry Guo <terry.guo@arm.com>
10608 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10610 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10611 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10612 (cortex_m4_fmacs): Use new reservations.
10613 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10615 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10617 PR rtl-optimization/56999
10618 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10620 (lra_coalesce): Remove split_origin_bitmap and related code.
10621 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10622 ranges if necessary.
10624 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
10626 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10628 (ix86_expand_call): Remove clobbered_registers array and use
10629 x86_64_ms_sysv_extra_clobbered_registers instead.
10630 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10632 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10634 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10635 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10636 (*call_value_rex64_ms_sysv): Ditto.
10638 2013-04-18 Cary Coutant <ccoutant@google.com>
10640 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10641 parent instead of die_mark.
10643 2013-04-18 Diego Novillo <dnovillo@google.com>
10645 * gimple.c (create_gimple_tmp): New.
10646 (get_expr_type): New.
10647 (build_assign): New.
10648 (build_type_cast): New.
10649 * gimple.h (enum ssa_mode): Define.
10650 (gimple_seq_set_location): New.
10651 * asan.c (build_check_stmt): Change some gimple_build_* calls
10652 to use build_assign and build_type_cast.
10654 2013-04-18 Richard Biener <rguenther@suse.de>
10656 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10657 handle negative step. Remove redundant checks.
10658 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10659 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10660 for negative step and grouped loads fail to vectorize.
10662 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10664 * emit-rtl.c (reset_insn_used_flags): New function.
10665 (reset_all_used_flags): Use it.
10666 (verify_insn_sharing): New function.
10667 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10669 2013-04-18 Jakub Jelinek <jakub@redhat.com>
10671 PR tree-optimization/56984
10672 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10673 and (x >> M) >= N don't register any assertion if N << M is the
10676 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10678 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10679 do not delete_insn insns that have not yet been emitted, only
10680 unlink them with remove_insn.
10681 * df-scan.c (df_insn_delete): Revert r197492.
10683 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10685 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10686 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10688 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10690 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10691 define_insn_and_split.
10692 (and_scc,ior_scc,negscc): Likewise.
10693 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10695 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10697 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10698 can benefit from using a sequence of LDRD instructions in epilogue
10699 instead of a single LDM instruction.
10701 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10704 * doc/extend.texi: Fix typo.
10706 2013-04-17 Richard Biener <rguenther@suse.de>
10708 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10709 (vect_build_slp_tree): ... here.
10710 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10711 match. Remove special-casing of mismatched complex loads.
10712 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10713 re-try the match with swapped commutative operands.
10714 (vect_supported_load_permutation_p): Remove special-casing of
10715 mismatched complex loads.
10716 (vect_analyze_slp_instance): Adjust.
10718 2013-04-17 Richard Biener <rguenther@suse.de>
10720 PR rtl-optimization/56921
10721 * cfgloop.h (struct loop): Add simple_loop_desc member.
10722 (struct niter_desc): Mark with GTY(()).
10723 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10724 * loop-iv.c (get_simple_loop_desc): Likewise.
10725 (free_simple_loop_desc): Likewise.
10728 2013-04-16 Richard Biener <rguenther@suse.de>
10730 PR rtl-optimization/56921
10731 * loop-init.c (pass_rtl_move_loop_invariants): Add
10732 TODO_do_not_ggc_collect to todo_flags_finish.
10733 (pass_rtl_unswitch): Same.
10734 (pass_rtl_unroll_and_peel_loops): Same.
10735 (pass_rtl_doloop): Same.
10737 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10739 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10740 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10741 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10743 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10744 * tree-streamer.c (record_common_node): Adjust reference in comment.
10746 2013-04-17 Terry Guo <terry.guo@arm.com>
10748 * config/arm/cortex-m4.md: Add a new bypass.
10750 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10752 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10754 (*subs_<optab><mode>_multp2): New pattern.
10755 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10756 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10758 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10760 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10761 (*subs_mul_imm_<mode>): New pattern.
10763 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
10766 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10767 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10768 (vsx_movti_32bit): Change j->wa to O->wa.
10770 2013-04-16 Richard Biener <rguenther@suse.de>
10772 PR rtl-optimization/56921
10773 * loop-init.c (pass_rtl_move_loop_invariants): Add
10774 TODO_do_not_ggc_collect to todo_flags_finish.
10775 (pass_rtl_unswitch): Same.
10776 (pass_rtl_unroll_and_peel_loops): Same.
10777 (pass_rtl_doloop): Same.
10779 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
10781 * config/arm/arm.c (emit_multi_reg_push): New declaration
10782 for an existing function.
10783 (arm_emit_strd_push): New function.
10784 (arm_expand_prologue): Used here.
10785 (arm_emit_ldrd_pop): New function.
10786 (arm_expand_epilogue): Used here.
10787 (arm_get_frame_offsets): Update condition.
10788 (arm_emit_multi_reg_pop): Add a special case for load of a single
10789 register with writeback.
10791 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
10793 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10796 2013-04-16 Richard Biener <rguenther@suse.de>
10798 PR tree-optimization/56756
10799 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10800 (first_mem_ref_loc): New.
10801 (execute_sm): Place the load temporarily before a previous
10802 access instead of in the latch edge to ensure its SSA dependencies
10803 are defined at points dominating the load.
10805 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10807 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10808 correct fix by moving header and footer insn to the footer of
10809 the merged basic block. Clear BB_END of the merged-away block.
10811 PR middle-end/43631
10812 * emit-rtl.c (make_note_raw): New function.
10813 (link_insn_into_chain): New static inline function.
10814 (add_insn): Use it.
10815 (add_insn_before, add_insn_after): Factor insn chain linking code...
10816 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10817 using link_insn_into_chain.
10818 (note_outside_basic_block_p): New helper function for emit_note_after
10819 and emit_note_before.
10820 (emit_note_after): Use nobb variant of add_insn_after if the note
10821 should not be contained in a basic block.
10822 (emit_note_before): Use nobb variant of add_insn_before if the note
10823 should not be contained in a basic block.
10824 (emit_note_copy): Use make_note_raw.
10825 (emit_note): Likewise.
10826 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10827 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10828 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10829 the moved barrier the tail of the basic block it follows.
10830 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10832 2013-04-15 Jakub Jelinek <jakub@redhat.com>
10834 PR tree-optimization/56962
10835 * gimple-ssa-strength-reduction.c (record_increment): Only set
10836 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10837 either rhs1 or rhs2 is equal to c->base_expr.
10839 2013-04-15 Richard Biener <rguenther@suse.de>
10841 PR tree-optimization/56933
10842 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10844 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10845 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10846 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10847 dependence check ...
10848 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10850 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10851 GROUP_READ_WRITE_DEPENDENCE.
10853 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10855 * emit-rtl.c (reset_all_used_flags): New function.
10856 (verify_rtl_sharing): Call reset_all_used_flags before and after
10857 performing the checks.
10859 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10861 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10862 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10863 * config/arm/constraints.md (De): New constraint.
10864 * config/arm/neon.md (anddi3_neon): Delete.
10865 (neon_vand<mode>): Expand to standard anddi3 pattern.
10866 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10867 Move earlier in the file.
10868 (neon_inv_logic_op2): Likewise.
10869 (arm_anddi_operand_neon): New predicate.
10871 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10873 * configure.ac (gcc_cv_ld_as_needed): Set
10874 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10875 Use -z ignore, -z record on *-*-solaris2*.
10876 (HAVE_LD_AS_NEEDED): Update comment.
10877 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10878 * configure: Regenerate.
10879 * config.in: Regenerate.
10880 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10881 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10882 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10883 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10884 equivalents. Fix markup.
10885 * doc/tm.texi: Regenerate.
10887 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10889 * config/i386/i386.opt: New option mstack-protector-guard=.
10890 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10891 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10892 TARGET_SSP_TLS_GUARD.
10893 * config/i386/i386.c (ix86_option_override_internal): Set
10894 ix86_stack_protector_guard.
10895 * config/i386/i386.md (stack_protect_set): Enable for
10896 TARGET_SSP_TLS_GUARD only.
10897 (stack_protect_set_<mode>): Ditto.
10898 (stack_protect_test): Ditto.
10899 (stack_protect_test_<mode>): Ditto.
10900 * doc/invoke.texi (i386 Option): Document.
10902 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10905 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10906 (S_MODES): Set H_MODE bit.
10907 (SF_MODES): Set only S_MODE and SF_MODE bits.
10908 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10909 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10910 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10911 <MODE_FLOAT>: Likewise.
10913 2013-04-15 Joey Ye <joey.ye@arm.com>
10915 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10917 2013-04-15 Joey Ye <joey.ye@arm.com>
10919 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10921 (thumb_far_jump_used_p): Count instruction size and set
10924 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10926 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10927 * resource.c (mark_target_live_regs): Reformat conditional block.
10929 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10931 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10932 notes, they are emitted only just before final.
10933 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10935 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10937 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10938 * cfgrtl.c (delete_insn): Call it here instead.
10939 * lra-spills.c (lra_final_code_change): Use delete_insn.
10940 * haifa-sched.c (sched_remove_insn): Likewise.
10941 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10942 returning to the nop pool.
10943 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10944 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10946 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10948 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10949 * doc/tm.texi: Regenerated.
10951 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
10953 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10956 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10958 * df-core.c (df_find_def): Compare register numbers.
10959 (df_find_use): Likewise.
10961 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10964 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
10965 lra_in_progress for return.
10967 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10969 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
10970 define_insn into define_insn_and_split and emit movsicc patterns.
10972 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
10974 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
10976 2013-04-12 Richard Biener <rguenther@suse.de>
10978 * tree-pass.h (TODO_do_not_ggc_collect): New.
10979 * passes.c (execute_one_ipa_transform_pass): Honor
10980 TODO_do_not_ggc_collect.
10981 (execute_one_pass): Likewise.
10984 2013-04-10 Richard Biener <rguenther@suse.de>
10986 * passes.c (init_optimization_passes): Remove reload pass.
10987 * ira.c (do_reload): Merge into ...
10989 (rest_of_handle_reload): Remove.
10990 (pass_reload): Likewise.
10991 * config/i386/i386.c (ix86_option_override): Refer to ira instead
10992 of reload for vzeroupper pass placement.
10994 2013-04-12 Jakub Jelinek <jakub@redhat.com>
10996 PR tree-optimization/56918
10997 PR tree-optimization/56920
10998 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
10999 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11000 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11001 use wide_mul_with_sign method.
11003 2013-04-12 Richard Biener <rguenther@suse.de>
11005 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11006 not be considered a gimple constant.
11008 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11010 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11011 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11014 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11015 Jakub Jelinek <jakub@redhat.com>
11017 * opts.c: Include diagnostic-color.h.
11018 (common_handle_option): Handle OPT_fdiagnostics_color_.
11019 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11020 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11021 (diagnostic-color.o): New.
11022 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11023 (diagnostic_color_rule): New enum.
11024 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11025 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11026 the location string.
11027 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11028 either NULL, or color kind.
11029 * diagnostic-color.c: New file.
11030 * diagnostic-color.h: New file.
11031 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11033 * doc/invoke.texi (-fdiagnostics-color): Document.
11034 * pretty-print.h (pp_show_color): Define.
11035 (struct pretty_print_info): Add show_color field.
11036 * diagnostic.c: Include diagnostic-color.h.
11037 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11038 macros. Colorize error:, warning: etc. strings and also the location
11040 (diagnostic_show_locus): Colorize the caret line.
11041 * pretty-print.c: Include diagnostic-color.h.
11042 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11043 inside of %< %> quotes or quoted through q format modifier.
11045 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11047 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11049 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11051 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11052 code in CC_NZ mode.
11053 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11056 2013-04-11 Marek Polacek <polacek@redhat.com>
11058 PR tree-optimization/48184
11059 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11061 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11063 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11064 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11065 (skip_simple_arithmetic): Tidy up.
11066 * tree.h (skip_simple_constant_arithmetic): Declare.
11068 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11070 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11072 2013-04-11 Richard Biener <rguenther@suse.de>
11074 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11075 generate vector constants.
11077 2013-04-11 Richard Biener <rguenther@suse.de>
11079 PR tree-optimization/56878
11080 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11081 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11083 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11084 Prefer to align the DR with the most invariant base address.
11086 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11088 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11091 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11093 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11094 floating-point vector comparisons against 0.
11096 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11098 PR tree-optimization/56899
11099 * fold-const.c (extract_muldiv_1): Apply distributive law
11100 only if TYPE_OVERFLOW_WRAPS (ctype).
11102 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11105 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11106 loading has side effect.
11108 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
11110 * config/sparc/sparc.c: Include tree-pass.h.
11111 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
11112 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
11113 head of file. Change return type. Split off gate function.
11114 (sparc_gate_work_around_errata): New function.
11115 (pass_work_around_errata): New pass definition.
11116 (insert_pass_work_around_errata) New pass insert definition to
11117 insert pass_work_around_errata just after delayed-branch scheduling.
11118 (sparc_option_override): Insert the pass.
11119 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
11121 2013-04-10 David S. Miller <davem@davemloft.net>
11123 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
11124 or -mcpu=hypersparc.
11126 * target.def (cstore_mode): New hook.
11127 * target.h: Include insn-codes.h
11128 * targhooks.c: Likewise.
11129 (default_cstore_mode): New function.
11130 * targhooks.h: Declare it.
11131 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
11132 * doc/tm.texi: Rebuild.
11133 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
11134 target hook, rather than inspecting the insn_data.
11135 * config/sparc/sparc.c (sparc_cstore_mode): New function.
11136 (TARGET_CSTORE_MODE): Redefine.
11137 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
11139 * config/sparc/predicates.md (cstore_result_operand): New special
11141 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
11142 Use it for operand 0.
11143 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
11144 (*snesi_special): Likewise.
11145 (*snesi_zero): Likewise.
11146 (*seqsi_zero): Likewise.
11147 (*sltu_insn): Likewise.
11148 (*sgeu_insn): Likewise.
11149 (*seqdi_special): Make operand 0 and comparison operation be of
11151 (*snedi_special): Likewise.
11152 (*snedi_special_vis3): Likewise.
11153 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
11154 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
11155 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
11156 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
11157 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
11158 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
11159 (*sltu_extend_sp64): Likewise.
11160 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
11161 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
11162 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
11163 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
11164 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
11166 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
11168 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
11169 (aarch64_start_file): Use the new function.
11171 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11172 Jason Merrill <jason@redhat.com>
11174 * common.opt: Add -gdwarf.
11175 * opts.c (common_handle_option): Handle it.
11176 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
11178 2013-04-10 Richard Biener <rguenther@suse.de>
11180 * passes.c (execute_todo): Do not call ggc_collect conditional here.
11181 (execute_one_ipa_transform_pass): But unconditionally here.
11182 (execute_one_pass): And here.
11183 (init_optimization_passes): Remove reload pass.
11184 * tree-pass.h (TODO_ggc_collect): Remove.
11185 (pass_reload): Likewise.
11186 * ira.c (do_reload): Merge into ...
11188 (rest_of_handle_reload): Remove.
11189 (pass_reload): Likewise.
11190 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11191 of reload for vzeroupper pass placement.
11192 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
11193 and todo_flags_finish of all passes.
11195 2013-04-10 Richard Biener <rguenther@suse.de>
11197 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
11198 first_const_oprnd field, rename first_def_type to first_op_type.
11199 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
11200 (vect_get_and_check_slp_defs): Always use the type of the
11201 operand. Allow mixed vect_external_def, vect_constant_def types.
11202 (vect_get_constant_vectors): Handle mixed vect_external_def,
11203 vect_constant_def types.
11205 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
11207 PR tree-optimization/55524
11208 * tree-ssa-math-opts.c
11209 (convert_mult_to_fma): Don't use an fms construct
11210 when we don't have an fms operation, but fnma, and it looks
11211 likely that we'll be able to use the latter.
11213 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11215 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
11217 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
11218 inline fail caused by overwritable functions.
11220 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
11222 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
11223 unnecessary bits in the constant power of two case.
11225 2013-04-10 Richard Biener <rguenther@suse.de>
11227 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
11228 broken code swapping operands.
11229 (vect_build_slp_tree): Do not compute load permutations here.
11230 (vect_analyze_slp_instance): Compute load permutations here,
11231 after building the SLP tree.
11233 2013-04-09 Christian Bruel <christian.bruel@st.com>
11235 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
11236 of next/prev_real_insn.
11238 2013-04-09 Jan Hubicka <jh@suse.cz>
11240 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11241 Drop aliased parameter.
11242 (function_and_variable_visibility): Do not handle alias pairs.
11243 * cgraph.c (varpool_externally_visible_p): Update prototype.
11244 * varpool.c (varpool_add_new_variable): Update.
11246 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11248 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
11250 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
11252 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
11254 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
11256 2013-04-09 Marek Polacek <polacek@redhat.com>
11258 PR tree-optimization/48762
11259 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
11261 2013-04-09 Richard Biener <rguenther@suse.de>
11263 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
11265 (vect_build_slp_tree): Likewise.
11266 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
11267 calculating the cost of a SLP instance.
11268 (vect_analyze_slp_instance): Use it from here, after building
11271 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11273 PR middle-end/56883
11274 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
11275 expand_omp_for_static_chunk): Use simple_p = true in
11276 force_gimple_operand_gsi calls when assigning to addressable decls.
11278 2013-04-09 Jeff Law <law@redhat.com>
11280 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
11281 when the boolean was created by converting a wider object which
11282 had a boolean range.
11284 2013-04-09 Richard Biener <rguenther@suse.de>
11286 * tree-vectorizer.h (slp_void_p): Remove.
11287 (slp_tree): Typedef before _slp_tree declaration.
11288 (struct _slp_tree): Use a vector of slp_tree as children.
11289 (vect_get_place_in_interleaving_chain): Remove.
11290 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
11292 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
11294 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
11295 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
11296 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
11297 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
11298 Use slp_node instead of slp_void_p and adjust.
11300 2013-04-09 Richard Biener <rguenther@suse.de>
11302 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
11303 work that is not necessary.
11305 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11307 PR tree-optimization/56854
11308 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
11309 forward into clobber stmts if it would change MEM_REF lhs into
11312 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
11314 * tree.c (type_hash_lookup, type_hash_add): Make static.
11315 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
11317 2013-04-09 Richard Biener <rguenther@suse.de>
11319 * tree.h (unsave_expr_now): Remove.
11320 * tree-inline.c (mark_local_for_remap_r): Remove.
11321 (unsave_expr_1): Likewise.
11322 (unsave_r): Likewise.
11323 (unsave_expr_now): Likewise.
11324 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
11325 (propagate_tree_value): Likewise.
11327 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
11329 * doc/rtl.texi (sequence): Rewrite documentation to match the
11330 current use of SEQUENCE rtl objects.
11331 * rtl.def (SEQUENCE): Likewise.
11333 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
11334 Update documentation.
11335 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
11336 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
11338 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
11340 2013-04-08 Teresa Johnson <tejohnson@google.com>
11342 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
11343 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
11345 (estimate_edge_size_and_time): Add comment to suggest using rounding
11347 (estimate_node_size_and_time): Ditto.
11348 (remap_edge_change_prob): Use helper rounding divide methods.
11349 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
11350 (gimple_mod_pow2_value_transform): Ditto.
11351 (gimple_mod_subtract_transform): Ditto.
11352 (gimple_ic_transform): Ditto.
11353 (gimple_stringops_transform): Ditto.
11354 * stmt.c (conditional_probability): Ditto.
11355 (emit_case_dispatch_table): Ditto.
11356 * lto-cgraph.c (merge_profile_summaries): Ditto.
11357 * tree-optimize.c (execute_fixup_cfg): Ditto.
11358 * cfgcleanup.c (try_forward_edges): Ditto.
11359 * cfgloopmanip.c (scale_loop_profile): Ditto.
11361 (duplicate_loop_to_header_edge): Ditto.
11362 (lv_adjust_loop_entry_edge): Ditto.
11363 * tree-vect-loop.c (vect_transform_loop): Ditto.
11364 * profile.c (compute_branch_probabilities): Ditto.
11365 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11366 * lto-streamer-in.c (input_cfg): Ditto.
11367 * gimple-streamer-in.c (input_bb): Ditto.
11368 * ipa-cp.c (update_profiling_info): Ditto.
11369 (update_specialized_profile): Ditto.
11370 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11371 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11373 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11374 (compute_trg_info): Ditto.
11375 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11376 (purge_dead_edges): Ditto.
11377 * loop-unswitch.c (unswitch_loop): Ditto.
11378 * cgraphclones.c (cgraph_clone_edge): Ditto.
11379 (cgraph_clone_node): Ditto.
11380 * tree-inline.c (copy_bb): Ditto.
11381 (copy_edges_for_bb): Ditto.
11382 (initialize_cfun): Ditto.
11383 (copy_cfg_body): Ditto.
11384 (expand_call_inline): Ditto.
11386 2013-04-08 Kai Tietz <ktietz@redhat.com>
11388 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11389 TARGET_CYGWIN64 by TARGET_64BIT.
11391 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11393 * config/epiphany/epiphany.md (GPR_1): New constant.
11394 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11395 * config/epiphany/epiphany.c (gen_compare_reg):
11396 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11397 is already in place.
11398 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11399 Don't require being called during rtl expansion; If y operlaps r0,
11401 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11402 (epiphany_expand_epilogue): Likewise.
11404 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11405 Don't use CC_FPmode for ORDERED / UNORDERED.
11406 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11408 * config/epiphany/constraints.md (CnL): New constraint.
11409 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11410 * config/epiphany/predicates.md (add_operand): Allow 1024.
11412 * config/epiphany/epiphany.md (logical_op): New code iterator.
11413 (op_mnc): New code attribute.
11414 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11415 (mov_f+1, mov_f+2): New peephole2 patterns.
11417 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11418 (cstoresi4): Also allow re-use of zero result when doing a NE
11419 comparison to a non-zero operand.
11420 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11422 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11423 Use gen_addsi3_i / gen_subsi3_i.
11425 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11429 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11430 contain anything but clobbers, at most one __builtin_stack_restore,
11431 optionally debug stmts and final resx, and if it has at least one
11432 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11433 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11434 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11435 which isn't defaut definition, remove them.
11436 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11437 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11438 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11439 with MEM_REF LHS with SSA_NAME address.
11441 2013-04-08 Jeff Law <law@redhat.com>
11443 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11445 2013-04-08 Richard Biener <rguenther@suse.de>
11447 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11449 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11450 determined vector type.
11451 (vect_analyze_data_refs): Likewise.
11452 (vect_get_new_vect_var): Adjust.
11453 (vect_create_destination_var): Preserve SSA name versions.
11454 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11455 not dump anything here.
11457 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11459 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11460 Add member lr_slot_known.
11461 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11463 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11464 Remove code that sets lr_slot_offset according to what a previous
11465 version of epiphany_emit_save_restore used to do.
11466 (epiphany_emit_save_restore): When doing an lr save or restore,
11467 set/verify lr_slot_known and lr_slot_offset.
11469 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11472 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11475 2013-04-08 Richard Biener <rguenther@suse.de>
11477 * alias.c (find_base_term): Fix thinko in previous change.
11479 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11481 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11482 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11483 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11484 if possible to compute val.
11485 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11486 For QImode integers don't require anything about precision. Use
11487 const_with_all_bytes_same to find out if the constant doesn't have
11488 repeated bytes in it.
11490 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11492 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11495 2013-04-08 Marek Polacek <polacek@redhat.com>
11497 PR rtl-optimization/48182
11498 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11501 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11504 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11505 nuses, make sure we have a label.
11507 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11510 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11511 (rs6000_emit_swdiv_low_precision): Remove.
11512 (rs6000_emit_swdiv): Rewrite to handle between one and four
11513 iterations of Newton-Raphson generally; modify required number of
11514 iterations for some cases.
11515 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11517 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11519 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11520 set-but-unused variable.
11522 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11523 basic blocks of released function bodies garbage-collectable.
11525 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11526 (struct rtl_opt_pass): Add TODO_df_finish.
11528 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11530 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11532 * config/arm/constraints.md (q): New constraint.
11533 * config/arm/ldrdstrd.md: New file.
11534 * config/arm/arm.md (ldrdstrd.md) New include.
11535 (arm_movdi): Use "q" instead of "r" constraint
11536 for double-word memory access.
11537 (movdf_soft_insn): Likewise.
11538 * config/arm/vfp.md (movdi_vfp): Likewise.
11539 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
11540 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
11541 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11542 (mem_ok_for_ldrd_strd): Likewise.
11543 (output_move_double): Update assertion.
11545 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11547 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11549 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11551 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11552 define_insn_and_split.
11553 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11555 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11557 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11558 define_insn_and_split.
11559 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
11560 (shiftsi3_compare): New pattern.
11561 (rrx): New pattern.
11562 * config/arm/unspecs.md (UNSPEC_RRX): New.
11564 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11566 * config/arm/arm.md (negdi_extendsidi): New pattern.
11567 (negdi_zero_extendsidi): Likewise.
11569 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11571 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11572 define_insn_and_split.
11573 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11574 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11576 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11578 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11579 define_insn_and_split.
11580 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11581 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11583 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11585 * config/arm/arm.md (subsi3_carryin): New pattern.
11586 (subsi3_carryin_const): Likewise.
11587 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11588 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11590 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11592 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11594 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11596 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
11597 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
11599 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11601 * config/arm/arm.c (arm_expand_builtin): Change fcode
11602 type to unsigned int.
11604 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11606 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11608 2013-04-04 Ian Lance Taylor <iant@google.com>
11610 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11613 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11615 PR middle-end/56729
11616 * df-scan.c (df_insn_delete): Disable failing assert.
11618 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11620 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11621 New function prototype.
11622 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11623 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11624 (arm_builtin_vectorized_function): New function.
11626 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11628 * config/arm/arm_neon_builtins.def: New file.
11629 * config/arm/arm.c (neon_builtin_data): Move contents to
11630 arm_neon_builtins.def.
11631 (enum arm_builtins): Include neon builtin definitions.
11632 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
11633 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
11635 2013-04-04 Marek Polacek <polacek@redhat.com>
11637 PR tree-optimization/48186
11638 * predict.c (maybe_hot_frequency_p): Return false if
11639 HOT_BB_FREQUENCY_FRACTION is 0.
11640 (cgraph_maybe_hot_edge_p): Likewise.
11642 2013-04-04 Richard Biener <rguenther@suse.de>
11644 PR tree-optimization/56826
11645 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11648 2013-04-04 Richard Biener <rguenther@suse.de>
11650 PR tree-optimization/56213
11651 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
11652 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
11654 2013-04-04 Richard Biener <rguenther@suse.de>
11656 PR tree-optimization/56837
11657 * tree-loop-distribution.c (classify_partition): For non-zero
11658 values require that the value has the same precision as its
11659 mode to be useful as memset value.
11661 2013-04-03 Nick Clifton <nickc@redhat.com>
11663 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
11664 (fmssf4): Use fmsf.s on E3V5 architectures.
11665 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11666 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11668 2013-04-03 Jeff Law <law@redhat.com>
11670 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11671 (lra-eliminations.o): Likewise.
11673 2013-04-03 Teresa Johnson <tejohnson@google.com>
11675 * gcov-io.c (compute_working_sets): Moved most of body of old
11676 compute_working_sets here from profile.c.
11677 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
11678 (gcov_working_set_t): Moved typedef here from basic-block.h
11679 (compute_working_set): Declare.
11680 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11681 (get_working_sets): Renamed from compute_working_set,
11682 replace most of body with call to new compute_working_sets.
11683 (get_exec_counts): Replace call to compute_working_sets
11684 to get_working_sets.
11685 * profile.h (get_working_sets): Renamed from compute_working_set.
11686 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
11687 to get_working_sets.
11688 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11689 * gcov-dump.c (dump_working_sets): New function.
11691 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11693 * hwint.c (sext_hwi, zext_hwi): New functions.
11694 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11695 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11696 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11697 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11698 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11699 (sext_hwi, zext_hwi): New functions.
11701 2013-04-03 Jeff Law <law@redhat.com>
11703 PR tree-optimization/56799
11704 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11705 back test for widening conversion erroneously dropped in prior change.
11707 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11710 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11711 instead of next_real_insn.
11713 2013-04-03 Marek Polacek <polacek@redhat.com>
11716 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
11718 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11721 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11723 (thumb1_output_casesi): Likewise.
11724 (thumb2_output_casesi): Likewise.
11726 2013-04-03 Richard Biener <rguenther@suse.de>
11728 PR tree-optimization/56817
11729 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11731 (tree_unroll_loops_completely_1): ... new function to manually
11732 walk the loop tree, properly defering outer loops of unrolled
11733 loops to later iterations.
11735 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11737 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11738 (vectorizable_load): Likewise.
11739 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11740 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11742 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11744 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11747 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11749 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11751 2013-04-03 Bin Cheng <bin.cheng@arm.com>
11753 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11755 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11757 PR tree-optimization/56790
11758 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11761 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11763 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11765 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11766 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11769 2013-04-03 Jakub Jelinek <jakub@redhat.com>
11772 * tree.h (force_folding_builtin_constant_p): New decl.
11773 * builtins.c (force_folding_builtin_constant_p): New variable.
11774 (fold_builtin_constant_p): Fold immediately also if
11775 force_folding_builtin_constant_p.
11777 2013-04-03 Richard Biener <rguenther@suse.de>
11779 PR tree-optimization/56812
11780 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11781 DRs of the same interleaving chain are independent.
11783 2013-04-02 Jason Merrill <jason@redhat.com>
11785 * gdbinit.in (pbb): Use debug fn.
11787 2013-04-02 Lawrence Crowl <crowl@google.com>
11789 * sese.h (struct ivtype_map_elt_s): Remove unused.
11790 (extern debug_ivtype_map): Remove unused.
11791 (extern eq_ivtype_map_elts): Remove unused.
11792 * sese.c (debug_ivtype_map): Removed unused.
11793 (debug_ivtype_map_1): Removed unused.
11794 (debug_ivtype_elt): Remove unused.
11795 (eq_ivtype_map_elts): Remove unused.
11798 2013-04-02 Kai Tietz <ktietz@redhat.com>
11801 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11802 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
11803 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11805 (legitimize_pe_coff_symbol): Likewise.
11806 (is_imported_p): New helper-function.
11807 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11808 for Windows x64 targets.
11809 (ix86_expand_prologue): Optimize for pe-coff targets.
11810 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11811 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11812 medium/large code-model.
11813 (legitimize_pic_address): Likewise.
11814 (legitimize_tls_address): Likewise.
11815 (ix86_expand_call): Likewise.
11816 (x86_output_mi_thunk): Likewise.
11817 (get_dllimport_decl): Add new beimport argument.
11818 (construct_plt_address): Don't assert for x64 pe-coff targets.
11819 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11821 (SYMBOL_FLAG_STUBVAR): New macro.
11822 (SYMBOL_REF_STUBVAR_P): Likewise.
11823 * config/i386/winnt.c (stub_list): New structure.
11824 (stub_head): New local variable.
11825 (i386_pe_record_stub): New function.
11826 (i386_pe_file_end): Emit refptr-stubs.
11828 2013-04-02 Jakub Jelinek <jakub@redhat.com>
11830 PR rtl-optimization/56745
11831 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11832 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11835 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11836 and both of them are MEM_REFs, just compare first argument for
11837 equality and attempt to deal even with differing offsets.
11840 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11841 of gimple_clobber_p to be MEM_REF.
11842 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11843 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11844 after gimplification.
11845 * asan.c (get_mem_ref_of_assignment): Don't instrument
11846 gimple_clobber_p stmts.
11847 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11848 gimple_clobber_p stmt if they have MEM_REF lhs and
11849 are dead because of another gimple_clobber_p stmt.
11850 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11851 gimple_clobber_p stmts like debug stmts.
11852 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11853 that refer to unused VAR_DECLs or uninitialized values.
11854 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11855 gimple_clobber_p stmts if they refer to removed parameters.
11856 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11859 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
11861 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11862 using SWI48 mode attribute.
11864 2013-04-02 Wei Mi <wmi@google.com>
11866 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11867 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11868 *<rotate_insn><mode>3_mask in i386.md.
11870 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11872 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11874 2013-04-02 Richard Biener <rguenther@suse.de>
11876 PR tree-optimization/56778
11877 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11878 Runtime alias tests are not supported for gather loads.
11879 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11880 stmts referenced from SSA operands before updating SSA form.
11882 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11883 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11885 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
11886 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11887 * config/arm/cortex-a53.md: New file.
11888 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11889 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11890 * config/arm/arm.c (arm_issue_rate): Likewise.
11891 * config/arm/arm-tune.md: Regenerate
11892 * config/arm/arm-tables.opt: Regenerate.
11893 * config/arm/arm-cores.def: Add cortex-a53.
11895 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11897 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11900 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11902 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11903 scalar load/store operations using B/H registers.
11904 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11906 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11908 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11910 * config/aarch64/aarch64.c
11911 (aarch64_simd_scalar_immediate_valid_for_move): New.
11912 * config/aarch64/aarch64-protos.h
11913 (aarch64_simd_scalar_immediate_valid_for_move): New.
11914 * config/aarch64/constraints.md (Dh, Dq): New.
11915 * config/aarch64/iterators.md (hq): New.
11917 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11919 * reorg.c (get_branch_condition): Deal with conditional returns.
11920 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11922 2013-04-01 Wei Mi <wmi@google.com>
11924 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11925 Truncate operand 2 using %b asm operand modifier.
11926 (*<shift_insn><mode>3_mask): Ditto.
11927 (*<rotate_insn><mode>3_mask): Ditto.
11929 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11931 PR middle-end/56798
11932 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11934 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11936 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11938 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11940 2013-03-30 Lawrence Crowl <crowl@google.com>
11942 * dse.c (clear_alias_sets): Remove never set.
11943 (disqualified_clear_alias_sets): Remove never set.
11944 (clear_alias_mode_pool): Remove never set.
11945 (dse_step0): Remove condition that is never true.
11946 (canon_address): Remove condition that is never true.
11947 (dse_step7): Remove condition that is never true.
11948 (rest_of_handle_dse): Remove condition that is never true.
11949 (rest_of_handle_dse::did_global): Remove never read from above.
11950 (dse_step2_spill): Remove never called from above.
11951 (dse_step5_spill): Remove never called from above.
11953 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11955 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11956 JUMP_TABLE_DATA changes.
11957 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11958 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11959 (Insns) <jump_table_data>: New entry.
11960 * doc/tm.texi: Regenerate.
11962 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11964 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
11965 for table jump at the end of a basic block using tablejump_p.
11966 * targhooks.c (default_invalid_within_doloop): Likewise.
11967 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
11968 target hook implementation that is identical to the default hook.
11969 (rs6000_invalid_within_doloop): Remove.
11971 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
11972 unused variable from tablejump_p call.
11974 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
11975 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
11976 (INSN_DELETED_P): Likewise.
11977 (emit_jump_table_data): New prototype.
11978 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
11979 after 4th as unused.
11980 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
11981 * sched-vis.c (print_insn): Likewise.
11982 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
11983 insn for compatibility with back ends that use next_active_insn to
11984 identify jump table data.
11985 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
11986 (remove_insn): Likewise.
11987 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
11989 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
11990 (emit_jump_table_data): New function.
11992 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
11993 basic block, a JUMP_TABLE_DATA never is.
11994 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
11995 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
11996 off from code handling real insns.
11997 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
11998 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
11999 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12000 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12001 is not a NONDEBUG_INSN_P.
12002 * ira-costs.c (scan_one_insn): Likewise.
12003 * jump.c (mark_all_labels): Likewise.
12004 (mark_jump_label_1): Likewise.
12005 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12006 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12008 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12009 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12010 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12011 JUMP_TABLE_DATA_P insns.
12012 (calculate_elim_costs_all_insns): Likewise.
12013 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12014 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12015 (delete_output_reload): Code style fixups.
12016 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12017 insn flags on this non-insn.
12018 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12019 as scheduling barriers, for pre-change compatibility.
12020 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12021 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12023 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12024 redundant JUMP_TABLE_DATA_P test.
12025 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12026 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12027 (frv_for_each_packet): Likewise.
12028 * config/i386/i386.c (min_insn_size): Likewise.
12029 (ix86_avoid_jump_mispredicts): Likewise.
12030 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12031 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12032 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12033 (mips16_insn_length): Robustify.
12034 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12035 (mips16_split_long_branches): Likewise.
12036 * config/pa/pa.c (pa_combine_instructions): Likewise.
12037 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12038 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12039 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12040 as contributing to pool range lengths.
12041 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12042 Remove redundant JUMP_TABLE_DATA_P test.
12043 (sh_loop_align): Likewise.
12044 (split_branches): Likewise.
12045 (sh_insn_length_adjustment): Likewise.
12046 * config/spu/spu.c (get_branch_target): Likewise.
12048 2013-03-29 Jan Hubicka <jh@suse.cz>
12050 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12051 gcov streaming; stream hot bb threshold to ltrans.
12052 * predict.c (get_hot_bb_threshold): Break out from ....
12053 (maybe_hot_count_p): ... here.
12054 (set_hot_bb_threshold): New function.
12055 * lto-section-in.c (lto_section_name): Add profile.
12056 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12057 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12058 and data-streamer.h
12059 (histogram_entry): New structure.
12060 (histogram, histogram_pool): New global vars.
12061 (histogram_hash): New structure.
12062 (histogram_hash::hash): New method.
12063 (histogram_hash::equal): Likewise.
12064 (account_time_size): New function.
12065 (cmp_counts): New function.
12066 (dump_histogram): New function.
12067 (ipa_profile_generate_summary): New function.
12068 (ipa_profile_write_summary): New function.
12069 (ipa_profile_read_summary): New function.
12070 (ipa_profile): Decide on threshold.
12071 (pass_ipa_profile): Add ipa_profile_write_summary and
12072 ipa_profile_read_summary.
12073 * Makefile.in (ipa.o): Update dependencies.
12074 * lto-streamer.h (LTO_section_ipa_profile): New section.
12076 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12078 * tree.h (VAR_P): New.
12080 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12083 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12085 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12087 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12088 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12089 (control_flow_insn_p): Likewise.
12090 * cfgrtl.c (duplicate_insn_chain): Likewise.
12091 * final.c (get_attr_length_1): Likewise.
12092 (shorten_branches): Likewise.
12093 (final_scan_insn): Likewise.
12094 * function.c (instantiate_virtual_regs): Likewise.
12095 * gcse.c (insert_insn_end_basic_block): Likewise.
12096 * ira-costs.c (scan_one_insn): Likewise.
12097 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12098 * lra.c (check_rtl): Likewise.
12099 * reload1.c (elimination_costs_in_insn): Likewise.
12100 * reorg.c (follow_jumps): Likewise.
12102 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12103 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12104 (thumb_far_jump_used_p): Likewise.
12105 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12106 (workaround_speculation): Likewise.
12107 (add_sched_insns_for_speculation): Likewise.
12108 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
12109 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12110 (frv_for_each_packet): Likewise.
12111 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
12112 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
12113 (final_emit_insn_group_barriers): Likewise.
12114 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12115 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12116 (mips16_insn_length): Likewise.
12117 * config/pa/pa.c (pa_reorg): Likewise.
12118 (pa_combine_instructions): Likewise.
12119 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
12120 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
12121 (sh_reorg): Likewise.
12122 (split_branches): Likewise.
12123 * config/spu/spu.c (get_branch_target): Likewise.
12125 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
12128 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
12130 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
12131 Fix declaration name.
12133 2013-03-28 Lawrence Crowl <crowl@google.com>
12135 * graphds.h (struct graph.indicies): Remove unused.
12136 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
12137 (SCOP_ORIGINAL_PDDRS): Remove unused.
12138 * sese.h (extern insert_loop_close_phis): Removed unused.
12139 (extern insert_guard_phis): Removed unused.
12140 (extern ivtype_map_elt_info): Removed unused.
12141 (new_ivtype_map_elt): Removed unused.
12142 * sese.c (ivtype_map_elt_info): Removed unused.
12144 2013-03-28 Lawrence Crowl <crowl@google.com>
12146 * Makefile.in: Add several missing include dependences.
12148 (test-dump.o): New. This object is not added to any executable,
12149 but is present for ad-hoc testing.
12151 (debug (const bitmap_head_def &)): New.
12152 (debug (const bitmap_head_def *)): New.
12154 (extern debug (const bitmap_head_def &)): New.
12155 (extern debug (const bitmap_head_def *)): New.
12157 (debug (edge_def &)): New.
12158 (debug (edge_def *)): New.
12160 (debug (basic_block_def &)): New.
12161 (debug (basic_block_def *)): New.
12163 (dump_node (const_tree, int, FILE *)): Correct source file.
12165 (debug (die_struct &)): New.
12166 (debug (die_struct *)): New.
12168 (extern debug (die_struct &)): New.
12169 (extern debug (die_struct *)): New.
12170 * gimple-pretty-print.c
12171 (debug (gimple_statement_d &)): New.
12172 (debug (gimple_statement_d *)): New.
12173 * gimple-pretty-print.h
12174 (extern debug (gimple_statement_d &)): New.
12175 (extern debug (gimple_statement_d *)): New.
12177 (debug (ira_allocno_copy &)): New.
12178 (debug (ira_allocno_copy *)): New.
12179 (debug (ira_allocno &)): New.
12180 (debug (ira_allocno *)): New.
12182 (extern debug (ira_allocno_copy &)): New.
12183 (extern debug (ira_allocno_copy *)): New.
12184 (extern debug (ira_allocno &)): New.
12185 (extern debug (ira_allocno *)): New.
12187 (debug (live_range &)): New.
12188 (debug (live_range *)): New.
12190 (debug (lra_live_range &)): New.
12191 (debug (lra_live_range *)): New.
12193 (debug (lra_live_range &)): New.
12194 (debug (lra_live_range *)): New.
12196 (debug (omega_pb_d &)): New.
12197 (debug (omega_pb_d *)): New.
12199 (extern debug (omega_pb_d &)): New.
12200 (extern debug (omega_pb_d *)): New.
12202 (debug (const rtx_def &)): New.
12203 (debug (const rtx_def *)): New.
12205 (debug_tree (tree): Move within file.
12206 (debug_raw (const tree_node &)): New.
12207 (debug_raw (const tree_node *)): New.
12208 (dump_tree_via_hooks (const tree_node *, int)): New.
12209 (debug (const tree_node &)): New.
12210 (debug (const tree_node *)): New.
12211 (debug_verbose (const tree_node &)): New.
12212 (debug_verbose (const tree_node *)): New.
12213 (debug_head (const tree_node &)): New.
12214 (debug_head (const tree_node *)): New.
12215 (debug_body (const tree_node &)): New.
12216 (debug_body (const tree_node *)): New.
12217 (debug_vec_tree (tree): Move and reimplement in terms of dump.
12218 (debug (vec<tree, va_gc> &)): New.
12219 (debug (vec<tree, va_gc> *)): New.
12221 (extern debug (const rtx_def &)): New.
12222 (extern debug (const rtx_def *)): New.
12224 (debug_raw (simple_bitmap_def &)): New.
12225 (debug_raw (simple_bitmap_def *)): New.
12226 (debug (simple_bitmap_def &)): New.
12227 (debug (simple_bitmap_def *)): New.
12229 (extern debug (simple_bitmap_def &)): New.
12230 (extern debug (simple_bitmap_def *)): New.
12231 (extern debug_raw (simple_bitmap_def &)): New.
12232 (extern debug_raw (simple_bitmap_def *)): New.
12234 (debug (vinsn_def &)): New.
12235 (debug (vinsn_def *)): New.
12236 (debug_verbose (vinsn_def &)): New.
12237 (debug_verbose (vinsn_def *)): New.
12238 (debug (expr_def &)): New.
12239 (debug (expr_def *)): New.
12240 (debug_verbose (expr_def &)): New.
12241 (debug_verbose (expr_def *)): New.
12242 (debug (vec<rtx> &)): New.
12243 (debug (vec<rtx> *)): New.
12245 (extern debug (vinsn_def &)): New.
12246 (extern debug (vinsn_def *)): New.
12247 (extern debug_verbose (vinsn_def &)): New.
12248 (extern debug_verbose (vinsn_def *)): New.
12249 (extern debug (expr_def &)): New.
12250 (extern debug (expr_def *)): New.
12251 (extern debug_verbose (expr_def &)): New.
12252 (extern debug_verbose (expr_def *)): New.
12253 (extern debug (vec<rtx> &)): New.
12254 (extern debug (vec<rtx> *)): New.
12256 (_list_iter_cond_expr): Make inline instead of static.
12258 (debug (sreal &)): New.
12259 (debug (sreal *)): New.
12261 (extern debug (sreal &)): New.
12262 (extern debug (sreal *)): New.
12264 (extern debug_raw (const tree_node &)): New.
12265 (extern debug_raw (const tree_node *)): New.
12266 (extern debug (const tree_node &)): New.
12267 (extern debug (const tree_node *)): New.
12268 (extern debug_verbose (const tree_node &)): New.
12269 (extern debug_verbose (const tree_node *)): New.
12270 (extern debug_head (const tree_node &)): New.
12271 (extern debug_head (const tree_node *)): New.
12272 (extern debug_body (const tree_node &)): New.
12273 (extern debug_body (const tree_node *)): New.
12274 (extern debug (vec<tree, va_gc> &)): New.
12275 (extern debug (vec<tree, va_gc> *)): New.
12277 (debug (struct loop &)): New.
12278 (debug (struct loop *)): New.
12279 (debug_verbose (struct loop &)): New.
12280 (debug_verbose (struct loop *)): New.
12281 * tree-dump.c: Add header dependence.
12283 (extern debug (struct loop &)): New.
12284 (extern debug (struct loop *)): New.
12285 (extern debug_verbose (struct loop &)): New.
12286 (extern debug_verbose (struct loop *)): New.
12288 (debug (data_reference &)): New.
12289 (debug (data_reference *)): New.
12290 (debug (vec<data_reference_p> &)): New.
12291 (debug (vec<data_reference_p> *)): New.
12292 (debug (vec<ddr_p> &)): New.
12293 (debug (vec<ddr_p> *)): New.
12295 (extern debug (data_reference &)): New.
12296 (extern debug (data_reference *)): New.
12297 (extern debug (vec<data_reference_p> &)): New.
12298 (extern debug (vec<data_reference_p> *)): New.
12299 (extern debug (vec<ddr_p> &)): New.
12300 (extern debug (vec<ddr_p> *)): New.
12302 (debug (pt_solution &)): New.
12303 (debug (pt_solution *)): New.
12305 (extern debug (pt_solution &)): New.
12306 (extern debug (pt_solution *)): New.
12308 (debug (_var_map &)): New.
12309 (debug (_var_map *)): New.
12310 (debug (tree_live_info_d &)): New.
12311 (debug (tree_live_info_d *)): New.
12313 (extern debug (_var_map &)): New.
12314 (extern debug (_var_map *)): New.
12315 (extern debug (tree_live_info_d &)): New.
12316 (extern debug (tree_live_info_d *)): New.
12318 2013-03-28 Jan Hubicka <jh@suse.cz>
12320 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
12322 2013-03-28 Ian Bolton <ian.bolton@arm.com>
12324 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
12325 record only when desired or required.
12327 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
12329 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
12330 *vec_extractv2di_1_rex64. Use x64 isa attribute.
12332 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12334 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
12335 (*andsi3_compare0_uxtw): New pattern.
12336 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
12337 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
12339 2013-03-28 Jan Hubicka <jh@suse.cz>
12341 * data-streamer-in.c (streamer_read_gcov_count): New function.
12342 * gimple-streamer-out.c: Include value-prof.h.
12343 (output_gimple_stmt): Output histogram.
12344 (output_bb): Use streamer_write_gcov_count.
12345 * value-prof.c: Include data-streamer.h
12346 (dump_histogram_value): Add HIST_TYPE_MAX.
12347 (stream_out_histogram_value): New function.
12348 (stream_in_histogram_value): New function.
12349 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
12350 (stream_out_histogram_value, stream_in_histogram_value): Declare.
12351 * data-streamer-out.c (streamer_write_gcov_count): New function.
12352 (streamer_write_gcov_count_stream): New function.
12353 * lto-cgraph.c (lto_output_edge): Update counter streaming.
12354 (lto_output_node): Likewise.
12355 (input_node, input_edge): Likewise.
12356 * lto-streamer-out.c (output_cfg): Update streaming.
12357 * lto-streamer-in.c (input_cfg): Likewise.
12358 * data-streamer.h (streamer_write_gcov_count,
12359 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
12360 * gimple-streamer-in.c: Include value-prof.h
12361 (input_gimple_stmt): Input histograms.
12362 (input_bb): Update profile streaming.
12364 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12366 * genmodes.c (emit_max_int): New function.
12367 (emit_insn_modes_h): Added call to emit_max_function.
12368 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12370 * machmode.def: Fixed comment.
12372 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12374 * combine.c (try_combine): Removed useless assert.
12375 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12377 2013-03-28 Marek Polacek <polacek@redhat.com>
12378 Richard Biener <rguenther@suse.de>
12380 PR tree-optimization/56695
12381 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12382 build signed result of a vector comparison.
12383 * tree-cfg.c (verify_gimple_comparison): Check that a result
12384 of a vector comparison has signed type.
12386 2013-03-28 Richard Biener <rguenther@suse.de>
12388 PR tree-optimization/37021
12389 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12390 do not restrict gaps between groups.
12391 * tree-vect-stmts.c (vectorizable_load): Properly account for
12392 a gap between groups.
12394 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12396 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12397 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12400 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12402 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12403 * doc/extend.texi (Named Address Spaces): Ditto.
12404 (Variable Attributes): Ditto.
12406 2013-03-27 Kai Tietz <ktietz@redhat.com>
12408 * config.build: Add support for cygwin x64 target.
12409 * config.gcc: Likewise.
12410 * config.host: Likewise.
12411 * configure.ac: Likewise
12412 * configure: Regenerated.
12414 2013-03-27 Kai Tietz <ktietz@redhat.com>
12416 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12417 * config/i386/t-cygwin-w64: New file.
12418 * config/i386/cygwin-w64.h: New file.
12419 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12420 and add support for x64-cygwin target.
12421 (CPP_SPEC): Likewise.
12422 (CXX_WRAP_SPEC_LIST): Undefine before define.
12423 (LIBGCJ_SONAME): Use 15 as version.
12425 2013-03-27 Richard Biener <rguenther@suse.de>
12427 PR tree-optimization/56716
12428 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12429 dumping for ref nodes.
12431 2013-03-27 Martin Jambor <mjambor@suse.cz>
12433 PR tree-optimization/55334
12434 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12435 restricted pointers to arrays.
12437 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12439 * Makefile.in (.SUFFIXES): Add .cc.
12440 (.c.o): Apply same recipe for implicit rule .cc.o.
12442 2013-03-27 Richard Biener <rguenther@suse.de>
12444 PR tree-optimization/37021
12445 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12446 REALPART/IMAGPART_EXPRs around the supported refs.
12447 * tree-ssa-structalias.c (find_func_aliases): Assume that
12448 floating-point values are not used to transfer pointers.
12450 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12452 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12453 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12454 * doc/tm.texi: Regenerate.
12455 * targhooks.h (default_has_ifunc_p): New.
12456 * targhooks.c (default_has_ifunc_p): Ditto.
12457 * config/linux-protos.h: New file.
12458 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12459 this hook for linux which disables support of indirect functions in
12461 * config/linux-android.c: New file.
12462 * config/t-linux-android.c: Ditto.
12463 * config.gcc: Added new object file linux-android.o.
12464 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
12465 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12466 * varasm.c (do_assemble_alias): Likewise.
12467 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12468 doesn't support indirect functions.
12469 * configure: Regenerate.
12471 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12474 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12475 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12478 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12480 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12482 2013-03-27 Terry Guo <terry.guo@arm.com>
12484 * config/arm/arm-cores.def: Added core cortex-r7.
12485 * config/arm/arm-tune.md: Regenerated.
12486 * config/arm/arm-tables.opt: Regenerated.
12487 * doc/invoke.texi: Added entry for core cortex-r7.
12489 2013-03-27 Walter Lee <walt@tilera.com>
12491 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12492 double-decrement of next_scratch_regno.
12494 2013-03-27 Walter Lee <walt@tilera.com>
12496 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12498 (insn_v1mulus): Ditto.
12499 (insn_v2muls): Ditto.
12501 2013-03-27 Walter Lee <walt@tilera.com>
12503 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12504 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12506 2013-03-27 Walter Lee <walt@tilera.com>
12508 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
12509 (*sibcall_value): Ditto.
12511 2013-03-27 Walter Lee <walt@tilera.com>
12513 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12514 (insn_mnz_v8qi): ... this ...
12515 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12517 (insn_v<n>mnz): Replaced by ...
12518 (insn_v1mnz): ... this ...
12519 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12521 (insn_mz_<mode>): Replaced by ...
12522 (insn_mz_v8qi): ... this ...
12523 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12525 (insn_v<n>mz): Replaced by ...
12526 (insn_v1mz): ... this ...
12527 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12530 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12532 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12534 2013-03-26 Roland McGrath <mcgrathr@google.com>
12536 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12537 than fprintf with a non-constant, non-format string.
12539 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12541 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12542 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12543 operand 0 predicate.
12544 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12545 attribute. Use general_x64nomem_operand as operand 1 predicate.
12546 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12547 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12548 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12549 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12550 (mov<mode>_insv_1): Remove expander. Merge insn with
12551 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12552 Use general_x64nomem_operand as operand 1 predicate.
12553 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12554 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12555 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12556 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12557 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12558 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12559 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12560 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12561 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12562 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12563 (general_x64nomem_operand): Ditto.
12565 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
12567 * config/rtems.opt: Add -pthread option.
12569 2013-03-26 Richard Biener <rguenther@suse.de>
12571 * alias.c (find_base_term): Avoid redundant and not used recursion.
12572 (base_alias_check): Get the initial base term from the caller.
12573 (true_dependence_1): Compute and pass base terms to base_alias_check.
12574 (write_dependence_p): Likewise.
12575 (may_alias_p): Likewise.
12577 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12579 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12580 PC-relative load in SI modes and above only.
12582 2013-03-26 Xinyu Qi <xyqi@marvell.com>
12584 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
12585 * config/arm/iwmmxt.md (WCGR0): Update.
12586 (WCGR1, WCGR2, WCGR3): Likewise.
12588 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12590 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12591 Use x64 and nox64 isa attributes.
12593 2013-03-26 Richard Biener <rguenther@suse.de>
12595 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12596 alignment computations and rely on get_object_alignment_1
12597 for the !TYPE_P case.
12598 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12600 2013-03-26 Walter Lee <walt@tilera.com>
12602 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12603 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12605 2013-03-25 Jeff Law <law@redhat.com>
12607 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12608 check for INTEGRAL_TYPE_P that was missing due to checking in
12609 wrong version of prior patch.
12611 2013-03-25 Walter Lee <walt@tilera.com>
12613 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12614 TILEGX_INSN_SHUFFLEBYTES1.
12615 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12617 (tilegx_builtins): Ditto.
12618 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12620 2013-03-25 Walter Lee <walt@tilera.com>
12622 * config/tilegx/tilegx.md (floatsisf2): New pattern.
12623 (floatunssisf2): New pattern.
12624 (floatsidf2): New pattern.
12625 (floatunssidf2): New pattern.
12627 2013-03-25 Walter Lee <walt@tilera.com>
12629 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12630 tests for constraint J, K, N, P.
12632 2013-03-25 Walter Lee <walt@tilera.com>
12634 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12635 Use indirect/pcrel encoding.
12636 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12639 2013-03-25 Steve Ellcey <sellcey@mips.com>
12641 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12642 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12643 * config/mips/mips.c (mips_option_override): Set IMADD default.
12644 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12645 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12646 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12647 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12648 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12650 2013-03-25 Jeff Law <law@redhat.com>
12652 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12653 slightly to avoid creating and folding useless trees. Simplify
12654 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12656 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
12658 * config/i386/i386.md (*zero_extendsidi2): Merge with
12659 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12660 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12661 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12662 targets. Clarify comment.
12664 2013-03-25 Martin Jambor <mjambor@suse.cz>
12666 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12667 pass-through jump functions differently.
12668 (ipa_read_jump_function): Likewise. Also use setter functions to set
12671 2013-03-25 Martin Jambor <mjambor@suse.cz>
12673 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12674 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12676 (ipa_get_indirect_edge_target): New function.
12677 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12678 ipa_get_indirect_edge_target. Update all callers.
12679 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12680 ipa_get_indirect_edge_target_1 instead of calling
12681 ipa_get_indirect_edge_target.
12682 (create_specialized_node): Pass aggvlas to
12683 ipcp_discover_new_direct_edges.
12685 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12687 * config/arm/arm.md (f_sels, f_seld): New types.
12688 (*cmov<mode>): New pattern.
12689 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12692 2013-03-25 Kai Tietz <ktietz@redhat.com>
12694 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12695 POSIX-printf for mingw-hosted builds.
12697 2013-03-25 Richard Biener <rguenther@suse.de>
12699 PR middle-end/56694
12700 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12701 must-not-throw stmt location.
12703 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12705 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12706 Emit load-acquire versions when acq is true.
12707 (arm_emit_store_exclusive): Add rel parameter.
12708 Emit store-release versions when rel is true.
12709 (arm_split_compare_and_swap): Use acquire-release instructions
12711 of barriers when appropriate.
12712 (arm_split_atomic_op): Likewise.
12713 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12714 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12715 (VUNSPEC_SLX): Likewise.
12716 (VUNSPEC_LDA): Likewise.
12717 (VUNSPEC_STL): Likewise.
12718 * config/arm/sync.md (atomic_load<mode>): New pattern.
12719 (atomic_store<mode>): Likewise.
12720 (arm_load_acquire_exclusive<mode>): Likewise.
12721 (arm_load_acquire_exclusivesi): Likewise.
12722 (arm_load_acquire_exclusivedi): Likewise.
12723 (arm_store_release_exclusive<mode>): Likewise.
12725 2013-03-25 Catherine Moore <clm@codesourcery.com>
12727 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12728 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12729 * config/mip/predicates.md (lwsp_swsp_operand,
12730 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12731 sb16_operand, db4_operand, db7_operand, ib3_operand,
12732 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12733 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12734 andi16_operand): New predicates.
12735 * config/mips/mips.md (compression): New attribute.
12736 (enabled): New attribute.
12737 (length): Consider compression in computing length.
12738 (shift_compression): New code attribute.
12739 (*add<mode>3): New operands. Record compression.
12740 (sub<mode>3): Likewise.
12741 (one_cmpl<mode>2): Likewise.
12742 (*and<mode>3): Likewise.
12743 (*ior<mode>3): Likewise.
12744 (unnamed pattern for xor): Likewise.
12745 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12746 (*<optab><mode>3): Likewise.
12747 (*mov<mode>_internal: Likewise.
12748 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12749 (mips_unsigned_immediate_p): New.
12750 (umips_lwsp_swsp_address_p): New.
12751 (m16_based_address_p): New.
12752 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12753 (mips_unsigned_immediate_p): New prototype.
12754 (lwsp_swsp_address_p): New prototype.
12755 (m16_based_address_p): New prototype.
12756 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12757 (mips_signed_immediate_p): New function.
12758 (m16_based_address_p): New function.
12759 (lwsp_swsp_address_p): New function.
12760 (mips_print_operand_punctuation): Recognize short delay slot insns
12761 for microMIPS.add<mode>3"
12763 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12766 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12767 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12769 2013-03-25 Richard Biener <rguenther@suse.de>
12771 PR tree-optimization/56689
12772 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12775 2013-03-25 Richard Biener <rguenther@suse.de>
12777 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12779 (memory_references): Likewise.
12780 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12781 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12782 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12783 (gather_mem_refs_in_loops): Fold into ...
12784 (analyze_memory_references): ... this. Move initialization
12785 to tree_ssa_lim_initialize.
12786 (fill_always_executed_in): Rename to ...
12787 (fill_always_executed_in_1): ... this.
12788 (fill_always_executed_in): Move contains_call computation to
12789 this new function from ...
12790 (tree_ssa_lim_initialize): ... here.
12791 (tree_ssa_lim): Call fill_always_executed_in.
12793 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12795 * postreload.c (reload_combine): Fix code detecting returns.
12797 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12799 * function.c (emit_use_return_register_into_block): On cc0 targets,
12800 do not emit the sequence between cc0 setter and user.
12802 2013-03-25 Kai Tietz <ktietz@redhat.com>
12804 * config/i386/predicates.md (local_symbolic_operand): Interpret
12805 dll-imported symbols as none-local.
12807 2013-03-25 Richard Biener <rguenther@suse.de>
12809 * tree-ssa-loop-im.c (struct depend): Remove.
12810 (struct lim_aux_data): Make depends a vec of gimples.
12811 (free_lim_aux_data): Adjust.
12812 (add_dependency): Likewise.
12813 (set_level): Likewise.
12815 2013-03-25 Richard Biener <rguenther@suse.de>
12817 PR middle-end/56434
12818 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12819 the pointer returned by calls with ECF_MALLOC set.
12821 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12823 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12825 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12827 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
12828 using MMXMODE mode iterator.
12829 (*move<mode>_internal): Merge with *movv2sf_internal and
12830 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12832 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12834 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12835 (record_last_mem_set_info): Likewise.
12837 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12838 of XNEWVEC followed by memset.
12839 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12841 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12843 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12844 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12845 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12846 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12847 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12848 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12849 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12850 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12851 BARRIER_P instead of GET_CODE.
12853 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12855 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12856 inaccuracy in the probing code.
12858 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12859 (ctrapdi4): Likewise.
12861 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12863 * calls.c (expand_call): Add missing guard to code handling return
12864 of non-BLKmode structures in MSB.
12865 * function.c (expand_function_end): Likewise.
12867 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12869 * combine.c (try_combine): Adjust comment. Do not add the set of
12870 insn #0 if the destination indirectly is set or dies in insn #2.
12871 Tidy up code to distribute a new note.
12873 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12875 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12876 also for alternatives 16 and 17.
12878 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12880 * config/i386/sse.md (*mov<mode>_internal): Merge with
12881 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12882 Emit insn template depending on type attribute. Use
12883 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12884 movd instead of movq mnemonic for interunit moves. Rewrite mode
12885 attribute calculation. Remove unit attribute calculation.
12886 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12887 Set prefix_data16 attribute for DImode ssemov types.
12888 Use Ym instead of y for SSE-MMX conversion alternatives.
12889 Reorder operand constraints.
12891 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12893 * df.h (df_insn_delete): Adjust prototype.
12894 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12895 and let it decide whether mark the basic block dirty.
12896 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12897 * df-scan.c (df_insn_info_delete): New helper function, split
12898 off from df_insn_delete.
12899 (df_scan_free_bb_info): Use it.
12900 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12902 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12903 that the insn is actually an insn and it has a non-NULL basic block.
12904 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12906 2013-03-22 Richard Biener <rguenther@suse.de>
12908 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12910 (mem_ref_alloc): Do not allocate them.
12911 (refs_independent_p): Do not query or maintain a cache.
12913 2013-03-22 Richard Biener <rguenther@suse.de>
12915 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12916 (gather_mem_refs_in_loops): Do not compute it.
12917 (analyze_memory_references): Do not allocate it.
12918 (tree_ssa_lim_finalize): Do not free it.
12919 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12921 2013-03-22 Richard Biener <rguenther@suse.de>
12923 * is-a.h (as_a): Use gcc_checking_assert.
12925 2013-03-22 Ian Bolton <ian.bolton@arm.com>
12927 * config/aarch64/aarch64.c (aarch64_print_operand): New
12928 format specifier for printing a constant in hex.
12929 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12930 format specifier for printing second operand.
12932 2013-03-22 Richard Biener <rguenther@suse.de>
12934 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12936 (gather_mem_refs_in_loops): Perform store accumulation here.
12937 (create_vop_ref_mapping_loop): Remove.
12938 (create_vop_ref_mapping): Likewise.
12939 (analyze_memory_references): Initialize refs_stored_in_loop.
12940 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
12941 (record_indep_loop): Remove.
12942 (record_dep_loop): New function.
12943 (ref_indep_loop_p_1): Adjust to only walk over references
12944 in the loop, not its subloops.
12945 (ref_indep_loop_p): Rename to ...
12946 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12947 maintaining a more fine-grained cache.
12948 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12949 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12951 2013-03-22 Richard Biener <rguenther@suse.de>
12953 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12954 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12955 aggregate mem_ref_loc.
12956 (free_mem_ref_locs): Inline into ...
12957 (memref_free): ... this and adjust.
12958 (mem_ref_alloc): Adjust.
12959 (mem_ref_locs_alloc): Remove.
12960 (record_mem_ref_loc): Adjust.
12961 (get_all_locs_in_loop): Rewrite into ...
12962 (for_all_locs_in_loop): ... this iterator.
12963 (rewrite_mem_ref_loc): New functor.
12964 (rewrite_mem_refs): Use for_all_locs_in_loop.
12965 (sm_set_flag_if_changed): New functor.
12966 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
12967 (ref_always_accessed): New functor.
12968 (ref_always_accessed_p): Use for_all_locs_in_loop.
12970 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
12972 * tree-pass.h (PROP_gimple_lvec): New.
12973 * passes.c (dump_properties): Handle PROP_gimple_lvec.
12974 (init_optimization_passes): Move pass_lower_vector.
12975 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
12977 (pass_lower_vector): Provide PROP_gimple_lvec.
12978 (pass_lower_vector_ssa): Likewise.
12979 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
12981 2013-03-21 Mark Wielaard <mjw@redhat.com>
12983 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
12985 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
12987 * config/i386/i386.md (*movdi_internal): Disparage slightly
12988 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
12989 conversion alternatives.
12991 2013-03-21 Jakub Jelinek <jakub@redhat.com>
12993 PR middle-end/48087
12994 * diagnostic.def (DK_WERROR): New kind.
12995 * diagnostic.h (werrorcount): Define.
12996 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
12997 promoted to DK_ERROR, increment DK_WERROR counter instead of
12999 * toplev.c (toplev_main): Call print_ignored_options even if
13000 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13001 even if just werrorcount is non-zero.
13004 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13006 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13007 (string_cst_pool_decl): New function.
13008 (optimize_one_addr_into_implicit_ptr): New function.
13009 (resolve_addr_in_expr): Optimize DWARF location expression
13010 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13011 which doesn't live in memory, but has DW_AT_location or
13012 DW_AT_const_value, or refers to a string literal, into
13013 DW_OP_GNU_implicit_pointer.
13014 (optimize_location_into_implicit_ptr): New function.
13015 (resolve_addr): If removing DW_AT_location of a variable because
13016 it was DW_OP_addr of address of the variable, but the variable doesn't
13017 live in memory, try to emit const value attribute for the initializer.
13019 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13021 * tree.h (VECTOR_TYPE_P): New macro.
13022 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13023 TYPE_MODE): Use it.
13024 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13025 VEC_COND_EXPR cannot be lvalues.
13026 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13028 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13030 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13031 Restrict the transformation to equal modes.
13033 2013-03-21 Richard Biener <rguenther@suse.de>
13035 PR tree-optimization/39326
13036 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13037 (MEM_ANALYZABLE): Adjust.
13038 (record_mem_ref_loc): Move bitmap ops ...
13039 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13040 unanalyzable refs, do not record locations for it.
13041 (analyze_memory_references): Allocate ref zero as shared
13043 (refs_independent_p): Do not test for unanalyzed mems here.
13044 (ref_indep_loop_p_1): Special-case disambiguation against
13045 the unanalyzed ref.
13046 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13048 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13050 * config/arm/arm-protos.h (tune_params): Add
13051 prefer_neon_for_64bits field.
13052 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13053 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13054 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13055 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13056 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13057 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13058 (arm_option_override): Handle -mneon-for-64bits new option.
13059 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13060 (prefer_neon_for_64bits): Declare new variable.
13061 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13062 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13063 (arch_enabled): Handle new arch types. Remove support for onlya8
13065 (one_cmpldi2): Use new arch names.
13066 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13067 * config/arm/arm.opt (mneon-for-64bits): Add option.
13068 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13069 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13070 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13072 * doc/invoke.texi (-mneon-for-64bits): Document.
13074 2013-03-21 Richard Biener <rguenther@suse.de>
13076 PR tree-optimization/39326
13077 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13078 (sort_bbs_in_loop_postorder_cmp): New function.
13079 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13081 2013-03-21 Richard Biener <rguenther@suse.de>
13083 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13084 (vect_insert_into_interleaving_chain): Likewise.
13085 (vect_drs_dependent_in_basic_block): Inline ...
13086 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13088 (vect_analyze_data_ref_dependence): ... here. Simplify.
13089 (vect_check_interleaving): Simplify.
13090 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13091 (vect_slp_analyze_data_ref_dependences): ... this new function.
13092 (dr_group_sort_cmp): New function.
13093 (vect_analyze_data_ref_accesses): Compute data-reference groups
13094 here instead of in vect_analyze_data_ref_dependence. Use
13095 a more efficient algorithm.
13096 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13097 vect_slp_analyze_data_ref_dependences. Call
13098 vect_analyze_data_ref_accesses earlier.
13099 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13100 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13101 (vect_slp_analyze_data_ref_dependences): New prototype.
13103 2013-03-21 Richard Biener <rguenther@suse.de>
13105 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13106 ref is stored in the loop.
13107 (find_refs_for_sm): Walk only over all stores.
13108 (store_motion_loop): Allocate from lim_bitmap_obstack.
13109 (store_motion): Likewise.
13111 2013-03-21 Richard Biener <rguenther@suse.de>
13113 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
13114 Update virtual SSA form.
13116 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13118 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
13119 * configure: Regenerate.
13120 * config.in: Regenerate.
13121 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
13122 if !HAVE_LD_EH_FRAME_CIEV3.
13124 2013-03-21 Richard Biener <rguenther@suse.de>
13126 * tree-cfg.c (verify_expr_no_block): New function.
13127 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
13128 nor DECL_VALUE_EXPR have locations with associated blocks.
13129 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
13130 (clear_unused_block_pointer): Remove code dealing with
13131 blocks in DECL_DEBUG_EXPR locations.
13133 2013-03-21 Richard Biener <rguenther@suse.de>
13135 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
13136 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
13137 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
13138 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
13139 instead of DECL_DEBUG_EXPR_IS_FROM.
13140 * gimplify.c (gimplify_modify_expr): Likewise.
13141 * tree-cfg.c (verify_expr_location_1): Likewise.
13142 * tree-complex.c (create_one_component_var): Likewise.
13143 * tree-sra.c (create_access_replacement): Likewise.
13144 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
13145 (clear_unused_block_pointer): Likewise.
13146 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
13147 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
13148 * var-tracking.c (var_debug_decl): Likewise.
13149 (track_expr_p): Likewise.
13150 * tree-inline.c (add_local_variables): Likewise. Set
13151 DECL_HAS_DEBUG_EXPR_P after copying it.
13152 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
13153 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
13155 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13158 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
13159 * configure: Regenerate.
13160 * config.in: Regenerate.
13161 * config/i386/i386.md (*movdf_internal): Use
13162 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13163 movd instead of movq mnemonic for interunit moves.
13164 (*movdi_internal): Ditto.
13166 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13168 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
13169 (abd<mode>_3): New pattern.
13170 (aba<mode>_3): New pattern.
13171 (fabd<mode>_3): New pattern.
13173 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13175 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
13176 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
13177 occurrence of REGISTER_PREFIX as its empty string.
13179 2013-03-20 Jeff Law <law@redhat.com>
13181 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
13182 addititional equivalences for equality comparisons between an SSA_NAME
13183 and a constant where the SSA_NAME was set from a widening conversion.
13185 2013-03-20 Walter Lee <walt@tilera.com>
13187 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
13189 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13191 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
13192 depending on type attribute.
13193 (*movti_internal): Ditto.
13194 (*movtf_internal): Ditto.
13195 (*movxf_internal): Ditto.
13196 (*movdf_internal): Ditto.
13197 (*movsf_internal): Ditto.
13199 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13201 * config/i386/i386.md (*movti_internal): Set prefix attribute to
13202 maybe_vex for sselog1 and ssemov types.
13203 (*movdi_internal): Reorder operand constraints.
13204 (*movsi_internal): Ditto. Set prefix attribute to
13205 maybe_vex for sselog1 and ssemov types.
13206 (*movtf_internal): Set prefix attribute to maybe_vex
13207 for sselog1 and ssemov types.
13208 (*movdf_internal): Ditto. Set prefix_data16 attribute for
13209 DImode ssemov types. Reorder operand constraints.
13210 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
13211 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
13212 attribute for SImode ssemov types. Reorder operand constraints.
13214 2013-03-20 Martin Jambor <mjambor@suse.cz>
13216 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
13217 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
13219 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
13221 * config/rs6000/predicates.md (indexed_address, update_address_mem
13222 update_indexed_address_mem): New predicates.
13223 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
13224 attribute for load/store instructions.
13225 * config/rs6000/dfp.md (movsd_store): Likewise.
13226 (movsd_load): Likewise.
13227 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
13228 (unnamed HI->DI extend define_insn): Likewise.
13229 (unnamed SI->DI extend define_insn): Likewise.
13230 (unnamed QI->SI extend define_insn): Likewise.
13231 (unnamed QI->HI extend define_insn): Likewise.
13232 (unnamed HI->SI extend define_insn): Likewise.
13233 (unnamed HI->SI extend define_insn): Likewise.
13234 (extendsfdf2_fpr): Likewise.
13235 (movsi_internal1): Likewise.
13236 (movsi_internal1_single): Likewise.
13237 (movhi_internal): Likewise.
13238 (movqi_internal): Likewise.
13239 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
13240 attribute for load/store instructions.
13241 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
13243 (mov<mode>_softfloat): Likewise.
13244 (mov<mode>_hardfloat32): Likewise.
13245 (mov<mode>_hardfloat64): Likewise.
13246 (mov<mode>_softfloat64): Likewise.
13247 (movdi_internal32): Likewise.
13248 (movdi_internal64): Likewise.
13249 (probe_stack_<mode>): Likewise.
13251 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
13253 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
13254 floating point, and decimal floating point to reload iterator.
13256 * config/rs6000/constraints.md (wl constraint): New constraints to
13257 return FLOAT_REGS if certain options are used to reduce the number
13258 of separate patterns that exist in the file.
13259 (wx constraint): Likewise.
13260 (wz constraint): Likewise.
13262 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13263 -mdebug=reg, print wg, wl, wx, and wz constraints.
13264 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13265 Initialize the reload functions for 64-bit binary/decimal floating
13267 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13268 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13269 create the buffer on the stack to overcome not having a 32-bit
13271 (rs6000_emit_move): Likewise.
13272 (rs6000_secondary_memory_needed_rtx): Likewise.
13273 (rs6000_alloc_sdmode_stack_slot): Likewise.
13274 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13275 via xxlxor, just like DFmode 0.0.
13277 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
13278 define as 1 if we are running on a power7 or newer.
13279 (enum r6000_reg_class_enum): Add new constraints.
13281 * config/rs6000/dfp.md (movsd): Delete, combine with binary
13282 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13283 with other moves by using conditional constraits (wg). Use LFIWZX
13284 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
13285 (movsd splitter): Likewise.
13286 (movsd_hardfloat): Likewise.
13287 (movsd_softfloat): Likewise.
13289 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
13290 binary and decimal floating point moves.
13291 (fmove_ok): New attributes to combine binary and decimal floating
13292 point moves, and to combine power6x (mfpgpr) moves along normal
13294 (real_value_to_target): Likewise.
13295 (f32_lr): Likewise.
13296 (f32_lm): Likewise.
13297 (f32_li): Likewise.
13298 (f32_sr): Likewise.
13299 (f32_sm): Likewise.
13300 (f32_si): Likewise.
13301 (movsf): Combine binary and decimal floating point moves. Combine
13302 power6x (mfpgpr) moves with other moves by using conditional
13303 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
13304 (mov<mode> for SFmode/SDmode); Likewise.
13305 (SFmode/SDmode splitters): Likewise.
13306 (movsf_hardfloat): Likewise.
13307 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
13308 (movsf_softfloat): Likewise.
13309 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
13311 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
13312 wx and wz constraints.
13314 * config/rs6000/constraints.md (wg constraint): New constraint to
13315 return FLOAT_REGS if -mmfpgpr (power6x) was used.
13317 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
13320 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13321 -mdebug=reg, print wg, wl, wx, and wz constraints.
13322 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13323 Initialize the reload functions for 64-bit binary/decimal floating
13325 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13326 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13327 create the buffer on the stack to overcome not having a 32-bit
13329 (rs6000_emit_move): Likewise.
13330 (rs6000_secondary_memory_needed_rtx): Likewise.
13331 (rs6000_alloc_sdmode_stack_slot): Likewise.
13332 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13333 via xxlxor, just like DFmode 0.0.
13335 * config/rs6000/dfp.md (movdd): Delete, combine with binary
13336 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13337 with other moves by using conditional constraits (wg). Use LFIWZX
13338 and STFIWX for loading SDmode on power7.
13339 (movdd splitters): Likewise.
13340 (movdd_hardfloat32): Likewise.
13341 (movdd_softfloat32): Likewise.
13342 (movdd_hardfloat64_mfpgpr): Likewise.
13343 (movdd_hardfloat64): Likewise.
13344 (movdd_softfloat64): Likewise.
13346 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
13347 64-bit binary and decimal floating point moves.
13348 (FMOVE64X): Likewise.
13349 (movdf): Combine 64-bit binary and decimal floating point moves.
13350 Combine power6x (mfpgpr) moves with other moves by using
13351 conditional constraits (wg).
13352 (mov<mode> for DFmode/DDmode): Likewise.
13353 (DFmode/DDmode splitters): Likewise.
13354 (movdf_hardfloat32): Likewise.
13355 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
13356 (movdf_softfloat32): Likewise.
13357 (movdf_hardfloat64_mfpgpr): Likewise.
13358 (movdf_hardfloat64): Likewise.
13359 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
13360 (movdf_softfloat64): Likewise.
13361 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13362 (reload_<mode>_load): Move to later in the file so they aren't in
13363 the middle of the floating point move insns.
13364 (reload_<mode>_store): Likewise.
13366 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13369 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13370 constraint if -mdebug=reg.
13371 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13372 Enable using dd reload support if needed.
13374 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13375 binary and decimal floating point moves in rs6000.md.
13376 (movtd_internal): Likewise.
13378 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13379 decimal floating point moves.
13381 (movtf_internal): Likewise.
13382 (mov<mode>_internal, TDmode/TFmode): Likewise.
13383 (movtf_softfloat): Likewise.
13384 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13386 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13387 movdi_internal64, using wg constraint for move direct operations.
13388 (movdi_internal64): Likewise.
13390 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13391 MODES_TIEABLE_P for selected modes. Print the numerical value of
13392 the various virtual registers. Use GPR/FPR first/last values,
13393 instead of hard coding the register numbers. Print which modes
13394 have reload functions registered.
13395 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13396 settings before/after setting cpu, target and subtarget settings.
13397 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13398 and for secondary reload failures in rs6000_secondary_reload_inner.
13399 (rs6000_secondary_reload_fail): Likewise.
13400 (rs6000_secondary_reload_inner): Likewise.
13402 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13403 macros for first/last GPR and FPR registers.
13404 (LAST_GPR_REGNO): Likewise.
13405 (FIRST_FPR_REGNO): Likewise.
13406 (LAST_FPR_REGNO): Likewise.
13408 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13409 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13410 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13411 (vcond<mode><mode>): Likewise.
13412 (vcondu<mode><mode>): Likewise.
13413 (vector_gtu<mode>): Likewise.
13414 (vector_gte<mode>): Likewise.
13415 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13416 to prevent the compiler from converting DImode operations to TImode.
13417 (ior<mode>3): Likewise.
13418 (and<mode>3): Likewise.
13419 (one_cmpl<mode>2): Likewise.
13420 (nor<mode>3): Likewise.
13421 (andc<mode>3): Likewise.
13423 * config/rs6000/constraints.md (wt constraint): New constraint
13424 that returns VSX_REGS if TImode is allowed in VSX registers.
13426 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13427 constant under VSX.
13429 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13430 similar to TImode, but it is restricted to being in the GPRs.
13432 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13433 TImode to occupy a single VSX register.
13435 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13436 -mvsx-timode for power7/power8.
13437 (power7 cpu): Likewise.
13438 (power8 cpu): Likewise.
13440 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13441 sure that TFmode/TDmode take up two registers if they are ever
13442 allowed in the upper VSX registers.
13443 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13445 (rs6000_init_hard_regno_mode_ok): Likewise.
13446 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13447 constraint. Print if LRA is turned on.
13448 (rs6000_option_override_internal): Give an error if -mvsx-timode
13449 and VSX is not enabled.
13450 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13451 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13452 to reg+offset addressing. Use PTImode when checking offset
13453 addresses for validity.
13454 (reg_offset_addressing_ok_p): Likewise.
13455 (rs6000_legitimate_offset_address_p): Likewise.
13456 (rs6000_legitimize_address): Likewise.
13457 (rs6000_legitimize_reload_address): Likewise.
13458 (rs6000_legitimate_address_p): Likewise.
13459 (rs6000_eliminate_indexed_memrefs): Likewise.
13460 (rs6000_emit_move): Likewise.
13461 (rs6000_secondary_reload): Likewise.
13462 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13463 reloads to fpr registers to continue to use reg+offset addressing,
13464 but 64-bit reloads to altivec registers need reg+reg addressing.
13465 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13466 it. Treat LO_SUM like a PLUS operation.
13467 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13468 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13469 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13470 registers to share a register with a smaller sized type, since VSX
13471 puts scalars in the upper 64-bits.
13472 (print_operand): Add support for PTImode.
13473 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13474 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13475 registers, but don't have arithmetic support.
13476 (rs6000_memory_move_cost): Add test for VSX.
13477 (rs6000_opt_masks): Add -mvsx-timode.
13479 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13482 (VSr): Use wt constraint for TImode.
13483 (VSv): Drop TImode support.
13484 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13485 (vsx_movti_64bit): Likewise.
13486 (vsx_movti_32bit): Likewise.
13487 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13488 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13489 one '?' on the appropriate output constraint. Do not allow TImode
13490 logical operations on 32-bit systems.
13491 (vsx_ior<mode>3): Likewise.
13492 (vsx_xor<mode>3): Likewise.
13493 (vsx_one_cmpl<mode>2): Likewise.
13494 (vsx_nor<mode>3): Likewise.
13495 (vsx_andc<mode>3): Likewise.
13496 (vsx_concat_<mode>): Likewise.
13497 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13499 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13500 OPTION_MASK_VSX_TIMODE.
13501 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13502 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13504 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13505 (TI2 iterator): New iterator for TImode, PTImode.
13506 (wd mode attribute): Add values for vector types.
13507 (movti_string): Replace TI move operations with operations for TImode
13508 and PTImode. Add support for TImode being allowed in VSX registers.
13509 (mov<mode>_string, TImode/PTImode): Likewise.
13510 (movti_ppc64): Likewise.
13511 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13512 (TI mode splitters): Likewise.
13514 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13517 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
13519 PR tree-optimization/56355
13520 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13521 Also handle integers with undefined overflow.
13523 2013-03-20 Catherine Moore <clm@codesourcery.com>
13524 Maciej W. Rozycki <macro@codesourcery.com>
13525 Tom de Vries <tom@codesourcery.com>
13526 Nathan Sidwell <nathan@codesourcery.com>
13527 Iain Sandoe <iain@codesourcery.com>
13528 Nathan Froyd <froydnj@codesourcery.com>
13529 Chao-ying Fu <fu@mips.com>
13531 * doc/extend.texi: (micromips, nomicromips, nocompression):
13532 Document new function attributes.
13533 * doc/invoke.texi (minterlink-compressed, mmicromips,
13534 m14k, m14ke, m14kec): Document new options.
13535 (minterlink-mips16): Update documentation.
13536 * doc/md.texi (ZC, ZD): Document new constraints.
13537 * configure.ac (gcc_cv_as_micromips): Check if linker
13538 supports the .set micromips directive.
13539 * configure: Regenerate.
13540 * config.in: Regenerate.
13541 * config/mips/mips-tables.opt: Regenerate.
13542 * config/mips/micromips.md: New file.
13543 * constraints.md (ZC, ZD): New constraints.
13544 * config/mips/predicates.md (movep_src_register): New predicate.
13545 (movep_src_operand): New predicate.
13546 (non_volatile_mem_operand): New predicate.
13547 * config/mips/mips.md (multimem): New type.
13548 (length): Differentiate between 17-bit and 18-bit branch offsets.
13549 (MOVEP1, MOVEP2): New mode iterator.
13550 (mov_<load>l): Use ZC constraint.
13551 (mov_<load>r): Likewise.
13552 (mov_<store>l): Likewise.
13553 (mov_<store>r): Likewise.
13554 (*branch_equality<mode>_inverted): Add microMIPS support.
13555 (*branch_equality<mode>): Likewise.
13556 (*jump_absolute): Likewise.
13557 (indirect_jump_<mode>): Likewise.
13558 (tablejump_<mode>): Likewise.
13559 (<optab>_internal): Likewise.
13560 (sibcall_internal): Likewise.
13561 (sibcall_value_internal): Likewise.
13562 (prefetch): Use constraint ZD.
13563 * config/mips/mips.opt (minterlink-compressed): New option.
13564 (minterlink-mips16): Now an alias for minterlink-compressed.
13565 (mmicromips): New option.
13566 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13567 (compare_and_swap_12): Likewise.
13568 (sync_add<mode>): Likewise.
13569 (sync_<optab>_12): Likewise.
13570 (sync_old_<optab>_12): Likewise.
13571 (sync_new_<optab>_12): Likewise.
13572 (sync_nand_12): Likewise.
13573 (sync_old_nand_12): Likewise.
13574 (sync_new_nand_12): Likewise.
13575 (sync_sub<mode>): Likewise.
13576 (sync_old_add<mode>): Likewise.
13577 (sync_old_sub<mode>): Likewise.
13578 (sync_new_add<mode>): Likewise.
13579 (sync_new_sub<mode>): Likewise.
13580 (sync_<optab><mode>): Likewise.
13581 (sync_old_<optab><mode>): Likewise.
13582 (sync_new_<optab><mode>): Likewise.
13583 (sync_nand<mode>): Likewise.
13584 (sync_old_nand<mode>): Likewise.
13585 (sync_new_nand<mode>): Likewise.
13586 (sync_lock_test_and_set<mode>): Likewise.
13587 (test_and_set_12): Likewise.
13588 (atomic_compare_and_swap<mode>): Likewise.
13589 (atomic_exchange<mode>_llsc): Likewise.
13590 (atomic_fetch_add<mode>_llsc): Likewise.
13591 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13592 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13593 (umips_save_restore_pattern_p): Likewise.
13594 (umips_load_store_pair_p): Likewise.
13595 (umips_output_load_store_pair): Likewise.
13596 (umips_movep_target_p): Likewise.
13597 (umips_12bit_offset_address_p): Likewise.
13598 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13599 (mips_base_mips16): Rename this...
13600 (mips_base_compression_flags): ...to this. Update all uses.
13601 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13602 (mips_mips16_decl_p): Delete.
13603 (mips_nomips16_decl_p): Delete.
13604 (mips_get_compress_on_flags): New function.
13605 (mips_get_compress_off_flags): New function.
13606 (mips_get_compress_mode): New function.
13607 (mips_get_compress_on_name): New function.
13608 (mips_get_compress_off_name): New function.
13609 (mips_insert_attributes): Support multiple compression types.
13610 (mips_merge_decl_attributes): Likewise.
13611 (umips_12bit_offset_address_p): New function.
13612 (mips_start_function_definition): Emit .set micromips directive.
13613 (mips_call_may_need_jalx_p): New function.
13614 (mips_function_ok_for_sibcall): Add microMIPS support.
13615 (mips_print_operand_punctuation): Support short delay slots and
13617 (umips_swm_mask, umips_swm_encoding): New.
13618 (umips_build_save_restore): New function.
13619 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13620 (was_mips16_p): Remove.
13621 (old_compression_mode): New.
13622 (mips_set_compression_mode): New function.
13623 (mips_set_current_function): Add microMIPS support.
13624 (mips_option_override): Likewise.
13625 (umips_save_restore_pattern_p): New function.
13626 (umips_output_save_restore): New function.
13627 (umips_load_store_pair_p_1): New function.
13628 (umips_load_store_pair_p): New function.
13629 (umips_output_load_store_pair_1): New function.
13630 (umips_output_load_store_pair): New function.
13631 (umips_movep_target_p) New function.
13632 (mips_prepare_pch_save): Add microMIPS support.
13633 * config/mips/mips.h (TARGET_COMPRESSION): New.
13634 (TARGET_CPU_CPP_BUILTINS): Update macro
13635 to use new compression flags and to support microMIPS.
13636 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13637 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13638 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13639 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13640 (ASM_SPEC): Support mmicromips and mno-micromips.
13641 (M16STORE_REG_P): New macro.
13642 (MIPS_CALL): Support TARGET_MICROMIPS.
13643 (MICROMIPS_J): New macro.
13644 (mips_base_mips16): Rename this...
13645 (mips_base_compression_flags): ...to this.
13646 (UMIPS_12BIT_OFFSET_P): New macro.
13647 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13648 (MULTILIB_DIRNAMES): Likewise.
13649 2013-03-20 Richard Biener <rguenther@suse.de>
13651 PR tree-optimization/56661
13652 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13653 the result does not have to be distinct.
13655 2013-03-20 Richard Biener <rguenther@suse.de>
13657 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13660 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13661 Steven Bosscher <steven@gcc.gnu.org>
13663 PR rtl-optimization/56605
13664 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13666 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13669 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13670 that require movd instead of movq.
13672 2013-03-20 Richard Biener <rguenther@suse.de>
13674 * tree-ssa-structalias.c (struct variable_info): Add pointer
13675 to the first field of an aggregate with sub-vars. Make
13676 this and the pointer to the next subfield its ID.
13677 (vi_next): New function.
13678 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13679 storedanything_id, integer_id): Increment by one.
13680 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13681 get_call_clobber_vi): Adjust.
13682 (solution_set_expand): Simplify and speedup.
13683 (solution_set_add): Inline into ...
13684 (set_union_with_increment): ... this. Adjust accordingly.
13685 (do_sd_constraint): Likewise.
13686 (do_ds_constraint): Likewise.
13687 (do_complex_constraint): Simplify.
13688 (build_pred_graph): Adjust.
13689 (solve_graph): Likewise. Simplify and speedup.
13690 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13691 get_constraint_for_component_ref, get_constraint_for_1,
13692 first_vi_for_offset, first_or_preceding_vi_for_offset,
13693 create_function_info_for, create_variable_info_for_1,
13694 create_variable_info_for, intra_create_variable_infos): Adjust.
13695 (init_base_vars): Push NULL for ID zero.
13696 (compute_points_to_sets): Adjust.
13698 2013-03-20 Richard Biener <rguenther@suse.de>
13700 * cfgloop.c (verify_loop_structure): Streamline and avoid
13701 ICEing on corrupt loop tree.
13702 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13705 2013-03-20 Richard Biener <rguenther@suse.de>
13707 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13708 check whether an SSA update is needed.
13710 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13712 * config/mips/constraints.md (T): Rename to...
13716 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13717 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13719 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13721 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13722 (*subsi3_carryin_uxtw): Likewise.
13724 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13726 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13727 (*rorsi3_insn_uxtw): Likewise.
13729 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13731 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13732 (*extrsi5_insn_uxtw): Likewise.
13734 2013-03-19 Richard Biener <rguenther@suse.de>
13736 PR tree-optimization/56273
13737 * passes.c (init_optimization_passes): Move second VRP after DOM.
13739 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
13741 * config/i386/i386.md (*movti_internal): Merge from
13742 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13743 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13744 nox64 isa attributes.
13746 2013-03-18 Richard Biener <rguenther@suse.de>
13748 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13750 (merge_node_constraints): Likewise.
13751 (build_succ_graph): Likewise.
13752 (valid_graph_edge): Inline into single caller.
13753 (unify_nodes): Likewise. Use bitmap_set_bit return value
13755 (scc_visit): Fix formatting and variable use.
13756 (do_sd_constraint): Use gcc_checking_assert.
13757 (do_ds_constraint): Likewise.
13758 (do_complex_constraint): Likewise.
13759 (condense_visit): Likewise. Cleanup.
13760 (dump_pred_graph): New function.
13761 (perform_var_substitution): Dump the pred-graph before
13762 variable substitution.
13763 (find_equivalent_node): Use gcc_checking_assert.
13764 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13766 2013-03-18 Richard Biener <rguenther@suse.de>
13768 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13769 Remove cond_expr_stmt_list argument and do not gimplify the
13771 (vect_loop_versioning): Adjust.
13772 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13773 Cleanup to use less temporaries.
13774 (vect_create_data_ref_ptr): Cleanup.
13776 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13778 PR tree-optimization/56635
13779 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13780 require types_compatible_p types.
13782 2013-03-18 Nick Clifton <nickc@redhat.com>
13784 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13785 spurious backslash.
13787 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13788 Add missing line to comment describing function.
13790 2013-03-18 Richard Biener <rguenther@suse.de>
13792 PR tree-optimization/56210
13793 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13794 Handle string / character search functions.
13795 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13797 2013-03-18 Richard Biener <rguenther@suse.de>
13799 PR middle-end/56483
13800 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13801 and implement properly.
13802 * gimple.h (gimple_cond_single_var_p): Remove.
13804 2013-03-18 Richard Biener <rguenther@suse.de>
13806 * tree-data-ref.h (find_data_references_in_loop): Declare.
13807 * tree-data-ref.c (get_references_in_stmt): Use a stack
13808 vector pre-allocated in the callers.
13809 (find_data_references_in_stmt): Adjust.
13810 (graphite_find_data_references_in_stmt): Likewise.
13811 (create_rdg_vertices): Likewise.
13812 (find_data_references_in_loop): Export.
13813 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13814 Compute dependences here...
13815 (vect_analyze_data_refs): ...not here. When we encounter
13816 a non-vectorizable data reference in basic-block vectorization
13817 truncate the data reference vector. Do not bother to
13818 fixup data-dependence information for gather loads.
13819 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13820 of data references, as reported.
13822 2013-03-18 Richard Biener <rguenther@suse.de>
13824 PR tree-optimization/3713
13825 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13826 has_constants and expr.
13827 (stmt_has_constants): Properly valueize SSA names when deciding
13828 whether the stmt has constants.
13830 2013-03-18 Richard Biener <rguenther@suse.de>
13832 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13833 whole function when there is nothing to do.
13834 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13835 * tree-vectorizer.c (vectorize_loops): Update virtual and
13836 loop-closed SSA once.
13837 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13839 2013-03-18 Richard Biener <rguenther@suse.de>
13841 PR middle-end/56113
13842 * domwalk.c (bb_postorder): New global static.
13843 (cmp_bb_postorder): New function.
13844 (walk_dominator_tree): Replace scheme imposing an order for
13845 visiting dominator sons by one sorting them at the time they
13846 are pushed on the stack.
13848 2013-03-18 Richard Biener <rguenther@suse.de>
13850 PR tree-optimization/39326
13851 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13852 (struct mem_ref): Replace mem member with ao_ref typed member.
13853 (MEM_ANALYZABLE): Adjust.
13854 (memref_eq): Likewise.
13855 (mem_ref_alloc): Likewise.
13856 (gather_mem_refs_stmt): Likewise.
13857 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13858 (execute_sm_if_changed_flag_set): Adjust.
13859 (execute_sm): Likewise.
13860 (ref_always_accessed_p): Likewise.
13861 (refs_independent_p): Likewise.
13862 (can_sm_ref_p): Likewise.
13864 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13867 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13868 return 1 even for !unsignedp.
13870 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13872 * config/i386/i386.md (isa): Add x64 and nox64.
13873 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13874 (*pushtf): Enable *roF alternative for x64 isa only.
13875 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13876 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13877 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13878 (*movtf_internal): Merge from *movtf_internal_rex64 and
13879 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13880 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13881 nox64 isa attributes.
13882 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13883 nox64 isa attributes.
13884 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13886 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13888 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13890 2013-03-17 Jakub Jelinek <jakub@redhat.com>
13893 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13894 class names. Remove trailing comma after "ALL_REGS".
13896 2013-03-16 Jan Hubicka <jh@suse.cz>
13898 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13899 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13900 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13901 of cgraph_get_create_node.
13902 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13904 2013-03-16 Jason Merrill <jason@redhat.com>
13907 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13908 with DW_AT_default_value.
13910 2013-03-16 Jakub Jelinek <jakub@redhat.com>
13912 * BASE-VER: Set to 4.9.0.
13914 2013-03-14 Andi Kleen <ak@linux.intel.com>
13917 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13918 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13919 Document _x* TSX intrinsics.
13921 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13922 David Holsgrove <david.holsgrove@xilinx.com>
13924 * configure.ac: Add MicroBlaze TLS support detection.
13925 * configure: Regenerate.
13926 * config/microblaze/microblaze-protos.h
13927 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13928 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13929 * config/microblaze/microblaze.c (microblaze_address_type): Add
13930 ADDRESS_TLS and tls_reloc address types.
13931 (microblaze_address_info): Add tls_reloc.
13932 (TARGET_HAVE_TLS): Define.
13933 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13934 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
13935 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13936 load_tls_operand, microblaze_call_tls_get_addr,
13937 microblaze_legitimize_tls_address): New functions.
13938 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13939 (get_base_reg): Use microblaze_tls_symbol_p.
13940 (microblaze_classify_address): Handle TLS.
13941 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13942 label_mentioned_p and microblaze_tls_referenced_p.
13943 (microblaze_legitimize_address): Handle TLS.
13944 (microblaze_address_insns): Handle ADDRESS_TLS.
13945 (pic_address_needs_scratch): Handle TLS.
13946 (print_operand_address): Handle TLS.
13947 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13948 (microblaze_expand_move): Handle TLS.
13949 (microblaze_legitimate_constant_p): Check
13950 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
13951 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13952 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13953 (PIC_OFFSET_TABLE_REGNUM): Set.
13954 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13955 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13956 (addsi3, movsi_internal2, movdf_internal): Update constraints
13957 * config/microblaze/predicates.md (arith_plus_operand): Define
13958 (move_operand): Redefine as move_src_operand,
13959 check microblaze_tls_referenced_p.
13961 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13963 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13964 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13966 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13968 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
13971 2013-03-14 Jakub Jelinek <jakub@redhat.com>
13973 PR tree-optimization/53265
13974 * common.opt (Waggressive-loop-optimizations): New option.
13975 * tree-ssa-loop-niter.c: Include tree-pass.h.
13976 (do_warn_aggressive_loop_optimizations): New function.
13977 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
13978 if number_of_latch_executions returned constant.
13979 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
13980 early. If number_of_latch_executions returned constant, set
13981 nb_iterations_upper_bound back to it.
13982 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
13984 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
13985 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
13987 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
13988 (MULTILIB_OSDIRNAMES): Set.
13989 * genmultilib: If defaultosdirname doesn't start with :: , set
13990 defaultosdirname2 instead, clear it and emit two . multilib_raw
13991 entries instead of just one.
13993 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
13995 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13996 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
13997 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
13998 (SUBTARGET_OVERRIDE_OPTIONS): New.
14000 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14003 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14004 (musermode): Convert to Var(TARGET_USERMODE).
14005 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14006 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14007 * config/sh/sh.c (sh_option_override): Use
14008 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14009 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14011 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14013 (udivsi3_i4_single, divsi3_i4_single): Use
14014 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14016 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14018 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14021 2013-03-13 Richard Biener <rguenther@suse.de>
14023 PR tree-optimization/56608
14024 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14025 calls when vectorizing basic-blocks.
14027 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14030 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14033 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14035 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14037 2013-03-11 Jan Hubicka <jh@suse.cz>
14040 * lto-streamer-out.c (output_symbol_p): Skip references from
14041 constructors of external variables.
14043 2013-03-11 Jan Hubicka <jh@suse.cz>
14045 PR middle-end/56571
14046 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14048 * emit-rtl.c (verify_rtx_sharing): Likewise.
14049 (copy_insn_1): Likewise.
14050 * rtl.c (copy_rtx): Likewise.
14052 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14055 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14056 output_operand_lossage message.
14058 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14061 * arm.c (shift_op): Validate RTL pattern on the fly.
14062 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14065 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14068 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14069 new call patterns that clobber %fr12.
14070 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14071 split and postreload patterns.
14072 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14073 registers %fr12 and %fr12R as call used.
14075 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14077 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14078 (canon_address, record_store, replace_read, check_mem_read_rtx,
14079 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14080 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14081 rest_of_handle_dse): Likewise.
14083 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14085 PR middle-end/56524
14086 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14088 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14089 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14090 (save_optabs_if_changed): Replace with...
14091 (init_tree_optimization_optabs): ...this.
14092 * optabs.c (save_optabs_if_changed): Rename to...
14093 (init_tree_optimization_optabs): ...this. Take the optimization node
14094 as argument. Do nothing if the base optabs are already correct.
14095 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14096 to recompute optabs.
14097 * function.h (function): Remove optabs field.
14098 * function.c (invoke_set_current_function_hook): Call
14099 init_tree_optimization_optabs. Use the result to initialize
14102 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14104 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14105 if GTMA_HAS_NO_INSTRUMENTATION.
14106 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14107 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14108 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
14109 * gimple-pretty-print.c (dump_gimple_transaction): Handle
14110 GTMA_HAS_NO_INSTRUMENTATION.
14112 2013-03-08 Jakub Jelinek <jakub@redhat.com>
14114 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
14117 2013-03-08 Marek Polacek <polacek@redhat.com>
14118 Jakub Jelinek <jakub@redhat.com>
14120 PR tree-optimization/56478
14121 * predict.c (is_comparison_with_loop_invariant_p): Change the
14122 type of loop_step to tree.
14123 (predict_loops): Adjust.
14124 (predict_iv_comparison): Perform the computations on double_ints.
14126 2013-03-08 Richard Biener <rguenther@suse.de>
14128 PR tree-optimization/56570
14129 * tree-cfg.c (verify_expr_location_1): Verify locations for
14131 * tree-sra.c (create_access_replacement): Strip locations
14132 from DECL_DEBUG_EXPRs.
14134 2013-03-08 Richard Biener <rguenther@suse.de>
14136 * tree-inline.c (expand_call_inline): Do not associate
14137 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
14138 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
14140 2013-03-08 Richard Biener <rguenther@suse.de>
14142 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
14143 or block changes with -Og. Fix for location / block encoding
14144 changes and PHI arguments with locations.
14146 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
14148 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
14150 (struct output_info): Likewise.
14151 (register_overhead): Remove bad gcc_assert.
14152 (bitmap_find_bit): If there is only a single bitmap element, do not
14153 count a miss as a search.
14154 (print_statistics): Update for counter type changes.
14155 (dump_bitmap_statistics): Likewise. Print headers such that they
14156 are properly lined up with the printed counters.
14158 2013-03-07 Jakub Jelinek <jakub@redhat.com>
14160 PR tree-optimization/56559
14161 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
14162 check that it has only a single use.
14164 2013-03-07 Richard Biener <rguenther@suse.de>
14166 * doc/invoke.texi (fwhole-program): Discourage use in combination
14169 2013-03-06 Jakub Jelinek <jakub@redhat.com>
14171 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
14173 PR tree-optimization/56539
14174 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
14175 instead of GSI_CONTINUE_LINKING as last argument to
14176 force_gimple_operand_gsi. Adjust function comment.
14178 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
14181 PR middle-end/56548
14182 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
14183 promoted mode, convert the result back to the original mode.
14185 2013-03-06 Richard Biener <rguenther@suse.de>
14187 PR middle-end/56294
14188 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
14189 (insert_updated_phi_nodes_compare_uids): New function.
14190 (update_ssa): Sort symbols_to_rename after UID before
14191 traversing it to insert PHI nodes.
14193 2013-03-06 Richard Biener <rguenther@suse.de>
14195 PR middle-end/50494
14196 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
14197 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
14200 2013-02-13 Richard Biener <rguenther@suse.de>
14203 * varasm.c (output_constant_def_1): Get the decl representing
14204 the constant as argument.
14205 (output_constant_def): Wrap output_constant_def_1.
14206 (make_decl_rtl): Use output_constant_def_1 with the decl
14207 representing the constant.
14208 (build_constant_desc): Optionally re-use a decl already
14209 representing the constant.
14210 (tree_output_constant_def): Adjust.
14212 2013-03-06 Joey Ye <joey.ye@arm.com>
14215 * gcc.c (convert_white_space): New function.
14216 (main): Handles white space in function name.
14218 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
14221 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
14222 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
14223 to SH_DIV_CALL_TABLE for TARGET_SH2.
14224 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
14226 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
14227 call-table options.
14229 2013-03-05 Sterling Augustine <saugustine@google.com>
14230 Cary Coutant <ccoutant@google.com>
14233 * dwarf2out.c (resolve_addr): Don't call
14234 remove_loc_list_addr_table_entries a second time for the same
14237 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14240 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
14241 (avoid_complex_debug_insns): New function.
14242 (expand_debug_locations): Call it.
14244 PR rtl-optimization/56484
14245 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
14246 lifetimes of hard registers on small register class machines.
14248 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14250 * config/microblaze/microblaze-protos.h: Rename
14251 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
14252 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
14254 (microblaze_fast_interrupt_function_p): New function.
14255 (microblaze_is_interrupt_handler): Rename to
14256 microblaze_is_interrupt_variant and add fast_interrupt check.
14257 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
14258 (save_restore_insns): Likewise.
14259 (compute_frame_size): Likewise.
14260 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
14261 (microblaze_globalize_label): Likewise.
14262 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
14263 * config/microblaze/microblaze.md: Use wrapper
14264 microblaze_is_interrupt_variant.
14266 2013-03-05 Kai Tietz <ktietz@redhat.com>
14268 * sdbout.c (sdbout_one_type): Switch to current function's section
14269 supporting cold/hot.
14271 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14273 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
14276 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14278 PR middle-end/56461
14279 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
14280 if VALGRIND_GET_VBITS is defined, temporarily make object
14281 memory all defined, and restore previous valgrind addressability
14282 and definability afterwards. Free this_object at the end.
14284 PR middle-end/56461
14285 * lra.c (lra): Call lra_clear_live_ranges if live_p,
14286 right before calling lra_create_live_ranges, also call it
14287 when clearing live_p. Only call lra_clear_live_ranges
14288 at the end if live_p.
14290 PR middle-end/56461
14291 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
14293 2013-03-05 Richard Biener <rguenther@suse.de>
14295 PR tree-optimization/56521
14296 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
14299 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
14302 * except.h (remove_unreachable_eh_regions): New prototype.
14303 * except.c (remove_eh_handler_splicer): New function, split out
14304 of remove_eh_handler.
14305 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
14306 warning about running it on many EH regions one at a time.
14307 (remove_unreachable_eh_regions_worker): New function, walk the
14308 EH tree in depth-first order and remove non-marked regions.
14309 (remove_unreachable_eh_regions): New function.
14310 * tree-eh.c (mark_reachable_handlers): New function, split out
14311 from remove_unreachable_handlers.
14312 (remove_unreachable_handlers): Use mark_reachable_handlers and
14313 remove_unreachable_eh_regions.
14314 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
14315 and remove_unreachable_eh_regions.
14317 2013-03-05 Richard Biener <rguenther@suse.de>
14319 PR middle-end/56525
14320 * loop-init.c (fix_loop_structure): Remove loops in two stages,
14321 not freeing them until the end.
14323 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14325 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
14327 2013-03-05 Richard Biener <rguenther@suse.de>
14329 PR tree-optimization/56270
14330 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
14331 of loads after scheduling an SLP instance.
14333 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14335 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
14337 (check_gcc_parallelize): Run guality.exp as a separate job from
14338 vect.exp with unsorted.exp and $(dg_target_exps) separately from
14339 struct-layout-1.exp with stackalign.exp.
14341 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
14343 PR middle-end/56461
14344 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
14345 load_index sbitmap even if some bit in it isn't set.
14347 PR middle-end/56461
14348 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
14349 (discover_iteration_bound_by_body_walk): Change queues to
14350 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
14351 spelling in comment. Call safe_push on queues[bound_index] directly.
14352 Release queues[queue_index] in every iteration unconditionally.
14353 Release bounds vector.
14355 PR middle-end/56461
14356 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
14357 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
14358 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14361 2013-03-05 Richard Biener <rguenther@suse.de>
14364 * tree-inline.c (remap_blocks_to_null): New function.
14365 (expand_call_inline): When expanding a call stmt without
14366 an associated block inline remap all callee blocks to NULL.
14368 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14370 PR rtl-optimization/56494
14371 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14372 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14373 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14375 PR middle-end/56461
14376 * sel-sched-ir.c (free_sched_pools): Release
14377 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14378 if succs_info_pool.max_top isn't -1.
14381 * opts.c (opts_obstack, opts_concat): Moved to...
14382 * opts-common.c (opts_obstack, opts_concat): ... here.
14384 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14386 PR middle-end/56461
14387 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14389 2013-03-04 Martin Jambor <mjambor@suse.cz>
14391 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14392 all appropriate places.
14394 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14396 PR tree-optimization/56424
14397 * ipa-split.c (split_function): Do not set the RSO flag if result is
14398 not by reference and its type is a register type.
14400 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14402 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14403 (microblaze_legitimate_pic_operand): Likewise
14404 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14405 new function microblaze_legitimate_pic_operand
14406 * config/microblaze/microblaze-protos.h
14407 (microblaze_legitimate_pic_operand): Declare.
14409 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14411 * config/microblaze/predicates.md (call_insn_simple_operand):
14412 New predicate for supported rtx code types.
14413 * config/microblaze/microblaze.md (call_internal1): Use
14414 call_insn_simple_operand predicate.
14416 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14418 PR middle-end/56461
14419 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14420 partitions.ordered_remove.
14422 PR middle-end/56461
14423 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14424 vec_oprnds0.create (1) for modifier == NONE.
14426 PR middle-end/56461
14427 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14428 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14429 vec_oprnds1 right before pushing anything to it for
14432 PR middle-end/56461
14433 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14434 set nbbs to 0 instead of having separate code path.
14435 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14436 instead of false as last argument if returning NULL.
14438 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14440 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14441 the attribute is now called "target" instead of "option".
14442 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14443 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14444 attribute/pragma name for TARGET_OPTION_VALID_P and
14445 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14446 * doc/tm.texi: Regenerated.
14448 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14450 * config/microblaze/microblaze.c:
14451 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14452 * config/microblaze/microblaze.h: Add -mxl-reorder to
14454 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14455 instructions emitted if TARGET_REORDER.
14456 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14457 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14460 2013-03-01 Xinliang David Li <davidxl@google.com>
14462 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14465 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14467 PR middle-end/56461
14468 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14469 vector even when returning true. Fix up function comment formatting.
14471 PR middle-end/56461
14472 * ira-build.c (ira_loop_nodes_count): New variable.
14473 (create_loop_tree_nodes): Initialize it.
14474 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14476 PR middle-end/56461
14477 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14478 method on dr_chain and result_chain.
14479 * tree-vect-stmts.c (vectorizable_store): Only call
14480 result_chain.create if j == 0.
14482 PR middle-end/56461
14483 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14484 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14485 before overwriting it.
14487 2013-03-01 Tobias Burnus <burnus@net-b.de>
14489 * doc/extended.texi (C Extensions): Change order in @menu
14491 (Other MIPS Built-in Functions): Move last MIPS entry before
14492 "picoChip Built-in Functions".
14493 (SH Built-in Functions): Move after RX Built-in Functions.
14494 * doc/gcc.texi (Introduction): Change order in @menu
14496 * doc/md.texi (Constraints): Ditto.
14497 * gty.texi (Type Information): Ditto.
14498 (User-provided marking routines for template types): Make
14500 * doc/invoke.texi (AArch64 Options): Move before
14501 "Adapteva Epiphany Options".
14503 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14504 Jakub Jelinek <jakub@redhat.com>
14507 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14508 no_address_safety_analysis attribute.
14509 * doc/extend.texi (no_address_safety_attribute): Rename to
14510 no_sanitize_address attribute, mention no_address_safety_analysis
14511 attribute as deprecated alias.
14513 2013-02-28 Jakub Jelinek <jakub@redhat.com>
14515 PR middle-end/56461
14516 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14517 type to vec<vec<tree> > *.
14518 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14519 to be vec<tree> instead of vec<tree> *, set vec_defs
14520 to vNULL and call vec_defs.create (number_of_vects), adjust other
14522 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14523 vectorizable_condition): Adjust vect_get_slp_defs callers.
14525 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14527 * config/aarch64/aarch64.c
14528 (aarch64_float_const_representable): Remove unused variable.
14530 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14532 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14534 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14536 * config/aarch64/aarch64-builtins.c
14537 (aarch64_init_simd_builtins): Make static.
14539 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14541 * config/aarch64/aarch64.c
14542 (aarch64_simd_make_constant): Make static.
14544 2013-02-28 Martin Jambor <mjambor@suse.cz>
14546 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14547 with no initialization to the RHS of debug statements.
14549 2013-02-28 Martin Jambor <mjambor@suse.cz>
14551 PR tree-optimization/56294
14552 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14554 (get_access_replacement): Do not call create_access_replacement.
14555 Assert a replacement exists.
14556 (get_repl_default_def_ssa_name): Create the replacement declaration
14559 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14561 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14562 final_end_function.
14564 2013-02-28 Marek Polacek <polacek@redhat.com>
14566 PR rtl-optimization/56466
14567 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14568 if we're changing a loop.
14569 (peel_loops_completely): Likewise.
14571 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14574 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14576 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14579 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14580 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14581 INTX_FTYPE_FX, FX_FTYPE_INTX.
14582 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14584 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14586 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14587 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14588 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14589 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14590 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14591 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14592 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14593 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14594 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14595 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14596 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14597 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14598 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14599 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14600 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14601 (avrxmega6): Increase max flash segments from 5 to 6.
14602 * config/avr/t-multilib: Regenerate.
14603 * config/avr/avr-tables.opt: Regenerate.
14604 * doc/avr-mmcu.texi: Regenerate.
14606 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14608 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14609 (avr_device_to_arch): Rename to avr_device_to_ld.
14610 (avr_device_to_as): New prototype.
14611 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14612 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14613 * config/avr/driver-avr.c (avr_device_to_as): New.
14614 (avr_device_to_arch): Rename to avr_device_to_ld.
14616 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14618 PR middle-end/56461
14619 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14620 method on dr_chain and result_chain.
14622 PR middle-end/56461
14623 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14624 pointer_set_destroy on not_executed_last_iteration.
14626 PR middle-end/56461
14627 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
14629 PR middle-end/56461
14630 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14631 FOR_EACH_DEFINED_FUNCTION when freeing state.
14633 PR middle-end/56461
14634 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14636 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14639 PR middle-end/56461
14640 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14641 known_aggs[i].items and release known_aggs vector.
14643 PR middle-end/56461
14644 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14646 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14648 * config/microblaze/microblaze.c (microblaze_emit_compare):
14649 Use xor for EQ/NE comparisions.
14650 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14651 (cbranchsf4): Adjust operator to comparison_operator.
14653 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14655 PR middle-end/56461
14656 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14658 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14659 vec_safe_push, always update *slot.
14660 (redirect_edge_var_map_clear): Use vec_free.
14661 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14662 (free_var_map_entry): Use vec_free.
14663 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14664 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14666 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
14668 PR middle-end/45472
14669 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14670 when the may_trap_p bit of the exprs being merged differs.
14671 Reorder tests for speculativeness in the logical and operator.
14673 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14675 * incpath.c (add_standard_paths): Use reconcat instead of concat
14676 where appropriate and avoid leaking memory.
14678 * opts.h: Include obstack.h.
14679 (opts_concat): New prototype.
14680 (opts_obstack): New declaration.
14681 * opts.c (opts_concat): New function.
14682 (opts_obstack): New variable.
14683 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14684 (finish_options): Use opts_concat instead of concat
14685 and XOBNEWVEC instead of XNEWVEC.
14686 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14687 generate_option): Likewise.
14688 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14689 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14692 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14693 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14695 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14697 PR middle-end/56461
14698 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14700 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14702 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14703 (arm_block_move_unaligned_straight): Likewise.
14704 (arm_adjust_block_mem): Likewise.
14706 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14709 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14710 temp, cond and label.
14711 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14714 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14715 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14718 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14719 of prologue/epilogue functions.
14722 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14723 Remove unused variable cfa_offset.
14724 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14727 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14728 type promotion to unsigned.
14731 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14732 for HOST_WIDE_INT of 32 bit / same size as int.
14733 (arm_block_move_unaligned_straight): Likewise.
14734 (arm_adjust_block_mem): Likewise.
14737 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14740 2013-02-26 Marek Polacek <polacek@redhat.com>
14742 PR tree-optimization/56426
14743 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
14745 2013-02-26 Richard Biener <rguenther@suse.de>
14748 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14749 unused variable loops.
14751 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14753 PR tree-optimization/56448
14754 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14755 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14756 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14757 later operands of the references, or even first operand for
14758 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14760 PR tree-optimization/56443
14761 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14762 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14763 to type_for_mode langhook.
14765 2013-02-25 Matt Turner <mattst88@gmail.com>
14767 * doc/invoke.texi: Document r4700.
14769 2013-02-25 Richard Biener <rguenther@suse.de>
14771 PR tree-optimization/56175
14772 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14774 (simplify_bitwise_binary): ... here. Also guard the conversion
14775 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14777 2013-02-25 Catherine Moore <clm@codesourcery.com>
14780 2013-02-24 Catherine Moore <clm@codesourcery.com>
14781 Maciej W. Rozycki <macro@codesourcery.com>
14782 Tom de Vries <tom@codesourcery.com>
14783 Nathan Sidwell <nathan@codesourcery.com>
14784 Iain Sandoe <iain@codesourcery.com>
14785 Nathan Froyd <froydnj@codesourcery.com>
14786 Chao-ying Fu <fu@mips.com>
14788 * doc/extend.texi: (micromips, nomicromips, nocompression):
14789 Document new function attributes.
14790 * doc/invoke.texi (minterlink-compressed, mmicromips,
14791 m14k, m14ke, m14kec): Document new options.
14792 (minterlink-mips16): Update documentation.
14793 * doc/md.texi (ZC, ZD): Document new constraints.
14794 * configure.ac (gcc_cv_as_micromips): Check if linker
14795 supports the .set micromips directive.
14796 * configure: Regenerate.
14797 * config.in: Regenerate.
14798 * config/mips/mips-tables.opt: Regenerate.
14799 * config/mips/micromips.md: New file.
14800 * constraints.md (ZC, AD): New constraints.
14801 * config/mips/predicates.md (movep_src_register): New predicate.
14802 (movep_src_operand): New predicate.
14803 (non_volatile_mem_operand): New predicate.
14804 * config/mips/mips.md (multimem): New type.
14805 (length): Differentiate between 17-bit and 18-bit branch offsets.
14806 (MOVEP1, MOVEP2): New mode iterator.
14807 (mov_<load>l): Use ZC constraint.
14808 (mov_<load>r): Likewise.
14809 (mov_<store>l): Likewise.
14810 (mov_<store>r): Likewise.
14811 (*branch_equality<mode>_inverted): Add microMIPS support.
14812 (*branch_equality<mode>): Likewise.
14813 (*jump_absolute): Likewise.
14814 (indirect_jump_<mode>): Likewise.
14815 (tablejump_<mode>): Likewise.
14816 (<optab>_internal): Likewise.
14817 (sibcall_internal): Likewise.
14818 (sibcall_value_internal): Likewise.
14819 (prefetch): Use constraint ZD.
14820 * config/mips/mips.opt (minterlink-compressed): New option.
14821 (minterlink-mips16): Now an alias for minterlink-compressed.
14822 (mmicromips): New option.
14823 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14824 (compare_and_swap_12): Likewise.
14825 (sync_add<mode>): Likewise.
14826 (sync_<optab>_12): Likewise.
14827 (sync_old_<optab>_12): Likewise.
14828 (sync_new_<optab>_12): Likewise.
14829 (sync_nand_12): Likewise.
14830 (sync_old_nand_12): Likewise.
14831 (sync_new_nand_12): Likewise.
14832 (sync_sub<mode>): Likewise.
14833 (sync_old_add<mode>): Likewise.
14834 (sync_old_sub<mode>): Likewise.
14835 (sync_new_add<mode>): Likewise.
14836 (sync_new_sub<mode>): Likewise.
14837 (sync_<optab><mode>): Likewise.
14838 (sync_old_<optab><mode>): Likewise.
14839 (sync_new_<optab><mode>): Likewise.
14840 (sync_nand<mode>): Likewise.
14841 (sync_old_nand<mode>): Likewise.
14842 (sync_new_nand<mode>): Likewise.
14843 (sync_lock_test_and_set<mode>): Likewise.
14844 (test_and_set_12): Likewise.
14845 (atomic_compare_and_swap<mode>): Likewise.
14846 (atomic_exchange<mode>_llsc): Likewise.
14847 (atomic_fetch_add<mode>_llsc): Likewise.
14848 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14849 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14850 (umips_save_restore_pattern_p): Likewise.
14851 (umips_load_store_pair_p): Likewise.
14852 (umips_output_load_store_pair): Likewise.
14853 (umips_movep_target_p): Likewise.
14854 (umips_12bit_offset_address_p): Likewise.
14855 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14856 (mips_base_mips16): Rename this...
14857 (mips_base_compression_flags): ...to this. Update all uses.
14858 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14859 (mips_mips16_decl_p): Delete.
14860 (mips_nomips16_decl_p): Delete.
14861 (mips_get_compress_on_flags): New function.
14862 (mips_get_compress_off_flags): New function.
14863 (mips_get_compress_mode): New function.
14864 (mips_get_compress_on_name): New function.
14865 (mips_get_compress_off_name): New function.
14866 (mips_insert_attributes): Support multiple compression types.
14867 (mips_merge_decl_attributes): Likewise.
14868 (umips_12bit_offset_address_p): New function.
14869 (mips_start_function_definition): Emit .set micromips directive.
14870 (mips_call_may_need_jalx_p): New function.
14871 (mips_function_ok_for_sibcall): Add microMIPS support.
14872 (mips_print_operand_punctuation): Support short delay slots and
14874 (umips_swm_mask, umips_swm_encoding): New.
14875 (umips_build_save_restore): New function.
14876 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14877 (was_mips16_p): Remove.
14878 (old_compression_mode): New.
14879 (mips_set_compression_mode): New function.
14880 (mips_set_current_function): Add microMIPS support.
14881 (mips_option_override): Likewise.
14882 (umips_save_restore_pattern_p): New function.
14883 (umips_output_save_restore): New function.
14884 (umips_load_store_pair_p_1): New function.
14885 (umips_load_store_pair_p): New function.
14886 (umips_output_load_store_pair_1): New function.
14887 (umips_output_load_store_pair): New function.
14888 (umips_movep_target_p) New function.
14889 (mips_prepare_pch_save): Add microMIPS support.
14890 * config/mips/mips.h (TARGET_COMPRESSION): New.
14891 (TARGET_CPU_CPP_BUILTINS): Update macro
14892 to use new compression flags and to support microMIPS.
14893 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14894 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14895 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14896 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14897 (ASM_SPEC): Support mmicromips and mno-micromips.
14898 (M16STORE_REG_P): New macro.
14899 (MIPS_CALL): Support TARGET_MICROMIPS.
14900 (MICROMIPS_J): New macro.
14901 (mips_base_mips16): Rename this...
14902 (mips_base_compression_flags): ...to this.
14903 (UMIPS_12BIT_OFFSET_P): New macro.
14904 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14905 (MULTILIB_DIRNAMES): Likewise.
14907 2013-02-25 Tom de Vries <tom@codesourcery.com>
14909 PR rtl-optimization/56131
14910 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14911 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14912 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14914 2013-02-25 Tobias Burnus <burnus@net-b.de>
14916 * doc/invoke.texi (-fsanitize=): Move from optimization
14917 to debugging options.
14919 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14921 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14923 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14924 Alexander Monakov <amonakov@ispras.ru>
14926 PR middle-end/56077
14927 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14928 flush pending lists also on non-jumps. Adjust comment.
14930 2013-02-24 Catherine Moore <clm@codesourcery.com>
14931 Maciej W. Rozycki <macro@codesourcery.com>
14932 Tom de Vries <tom@codesourcery.com>
14933 Nathan Sidwell <nathan@codesourcery.com>
14934 Iain Sandoe <iain@codesourcery.com>
14935 Nathan Froyd <froydnj@codesourcery.com>
14936 Chao-ying Fu <fu@mips.com>
14938 * doc/extend.texi: (micromips, nomicromips, nocompression):
14939 Document new function attributes.
14940 * doc/invoke.texi (minterlink-compressed, mmicromips,
14941 m14k, m14ke, m14kec): Document new options.
14942 (minterlink-mips16): Update documentation.
14943 * doc/md.texi (ZC, ZD): Document new constraints.
14944 * configure.ac (gcc_cv_as_micromips): Check if linker
14945 supports the .set micromips directive.
14946 * configure: Regenerate.
14947 * config.in: Regenerate.
14948 * config/mips/mips-tables.opt: Regenerate.
14949 * config/mips/micromips.md: New file.
14950 * constraints.md (ZC, AD): New constraints.
14951 * config/mips/predicates.md (movep_src_register): New predicate.
14952 (movep_src_operand): New predicate.
14953 (non_volatile_mem_operand): New predicate.
14954 * config/mips/mips.md (multimem): New type.
14955 (length): Differentiate between 17-bit and 18-bit branch offsets.
14956 (MOVEP1, MOVEP2): New mode iterator.
14957 (mov_<load>l): Use ZC constraint.
14958 (mov_<load>r): Likewise.
14959 (mov_<store>l): Likewise.
14960 (mov_<store>r): Likewise.
14961 (*branch_equality<mode>_inverted): Add microMIPS support.
14962 (*branch_equality<mode>): Likewise.
14963 (*jump_absolute): Likewise.
14964 (indirect_jump_<mode>): Likewise.
14965 (tablejump_<mode>): Likewise.
14966 (<optab>_internal): Likewise.
14967 (sibcall_internal): Likewise.
14968 (sibcall_value_internal): Likewise.
14969 (prefetch): Use constraint ZD.
14970 * config/mips/mips.opt (minterlink-compressed): New option.
14971 (minterlink-mips16): Now an alias for minterlink-compressed.
14972 (mmicromips): New option.
14973 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14974 (compare_and_swap_12): Likewise.
14975 (sync_add<mode>): Likewise.
14976 (sync_<optab>_12): Likewise.
14977 (sync_old_<optab>_12): Likewise.
14978 (sync_new_<optab>_12): Likewise.
14979 (sync_nand_12): Likewise.
14980 (sync_old_nand_12): Likewise.
14981 (sync_new_nand_12): Likewise.
14982 (sync_sub<mode>): Likewise.
14983 (sync_old_add<mode>): Likewise.
14984 (sync_old_sub<mode>): Likewise.
14985 (sync_new_add<mode>): Likewise.
14986 (sync_new_sub<mode>): Likewise.
14987 (sync_<optab><mode>): Likewise.
14988 (sync_old_<optab><mode>): Likewise.
14989 (sync_new_<optab><mode>): Likewise.
14990 (sync_nand<mode>): Likewise.
14991 (sync_old_nand<mode>): Likewise.
14992 (sync_new_nand<mode>): Likewise.
14993 (sync_lock_test_and_set<mode>): Likewise.
14994 (test_and_set_12): Likewise.
14995 (atomic_compare_and_swap<mode>): Likewise.
14996 (atomic_exchange<mode>_llsc): Likewise.
14997 (atomic_fetch_add<mode>_llsc): Likewise.
14998 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14999 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15000 (umips_save_restore_pattern_p): Likewise.
15001 (umips_load_store_pair_p): Likewise.
15002 (umips_output_load_store_pair): Likewise.
15003 (umips_movep_target_p): Likewise.
15004 (umips_12bit_offset_address_p): Likewise.
15005 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15006 (mips_base_mips16): Rename this...
15007 (mips_base_compression_flags): ...to this. Update all uses.
15008 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15009 (mips_mips16_decl_p): Delete.
15010 (mips_nomips16_decl_p): Delete.
15011 (mips_get_compress_on_flags): New function.
15012 (mips_get_compress_off_flags): New function.
15013 (mips_get_compress_mode): New function.
15014 (mips_get_compress_on_name): New function.
15015 (mips_get_compress_off_name): New function.
15016 (mips_insert_attributes): Support multiple compression types.
15017 (mips_merge_decl_attributes): Likewise.
15018 (umips_12bit_offset_address_p): New function.
15019 (mips_start_function_definition): Emit .set micromips directive.
15020 (mips_call_may_need_jalx_p): New function.
15021 (mips_function_ok_for_sibcall): Add microMIPS support.
15022 (mips_print_operand_punctuation): Support short delay slots and
15024 (umips_swm_mask, umips_swm_encoding): New.
15025 (umips_build_save_restore): New function.
15026 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15027 (was_mips16_p): Remove.
15028 (old_compression_mode): New.
15029 (mips_set_compression_mode): New function.
15030 (mips_set_current_function): Add microMIPS support.
15031 (mips_option_override): Likewise.
15032 (umips_save_restore_pattern_p): New function.
15033 (umips_output_save_restore): New function.
15034 (umips_load_store_pair_p_1): New function.
15035 (umips_load_store_pair_p): New function.
15036 (umips_output_load_store_pair_1): New function.
15037 (umips_output_load_store_pair): New function.
15038 (umips_movep_target_p) New function.
15039 (mips_prepare_pch_save): Add microMIPS support.
15040 * config/mips/mips.h (TARGET_COMPRESSION): New.
15041 (TARGET_CPU_CPP_BUILTINS): Update macro
15042 to use new compression flags and to support microMIPS.
15043 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15044 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15045 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15046 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15047 (ASM_SPEC): Support mmicromips and mno-micromips.
15048 (M16STORE_REG_P): New macro.
15049 (MIPS_CALL): Support TARGET_MICROMIPS.
15050 (MICROMIPS_J): New macro.
15051 (mips_base_mips16): Rename this...
15052 (mips_base_compression_flags): ...to this.
15053 (UMIPS_12BIT_OFFSET_P): New macro.
15054 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15055 (MULTILIB_DIRNAMES): Likewise.
15057 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15060 * target-globals.c (save_target_globals): For init_reg_sets and
15061 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15063 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15065 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15066 * config/aarch64/t-aarch64
15067 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15069 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15071 PR inline-asm/56148
15072 * lra-constraints.c (process_alt_operands): Reload operand
15073 conflicting with earlier clobber only if no more other conflicting
15076 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15079 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15080 if not linking a shared library.
15082 2013-02-22 Seth LaForge <sethml@google.com>
15084 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15086 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15088 * config/arm/arm.md (split for extendsidi): Update condition.
15089 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15090 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15091 (qhs_zextenddi_cstr): Likewise.
15093 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15095 PR middle-end/56420
15096 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15097 avoid signed wrapping.
15098 (expand_mult): Handle properly multiplication by
15099 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15100 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15101 in the compiler if coeff is HOST_WIDE_INT_MIN.
15102 (expand_divmod): Don't make ext_op1 static, change it's type to
15103 uhwi. Avoid undefined behavior in -INTVAL (op1).
15105 PR rtl-optimization/50339
15106 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15108 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
15109 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
15110 into splitting_ashiftrt field.
15111 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
15113 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
15116 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
15118 PR middle-end/56108
15119 * trans-mem.c (execute_tm_mark): Do not expand transactions that
15120 are sure to go irrevocable.
15122 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
15124 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
15125 scalars are valid operands.
15127 2013-02-21 Martin Jambor <mjambor@suse.cz>
15129 PR tree-optimization/56310
15130 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
15131 only matching indices and non-negative final offsets.
15132 (intersect_aggregates_with_edge): Pass src_idx to
15133 agg_replacements_to_vector. Pass src_idx insstead of index to
15134 intersect_with_agg_replacements.
15136 2013-02-21 Martin Jambor <mjambor@suse.cz>
15138 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
15139 instead of hard-wired defaults.
15141 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
15143 * doc/invoke.texi (MIPS Options): Update documentation of the
15144 floating-point multiply-accumulate instruction restrictions.
15146 2013-02-21 Kostya Serebryany <kcc@google.com>
15148 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
15149 asan_shadow_offset on x86_64 linux.
15151 2013-02-21 Richard Biener <rguenther@suse.de>
15153 PR tree-optimization/56415
15155 2013-02-11 Richard Biener <rguenther@suse.de>
15157 PR tree-optimization/56273
15158 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15161 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15164 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
15167 PR inline-asm/56405
15168 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
15169 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
15171 2013-02-20 Jan Hubicka <jh@suse.cz>
15173 PR tree-optimization/56265
15174 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
15175 when target is referenced for first time.
15177 2013-02-20 Richard Biener <rguenther@suse.de>
15179 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
15180 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
15181 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
15182 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
15183 not return anything.
15184 (rename_ssa_copies): Do not remove unused locals.
15185 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
15186 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
15187 * passes.c (execute_function_todo): Do not schedule unused locals
15188 removal if cleanup_tree_cfg did something.
15189 * tree-ssa-live.c (remove_unused_locals): Dump statistics
15190 about the number of removed locals.
15192 2013-02-20 Richard Biener <rguenther@suse.de>
15194 PR tree-optimization/56398
15195 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
15197 2013-02-20 Martin Jambor <mjambor@suse.cz>
15199 PR tree-optimization/55334
15200 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
15201 restricted pointers to arrays.
15203 2013-02-20 Richard Biener <rguenther@suse.de>
15204 Jakub Jelinek <jakub@redhat.com>
15206 PR tree-optimization/56396
15207 * tree-ssa-ccp.c (n_const_val): New static variable.
15208 (get_value): Return NULL for SSA names we don't have a lattice
15210 (ccp_initialize): Initialize n_const_val.
15211 * tree-ssa-copy.c (n_copy_of): New static variable.
15212 (init_copy_prop): Initialize n_copy_of.
15213 (get_value): Return NULL_TREE for SSA names we don't have a
15216 2013-02-20 Martin Jambor <mjambor@suse.cz>
15218 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
15220 2013-02-20 Richard Biener <rguenther@suse.de>
15222 * genpreds.c (write_lookup_constraint): Do not compare first
15223 letter of the constraint again.
15225 2013-02-20 Richard Biener <rguenther@suse.de>
15227 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
15229 (get_use_iv_cost): Terminate hashtable walk when coming across
15232 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
15234 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
15235 reassociation for avx2 targets.
15237 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15239 * config/microblaze/microblaze.c: microblaze_has_clz = 0
15240 Add version check for v8.10.a to enable microblaze_has_clz
15241 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
15242 version and TARGET_PATTERN_COMPARE check
15243 * config/microblaze/microblaze.md: New clzsi2 instruction
15245 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15247 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
15248 function before branching.
15250 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15252 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
15254 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
15256 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15258 PR middle-end/55889
15259 * sel-sched.c: Include ira.h.
15260 (implicit_clobber_conflict_p): New function.
15261 (moveup_expr): Use it.
15262 * Makefile.in (sel-sched.o): Depend on ira.h.
15264 2013-02-19 Richard Biener <rguenther@suse.de>
15266 PR tree-optimization/56384
15267 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
15268 (vn_hash_type): Split out from ...
15269 (vn_hash_constant_with_type): ... here.
15270 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
15271 (vn_phi_eq): Compare types from vn_phi_s structure.
15272 (vn_phi_lookup): Populate vn_phi_s type.
15273 (vn_phi_insert): Likewise.
15275 2013-02-19 Jakub Jelinek <jakub@redhat.com>
15277 PR tree-optimization/56350
15278 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
15279 if haven't found reduction or nested cycle operand, rather than
15280 asserting we must find it.
15282 PR tree-optimization/56381
15283 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
15286 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
15287 Jakub Jelinek <jakub@redhat.com>
15290 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
15291 (swap_optab_enable): Same.
15292 (init_all_optabs): Use argument instead of global.
15293 * tree.h (struct tree_optimization_option): New field target_optabs.
15294 * expr.h (init_all_optabs): Add argument to prototype.
15295 (TREE_OPTIMIZATION_OPTABS): New.
15296 (save_optabs_if_changed): Protoize.
15297 * optabs.h: Declare this_fn_optabs.
15298 * optabs.c (save_optabs_if_changed): New.
15299 Declare this_fn_optabs.
15300 (init_optabs): Add argument to init_all_optabs() call.
15301 * function.c (invoke_set_current_function_hook): Handle per
15303 * function.h (struct function): New field optabs.
15304 * config/mips/mips.c (mips_set_mips16_mode): Handle when
15305 optimization_current_node has changed.
15306 * target-globals.h (save_target_globals_default_opts): Protoize.
15307 * target-globals.c (save_target_globals_default_opts): New.
15309 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15312 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
15313 registers %fr12 and %fr12R as call used.
15316 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
15317 and HImode, require all displacements to be an integer multiple of
15319 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
15320 only allow QImode and HImode when reload is in progress and strict is
15321 true. Likewise for symbolic addresses. Use base14_operand to check
15322 displacements in REG+BASE addresses.
15324 2013-02-18 Richard Biener <rguenther@suse.de>
15326 PR tree-optimization/56366
15327 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15328 handle sign-conversion of outer-loop initial induction value.
15330 2013-02-18 Richard Biener <rguenther@suse.de>
15332 PR middle-end/56349
15333 * cfghooks.c (merge_blocks): If we merge a latch into another
15334 block adjust references to it.
15335 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
15336 (verify_loop_structure): Verify that a recorded latch is in fact
15339 2013-02-18 Richard Biener <rguenther@suse.de>
15341 PR tree-optimization/56321
15342 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
15343 order SSA name release and virtual operand unlinking.
15345 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15347 * config/microblaze/microblaze.md (save_stack_block): Define.
15348 (restore_stack_block): Likewise.
15350 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15352 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
15353 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
15354 * config/microblaze/microblaze.c (microblaze_option_override):
15355 Bail out early for PIC modes when target does not support PIC.
15357 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15359 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
15360 Replace with a microblaze version.
15361 (microblaze_trampoline_init): Adapt for microblaze.
15362 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15365 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15366 Dodji Seketeli <dodji@redhat.com>
15369 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15370 (instrument_mem_region_access): Do not forget to always put
15371 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15372 0) statement, even for cases where either 'base' or 'base + len'
15373 are not instrumented -- because they have been previously
15374 instrumented. Simplify the logic by putting all the statements
15375 instrument 'base + len' inside a sequence, and then insert that
15376 sequence right before the current insertion point. Then, to
15377 instrument 'base + len', just get an iterator on that statement.
15378 And do not forget to update the pointer to iterator the function
15379 received as argument.
15381 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15383 PR rtl-optimization/56348
15384 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15386 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15388 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15389 (clean_graph_dump_file): Pass base to start_graph_dump.
15391 2013-02-14 Richard Henderson <rth@redhat.com>
15394 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15396 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15398 * collect2-aix.h: Define F_LOADONLY.
15400 2013-02-14 Richard Biener <rguenther@suse.de>
15403 * varasm.c (output_constant_def_1): Get the decl representing
15404 the constant as argument.
15405 (output_constant_def): Wrap output_constant_def_1.
15406 (make_decl_rtl): Use output_constant_def_1 with the decl
15407 representing the constant.
15408 (build_constant_desc): Optionally re-use a decl already
15409 representing the constant.
15410 (tree_output_constant_def): Adjust.
15412 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15415 * asan.c (instrument_builtin_call): Really put the length of the
15416 second source argument into src1_len.
15418 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15420 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15421 argument. If it is false, don't create edge from then_bb to
15423 (insert_if_then_before_iter): Pass true to it.
15424 (build_check_stmt): Pass false to it.
15425 (transform_statements): Flush hash table only on extended basic
15426 block boundaries, rather than at the beginning of every bb.
15427 Don't flush hash table on nonfreeing_call_p calls.
15428 * tree-flow.h (nonfreeing_call_p): New prototype.
15429 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15431 2013-02-13 David S. Miller <davem@davemloft.net>
15433 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15435 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15438 * ira.c (max_regno_before_ira): Move from ...
15440 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15443 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15445 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15447 2013-02-13 Richard Biener <rguenther@suse.de>
15450 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15451 globals in MEM_REFs.
15453 2013-02-13 Richard Biener <rguenther@suse.de>
15455 * loop-init.c (loop_optimizer_init): Clear loop state when
15456 re-initializing preserved loops.
15457 * loop-unswitch.c (unswitch_single_loop): Return whether
15458 we unswitched the loop. Do not verify loop state here.
15459 (unswitch_loops): When we unswitched a loop discover new loops.
15461 2013-02-13 Kostya Serebryany <kcc@google.com>
15463 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15465 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15467 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15469 Avoid instrumenting duplicated memory access in the same basic block
15470 * Makefile.in (asan.o): Add new dependency on hash-table.h
15471 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15472 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15473 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15474 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15475 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15476 (get_mem_ref_of_assignment): New functions.
15477 (get_mem_refs_of_builtin_call): Extract from
15478 instrument_builtin_call and tweak a little bit to make it fit with
15480 (instrument_builtin_call): Use the new
15481 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15482 of is_gimple_builtin_call.
15483 (instrument_derefs, instrument_mem_region_access): Insert the
15484 instrumented memory reference into the hash table.
15485 (maybe_instrument_assignment): Renamed instrument_assignment into
15486 this, and change it to advance the iterator when instrumentation
15487 actually happened and return true in that case. This makes it
15488 homogeneous with maybe_instrument_assignment, and thus give a
15489 chance to callers to be more 'regular'.
15490 (transform_statements): Clear the memory reference hash table
15491 whenever we enter a new BB, when we cross a function call, or when
15492 we are done transforming statements. Use
15493 maybe_instrument_assignment instead of instrumentation. No more
15494 need to special case maybe_instrument_assignment and advance the
15495 iterator after calling it; it's now handled just like
15496 maybe_instrument_call. Update comment.
15498 2013-02-13 Richard Biener <rguenther@suse.de>
15500 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15501 Fix loop discovery code.
15503 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15505 PR inline-asm/56148
15506 * lra-constraints.c (process_alt_operands): Match early clobber
15507 operand with itself. Check conflicts with earlyclobber only if
15508 the operand is not reloaded. Prefer to reload conflicting operand
15509 if earlyclobber and matching operands are the same.
15511 2013-02-12 Richard Biener <rguenther@suse.de>
15514 * lto-streamer-out.c (write_symbol): Do not output symbols
15515 for hard register variables.
15517 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
15520 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15521 (umulsidi3_insn, mulsidi3_insn): New insns.
15523 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
15525 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15526 (struct tune_params): Add vec_costs field.
15527 * config/arm/arm.c (arm_builtin_vectorization_cost)
15528 (arm_add_stmt_cost): New functions.
15529 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15530 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15531 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15532 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15533 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15534 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15535 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15537 2013-02-12 Richard Biener <rguenther@suse.de>
15540 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15541 decls again if possible.
15543 2013-02-12 Richard Biener <rguenther@suse.de>
15545 PR middle-end/56288
15546 * tree-ssa.c (verify_ssa_name): Fix check, move
15547 SSA_NAME_IN_FREE_LIST check up.
15549 2013-02-12 Jakub Jelinek <jakub@redhat.com>
15550 Steven Bosscher <steven@gcc.gnu.org>
15552 PR rtl-optimization/56151
15553 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15554 equal to op0 or op1, and last_insn pattern is CODE operation
15555 with MEM dest and one of the operands matches that MEM.
15557 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15559 * doc/extend.texi: Document Function Multiversioning and "default"
15560 parameter string to target attribute.
15561 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15562 target attribute parameter is "default".
15563 (ix86_compare_version_priority): Remove checks for target attribute.
15564 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15565 Remove check for target attribute equal to NULL. Add assert.
15566 (ix86_generate_version_dispatcher_body): Change error to sorry.
15568 2013-02-11 Iain Sandoe <iain@codesourcery.com>
15569 Jack Howarth <howarth@bromo.med.uc.edu>
15570 Patrick Marlier <patrick.marlier@gmail.com>
15573 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15574 define ENDFILE_SPEC as TM_DESTRUCTOR.
15575 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15577 2013-02-11 Alexander Potapenko <glider@google.com>
15578 Jack Howarth <howarth@bromo.med.uc.edu>
15579 Jakub Jelinek <jakub@redhat.com>
15582 * config/darwin.c (cdtor_record): Rename ctor_record.
15583 (sort_cdtor_records): Rename sort_ctor_records.
15584 (finalize_dtors): New routine to sort destructors by
15585 priority before use in assemble_integer.
15586 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15588 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
15590 PR rtl-optimization/56275
15591 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15592 offset is non-negative and less than cmode size before
15593 calling simplify_subreg.
15595 2013-02-11 Richard Biener <rguenther@suse.de>
15597 PR tree-optimization/56264
15598 * cfgloop.h (fix_loop_structure): Adjust prototype.
15599 * loop-init.c (fix_loop_structure): Return the number of
15600 newly discovered loops.
15601 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15602 are discovered, do a full loop-closed SSA rewrite.
15604 2013-02-11 Richard Biener <rguenther@suse.de>
15606 PR tree-optimization/56273
15607 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15609 (check_array_ref): Fix missing newline in dumps.
15610 (search_for_addr_array): Likewise.
15612 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
15614 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15616 2013-02-09 Jakub Jelinek <jakub@redhat.com>
15619 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15621 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15623 PR rtl-optimization/56246
15624 * lra-constraints.c (simplify_operand_subreg): Try to reuse
15626 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15627 constraints are satisfied.
15629 2013-02-08 Jeff Law <law@redhat.com>
15632 * emit-rtl.c (reg_is_parm_p): New function.
15633 * regs.h (reg_is_parm_p): New prototype.
15634 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15635 callee-clobbered registers.
15637 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15640 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15641 If there is no implicit builtin declaration, just return NULL.
15643 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15645 * config/i386/sse.md (FMAMODEM): New mode iterator.
15646 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15647 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15649 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15651 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15652 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15653 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15655 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15657 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15658 (microblaze*-*-elf): Likewise.
15659 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15661 * config/microblaze/microblaze-c.c: Add builtin defines for
15662 _LITTLE_ENDIAN and _BIG_ENDIAN.
15663 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15664 add to TARGET_DEFAULT flags.
15665 Expand ASM_SPEC and LINK_SPEC.
15666 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15667 * config/microblaze/microblaze.md: Update extendsidi2 and
15668 movdi_internal instructions to use low-order / high-order reg
15670 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15671 options and inversemask / mask of LITTLE_ENDIAN.
15672 * config/microblaze/t-microblaze: Expand multilib options to
15673 include mlittle-endian (le) and update exceptions patterns.
15675 2013-02-08 Jakub Jelinek <jakub@redhat.com>
15677 PR rtl-optimization/56195
15678 * lra-constraints.c (get_reload_reg): Don't reuse regs
15679 if they have smaller mode than requested, if they have
15680 wider mode than requested, try to return a SUBREG.
15682 PR tree-optimization/56250
15683 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15684 if type is unsigned and code isn't MULT_EXPR.
15686 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15688 PR tree-optimization/56064
15689 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15690 bits according to mode.
15691 * fixed-value.h (fixed_from_double_int)
15692 (const_fixed_from_double_int): Adjust comments.
15694 2013-02-08 Richard Biener <rguenther@suse.de>
15697 * lto-streamer.h (struct data_in): Remove current_file, current_line
15698 and current_col members.
15699 * lto-streamer-out.c (lto_output_location): Stream changed bits
15700 en-block for efficiency.
15701 * lto-streamer-in.c (clear_line_info): Remove.
15702 (lto_input_location): Cache current file, line and column
15703 globally via local statics. Read changed bits en-block.
15704 (input_function): Do not call clear_line_info.
15705 (lto_read_body): Likewise.
15706 (lto_input_toplevel_asms): Likewise.
15708 2013-02-08 Michael Matz <matz@suse.de>
15710 PR tree-optimization/52448
15711 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15712 (nt_call_phase): New static.
15713 (add_or_mark_expr): Only mark accesses with newer phase than any
15715 (nonfreeing_call_p): New.
15716 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15717 (nt_fini_block): Keep blocks marked as visited.
15718 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15720 2013-02-08 Richard Biener <rguenther@suse.de>
15722 * ira.c (ira): Free broken dominator information.
15724 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15726 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15728 2013-02-08 Marek Polacek <polacek@redhat.com>
15730 * cfgloop.c (verify_loop_structure): Add more checking of headers.
15732 2013-02-08 Richard Biener <rguenther@suse.de>
15734 PR middle-end/56181
15735 * cfgloop.h (flow_loops_find): Adjust.
15736 (bb_loop_header_p): Declare.
15737 * cfgloop.c (bb_loop_header_p): New function split out from ...
15738 (flow_loops_find): ... here. Adjust function signature,
15739 support incremental loop structure update.
15740 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15741 * cfgloopmanip.c (fix_loop_structure): Move ...
15742 * loop-init.c (fix_loop_structure): ... here.
15743 (apply_loop_flags): Split out from ...
15744 (loop_optimizer_init): ... here.
15745 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15746 in incremental mode, only remove dead loops here.
15748 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15751 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15752 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15753 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15754 (*round<mode>3.libgcc): New insns for fixed-modes.
15755 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15756 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15757 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15758 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15759 implementations. Define to __builtin_avr_absFX,
15760 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15761 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15762 __builtin_avr_countlsFX, respectively.
15763 * config/avr/avr-c.c (target.h): Include it.
15764 (enum avr_builtin_id): New enum.
15765 (avr_resolve_overloaded_builtin): New static function.
15766 (avr_register_target_pragmas): Use it to set
15767 targetm.resolve_overloaded_builtin.
15768 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15769 tree nodes used by DEF_BUILTIN.
15770 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15771 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15772 <AVR_BUILTIN_xxBITS>: Same.
15774 2013-02-08 Richard Biener <rguenther@suse.de>
15776 * cfgloop.c (verify_loop_structure): Properly handle
15777 a loop exiting to another loop header.
15778 * ira-int.h (ira_loops): Remove.
15779 * ira.c (ira_loops): Remove.
15780 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15781 (do_reload): Use loop_optimizer_finalize.
15782 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15783 number_of_loops to access the loop tree.
15784 (more_one_region_p): Likewise.
15785 (finish_loop_tree_nodes): Likewise.
15786 (rebuild_regno_allocno_maps): Likewise.
15787 (mark_loops_for_removal): Likewise.
15788 (mark_all_loops_for_removal): Likewise.
15789 (remove_unnecessary_regions): Likewise.
15790 (ira_build): Likewise.
15791 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15793 2013-02-08 Richard Biener <rguenther@suse.de>
15795 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15796 * ipa-pure-const.c (analyze_function): Avoid calling
15797 mark_irreducible_loops twice.
15798 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
15800 2013-02-07 David S. Miller <davem@davemloft.net>
15802 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15804 * var-tracking.c (vt_add_function_parameter): Test the presence of
15805 HAVE_window_save properly and do not remap argument registers when
15806 we have a leaf function.
15808 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
15811 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15813 * config/i386/i386.c (ix86_print_operand): Ditto.
15815 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15817 * lra-constraints.c (process_alt_operands): Fix recently added comment.
15819 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15821 PR rtl-optimization/56225
15822 * lra-constraints.c (process_alt_operands): Check that reload hard
15823 reg can hold value for strict_low_part.
15825 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15828 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15829 dwarf2out_end_function.
15830 (in_first_function_p, maybe_at_text_label_p,
15831 first_loclabel_num_not_at_text_label): New variables.
15832 (dwarf2out_var_location): In the first function find out
15833 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
15834 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15837 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15839 PR rtl-optimization/56178
15840 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15841 SUBREG of a register. Tidy up related block of code.
15842 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15843 note if the source is a register or a SUBREG of a register.
15845 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15848 * config/rs6000/rs6000.md (ptrm): New mode attr.
15849 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15850 call_value_indirect_aix<pttrsize>,
15851 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15854 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15856 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15857 if -bnortl. Convert to strcmp and strncmp.
15859 2013-02-07 Alan Modra <amodra@gmail.com>
15862 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15863 addresses won't wrap when offsetting.
15864 (rs6000_secondary_reload): Provide secondary reloads needed for
15865 wrapping LO_SUM addresses.
15867 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15869 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15870 MACH, just __MACH__.
15872 2013-02-06 Richard Biener <rguenther@suse.de>
15874 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15875 instead of calling fix_loop_structure.
15877 2013-02-06 Jakub Jelinek <jakub@redhat.com>
15879 PR middle-end/56217
15880 * omp-low.c (use_pointer_for_field): Return false if
15881 lower_send_shared_vars doesn't generate any copy-out code.
15883 2013-02-06 Tom de Vries <tom@codesourcery.com>
15885 PR rtl-optimization/56131
15886 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15887 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15888 of the label is NULL. Add comment.
15890 2013-02-05 Jakub Jelinek <jakub@redhat.com>
15892 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15895 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15896 (STATIC_LIBTSAN_LIBS): Likewise.
15897 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15898 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15899 is defined, don't add anything else beyond that.
15900 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15901 (LINK_COMMAND_SPEC): Use them.
15903 PR tree-optimization/56205
15904 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15905 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15906 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15908 2013-02-05 Richard Biener <rguenther@suse.de>
15910 PR tree-optimization/53342
15911 PR tree-optimization/53185
15912 * tree-vectorizer.h (vect_check_strided_load): Remove.
15913 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15914 not disallow peeling for vectorized strided loads.
15915 (vect_check_strided_load): Make static and simplify.
15916 (vect_analyze_data_refs): Adjust.
15917 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15918 correctly when vectorizing strided loads.
15920 2013-02-05 Richard Biener <rguenther@suse.de>
15922 * doc/install.texi: Refer to ISL, not PPL.
15924 2013-02-05 Jan Hubicka <jh@suse.cz>
15926 PR tree-optimization/55789
15927 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15929 2013-02-05 Jan Hubicka <jh@suse.cz>
15931 PR tree-optimization/55789
15932 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15933 the dead call anyway.
15935 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15938 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15940 2013-02-04 Alexander Potapenko <glider@google.com>
15941 Jack Howarth <howarth@bromo.med.uc.edu>
15942 Jakub Jelinek <jakub@redhat.com>
15945 * config/darwin.c (sort_ctor_records): Stabilized qsort
15946 on constructor priority by using original position.
15947 (finalize_ctors): New routine to sort constructors by
15948 priority before use in assemble_integer.
15949 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15951 2013-02-04 Jakub Jelinek <jakub@redhat.com>
15954 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15955 about visibility on artificial decls.
15956 * config/sol2.c (solaris_assemble_visibility): Likewise.
15958 2013-02-04 Kai Tietz <ktietz@redhat.com>
15961 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15962 argument and improve checking of return-argument types for 16-byte
15964 (ix86_function_value_1): Add additional valtype argument on call
15965 of function_value_64.
15966 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
15967 handling infunction_value_64 function.
15969 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15971 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
15973 2013-02-04 Richard Biener <rguenther@suse.de>
15975 PR tree-optimization/56188
15976 * tree-ssa-structalias.c (label_visit): Consider case with
15977 initially non-empty points-to set.
15978 (perform_var_substitution): Dump node mapping and clean up.
15980 2013-02-04 Richard Guenther <rguenther@suse.de>
15983 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
15984 node prevail as last resort.
15985 (lto_symtab_merge_decls): Remove guard on LTRANS here.
15986 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
15988 2013-02-04 Richard Biener <rguenther@suse.de>
15990 PR tree-optimization/56113
15991 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
15993 (equiv_class_lookup_or_add): ... this.
15994 (label_visit): Adjust and fix error in previous patch.
15995 (perform_var_substitution): Adjust.
15997 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
15999 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16000 * config/sh/sh4-300.md: Likewise.
16001 * config/sh/sh4a.md: Likewise.
16002 * config/sh/constraints.md: Likewise.
16003 * config/sh/sh.md: Likewise.
16004 * config/sh/netbsd-elf.h: Likewise.
16005 * config/sh/predicates.md: Likewise.
16006 * config/sh/sh-protos.h: Likewise.
16007 * config/sh/ushmedia.h: Likewise.
16008 * config/sh/linux.h: Likewise.
16009 * config/sh/sh.c: Likewise.
16010 * config/sh/superh.h: Likewise.
16011 * config/sh/elf.h: Likewise.
16012 * config/sh/sh4.md: Likewise.
16013 * config/sh/sh.h: Likewise.
16015 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16017 * config/pa/constraints.md: Adjust unused letters. Change "T"
16018 constraint to match_test floating_point_store_memory_operand().
16019 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16020 (base14_operand): New.
16021 (floating_point_store_memory_operand): New.
16022 (integer_store_memory_operand): Revise to use base14_operand and
16023 reg_plus_base_memory_operand.
16024 (move_dest_operand): Allow symbolic_memory_operands.
16025 (symbolic_memory_operand): Check for LO_SOM.
16026 (symbolic_operand): Change default case to break.
16027 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16028 CONST_DOUBLE values to be reloaded by putting them into memory when
16029 the destination is a floating point register.
16030 (movdf): Remove code to handle CONST_DOUBLE.
16032 (reload_indf_r1): New.
16033 (reload_insf_r1): New.
16034 Consistently use "Q" and "T" constraints with integer and floating
16035 point move instructions, respectively.
16036 (movdi): Remove FAIL.
16037 Change predicate for source operand unamed DImode move from
16038 general_operand to move_src_operand.
16039 (umulsidi3): Change predicate for destination operand to
16041 Likewise for similar unamed patterns.
16042 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16043 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16044 (hppa_legitimize_address): Simplify mask calculation.
16045 (pa_emit_move_sequence): Revised handling of secondary reloads from
16046 REG+D addresses for floating point loads and stores. Directly handle
16047 loading CONST0_RTX (mode) to a floating point register.
16048 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16049 to floating point registers. Don't restrict secondary reloads to
16050 floating point registers to integer modes. Revise some comments and
16052 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16053 (pa_legitimate_address_p): New.
16054 (pa_legitimize_reload_address): New.
16055 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16056 (STRICT_REG_OK_FOR_BASE_P): New.
16057 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16058 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16060 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16061 Andrew Dixie <andrewd@gentrack.com>
16063 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16066 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16068 * expmed.c (extract_bit_field_1): Pass the full width of the
16069 structure to get_best_reg_extraction_insn.
16071 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16074 * configure.ac (use_cxa_atexit): Add AIX.
16075 * configure: Regenerate.
16077 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16079 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16082 * final.c (need_profile_function): New variable.
16083 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16084 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16085 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16086 notes, targetm.asm_out.function_prologue doesn't emit anything,
16087 HAVE_prologue and profiler should be emitted before prologue,
16088 set need_profile_function instead of emitting it.
16089 (final_scan_insn): If need_profile_function, emit
16090 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16091 NOTE_INSN_FUNCTION_BEG note.
16093 2013-02-01 Richard Henderson <rth@redhat.com>
16095 * config/rs6000/rs6000.md (smulditi3): New.
16098 * config/alpha/alpha.md (umulditi3): New.
16100 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16102 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16103 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16105 2013-02-01 Richard Biener <rguenther@suse.de>
16107 PR tree-optimization/56113
16108 * tree-ssa-structalias.c (label_visit): Reduce work for
16109 single-predecessor nodes.
16111 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
16113 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
16114 range isn't testing for zero.
16116 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
16118 PR middle-end/56113
16119 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
16121 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
16122 Nick Clifton <nickc@redhat.com>
16124 * config/v850/constraints.md (Q): Define as a memory constraint.
16125 * config/v850/predicates.md (label_ref_operand): New predicate.
16126 (e3v5_shift_operand): New predicate.
16127 (ior_operator): New predicate.
16128 * config/v850/t-v850: Add e3v5 multilib.
16129 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
16130 (v850_gen_movdi): Prototype.
16131 * config/v850/v850.c: Add support for e3v5 architecture.
16132 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
16134 (construct_save_jarl): Add e3v5 long JARL support.
16135 (v850_adjust_insn_length): New function. Adjust length of call
16136 insns when using e3v5 instructions.
16137 (v850_gen_movdi): New function: Generate instructions to move a
16139 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
16140 (CPP_SPEC): Define __v850e3v5__ as appropriate.
16141 (TARGET_USE_FPU): Enable for e3v5.
16142 (CONST_OK_FOR_W): New macro.
16143 (ADJUST_INSN_LENGTH): Define.
16144 * config/v850/v850.md (UNSPEC_LOOP): Define.
16145 (attr cpu): Add v850e3v5.
16146 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
16147 (movdi): New pattern.
16148 (movdi_internal): New pattern.
16149 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
16150 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
16151 (cstoresf4): Likewise.
16152 (cstoredf4): Likewise.
16153 (insv): New pattern.
16154 (rotlso3_a): New pattern.
16155 (rotlsi3_b): New pattern
16156 (rotlsi3_v850e3v5): New pattern.
16157 (doloop_begin): New pattern.
16158 (fix_loop_counter): New pattern.
16159 (doloop_end): New pattern.
16160 (branch_normal): Add e3v5 long branch support.
16161 (branch_invert): Likewise.
16162 (branch_z_normal): Likewise.
16163 (branch_z_invert): Likewise.
16164 (branch_nz_normal): Likewise.
16165 (branch_nz_invert): Likewise.
16166 (call_internal_short): Add e3v5 register-indirect JARL support.
16167 (call_internal_long): Likewise.
16168 (call_value_internal_short): Likewise.
16169 (call_value_internal_long): Likewise.
16170 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
16171 (mloop): New option.
16172 * config.gcc: Add support for configuring v840e3v5 target.
16173 * doc/invoke.texi: Document new v850 specific command line options.
16175 2013-01-31 Paul Koning <ni1d@arrl.net>
16179 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
16180 children if parent is a class.
16181 (prune_unused_types_prune): Don't add DW_AT_declaration.
16183 2013-01-31 Richard Biener <rguenther@suse.de>
16185 PR tree-optimization/56157
16186 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
16187 match up operand with SLP child.
16189 2013-01-31 Jason Merrill <jason@redhat.com>
16192 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
16193 parameters the first time.
16194 (gen_scheduled_generic_parms_dies): Check completeness here.
16196 2013-01-31 Richard Biener <rguenther@suse.de>
16198 PR middle-end/53073
16199 * common.opt (faggressive-loop-optimizations): New flag,
16200 enabled by default.
16201 * doc/invoke.texi (faggressive-loop-optimizations): Document.
16202 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
16203 infer_loop_bounds_from_undefined by it.
16205 2013-01-31 Richard Biener <rguenther@suse.de>
16207 PR tree-optimization/56150
16208 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
16209 visit virtual operands.
16210 (find_uses_to_rename_bb): Likewise.
16212 2013-01-31 Richard Biener <rguenther@suse.de>
16214 PR tree-optimization/56150
16215 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
16216 mixed store non-store stmts.
16218 2013-01-30 Jakub Jelinek <jakub@redhat.com>
16221 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
16222 LIBASAN_EARLY_SPEC is defined.
16223 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
16224 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
16226 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
16229 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
16230 invalid args instead of ICEing on it.
16231 (ix86_valid_target_attribute_tree): Return error_mark_node if
16232 ix86_valid_target_attribute_inner_p failed.
16233 (ix86_valid_target_attribute_p): Return false only if
16234 ix86_valid_target_attribute_tree returned error_mark_node. Allow
16235 target("default") attribute.
16236 (sorted_attr_string): Change argument from const char * to tree,
16237 merge in all target attribute arguments rather than just one.
16238 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
16239 instead of free. Avoid using strcat.
16240 (ix86_mangle_function_version_assembler_name): Mangle
16241 target("default") as if no target attribute is present. Adjust
16242 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
16243 instead of xmalloc and XDELETEVEC instead of free.
16244 (ix86_function_versions): Don't return true if one of the decls
16245 doesn't have target attribute. If they don't and one of the decls
16246 is DECL_FUNCTION_VERSIONED, report an error. Adjust
16247 sorted_attr_string caller. Use XDELETEVEC instead of free.
16248 (ix86_supports_function_versions): Remove.
16249 (make_name): Fix up formatting.
16250 (make_dispatcher_decl): Remove resolver_name and its initialization.
16251 Avoid leaking memory.
16252 (is_function_default_version): Return true if there is
16253 target("default") attribute rather than no target attribute at all.
16254 (make_resolver_func): Avoid leaking memory.
16255 (ix86_generate_version_dispatcher_body): Likewise.
16256 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
16257 * target.def (supports_function_versions): Remove.
16258 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
16259 * doc/tm.texi: Regenerated.
16261 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
16263 PR rtl-optimization/56144
16264 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
16265 for values with side effects.
16267 2013-01-30 Richard Biener <rguenther@suse.de>
16269 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
16270 (sparseset_pop): Likewise.
16271 * cfganal.c (compute_idf): Likewise. Increase work-stack size
16272 to be able to use quick_push in the worker loop.
16274 2013-01-30 Marek Polacek <polacek@redhat.com>
16276 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
16278 2013-01-30 Richard Biener <rguenther@suse.de>
16281 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
16283 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
16285 PR tree-optimization/56064
16286 * fixed-value.c (fixed_from_double_int): New function.
16287 * fixed-value.h (fixed_from_double_int): New prototype.
16288 (const_fixed_from_double_int): New static inline function.
16289 * fold-const.c (native_interpret_fixed): New static function.
16290 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
16291 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
16292 (native_encode_fixed): New static function.
16293 (native_encode_expr) <FIXED_CST>: Use it.
16294 (native_interpret_int): Move double_int worker code to...
16295 * double-int.c (double_int::from_buffer): ...this new static method.
16296 * double-int.h (double_int::from_buffer): Prototype it.
16298 2013-01-30 Richard Biener <rguenther@suse.de>
16300 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
16301 New pointer-map and obstack.
16302 (init_alias_vars): Allocate pointer-map and obstack.
16303 (delete_points_to_sets): Free them.
16304 (find_what_var_points_to): Cache result.
16305 (find_what_p_points_to): Adjust for changed interface of
16306 find_what_var_points_to.
16307 (compute_points_to_sets): Likewise.
16308 (ipa_pta_execute): Likewise.
16310 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16312 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
16313 * configure: Regenerate.
16314 * config.in: Regenerate.
16315 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
16316 #nobits/#progbits if supported.
16318 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
16321 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
16322 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
16323 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
16325 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16327 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
16328 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
16330 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16332 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
16333 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
16335 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16337 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
16339 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
16340 * config/arm/cortex-a7.md: New bypasses using
16341 arm_mac_accumulator_is_result.
16343 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16345 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
16346 (cortex_a7_neon_mla): Likewise.
16347 (cortex_a7_fpfmad): New reservation.
16348 (cortex_a7_fpmacs): Use ffmas and update required units.
16349 (cortex_a7_fpmuld): Update required units and latency.
16350 (cortex_a7_fpmacd): Likewise.
16351 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
16352 (cortex_a7_neon). Likewise.
16353 (bypass) Update participating units.
16355 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16357 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
16358 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
16360 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16361 (vfp_fmul): Use ffmad.
16362 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16363 (cortex_r4_fmacd): Use ffmad.
16364 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16365 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16366 (cortex_a9_fmacd): Use ffmad.
16367 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16368 (cortex_a8_vfp_macd): Use ffmad.
16369 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16370 (cortex_a5_fpmacd): Use ffmad.
16371 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16372 (cortex_a15_vfp_macd): Use ffmad.
16373 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16375 2013-01-29 Jason Merrill <jason@redhat.com>
16378 * varasm.c (default_assemble_visibility): Don't warn about
16379 visibility on artificial decls.
16381 2013-01-29 Richard Biener <rguenther@suse.de>
16383 PR tree-optimization/56113
16384 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16386 (label_visit): Free duplicate bitmaps and record the leader instead.
16387 (perform_var_substitution): Adjust.
16389 2013-01-29 Richard Biener <rguenther@suse.de>
16391 PR tree-optimization/55270
16392 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16393 the CFG, schedule loops for fixup.
16395 2013-01-29 Nick Clifton <nickc@redhat.com>
16397 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16400 2013-01-28 Leif Ekblad <leif@rdos.net>
16402 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16403 * config/i386/i386.h (TARGET_RDOS): New macro.
16404 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16405 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16406 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16407 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16408 DEFAULT_LARGE_SECTION_THRESHOLD.
16409 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16410 * config/i386/rdos.h: New file.
16411 * config/i386/rdos64.h: New file.
16413 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16416 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16419 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16421 PR rtl-optimization/56117
16422 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16423 call cselib_lookup_from_insn on the MEM before calling
16424 add_insn_mem_dependence.
16426 2013-01-28 Richard Biener <rguenther@suse.de>
16428 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16429 to a stmt that didn't have one.
16430 (copy_phis_for_bb): Likewise for PHI arguments.
16431 (copy_debug_stmt): Likewise for debug stmts.
16433 2013-01-28 Richard Biener <rguenther@suse.de>
16435 PR tree-optimization/56034
16436 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16437 (partition_builtin_p): Adjust.
16438 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16439 it is the last partition.
16440 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16441 up the vertex for the definition.
16442 (classify_partition): Classify whether a partition is a
16443 PKIND_REDUCTION, thus has uses outside of the loop.
16444 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16445 Merge all PKIND_REDUCTION partitions into the last partition.
16446 (tree_loop_distribution): Seed partitions from reductions as well.
16448 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16450 PR tree-optimization/56125
16451 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16452 pow(x,c) into sqrt(x) * powi(x, n/2) or
16453 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16454 optimizing for size.
16455 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16456 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16459 PR tree-optimization/56094
16460 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16461 to UNKNOWN_LOCATION while gimplifying expr.
16463 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16466 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16467 operand 0 in movabs insn template for -masm=intel asm alternative.
16468 (*movabs<mode>_2): Ditto for operand 1.
16470 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16473 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16476 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16478 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16481 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16483 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16484 Undef to avoid warning.
16486 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16488 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16489 * configure: Regenerate.
16491 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16493 PR tree-optimization/56098
16494 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16495 for stmts with volatile ops.
16496 (cond_store_replacement): Don't optimize if assign has volatile ops.
16497 (cond_if_else_store_replacement_1): Don't optimize if either
16498 then_assign or else_assign have volatile ops.
16499 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16502 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16504 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16506 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16508 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16509 missing ':' in asm example.
16511 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16513 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16514 entries into lane and laneq entries.
16515 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16516 Remove AdvSIMD scalar modes.
16517 (aarch64_sq<r>dmulh_laneq<mode>): New.
16518 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16520 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16521 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16522 * config/aarch64/iterators.md (VCOND): New.
16525 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16528 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16529 Add NULL LIBNAME argument to existing definitions.
16530 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16531 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16532 * config/avr/avr.c (DEF_BUILTIN): Same.
16533 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16534 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16535 implementation is available (DECL_ASSEMBLER_NAME is set).
16536 (avr_fold_absfx): New static function.
16537 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16538 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16539 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16540 AVR_BUILTIN_ABSLLK.
16541 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16542 (abshk, absk, abslk, absllk): Provide as static inline functions.
16544 2013-01-25 Marek Polacek <polacek@redhat.com>
16546 PR tree-optimization/56035
16547 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16549 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
16551 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16552 (*movtf_internal_rex64): Add (!o,C) alternative
16553 (*movxf_internal_rex64): Ditto.
16554 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16556 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
16558 * doc/invoke.texi: fix typo.
16559 * doc/objc.texi: fix typo.
16561 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16563 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16564 for the first two alternatives.
16566 2013-01-24 Diego Novillo <dnovillo@google.com>
16568 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
16569 (ggc-zone.o): Remove.
16570 * configure.ac: Remove option --with-gc.
16571 * configure: Re-generate.
16572 * doc/install.texi: Remove documentation for --with-gc.
16573 * gengtype.c (write_enum_defn): Remove. Update all users.
16574 (write_Types_process_field): Remove generation of gt_e_* argument.
16575 (output_type_enum): Remove. Update all users.
16576 (write_enum_defn): Remove. Update all users.
16577 (enum alloc_zone): Remove. Update all users.
16578 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
16579 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16580 Update all callers.
16581 (struct ptr_data): Remove field TYPE. Update all users.
16582 (gt_pch_note_object): Remove argument TYPE. Update all users.
16583 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16585 * ggc-none.c (ggc_alloc_typed_stat): Remove.
16586 (struct alloc_zone): Remove.
16587 (ggc_internal_alloc_zone_stat): Remove.
16588 (ggc_internal_cleared_alloc_zone_stat): Remove.
16589 * ggc-page.c (ggc_alloc_typed_stat): Remove.
16590 (ggc_pch_count_object): Remove last argument. Update all users.
16591 (ggc_pch_alloc_object): Remove last argument. Update all users.
16592 (struct alloc_zone): Remove.
16593 * ggc-zone.c: Remove.
16594 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
16595 (struct alloc_zone): Remove.
16596 (ggc_alloc_typed_stat): Remove.
16597 (ggc_alloc_typed): Remove.
16598 (ggc_splay_alloc): Remove first argument.
16599 (rtl_zone): Remove. Update all users.
16600 (tree_zone): Remove. Update all users.
16601 (tree_id_zone): Remove. Update all users.
16602 (ggc_internal_zone_alloc_stat): Remove. Update all users.
16603 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
16604 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
16605 * tree-ssanames.c: Remove references to zone allocator in comments.
16607 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
16609 * config/avr/avr.c (avr_out_fract): Make register numbers that
16610 might be outside of source operand signed.
16612 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
16614 * config/i386/constraints.md (Yf): New constraint.
16615 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16616 of f constraint to conditionaly disable x87 register preferences.
16617 (*movdf_internal): Ditto.
16618 (*movsf_internal): Ditto.
16620 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16622 PR inline-asm/55934
16623 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16624 that have operands with impossible constraints.
16625 Add a FIXME for a speed-up opportunity.
16626 * lra-constraints.c (process_alt_operands): Verify that a class
16627 selected from constraints on asms is valid for the operand mode.
16628 (curr_insn_transform): Remove incorrect comment.
16630 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
16632 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16633 TOC operand is a valid symbol ref in the constant pool.
16635 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16637 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
16639 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
16642 * config/avr/stdfix.h: New file.
16643 * t-avr (stdfix-gcc.h): New rule to build it.
16644 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16646 2013-01-23 Kostya Serebryany <kcc@google.com>
16648 * config/darwin.h: remove dependency on
16649 CoreFoundation (asan on Mac OS).
16651 2013-01-23 Jakub Jelinek <jakub@redhat.com>
16654 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16655 instead of cmpdi_operand for first comparison operand.
16656 Don't assert that comparison operands aren't both constants.
16658 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16660 * doc/install.texi (Downloading the Source): Update references to
16661 downloading separate components.
16663 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16665 * doc/extend.texi (__int128): Improve grammar.
16667 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
16670 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16671 alternative to (o,r).
16672 (*movdi_internal_rex64): Remove (!o,n) alternative.
16673 (DImode immediate->memory splitter): Remove.
16674 (DImode immediate->memory peephole2): Remove.
16675 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16676 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16677 alternative to (!o,*r).
16678 (*movtf_internal_sse): New pattern.
16679 (*movxf_internal_rex64): New pattern.
16680 (*movxf_internal): Disable for TARGET_64BIT.
16681 (*movdf_internal_rex64): Remove (!o,F) alternative.
16683 2013-01-22 Jakub Jelinek <jakub@redhat.com>
16685 PR middle-end/56074
16686 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16687 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16688 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16689 stmt locations where LOCATION_LOCUS of the stmt location is
16690 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16693 * config/i386/i386.md (UNSPEC_STOS): New.
16694 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16695 *strsetqi_1): Add UNSPEC_STOS.
16697 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16700 * doc/invoke.texi: Remove left over -Wsynth example.
16702 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16704 PR tree-optimization/56051
16705 * fold-const.c (fold_binary_loc): Don't fold
16706 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16707 a narrowing conversion, or widening conversion from signed
16710 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16712 PR rtl-optimization/56023
16713 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16714 dependent on debug instruction.
16716 2013-01-21 Martin Jambor <mjambor@suse.cz>
16718 PR middle-end/56022
16719 * function.c (allocate_struct_function): Call
16720 invoke_set_current_function_hook earlier.
16722 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16724 * reload1.c (init_reload): Only initialize reload_obstack
16725 during the first call.
16727 2013-01-21 Marek Polacek <polacek@redhat.com>
16729 * cfgloop.c (verify_loop_structure): Fix up grammar.
16731 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16733 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16734 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16736 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16739 * config/arm/marvell-pj4.md: Update copyright year.
16740 Fix up use of alu to alu_reg and simple_alu_imm.
16742 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16744 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16746 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16749 * lra-constraints.c (curr_insn_transform): Don't reuse original
16750 insn for secondary memory move when memory mode should be different.
16752 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16754 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16755 atomic_storedi_1): New patterns.
16757 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16759 btver2 pipeline descriptions.
16760 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16762 * config/i386/i386.md (btver2_decode): New type attributes.
16763 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
16765 * config/i386/btver2.md: New file describing btver2 pipelines.
16767 2013-01-19 Andrew Pinski <apinski@cavium.com>
16769 PR tree-optimization/52631
16770 * tree-ssa-sccvn (visit_use): Before looking up the original
16771 statement, try looking up the simplified expression.
16773 2013-01-19 Anthony Green <green@moxielogic.com>
16775 * config/moxie/moxie.c (moxie_expand_prologue): Set
16776 current_function_static_stack_size.
16778 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16780 PR tree-optimization/56029
16781 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16782 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16784 2013-01-18 Sharad Singhai <singhai@google.com>
16786 PR tree-optimization/55995
16787 * dumpfile.c (dump_loc): Print location only if available.
16788 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16790 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16793 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16794 secondary memory move.
16795 (inherit_reload_reg): Use rclass instead of cl for
16796 check_secondary_memory_needed_p.
16798 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16800 PR middle-end/56015
16801 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
16802 the case where writing real complex part of target modifies op1.
16804 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16806 * config/aarch64/aarch64-simd.md
16807 (aarch64_vcond_internal<mode>): Handle unordered cases.
16808 * config/aarch64/iterators.md (v_cmp_result): New.
16810 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16811 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16813 * config/arm/marvell-pj4.md: New file.
16814 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16815 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16816 (generic_vfp): Likewise.
16817 * config/arm/arm-cores.def: Add marvell-pj4.
16818 * config/arm/arm-tune.md: Regenerate.
16819 * config/arm/arm-tables.opt: Regenerate.
16820 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16821 * doc/invoke.texi: Document marvell-pj4.
16823 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16825 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16827 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16832 * var-tracking.c (negative_power_of_two_p): New.
16833 (global_get_addr_cache, local_get_addr_cache): New.
16834 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16835 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16837 (vt_stack_offset_p): Remove.
16838 (vt_canon_true_dep): Always canonicalize loc's address.
16839 (clobber_overlapping_mems): Make sure we have a MEM.
16840 (local_get_addr_clear_given_value): New.
16841 (val_reset): Clear local cached entries.
16842 (compute_bb_dataflow): Create and release the local cache.
16843 Disable duplicate MEMs clobbering.
16844 (emit_notes_in_bb): Clobber MEMs likewise.
16845 (vt_emit_notes): Create and release the local cache.
16846 (vt_initialize, vt_finalize): Create and release the global
16847 cache, respectively.
16848 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
16850 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16852 PR libmudflap/53359
16853 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16854 not found in the symtab.
16856 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16859 PR rtl-optimization/55547
16860 PR rtl-optimization/53827
16863 * alias.c (offset_overlap_p): New, factored out of...
16864 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16865 the conservative special case for symbolic constants. Don't
16866 adjust zero sizes on alignment.
16868 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16870 PR rtl-optimization/52573
16871 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16872 REG_UNUSED for the same register.
16874 2013-01-17 Richard Biener <rguenther@suse.de>
16875 Marek Polacek <polacek@redhat.com>
16877 PR rtl-optimization/55833
16878 * loop-unswitch.c (unswitch_loops): Move loop verification...
16879 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16880 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16881 Set it to true when we're removing a loop from hierarchy tree in
16882 an irreducible region.
16883 (fix_bb_placements): Adjust caller.
16884 (fix_loop_placements): Likewise.
16886 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
16888 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16889 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16890 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16891 Remove ID. Adjust comments.
16892 * config/avr/avr-c.c (avr_builtin_name): Remove.
16893 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16894 * config/avr/avr.c (avr_tolower): New static function.
16895 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16896 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16897 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16900 2013-01-17 Jan Hubicka <jh@suse.cz>
16902 PR tree-optimization/55273
16903 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16905 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
16908 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16909 store through atomic_store<mode>_1.
16910 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16912 2013-01-17 Martin Jambor <mjambor@suse.cz>
16914 PR tree-optimizations/55264
16915 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16916 for virtual methods.
16917 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16918 virtual methods before inlining is over.
16919 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16921 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16924 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16926 PR rtl-optimization/56005
16927 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16928 pending reads for prefetch.
16930 2013-01-16 Ian Bolton <ian.bolton@arm.com>
16932 * config/aarch64/aarch64.md
16933 (*cstoresi_neg_uxtw): New pattern.
16934 (*cmovsi_insn_uxtw): New pattern.
16935 (*<optab>si3_uxtw): New pattern.
16936 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16937 (*<optab>si3_insn_uxtw): New pattern.
16938 (*bswapsi2_uxtw): New pattern.
16940 2013-01-16 Richard Biener <rguenther@suse.de>
16942 * tree-inline.c (tree_function_versioning): Remove set but
16943 never used variable.
16945 2013-01-16 Richard Biener <rguenther@suse.de>
16947 PR tree-optimization/55964
16948 * tree-flow.h (rename_variables_in_loop): Remove.
16949 (rename_variables_in_bb): Likewise.
16950 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16951 (copy_loop_before): Adjust and delete update-ssa status.
16952 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
16953 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
16954 (rename_variables_in_loop): Remove.
16955 (slpeel_update_phis_for_duplicate_loop): Likewise.
16956 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16957 use available cfg machinery instead of duplicating it.
16958 Update PHI nodes and perform poor-mans SSA update here.
16959 (slpeel_tree_peel_loop_to_edge): Adjust.
16961 2013-01-16 Richard Biener <rguenther@suse.de>
16963 PR tree-optimization/54767
16964 PR tree-optimization/53465
16965 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
16966 (vrp_visit_phi_node): For PHI arguments coming via backedges
16967 drop all symbolical range information.
16968 (execute_vrp): Compute backedges.
16970 2013-01-16 Richard Biener <rguenther@suse.de>
16972 * doc/install.texi: Update CLooG and ISL requirements to
16975 2013-01-16 Christian Bruel <christian.bruel@st.com>
16978 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
16979 (broken_move): Handle UNSPECV_SP_SWITCH_B.
16980 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
16982 2013-01-16 DJ Delorie <dj@redhat.com>
16984 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
16985 (UNSPECV_SP_SWITCH_E): New.
16986 (sp_switch_1): Change to an unspec.
16987 (sp_switch_2): Change to an unspec. Don't use post-inc when we
16990 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
16992 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
16993 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
16994 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
16995 (expand_mem_thread_fence): Ditto.
16996 (expand_mem_signal_fence): Ditto.
16997 (expand_atomic_load): Ditto.
16998 (expand_atomic_store): Ditto.
17000 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17002 PR rtl-optimization/55547
17003 PR rtl-optimization/53827
17006 * alias.c (memrefs_conflict_p): Set sizes to negative after
17009 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17012 * function.c (thread_prologue_and_epilogue_insns): Always
17013 add crtl->drap_reg to set_up_by_prologue.set, even if
17014 stack_realign_drap is false.
17016 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17018 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17019 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17020 *call): Fix indention.
17022 2013-01-15 Tom de Vries <tom@codesourcery.com>
17025 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17028 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17030 PR rtl-optimization/55153
17031 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17033 2013-01-15 Martin Jambor <mjambor@suse.cz>
17035 PR tree-optimization/55920
17036 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17037 accesses as grp_to_be_debug_replaced.
17039 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17041 PR tree-optimization/55920
17042 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17043 there is non-useless type conversion needed from debug rhs to lhs,
17044 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17046 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17047 Mikael Pettersson <mikpe@it.uu.se>
17050 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17052 (ASM_OUTPUT_CASE_LABEL): Remove.
17053 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17054 * final.c (shorten_branches): Update alignment of labels before
17055 jump tables if CASE_VECTOR_SHORTEN_MODE.
17057 2013-01-15 Richard Biener <rguenther@suse.de>
17060 * system.h: Do not include gmp.h for building host tools.
17062 2013-01-15 Richard Biener <rguenther@suse.de>
17064 PR middle-end/55882
17065 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17066 account for bitpos when computing alignment.
17068 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17070 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17071 (ix86_target_macros_internal): Likewise.
17073 * config/i386/i386.c (m_CORE2I7): Removed.
17074 (m_CORE_HASWELL): New macro.
17075 (m_CORE_ALL): Likewise.
17076 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17077 (initial_ix86_arch_features): Likewise.
17078 (processor_target_table): Initializations for Core avx2.
17079 (cpu_names): New names "core-avx2".
17080 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17081 PROCESSOR_CORE_HASWELL.
17082 (ix86_issue_rate): New case.
17083 (ia32_multipass_dfa_lookahead): Likewise.
17084 (ix86_sched_init_global): Likewise.
17086 * config/i386/i386.h (TARGET_HASWELL): New macro.
17087 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17088 (processor_type): New PROCESSOR_HASWELL.
17090 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17092 PR tree-optimization/55955
17093 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17094 *SHIFT_EXPR and *ROTATE_EXPR codes.
17096 PR tree-optimization/48766
17097 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17098 -ftrapv disable -fwrapv.
17100 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17103 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17104 etc. to 1 and not to __flash.
17105 Use LL suffix for __INT24_MAX__ with -mint8.
17106 Use ULL suffix for __UINT24_MAX__ with -mint8.
17108 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17110 * config/avr/avr-arch.h
17111 (struct base_arch_s): Use typedef avr_arch_t instead.
17112 (struct arch_info_s): Use typedef avr_arch_info_t instead.
17113 (struct mcu_type_s): Use typedef avr_mcu_t instead.
17114 * config/avr/avr.c: Same.
17115 * config/avr/avr-devices.c: Same.
17116 * config/avr/driver-avr.c: Same.
17117 * config/avr/gen-avr-mmcu-texi.c: Same.
17118 * config/avr/avr-mcus.def: Adjust comment.
17120 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
17122 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
17123 * config/aarch64/iterators.md (VALLDI): New.
17125 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
17126 Andi Kleen <ak@linux.intel.com>
17129 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
17130 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
17133 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17135 * config/avr/avr-stdint.h: Remove trailing blanks.
17136 * config/avr/avr-log.h: Same.
17137 * config/avr/avr-arch.h: Same.
17138 * config/avr/avr-devices.c: Same.
17139 * config/avr/avr-dimode.md: Same.
17140 * config/avr/predicates.md: Same.
17141 * config/avr/avr-c.c: Same. And fix typo.
17143 * config/avr/avr-protos.h: Same. And:
17144 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
17145 (init_cumulative_args): Rename to avr_init_cumulative_args.
17146 (expand_prologue): Rename to avr_expand_prologue.
17147 (expand_epilogue): Rename to avr_expand_epilogue.
17148 (adjust_insn_length): Rename to avr_adjust_insn_length.
17149 (notice_update_cc): Rename to avr_notice_update_cc.
17150 (final_prescan_insn): Rename to avr_final_prescan_insn.
17151 * config/avr/avr.c: Same.
17152 * config/avr/avr.h: Same.
17153 * config/avr/avr.md: Remove trailing blanks.
17154 (prologue): Use avr_expand_prologue.
17155 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
17157 2013-01-14 Richard Biener <rguenther@suse.de>
17159 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
17160 verify_location, collect_subblocks): New functions.
17161 (verify_gimple_in_cfg): Verify that locations only reference
17162 BLOCKs in the functions BLOCK tree.
17164 2013-01-14 Richard Biener <rguenther@suse.de>
17166 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
17168 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
17170 (insert_out_of_ssa_copy_on_edge): Likewise.
17171 (rewrite_close_phi_out_of_ssa): Likewise.
17172 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
17174 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
17175 propagated constants.
17176 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
17179 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17181 * config/avr/avr-modes.def: Add GPL copyright notice.
17183 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
17185 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
17186 MEMMODEL_MASK to determine memory model.
17187 (atomic_store<mode>): Ditto from operands[2].
17188 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
17190 2013-01-13 Jakub Jelinek <jakub@redhat.com>
17193 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
17194 (fold_gimple_assign): Don't call unshare_expr here.
17195 (fold_ctor_reference): Call unshare_expr.
17197 2013-01-13 Terry Guo <terry.guo@arm.com>
17199 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
17200 * doc/fragments.texi: Document MULTILIB_REUSE.
17201 * gcc.c (multilib_reuse): New internal spec.
17202 (set_multilib_dir): Also search multilib from multilib_reuse.
17203 * genmultilib (tmpmultilib3): Refactor code.
17204 (tmpmultilib4): Ditto.
17205 (multilib_reuse): New multilib argument.
17207 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
17209 * Makefile.in: Update copyright.
17211 2013-01-12 Tom de Vries <tom@codesourcery.com>
17213 PR middle-end/55890
17214 * calls.c (expand_call): Check if arg_nr is valid.
17216 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17218 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
17219 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
17220 documentation. Add missing '__' in front of
17221 __builtin_ia32_packssdw256.
17223 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17226 * config/s390/s390.c (s390_preferred_reload_class): Do not return
17227 NO_REGS for larl operands.
17228 (s390_reload_larl_operand): Use s390_load_address instead of
17231 2013-01-11 Richard Biener <rguenther@suse.de>
17233 * tree-cfg.c (verify_node_sharing_1): Split out from ...
17234 (verify_node_sharing): ... here.
17235 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
17237 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
17239 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
17240 Substitute TREECHECKING.
17241 * configure: Regenerate.
17242 * Makefile.in (TREECHECKING): New.
17244 2013-01-11 Richard Guenther <rguenther@suse.de>
17246 PR tree-optimization/44061
17247 * tree-vrp.c (extract_range_basic): Compute zero as
17248 value-range for __builtin_constant_p of function parameters.
17250 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
17252 Update copyright years.
17254 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
17256 PR rtl-optimization/55672
17257 * lra-eliminations.c (mark_not_eliminable): Permit addition with
17258 const to be eliminable.
17260 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
17262 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
17263 * configure: Regenerate.
17265 2013-01-10 Richard Biener <rguenther@suse.de>
17267 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
17269 2013-01-10 Richard Biener <rguenther@suse.de>
17272 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
17273 locations for virtual PHI arguments.
17274 (rewrite_update_phi_arguments): Likewise.
17276 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
17278 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
17281 2013-01-10 Jakub Jelinek <jakub@redhat.com>
17283 PR tree-optimization/55921
17284 * tree-complex.c (expand_complex_asm): New function.
17285 (expand_complex_operations_1): Call it for GIMPLE_ASM.
17287 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17290 * config/s390/s390.c (s390_symref_operand_p)
17291 (s390_loadrelative_operand_p): Merge the two functions.
17292 (s390_check_qrst_address, print_operand_address): Add parameters
17293 to s390_loadrelative_operand_p invokation.
17294 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
17295 (s390_reload_larl_operand, s390_secondary_reload): Use
17296 s390_loadrelative_operand_p instead of s390_symref_operand_p.
17297 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
17299 2013-01-09 Mike Stump <mikestump@comcast.net>
17301 * dse.c (record_store): Remove unnecessary assert.
17303 2013-01-09 Jan Hubicka <jh@suse.cz>
17305 PR tree-optimization/55569
17306 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
17307 * cfgloop.h (scale_loop_profile): Likewise.
17309 2013-01-09 Jan Hubicka <jh@suse.cz>
17312 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
17314 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
17316 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
17318 PR middle-end/55114
17319 * expr.h (maybe_emit_group_store): Declare.
17320 * expr.c (maybe_emit_group_store): New function.
17321 * builtins.c (expand_builtin_int_roundingfn): Call it.
17322 (expand_builtin_int_roundingfn_2): Likewise.
17324 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
17326 PR rtl-optimization/55829
17327 * lra-constraints.c (match_reload): Add code for absent output.
17328 (curr_insn_transform): Add code for reloads of matched inputs
17331 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
17333 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
17334 attribute of movddup insn to DF.
17335 (*vec_interleave_lowv2df): Ditto.
17336 (vec_dupv2df): Ditto.
17338 2013-01-09 Jan Hubicka <jh@suse.cz>
17340 PR tree-optimiation/55875
17341 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
17342 EVERY_ITERATION parameter.
17343 (number_of_iterations_exit): Check if exit is executed every iteration.
17344 (idx_infer_loop_bounds): Similarly here.
17345 (n_of_executions_at_most): Simplify
17346 to only test for cases where statement is dominated by the
17347 particular bound; handle correctly the "postdominance" test.
17348 (scev_probably_wraps_p): Use max loop iterations info
17349 as a global bound first.
17351 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
17352 Nick Clifton <nickc@redhat.com>
17354 * config/v850/v850.md (cbranchsf4): New pattern.
17355 (cstoresf4): New pattern.
17356 (cbranchdf4): New pattern.
17357 (cstoredf4): New pattern.
17358 (movsicc): Disallow floating point comparisons.
17359 (cmpsf_le_insn): Fix order of operators.
17360 (cmpsf_lt_insn): Likewise.
17361 (cmpsf_eq_insn): Likewise.
17362 (cmpdf_le_insn): Likewise.
17363 (cmpdf_lt_insn): Likewise.
17364 (cmpdf_eq_insn): Likewise.
17365 (cmpsf_ge_insn): Use LE comparison.
17366 (cmpdf_ge_insn): Likewise.
17367 (cmpsf_gt_insn): Use LT comparison.
17368 (cmpdf_gt_insn): Likewise.
17369 (cmpsf_ne_insn): Delete pattern.
17370 (cmpdf_ne_insn): Delete pattern.
17371 * config/v850/v850.c (v850_gen_float_compare): Use
17372 gen_cmpdf_eq_insn for NE comparison.
17373 (v850_float_z_comparison_operator)
17374 (v850_float_nz_comparison_operator): Move from here ...
17375 * config/v850/predicates.md: ... to here. Move GT and GE
17376 comparisons into v850_float_z_comparison_operator.
17377 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17379 (v850_float_nz_comparison_operator): Likewise.
17381 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17383 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17384 with calls to gen_insvsi/gen_insvdi.
17386 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17388 * config/i386/i386.c (initial_ix86_tune_features): Set up
17389 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17391 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17392 Jakub Jelinek <jakub@redhat.com>
17394 PR tree-optimization/48189
17395 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17396 If nitercst is 0, don't predict the exit edge.
17398 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17400 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17401 in asm_fprintf with reg_names.
17402 (aarch64_print_operand_address): Likewise.
17403 (aarch64_return_addr): Likewise.
17404 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17406 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17408 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17409 (INT_U6_BITS): Likewise.
17410 * config/pa/predicates.md (uint6_operand): New predicate.
17411 (shift5_operand, shift6_operand): Likewise.
17412 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17414 (lshrdi3): Use shift6_operand.
17415 (shrpsi4, shrpdi4): New insn patterns.
17416 (extzv): Delete expander.
17417 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17418 predicates in unamed zero extract patterns. Tighten common constraint.
17419 (extv): Delete expander.
17420 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17421 predicates in unamed sign extract patterns. Tighten common constraint.
17422 (insv): Delete expander.
17423 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17424 predicates in unamed insert patterns. Tighten common constraint.
17425 Change uint32_operand predicate to uint6_operand predicate in unamed
17426 DImode pattern to insert constant values of type 1...1xxxx.
17428 2013-01-04 Jan Hubicka <jh@suse.cz>
17430 PR tree-optimization/55823
17431 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17434 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17435 Uros Bizjak <ubizjak@gmail.com>
17437 PR rtl-optimization/55845
17438 * df-problems.c (can_move_insns_across): Stop scanning at
17439 volatile_insn_p source instruction or give up if
17440 across_from .. across_to range contains any volatile_insn_p
17443 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17445 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17446 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17448 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17449 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17451 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17454 * asan.c (asan_clear_shadow): New function.
17455 (asan_emit_stack_protection): Use it.
17457 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17459 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17460 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17461 with tab instead of space.
17463 2013-01-08 Nick Clifton <nickc@redhat.com>
17465 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17466 register bank 0 at the start of an interrupt handler.
17467 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17470 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17472 * config/aarch64/aarch64-simd.md
17473 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17474 (aarch64_simd_bsl): Likewise.
17475 (aarch64_vcond_internal<mode>): Likewise.
17476 (vcond<mode><mode>): Likewise.
17477 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17478 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17480 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17482 * config/aarch64/aarch64-builtins.c
17483 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17485 2013-01-08 Martin Jambor <mjambor@suse.cz>
17488 * tree-sra.c (analyze_access_subtree): Return true also after
17489 potentially creating a debug-only replacement.
17491 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17493 PR middle-end/55890
17494 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17496 PR tree-optimization/54120
17497 * tree-vrp.c (range_fits_type_p): Don't allow
17498 src_precision < precision from signed vr to unsigned_p
17499 if vr->min or vr->max is negative.
17500 (simplify_float_conversion_using_ranges): Test can_float_p
17501 against CODE_FOR_nothing.
17503 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17504 Richard Biener <rguenther@suse.de>
17506 PR middle-end/55851
17507 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17508 types instead of just INTEGER_TYPE types.
17510 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
17512 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17515 2013-01-07 Steve Ellcey <sellcey@mips.com>
17518 * config/mips/mips.opt: Change mad to mmad to match documentation.
17520 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17523 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17524 .progmemx.data now.
17526 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17529 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17530 (avr_addrspace_t): Add .section_name field.
17531 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17533 (avr_addrspace): Same. Initialize .section_name. Remove last
17534 NULL entry. Put __memx into .progmemx.data.
17535 (progmem_section_prefix): Remove.
17536 (avr_asm_init_sections): No need to initialize progmem_section.
17537 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17538 section name prefix.
17539 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17540 retrieve the progmem section.
17541 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17542 boundary to run over avr_addrspace[].
17543 (avr_register_target_pragmas): Ditto.
17545 2013-01-06 Jakub Jelinek <jakub@redhat.com>
17547 * varasm.c (output_constant_def_contents): For asan_protect_global
17548 protected strings, adjust DECL_ALIGN if needed, before testing for
17550 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17551 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17553 (output_object_block): For asan protected decls, emit asan padding
17554 after their contents.
17555 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17556 (asan_finish_file): Test it here instead.
17558 2013-01-07 Nick Clifton <nickc@redhat.com>
17559 Matthias Klose <doko@debian.org>
17560 Doug Kwan <dougkwan@google.com>
17561 H.J. Lu <hongjiu.lu@intel.com>
17564 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17566 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17568 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17570 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
17572 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17574 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17577 * doc/install.texi (Cross-Compiler-Specific Options): Document
17580 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17582 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17583 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17584 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17585 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17586 vqmovun_high_s64): Fix source operand number and update copyright.
17588 2013-01-07 Richard Biener <rguenther@suse.de>
17590 PR middle-end/55890
17591 * gimple.h (gimple_call_builtin_p): New overload.
17592 * gimple.c (validate_call): New function.
17593 (gimple_call_builtin_p): Likewise.
17594 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17595 Use gimple_call_builtin_p.
17596 (find_func_clobbers): Likewise.
17597 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17598 (strlen_optimize_stmt): Likewise.
17600 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17602 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17603 (vld1q_dup_*): Likewise.
17604 (vld1_*): Likewise.
17605 (vld1q_*): Likewise.
17606 (vld1_lane_*): Likewise.
17607 (vld1q_lane_*): Likewise.
17609 2013-01-07 Richard Biener <rguenther@suse.de>
17611 * lto-streamer.h (LTO_minor_version): Bump to 2.
17613 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17615 * config/aarch64/aarch64-protos.h
17616 (aarch64_const_double_zero_rtx_p): Rename to...
17617 (aarch64_float_const_zero_rtx_p): ...this.
17618 (aarch64_float_const_representable_p): New.
17619 (aarch64_output_simd_mov_immediate): Likewise.
17620 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17621 move immediate case.
17622 * config/aarch64/aarch64.c
17623 (aarch64_const_double_zero_rtx_p): Rename to...
17624 (aarch64_float_const_zero_rtx_p): ...this.
17625 (aarch64_print_operand): Allow printing of new constants.
17626 (aarch64_valid_floating_const): New.
17627 (aarch64_legitimate_constant_p): Check for valid floating-point
17629 (aarch64_simd_valid_immediate): Likewise.
17630 (aarch64_vect_float_const_representable_p): New.
17631 (aarch64_float_const_representable_p): Likewise.
17632 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17633 (aarch64_output_simd_mov_immediate): New.
17634 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17635 (*movdf_aarch64): Likewise.
17636 * config/aarch64/constraints.md (Ufc): New.
17637 (Y): call aarch64_float_const_zero_rtx.
17638 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17640 2013-01-07 Richard Biener <rguenther@suse.de>
17642 PR tree-optimization/55888
17643 PR tree-optimization/55862
17644 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17645 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17646 not if it is contained therein.
17648 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17650 * config/avr/t-avr: Typo.
17652 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17655 * config/avr/t-avr: Don't automatically rebuild
17656 $(srcdir)/config/avr/t-multilib
17657 $(srcdir)/config/avr/avr-tables.opt
17658 $(srcdir)/doc/avr-mmcu.texi
17659 (avr-mcus): New phony target to build them on request.
17660 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17661 * avr/avr-mcus.def: Adjust comments.
17663 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
17665 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17667 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17669 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17671 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17673 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17675 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
17677 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17678 to generate profiling.
17679 * config/rs6000/aix64.h (LIB_SPEC): Same.
17681 2013-01-04 Andrew Pinski <apinski@cavium.com>
17683 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17685 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17687 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
17689 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17691 (ix86_expand_move): Ditto.
17692 (ix86_zero_extend_to_Pmode): Ditto.
17693 (ix86_expand_call): Ditto.
17694 (ix86_expand_special_args_builtin): Ditto.
17695 (ix86_expand_builtin): Ditto.
17697 2013-01-04 Richard Biener <rguenther@suse.de>
17699 PR tree-optimization/55862
17700 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17701 translating them through PHI nodes.
17703 2013-01-04 Martin Jambor <mjambor@suse.cz>
17705 PR tree-optimization/55755
17706 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17707 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17709 2013-01-04 Marek Polacek <polacek@redhat.com>
17711 PR middle-end/55859
17712 * opts.c (default_options_optimization): Clarify error message.
17714 2013-01-04 Richard Biener <rguenther@suse.de>
17716 PR middle-end/55863
17717 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17720 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17723 * config/pa/pa.md (movsi): Revert previous change.
17724 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17727 2013-01-03 Richard Henderson <rth@redhat.com>
17729 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17730 after eliminating TLS symbols.
17732 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
17735 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17736 * graphite-poly.c (debug_gmp_value): Likewise.
17738 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
17741 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17742 selected code model, define __code_mode_small__, __code_model_medium__,
17743 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17744 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17745 xchg temporary register with %k. Declare temporary register as
17747 [__x86_64__]: For medium and large code models, preserve %rbx register.
17749 2013-01-03 Richard Biener <rguenther@suse.de>
17751 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
17752 (dump_subscript): Adjust.
17753 (finalize_ddr_dependent): Do not dump redundant info.
17754 (analyze_siv_subscript): Adjust.
17755 (subscript_dependence_tester): Likewise.
17756 (compute_affine_dependence): Likewise.
17758 2013-01-03 Richard Biener <rguenther@suse.de>
17761 2013-01-03 Richard Biener <rguenther@suse.de>
17763 PR tree-optimization/55857
17764 * tree-vect-stmts.c (vectorizable_load): Do not setup
17765 re-alignment for invariant loads.
17767 2013-01-02 Richard Biener <rguenther@suse.de>
17769 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17770 invariant load do not generate a vector load from the scalar location.
17772 2013-01-03 Richard Biener <rguenther@suse.de>
17774 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17775 for not vectorizing.
17776 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17777 not build INDIRECT_REFs, call get_name once only.
17778 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17779 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17781 2013-01-03 Richard Biener <rguenther@suse.de>
17783 PR tree-optimization/55857
17784 * tree-vect-stmts.c (vectorizable_load): Do not setup
17785 re-alignment for invariant loads.
17787 2013-01-03 Richard Biener <rguenther@suse.de>
17790 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17791 prefer a built-in decl.
17793 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17795 * gcc.c (process_command): Update copyright notice dates.
17796 * gcov.c (print_version): Likewise.
17797 * gcov-dump.c (print_version): Likewise.
17799 PR rtl-optimization/55838
17800 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17801 iv0.step, iv1.step and step.
17803 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17804 Marc Glisse <marc.glisse@inria.fr>
17806 PR tree-optimization/55832
17807 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17808 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17809 integer_{one,zero}_node.
17811 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17814 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17815 * var-tracking.c (reverse_op): Don't add reverse ops to
17816 VALUEs that have already
17817 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
17819 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17821 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17823 2013-01-02 Teresa Johnson <tejohnson@google.com>
17825 * dumpfile.c (dump_loc): Print filename with location.
17826 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17827 new location_t parameter to emit complete unroll message with
17828 new dump framework.
17829 (canonicalize_loop_induction_variables): Compute loops location
17830 and pass to try_unroll_loop_completely.
17831 * loop-unroll.c (report_unroll_peel): New function.
17832 (peel_loops_completely): Use new dump format with location
17833 for main dumpfile message, and invoke report_unroll_peel on success.
17834 (decide_unrolling_and_peeling): Ditto.
17835 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17836 by report_unroll_peel.
17837 (decide_peel_completely): Ditto.
17838 (decide_unroll_constant_iterations): Ditto.
17839 (decide_unroll_runtime_iterations): Ditto.
17840 (decide_peel_simple): Ditto.
17841 (decide_unroll_stupid): Ditto.
17842 * cfgloop.c (get_loop_location): New function.
17843 * cfgloop.h (get_loop_location): Declare.
17845 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17847 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17850 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17852 PR middle-end/55198
17853 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17854 BLKmode objects when EXPAND_MEMORY is specified.
17856 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17858 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17860 (fold_builtin_cpu): Do not share cpu model decls across statements.
17862 2013-01-02 Jason Merrill <jason@redhat.com>
17865 * tree.c (build_array_type_1): Revert earlier change.
17867 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17869 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17871 * config/aarch64/aarch64-tune.md: Re-generate.
17873 2013-01-02 Richard Biener <rguenther@suse.de>
17875 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17876 invariant load do not generate a vector load from the scalar location.
17878 2013-01-02 Richard Biener <rguenther@suse.de>
17881 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17882 * configure: Regenerate.
17884 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17886 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17887 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17888 (expand_builtin_int_roundingfn_2): Keep the original target around
17889 for the fallback case.
17891 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17893 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17894 to be clear for sign changes.
17896 2013-01-01 Jan Hubicka <jh@suse.cz>
17898 * ipa-inline-analysis.c: Fix formatting.
17900 2013-01-01 Jakub Jelinek <jakub@redhat.com>
17902 PR tree-optimization/55831
17903 * tree-vect-loop.c (get_initial_def_for_induction): Use
17904 gsi_after_labels instead of gsi_start_bb.
17906 Copyright (C) 2013 Free Software Foundation, Inc.
17908 Copying and distribution of this file, with or without modification,
17909 are permitted in any medium without royalty provided the copyright
17910 notice and this notice are preserved.