1 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
3 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
4 iterator instead of X87MODEF.
6 2013-08-12 Perez Read <netfirewall@gmail.com>
9 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
10 operand 0 for intel asm alternative.
11 (*movabs<mode>_2): Ditto for operand 1.
13 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
15 * config/aarch64/arm_none.h
16 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
18 2013-08-12 Nick Clifton <nickc@redhat.com>
20 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
22 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
24 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
25 expand for QI/HImode operand to produce more effictive code for
26 unsigned char(short) --> float(double) conversion.
28 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
30 * doc/invoke.texi: Mention that -ftls-model does not force the final
33 2013-08-12 Marek Polacek <polacek@redhat.com>
34 Marc Glisse <marc.glisse@inria.fr>
36 PR tree-optimization/57980
37 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
38 when creating -1 constant.
40 2013-08-10 Jan Hubicka <jh@suse.cz>
42 Workaround binutils PR14342.
43 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
44 (gimple_init_edge_profiler): Likewise.
45 (gimple_gen_ic_func_profiler): Likewise.
47 2013-08-09 Jan Hubicka <jh@suse.cz>
49 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
51 2013-08-09 Xinliang David Li <davidxl@google.com>
53 * config/i386/stringop.def: New file.
54 * config/i386/stringop.opt: New file.
55 * config/i386/i386-opts.h: Include stringopt.def.
56 * config/i386/i386.opt: Include stringopt.opt.
57 * config/i386/i386.c (ix86_option_override_internal):
58 Override default size based stringop inline strategies with options.
59 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
62 2013-08-09 Jan Hubicka <jh@suse.cz>
64 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
66 2013-08-09 Jan Hubicka <jh@suse.cz>
68 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
70 (dump_cgraph_node): Dump profile-id.
71 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
72 and common_target_probability.
73 * lto-cgraph.c (lto_output_edge): Stream common targets.
74 (lto_output_node): Stream profile ids.
75 (input_node): Stream profile ids.
76 (input_edge): Stream common targets.
77 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
78 * ipa.c: Include value-prof.h
79 (ipa_profile_generate_summary): Turn indirect call statement histograms
81 (ipa_profile): Turn common targets into speculative edges.
83 2013-08-09 Jan Hubicka <jh@suse.cz>
85 * cgraph.h (cgraph_node): Add profile_id.
86 * value-prof.c (cgraph_node_map): Turn into pointer_map.
87 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
88 (del_node_map): Update.
89 (find_func_by_funcdef_no): Replace by ...
90 (find_func_by_profile_id): ... this one.
91 (gimple_ic_transform): Do not remove useful histograms when
92 speculation is not done; dump info when indirect call removal
94 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
95 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
96 (__gcov_indirect_call_profiler_v2): .. this one.
97 * profile.h (init_node_map): Update.
98 * coverage.c (coverage_compute_profile_id): New function.
99 * coverage.h (coverage_compute_profile_id): Declare.
100 * tree-profile.c (init_ic_make_global_vars): Make
101 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
102 (gimple_init_edge_profiler): Update prototype of
103 __gcov_indirect_call_profiler.
104 (gimple_gen_ic_func_profiler): Simplify.
105 (tree_profiling): Use init_node_map
107 2013-08-09 Jan Hubicka <jh@suse.cz>
109 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
110 non-speculative refs.
111 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
112 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
113 (cgraph_set_call_stmt): Likewise.
114 (cgraph_create_edge_1): Fix release checking compilatoin;
116 (cgraph_free_edge): Free indirect info.
117 (cgraph_turn_edge_to_speculative): New function.
118 (cgraph_speculative_call_info): New function.
119 (cgraph_make_edge_direct): Return direct edge; handle speculation.
120 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
121 (dump_cgraph_node): Dump speculation.
122 (verify_edge_count_and_frequency): Accept speculative edges.
123 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
124 (verify_cgraph_node): Handle speculation.
125 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
126 (cgraph_set_call_stmt): Update prototype.
127 (cgraph_make_edge_direct): Update prototype.
128 (cgraph_speculative_call_info): Declare.
129 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
130 to change; update call of ipa_find_references.
131 * ipa-ref.c (ipa_record_reference): Fix return value; clear
132 lto_stmt_uid and speculative flags.
133 (ipa_dump_references): Dump speculation.
134 (ipa_clone_references): Clone speculative flag.
135 (ipa_clone_referring): Likewise.
136 (ipa_clone_ref): New function.
137 (ipa_find_reference): Look into lto_stmt_uids
138 (ipa_clear_stmts_in_references): Do not clear speculative calls.
139 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
140 (ipa_find_reference): Update declaration.
141 (ipa_clone_ref): Declare.
142 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
143 stream speculative flag.
144 (lto_output_ref): Stream statements uids and speculation.
145 (input_ref): Likewise.
146 (input_edge): Stream speuclation.
147 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
148 (cgraph_set_call_stmt_including_clones): Handle speculation.
149 * ipa-inline.c (heap_edge_removal_hook): New function.
150 (inline_small_functions): Register it.
151 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
152 also initialize refs.
153 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
155 (try_make_edge_direct_simple_call): Likewise.
156 (try_make_edge_direct_simple_call): Likewise.
157 (update_indirect_edges_after_inlining): Likewise.
158 (remove_described_reference): Look proper lto_stmt_uid.
159 (propagate_controlled_uses): Likewise.
160 (propagate_controlled_uses): Liekwise.
161 * tree-inline.c (copy_bb): Copy speculative edges.
162 (redirect_all_calls): New function.
163 (copy_cfg_body): Do redirection after loop info is updated.
164 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
166 2013-08-09 Jan Hubicka <jh@suse.cz>
168 * lto-streamer-out.c (output_function): Renumber PHIs.
169 * lto-streamer-in.c (input_function): Likewise.
171 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
173 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
174 (get_lane_unsigned): Likewise.
175 (dup_lane_scalar): Likewise.
176 (get_lane): enable for VALL.
177 * config/aarch64/aarch64-simd.md
178 (aarch64_dup_lane_scalar<mode>): Remove.
179 (aarch64_get_lane_signed<mode>): Likewise.
180 (aarch64_get_lane_unsigned<mode>): Likewise.
181 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
182 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
183 (aarch64_get_lane<mode>): Enable for all vector modes.
184 (aarch64_get_lanedi): Remove misleading constraints.
185 * config/aarch64/arm_neon.h
186 (__aarch64_vget_lane_any): Define.
187 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
188 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
189 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
190 * config/aarch64/iterators.md (VDQQH): New.
194 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
196 * configure.ac: Add GAS check for LEON instructions on SPARC.
197 * configure: Regenerate.
198 * config.in: Likewise.
199 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
201 * config/sparc/sparc.opt (LEON, LEON3): New masks.
202 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
204 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
205 (AS_LEON_FLAG): New macro.
206 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
207 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
208 Deal with LEON and LEON3 for the memory model.
209 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
210 (atomic_compare_and_swap<mode>_1): Likewise.
211 (*atomic_compare_and_swap<mode>_1): Likewise.
213 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
215 * config/arm/neon.md (vcond): Fix floating-point vector
216 comparisons against 0.
218 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
220 * lra-constraints.c (emit_spill_move): Remove assert.
221 (process_alt_operands): Add more debugging
222 output. Increase reject for spilling into memory. Decrease
223 reject for reloading scratch.
224 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
226 2013-08-08 Steve Ellcey <sellcey@mips.com>
228 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
229 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
230 micromips incompatible. Add nan2008.
231 (MULTILIB_DIRNAMES): Add nan2008.
232 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
233 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
234 and micromips incompatible. Add nan2008.
235 (MULTILIB_DIRNAMES): Add nan2008.
236 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
238 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
240 PR rtl-optimization/58079
241 * combine.c (combine_simplify_rtx): Avoid using SUBST if
242 simplify_comparison has widened a comparison with an integer.
244 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
246 * config/arm/neon.md (movmisalign<mode>): Disable when we
247 don't allow unaligned accesses.
248 (*movmisalign<mode>_neon_store): Likewise.
249 (*movmisalign<mode>_neon_load): Likewise.
250 (*movmisalign<mode>_neon_store): Likewise.
251 (*movmisalign<mode>_neon_load): Likewise.
253 2013-08-08 Jan Hubicka <jh@suse.cz>
255 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
256 (make_pass_rebuild_cgraph_edges): Also clear references.
257 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
258 * ipa-inline-transform.c (inline_transform): Remove all references
260 * cgraphunit.c (expand_function): Remove all references after
262 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
263 (ipa_find_reference): Rewrite to iterator.
264 (remove_stmt_references): Likewise.
265 (ipa_clear_stmts_in_references): New function.
266 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
267 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
269 * ipa-split.c (split_function): Remove references in split function.
271 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
274 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
275 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
277 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
280 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
281 suggested mode for the assignment isn't compatible with the
284 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
287 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
289 2013-08-07 Xinliang David Li <davidxl@google.com>
291 * config/i386/i386.opt: New option -mtune-ctrl=.
292 * config/i386/x86-tune.def: New file.
293 * config/i386/i386.h: include x86-tune.def.
294 * config/i386/i386.c (ix86_option_override_internal):
295 Parsing -mtune-ctrl= option and set tune features.
297 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
300 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
301 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
304 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
306 * rtl.h (update_alignments): Declare.
307 * final.c (grow_label_align): New function extracted from...
308 (shorten_branches): ...here. Call it.
309 (update_alignments): New function.
310 * reorg.c (sibling_labels): New variable.
311 (get_label_before): Add SIBLING parameter. If it is non-zero, push
312 the new label along with it onto the sibling_labels vector.
313 (fill_simple_delay_slots): Adjust call to get_label_before.
314 (fill_slots_from_thread): Likewise.
315 (relax_delay_slots): Likewise.
316 (make_return_insns): Likewise.
317 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
319 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
321 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
322 document its semantics.
323 (diagnostic_report_diagnostic): Adjust accordingly.
325 2013-08-07 David Malcolm <dmalcolm@redhat.com>
327 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
328 (sparc_option_override): ...and port to new C++ pass API.
329 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
331 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
333 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
335 2013-08-06 Caroline Tice <cmtice@google.com>
337 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
338 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
339 * tree-pass.h: Add pass_vtable_verify.
340 * varasm.c (assemble_variable): Add code to properly set the comdat
341 section and name for the .vtable_map_vars section.
342 (assemble_vtyv_preinit_initializer): New function.
343 (default_sectin_type_flags): Make sure .vtable_map_vars section has
345 * output.h: Add function decl for assemble_vtv_preinit_initializer.
346 * vtable-verify.c: New file.
347 * vtable-verify.h: New file.
348 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
349 initialiation levels.
350 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
351 * passes.def: Insert pass_vtable_verify.
352 * aclocal.m4: Reorder includes.
353 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
354 -fvtv-counts options.
355 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
356 as appropriate, if -fvtable-verify=... is used.
357 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
358 -fvtable-verify=... is used.
359 * Makefile.in (OBJS): Add vtable-verify.o to list.
360 (vtable-verify.o): Add new build rule.
361 (GTFILES): Add vtable-verify.c to list.
362 * common.opt (fvtable-verify=): New flag.
363 (vtv_priority): Values for fvtable-verify= flag.
364 (fvtv-counts): New flag.
365 (fvtv-debug): New flag.
366 * tree.h (save_vtable_map_decl): New extern function decl.
368 2013-08-07 David Malcolm <dmalcolm@redhat.com>
370 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
371 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
372 (pass_data_rl78_devirt): ...new pass_data instance and...
373 (make_pass_rl78_devirt): ...new function.
374 (rl78_asm_file_start): Port pass registration to new C++ API.
376 2013-08-07 David Malcolm <dmalcolm@redhat.com>
378 * coretypes.h (rtl_opt_pass): Add.
380 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
381 (epiphany_init): Port to new C++ pass API.
382 (epiphany_optimize_mode_switching): Likewise.
383 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
384 (pass_manager::get_pass_mode_switching): New.
385 (pass_manager::get_pass_peephole2): New.
386 * mode-switching.c (pass_mode_switching): Add clone method.
387 * recog.c (pass_peephole2): Add clone method.
388 (pass_split_all_insns): Add clone method.
390 2013-08-06 David Malcolm <dmalcolm@redhat.com>
392 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
393 (mips_option_override): ...here, porting to new C++ API for passes.
395 2013-08-06 Jan Hubicka <jh@suse.cz>
397 * cgraph.c (cgraph_get_body): New function based on lto.c
399 * cgraph.h (cgraph_get_body): Declare.
400 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
402 * cgraphunit.c (expand_function): Get body prior expanding.
403 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
404 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
406 * passes.c (do_per_function_toporder): Get body.
407 * tree-inline.c (expand_call_inline): Get body prior inlining it.
408 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
410 2013-08-06 Martin Jambor <mjambor@suse.cz>
413 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
414 is not re-finalized. Rename second parameter to no_collect.
416 2013-08-06 Martin Jambor <mjambor@suse.cz>
419 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
420 MEM_REF has proper alignment information.
422 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
425 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
427 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
428 rtx (*) (rtx, ...) with insn_gen_fn.
429 * genoutput.c (output_insn_data): Cast gen_? function pointers to
430 insn_gen_fn::stored_funcptr. Add initializer braces.
432 2013-08-05 David Malcolm <dmalcolm@redhat.com>
434 Rewrite how instances of passes are cloned to remove assumptions
435 about their sizes (thus allowing pass subclasses to have
436 additional data fields, albeit non-GC-managed ones at this point).
438 * passes.c (make_pass_instance): Now that passes have clone
439 methods, rewrite this function to eliminate XNEW and memcpy
440 calls that used hardcoded sizes. Since this function no longer
441 creates pass instances, rename it to...
442 (add_pass_instance): ...this. Document the old way that passes were
443 numbered and flagged, and rework this function to continue using it.
444 (next_pass_1): Add an initial_pass argument for use by
446 (position_pass): When adding multiple instances of a pass, use
447 the pass's clone method, rather than relying on the XNEW/memcpy
448 within the former make_pass_instance (now add_pass_instance).
449 (pass_manager::pass_manager): When invoking next_pass_1, also supply
450 the initial instance of the current pass within the pass manager.
452 2013-08-05 David Malcolm <dmalcolm@redhat.com>
454 This is the automated part of the conversion of passes from C
455 structs to C++ classes.
457 Patch autogenerated by refactor_passes.py from
458 https://github.com/davidmalcolm/gcc-refactoring-scripts
459 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
461 * asan.c (pass_asan): Convert from a global struct to a subclass of
462 gimple_opt_pass along with...
463 (pass_data_asan): ...new pass_data instance and...
464 (make_pass_asan): ...new function.
465 (pass_asan_O0): Convert from a global struct to a subclass of
466 gimple_opt_pass along with...
467 (pass_data_asan_O0): ...new pass_data instance and...
468 (make_pass_asan_O0): ...new function.
469 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
470 subclass of rtl_opt_pass along with...
471 (pass_data_inc_dec): ...new pass_data instance and...
472 (make_pass_inc_dec): ...new function.
473 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
474 a subclass of rtl_opt_pass along with...
475 (pass_data_reorder_blocks): ...new pass_data instance and...
476 (make_pass_reorder_blocks): ...new function.
477 (pass_duplicate_computed_gotos): Convert from a global struct to a
478 subclass of rtl_opt_pass along with...
479 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
480 (make_pass_duplicate_computed_gotos): ...new function.
481 (pass_partition_blocks): Convert from a global struct to a subclass of
482 rtl_opt_pass along with...
483 (pass_data_partition_blocks): ...new pass_data instance and...
484 (make_pass_partition_blocks): ...new function.
485 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
486 struct to a subclass of rtl_opt_pass along with...
487 (pass_data_branch_target_load_optimize1): ...new pass_data instance
489 (make_pass_branch_target_load_optimize1): ...new function.
490 (pass_branch_target_load_optimize2): Convert from a global struct to a
491 subclass of rtl_opt_pass along with...
492 (pass_data_branch_target_load_optimize2): ...new pass_data instance
494 (make_pass_branch_target_load_optimize2): ...new function.
495 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
496 of rtl_opt_pass along with...
497 (pass_data_jump): ...new pass_data instance and...
498 (make_pass_jump): ...new function.
499 (pass_jump2): Convert from a global struct to a subclass of
500 rtl_opt_pass along with...
501 (pass_data_jump2): ...new pass_data instance and...
502 (make_pass_jump2): ...new function.
503 * cfgexpand.c (pass_expand): Convert from a global struct to a
504 subclass of rtl_opt_pass along with...
505 (pass_data_expand): ...new pass_data instance and...
506 (make_pass_expand): ...new function.
507 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
508 of rtl_opt_pass along with...
509 (pass_data_free_cfg): ...new pass_data instance and...
510 (make_pass_free_cfg): ...new function.
511 (pass_into_cfg_layout_mode): Convert from a global struct to a
512 subclass of rtl_opt_pass along with...
513 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
514 (make_pass_into_cfg_layout_mode): ...new function.
515 (pass_outof_cfg_layout_mode): Convert from a global struct to a
516 subclass of rtl_opt_pass along with...
517 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
518 (make_pass_outof_cfg_layout_mode): ...new function.
519 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
520 struct to a subclass of gimple_opt_pass along with...
521 (pass_data_build_cgraph_edges): ...new pass_data instance and...
522 (make_pass_build_cgraph_edges): ...new function.
523 (pass_rebuild_cgraph_edges): Convert from a global struct to a
524 subclass of gimple_opt_pass along with...
525 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
526 (make_pass_rebuild_cgraph_edges): ...new function.
527 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
528 subclass of gimple_opt_pass along with...
529 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
531 (make_pass_remove_cgraph_callee_edges): ...new function.
532 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
533 struct to a subclass of rtl_opt_pass along with...
534 (pass_data_stack_adjustments): ...new pass_data instance and...
535 (make_pass_stack_adjustments): ...new function.
536 * combine.c (pass_combine): Convert from a global struct to a subclass
537 of rtl_opt_pass along with...
538 (pass_data_combine): ...new pass_data instance and...
539 (make_pass_combine): ...new function.
540 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
541 global struct to a subclass of rtl_opt_pass along with...
542 (pass_data_compare_elim_after_reload): ...new pass_data instance
544 (make_pass_compare_elim_after_reload): ...new function.
545 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
546 of rtl_opt_pass along with...
547 (pass_data_rtl_cprop): ...new pass_data instance and...
548 (make_pass_rtl_cprop): ...new function.
549 * cse.c (pass_cse): Convert from a global struct to a subclass of
550 rtl_opt_pass along with...
551 (pass_data_cse): ...new pass_data instance and...
552 (make_pass_cse): ...new function.
553 (pass_cse2): Convert from a global struct to a subclass of
554 rtl_opt_pass along with...
555 (pass_data_cse2): ...new pass_data instance and...
556 (make_pass_cse2): ...new function.
557 (pass_cse_after_global_opts): Convert from a global struct to a
558 subclass of rtl_opt_pass along with...
559 (pass_data_cse_after_global_opts): ...new pass_data instance and...
560 (make_pass_cse_after_global_opts): ...new function.
561 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
562 of rtl_opt_pass along with...
563 (pass_data_ud_rtl_dce): ...new pass_data instance and...
564 (make_pass_ud_rtl_dce): ...new function.
565 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
566 rtl_opt_pass along with...
567 (pass_data_fast_rtl_dce): ...new pass_data instance and...
568 (make_pass_fast_rtl_dce): ...new function.
569 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
570 a subclass of rtl_opt_pass along with...
571 (pass_data_df_initialize_opt): ...new pass_data instance and...
572 (make_pass_df_initialize_opt): ...new function.
573 (pass_df_initialize_no_opt): Convert from a global struct to a
574 subclass of rtl_opt_pass along with...
575 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
576 (make_pass_df_initialize_no_opt): ...new function.
577 (pass_df_finish): Convert from a global struct to a subclass of
578 rtl_opt_pass along with...
579 (pass_data_df_finish): ...new pass_data instance and...
580 (make_pass_df_finish): ...new function.
581 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
582 rtl_opt_pass along with...
583 (pass_data_rtl_dse1): ...new pass_data instance and...
584 (make_pass_rtl_dse1): ...new function.
585 (pass_rtl_dse2): Convert from a global struct to a subclass of
586 rtl_opt_pass along with...
587 (pass_data_rtl_dse2): ...new pass_data instance and...
588 (make_pass_rtl_dse2): ...new function.
589 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
590 subclass of rtl_opt_pass along with...
591 (pass_data_dwarf2_frame): ...new pass_data instance and...
592 (make_pass_dwarf2_frame): ...new function.
593 * except.c (pass_set_nothrow_function_flags): Convert from a global
594 struct to a subclass of rtl_opt_pass along with...
595 (pass_data_set_nothrow_function_flags): ...new pass_data instance
597 (make_pass_set_nothrow_function_flags): ...new function.
598 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
599 subclass of rtl_opt_pass along with...
600 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
602 (make_pass_convert_to_eh_region_ranges): ...new function.
603 * final.c (pass_compute_alignments): Convert from a global struct to a
604 subclass of rtl_opt_pass along with...
605 (pass_data_compute_alignments): ...new pass_data instance and...
606 (make_pass_compute_alignments): ...new function.
607 (pass_final): Convert from a global struct to a subclass of
608 rtl_opt_pass along with...
609 (pass_data_final): ...new pass_data instance and...
610 (make_pass_final): ...new function.
611 (pass_shorten_branches): Convert from a global struct to a subclass of
612 rtl_opt_pass along with...
613 (pass_data_shorten_branches): ...new pass_data instance and...
614 (make_pass_shorten_branches): ...new function.
615 (pass_clean_state): Convert from a global struct to a subclass of
616 rtl_opt_pass along with...
617 (pass_data_clean_state): ...new pass_data instance and...
618 (make_pass_clean_state): ...new function.
619 * function.c (pass_instantiate_virtual_regs): Convert from a global
620 struct to a subclass of rtl_opt_pass along with...
621 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
622 (make_pass_instantiate_virtual_regs): ...new function.
623 (pass_leaf_regs): Convert from a global struct to a subclass of
624 rtl_opt_pass along with...
625 (pass_data_leaf_regs): ...new pass_data instance and...
626 (make_pass_leaf_regs): ...new function.
627 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
628 subclass of rtl_opt_pass along with...
629 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
631 (make_pass_thread_prologue_and_epilogue): ...new function.
632 (pass_match_asm_constraints): Convert from a global struct to a
633 subclass of rtl_opt_pass along with...
634 (pass_data_match_asm_constraints): ...new pass_data instance and...
635 (make_pass_match_asm_constraints): ...new function.
636 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
637 subclass of rtl_opt_pass along with...
638 (pass_data_rtl_fwprop): ...new pass_data instance and...
639 (make_pass_rtl_fwprop): ...new function.
640 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
641 rtl_opt_pass along with...
642 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
643 (make_pass_rtl_fwprop_addr): ...new function.
644 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
645 rtl_opt_pass along with...
646 (pass_data_rtl_pre): ...new pass_data instance and...
647 (make_pass_rtl_pre): ...new function.
648 (pass_rtl_hoist): Convert from a global struct to a subclass of
649 rtl_opt_pass along with...
650 (pass_data_rtl_hoist): ...new pass_data instance and...
651 (make_pass_rtl_hoist): ...new function.
652 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
653 subclass of gimple_opt_pass along with...
654 (pass_data_lower_cf): ...new pass_data instance and...
655 (make_pass_lower_cf): ...new function.
656 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
657 from a global struct to a subclass of gimple_opt_pass along with...
658 (pass_data_strength_reduction): ...new pass_data instance and...
659 (make_pass_strength_reduction): ...new function.
660 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
661 of rtl_opt_pass along with...
662 (pass_data_rtl_ifcvt): ...new pass_data instance and...
663 (make_pass_rtl_ifcvt): ...new function.
664 (pass_if_after_combine): Convert from a global struct to a subclass of
665 rtl_opt_pass along with...
666 (pass_data_if_after_combine): ...new pass_data instance and...
667 (make_pass_if_after_combine): ...new function.
668 (pass_if_after_reload): Convert from a global struct to a subclass of
669 rtl_opt_pass along with...
670 (pass_data_if_after_reload): ...new pass_data instance and...
671 (make_pass_if_after_reload): ...new function.
672 * init-regs.c (pass_initialize_regs): Convert from a global struct to
673 a subclass of rtl_opt_pass along with...
674 (pass_data_initialize_regs): ...new pass_data instance and...
675 (make_pass_initialize_regs): ...new function.
676 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
677 of ipa_opt_pass_d along with...
678 (pass_data_ipa_cp): ...new pass_data instance and...
679 (make_pass_ipa_cp): ...new function.
680 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
681 global struct to a subclass of gimple_opt_pass along with...
682 (pass_data_inline_parameters): ...new pass_data instance and...
683 (make_pass_inline_parameters): ...new function.
684 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
685 subclass of gimple_opt_pass along with...
686 (pass_data_early_inline): ...new pass_data instance and...
687 (make_pass_early_inline): ...new function.
688 (pass_ipa_inline): Convert from a global struct to a subclass of
689 ipa_opt_pass_d along with...
690 (pass_data_ipa_inline): ...new pass_data instance and...
691 (make_pass_ipa_inline): ...new function.
692 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
693 struct to a subclass of gimple_opt_pass along with...
694 (pass_data_local_pure_const): ...new pass_data instance and...
695 (make_pass_local_pure_const): ...new function.
696 (pass_ipa_pure_const): Convert from a global struct to a subclass of
697 ipa_opt_pass_d along with...
698 (pass_data_ipa_pure_const): ...new pass_data instance and...
699 (make_pass_ipa_pure_const): ...new function.
700 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
701 to a subclass of ipa_opt_pass_d along with...
702 (pass_data_ipa_reference): ...new pass_data instance and...
703 (make_pass_ipa_reference): ...new function.
704 * ipa-split.c (pass_split_functions): Convert from a global struct to
705 a subclass of gimple_opt_pass along with...
706 (pass_data_split_functions): ...new pass_data instance and...
707 (make_pass_split_functions): ...new function.
708 (pass_feedback_split_functions): Convert from a global struct to a
709 subclass of gimple_opt_pass along with...
710 (pass_data_feedback_split_functions): ...new pass_data instance and...
711 (make_pass_feedback_split_functions): ...new function.
712 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
713 global struct to a subclass of simple_ipa_opt_pass along with...
714 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
716 (make_pass_ipa_function_and_variable_visibility): ...new function.
717 (pass_ipa_free_inline_summary): Convert from a global struct to a
718 subclass of simple_ipa_opt_pass along with...
719 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
720 (make_pass_ipa_free_inline_summary): ...new function.
721 (pass_ipa_whole_program_visibility): Convert from a global struct to a
722 subclass of ipa_opt_pass_d along with...
723 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
725 (make_pass_ipa_whole_program_visibility): ...new function.
726 (pass_ipa_profile): Convert from a global struct to a subclass of
727 ipa_opt_pass_d along with...
728 (pass_data_ipa_profile): ...new pass_data instance and...
729 (make_pass_ipa_profile): ...new function.
730 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
731 ipa_opt_pass_d along with...
732 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
733 (make_pass_ipa_cdtor_merge): ...new function.
734 * ira.c (pass_ira): Convert from a global struct to a subclass of
735 rtl_opt_pass along with...
736 (pass_data_ira): ...new pass_data instance and...
737 (make_pass_ira): ...new function.
738 (pass_reload): Convert from a global struct to a subclass of
739 rtl_opt_pass along with...
740 (pass_data_reload): ...new pass_data instance and...
741 (make_pass_reload): ...new function.
742 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
743 subclass of rtl_opt_pass along with...
744 (pass_data_cleanup_barriers): ...new pass_data instance and...
745 (make_pass_cleanup_barriers): ...new function.
746 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
747 of rtl_opt_pass along with...
748 (pass_data_loop2): ...new pass_data instance and...
749 (make_pass_loop2): ...new function.
750 (pass_rtl_loop_init): Convert from a global struct to a subclass of
751 rtl_opt_pass along with...
752 (pass_data_rtl_loop_init): ...new pass_data instance and...
753 (make_pass_rtl_loop_init): ...new function.
754 (pass_rtl_loop_done): Convert from a global struct to a subclass of
755 rtl_opt_pass along with...
756 (pass_data_rtl_loop_done): ...new pass_data instance and...
757 (make_pass_rtl_loop_done): ...new function.
758 (pass_rtl_move_loop_invariants): Convert from a global struct to a
759 subclass of rtl_opt_pass along with...
760 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
761 (make_pass_rtl_move_loop_invariants): ...new function.
762 (pass_rtl_unswitch): Convert from a global struct to a subclass of
763 rtl_opt_pass along with...
764 (pass_data_rtl_unswitch): ...new pass_data instance and...
765 (make_pass_rtl_unswitch): ...new function.
766 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
767 subclass of rtl_opt_pass along with...
768 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
770 (make_pass_rtl_unroll_and_peel_loops): ...new function.
771 (pass_rtl_doloop): Convert from a global struct to a subclass of
772 rtl_opt_pass along with...
773 (pass_data_rtl_doloop): ...new pass_data instance and...
774 (make_pass_rtl_doloop): ...new function.
775 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
776 a subclass of rtl_opt_pass along with...
777 (pass_data_lower_subreg): ...new pass_data instance and...
778 (make_pass_lower_subreg): ...new function.
779 (pass_lower_subreg2): Convert from a global struct to a subclass of
780 rtl_opt_pass along with...
781 (pass_data_lower_subreg2): ...new pass_data instance and...
782 (make_pass_lower_subreg2): ...new function.
783 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
784 struct to a subclass of ipa_opt_pass_d along with...
785 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
786 (make_pass_ipa_lto_gimple_out): ...new function.
787 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
788 of ipa_opt_pass_d along with...
789 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
790 (make_pass_ipa_lto_finish_out): ...new function.
791 * mode-switching.c (pass_mode_switching): Convert from a global struct
792 to a subclass of rtl_opt_pass along with...
793 (pass_data_mode_switching): ...new pass_data instance and...
794 (make_pass_mode_switching): ...new function.
795 * modulo-sched.c (pass_sms): Convert from a global struct to a
796 subclass of rtl_opt_pass along with...
797 (pass_data_sms): ...new pass_data instance and...
798 (make_pass_sms): ...new function.
799 * omp-low.c (pass_expand_omp): Convert from a global struct to a
800 subclass of gimple_opt_pass along with...
801 (pass_data_expand_omp): ...new pass_data instance and...
802 (make_pass_expand_omp): ...new function.
803 (pass_lower_omp): Convert from a global struct to a subclass of
804 gimple_opt_pass along with...
805 (pass_data_lower_omp): ...new pass_data instance and...
806 (make_pass_lower_omp): ...new function.
807 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
808 of gimple_opt_pass along with...
809 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
810 (make_pass_diagnose_omp_blocks): ...new function.
811 * passes.c (pass_early_local_passes): Convert from a global struct to
812 a subclass of simple_ipa_opt_pass along with...
813 (pass_data_early_local_passes): ...new pass_data instance and...
814 (make_pass_early_local_passes): ...new function.
815 (pass_all_early_optimizations): Convert from a global struct to a
816 subclass of gimple_opt_pass along with...
817 (pass_data_all_early_optimizations): ...new pass_data instance and...
818 (make_pass_all_early_optimizations): ...new function.
819 (pass_all_optimizations): Convert from a global struct to a subclass
820 of gimple_opt_pass along with...
821 (pass_data_all_optimizations): ...new pass_data instance and...
822 (make_pass_all_optimizations): ...new function.
823 (pass_all_optimizations_g): Convert from a global struct to a subclass
824 of gimple_opt_pass along with...
825 (pass_data_all_optimizations_g): ...new pass_data instance and...
826 (make_pass_all_optimizations_g): ...new function.
827 (pass_rest_of_compilation): Convert from a global struct to a subclass
828 of rtl_opt_pass along with...
829 (pass_data_rest_of_compilation): ...new pass_data instance and...
830 (make_pass_rest_of_compilation): ...new function.
831 (pass_postreload): Convert from a global struct to a subclass of
832 rtl_opt_pass along with...
833 (pass_data_postreload): ...new pass_data instance and...
834 (make_pass_postreload): ...new function.
835 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
836 subclass of rtl_opt_pass along with...
837 (pass_data_gcse2): ...new pass_data instance and...
838 (make_pass_gcse2): ...new function.
839 * postreload.c (pass_postreload_cse): Convert from a global struct to
840 a subclass of rtl_opt_pass along with...
841 (pass_data_postreload_cse): ...new pass_data instance and...
842 (make_pass_postreload_cse): ...new function.
843 * predict.c (pass_profile): Convert from a global struct to a subclass
844 of gimple_opt_pass along with...
845 (pass_data_profile): ...new pass_data instance and...
846 (make_pass_profile): ...new function.
847 (pass_strip_predict_hints): Convert from a global struct to a subclass
848 of gimple_opt_pass along with...
849 (pass_data_strip_predict_hints): ...new pass_data instance and...
850 (make_pass_strip_predict_hints): ...new function.
851 * recog.c (pass_peephole2): Convert from a global struct to a subclass
852 of rtl_opt_pass along with...
853 (pass_data_peephole2): ...new pass_data instance and...
854 (make_pass_peephole2): ...new function.
855 (pass_split_all_insns): Convert from a global struct to a subclass of
856 rtl_opt_pass along with...
857 (pass_data_split_all_insns): ...new pass_data instance and...
858 (make_pass_split_all_insns): ...new function.
859 (pass_split_after_reload): Convert from a global struct to a subclass
860 of rtl_opt_pass along with...
861 (pass_data_split_after_reload): ...new pass_data instance and...
862 (make_pass_split_after_reload): ...new function.
863 (pass_split_before_regstack): Convert from a global struct to a
864 subclass of rtl_opt_pass along with...
865 (pass_data_split_before_regstack): ...new pass_data instance and...
866 (make_pass_split_before_regstack): ...new function.
867 (pass_split_before_sched2): Convert from a global struct to a subclass
868 of rtl_opt_pass along with...
869 (pass_data_split_before_sched2): ...new pass_data instance and...
870 (make_pass_split_before_sched2): ...new function.
871 (pass_split_for_shorten_branches): Convert from a global struct to a
872 subclass of rtl_opt_pass along with...
873 (pass_data_split_for_shorten_branches): ...new pass_data instance
875 (make_pass_split_for_shorten_branches): ...new function.
876 * ree.c (pass_ree): Convert from a global struct to a subclass of
877 rtl_opt_pass along with...
878 (pass_data_ree): ...new pass_data instance and...
879 (make_pass_ree): ...new function.
880 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
881 subclass of rtl_opt_pass along with...
882 (pass_data_stack_regs): ...new pass_data instance and...
883 (make_pass_stack_regs): ...new function.
884 (pass_stack_regs_run): Convert from a global struct to a subclass of
885 rtl_opt_pass along with...
886 (pass_data_stack_regs_run): ...new pass_data instance and...
887 (make_pass_stack_regs_run): ...new function.
888 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
889 subclass of rtl_opt_pass along with...
890 (pass_data_cprop_hardreg): ...new pass_data instance and...
891 (make_pass_cprop_hardreg): ...new function.
892 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
893 subclass of rtl_opt_pass along with...
894 (pass_data_reginfo_init): ...new pass_data instance and...
895 (make_pass_reginfo_init): ...new function.
896 * regmove.c (pass_regmove): Convert from a global struct to a subclass
897 of rtl_opt_pass along with...
898 (pass_data_regmove): ...new pass_data instance and...
899 (make_pass_regmove): ...new function.
900 * regrename.c (pass_regrename): Convert from a global struct to a
901 subclass of rtl_opt_pass along with...
902 (pass_data_regrename): ...new pass_data instance and...
903 (make_pass_regrename): ...new function.
904 * reorg.c (pass_delay_slots): Convert from a global struct to a
905 subclass of rtl_opt_pass along with...
906 (pass_data_delay_slots): ...new pass_data instance and...
907 (make_pass_delay_slots): ...new function.
908 (pass_machine_reorg): Convert from a global struct to a subclass of
909 rtl_opt_pass along with...
910 (pass_data_machine_reorg): ...new pass_data instance and...
911 (make_pass_machine_reorg): ...new function.
912 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
913 of rtl_opt_pass along with...
914 (pass_data_sched): ...new pass_data instance and...
915 (make_pass_sched): ...new function.
916 (pass_sched2): Convert from a global struct to a subclass of
917 rtl_opt_pass along with...
918 (pass_data_sched2): ...new pass_data instance and...
919 (make_pass_sched2): ...new function.
920 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
921 to a subclass of rtl_opt_pass along with...
922 (pass_data_stack_ptr_mod): ...new pass_data instance and...
923 (make_pass_stack_ptr_mod): ...new function.
924 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
925 to a subclass of rtl_opt_pass along with...
926 (pass_data_rtl_store_motion): ...new pass_data instance and...
927 (make_pass_rtl_store_motion): ...new function.
928 * tracer.c (pass_tracer): Convert from a global struct to a subclass
929 of gimple_opt_pass along with...
930 (pass_data_tracer): ...new pass_data instance and...
931 (make_pass_tracer): ...new function.
932 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
933 to a subclass of gimple_opt_pass along with...
934 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
935 (make_pass_diagnose_tm_blocks): ...new function.
936 (pass_lower_tm): Convert from a global struct to a subclass of
937 gimple_opt_pass along with...
938 (pass_data_lower_tm): ...new pass_data instance and...
939 (make_pass_lower_tm): ...new function.
940 (pass_tm_init): Convert from a global struct to a subclass of
941 gimple_opt_pass along with...
942 (pass_data_tm_init): ...new pass_data instance and...
943 (make_pass_tm_init): ...new function.
944 (pass_tm_mark): Convert from a global struct to a subclass of
945 gimple_opt_pass along with...
946 (pass_data_tm_mark): ...new pass_data instance and...
947 (make_pass_tm_mark): ...new function.
948 (pass_tm_edges): Convert from a global struct to a subclass of
949 gimple_opt_pass along with...
950 (pass_data_tm_edges): ...new pass_data instance and...
951 (make_pass_tm_edges): ...new function.
952 (pass_tm_memopt): Convert from a global struct to a subclass of
953 gimple_opt_pass along with...
954 (pass_data_tm_memopt): ...new pass_data instance and...
955 (make_pass_tm_memopt): ...new function.
956 (pass_ipa_tm): Convert from a global struct to a subclass of
957 simple_ipa_opt_pass along with...
958 (pass_data_ipa_tm): ...new pass_data instance and...
959 (make_pass_ipa_tm): ...new function.
960 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
961 subclass of gimple_opt_pass along with...
962 (pass_data_call_cdce): ...new pass_data instance and...
963 (make_pass_call_cdce): ...new function.
964 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
965 subclass of gimple_opt_pass along with...
966 (pass_data_build_cfg): ...new pass_data instance and...
967 (make_pass_build_cfg): ...new function.
968 (pass_split_crit_edges): Convert from a global struct to a subclass of
969 gimple_opt_pass along with...
970 (pass_data_split_crit_edges): ...new pass_data instance and...
971 (make_pass_split_crit_edges): ...new function.
972 (pass_warn_function_return): Convert from a global struct to a
973 subclass of gimple_opt_pass along with...
974 (pass_data_warn_function_return): ...new pass_data instance and...
975 (make_pass_warn_function_return): ...new function.
976 (pass_warn_function_noreturn): Convert from a global struct to a
977 subclass of gimple_opt_pass along with...
978 (pass_data_warn_function_noreturn): ...new pass_data instance and...
979 (make_pass_warn_function_noreturn): ...new function.
980 (pass_warn_unused_result): Convert from a global struct to a subclass
981 of gimple_opt_pass along with...
982 (pass_data_warn_unused_result): ...new pass_data instance and...
983 (make_pass_warn_unused_result): ...new function.
984 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
985 a subclass of gimple_opt_pass along with...
986 (pass_data_merge_phi): ...new pass_data instance and...
987 (make_pass_merge_phi): ...new function.
988 * tree-complex.c (pass_lower_complex): Convert from a global struct to
989 a subclass of gimple_opt_pass along with...
990 (pass_data_lower_complex): ...new pass_data instance and...
991 (make_pass_lower_complex): ...new function.
992 (pass_lower_complex_O0): Convert from a global struct to a subclass of
993 gimple_opt_pass along with...
994 (pass_data_lower_complex_O0): ...new pass_data instance and...
995 (make_pass_lower_complex_O0): ...new function.
996 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
997 subclass of gimple_opt_pass along with...
998 (pass_data_lower_eh): ...new pass_data instance and...
999 (make_pass_lower_eh): ...new function.
1000 (pass_refactor_eh): Convert from a global struct to a subclass of
1001 gimple_opt_pass along with...
1002 (pass_data_refactor_eh): ...new pass_data instance and...
1003 (make_pass_refactor_eh): ...new function.
1004 (pass_lower_resx): Convert from a global struct to a subclass of
1005 gimple_opt_pass along with...
1006 (pass_data_lower_resx): ...new pass_data instance and...
1007 (make_pass_lower_resx): ...new function.
1008 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1009 of gimple_opt_pass along with...
1010 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1011 (make_pass_lower_eh_dispatch): ...new function.
1012 (pass_cleanup_eh): Convert from a global struct to a subclass of
1013 gimple_opt_pass along with...
1014 (pass_data_cleanup_eh): ...new pass_data instance and...
1015 (make_pass_cleanup_eh): ...new function.
1016 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1017 to a subclass of simple_ipa_opt_pass along with...
1018 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1019 (make_pass_ipa_lower_emutls): ...new function.
1020 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1021 a subclass of gimple_opt_pass along with...
1022 (pass_data_if_conversion): ...new pass_data instance and...
1023 (make_pass_if_conversion): ...new function.
1024 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1025 subclass of gimple_opt_pass along with...
1026 (pass_data_build_ssa): ...new pass_data instance and...
1027 (make_pass_build_ssa): ...new function.
1028 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1029 global struct to a subclass of gimple_opt_pass along with...
1030 (pass_data_loop_distribution): ...new pass_data instance and...
1031 (make_pass_loop_distribution): ...new function.
1032 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1033 subclass of gimple_opt_pass along with...
1034 (pass_data_mudflap_1): ...new pass_data instance and...
1035 (make_pass_mudflap_1): ...new function.
1036 (pass_mudflap_2): Convert from a global struct to a subclass of
1037 gimple_opt_pass along with...
1038 (pass_data_mudflap_2): ...new pass_data instance and...
1039 (make_pass_mudflap_2): ...new function.
1040 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1041 subclass of gimple_opt_pass along with...
1042 (pass_data_mudflap_1): ...new pass_data instance and...
1043 (make_pass_mudflap_1): ...new function.
1044 (pass_mudflap_2): Convert from a global struct to a subclass of
1045 gimple_opt_pass along with...
1046 (pass_data_mudflap_2): ...new pass_data instance and...
1047 (make_pass_mudflap_2): ...new function.
1048 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1049 gimple_opt_pass along with...
1050 (pass_data_nrv): ...new pass_data instance and...
1051 (make_pass_nrv): ...new function.
1052 (pass_return_slot): Convert from a global struct to a subclass of
1053 gimple_opt_pass along with...
1054 (pass_data_return_slot): ...new pass_data instance and...
1055 (make_pass_return_slot): ...new function.
1056 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1057 to a subclass of gimple_opt_pass along with...
1058 (pass_data_object_sizes): ...new pass_data instance and...
1059 (make_pass_object_sizes): ...new function.
1060 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1061 global struct to a subclass of gimple_opt_pass along with...
1062 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1064 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1065 (pass_fixup_cfg): Convert from a global struct to a subclass of
1066 gimple_opt_pass along with...
1067 (pass_data_fixup_cfg): ...new pass_data instance and...
1068 (make_pass_fixup_cfg): ...new function.
1069 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1070 (make_pass_mudflap_1): ...new function.
1071 (pass_mudflap_2): Replace declaration with that of...
1072 (make_pass_mudflap_2): ...new function.
1073 (pass_asan): Replace declaration with that of...
1074 (make_pass_asan): ...new function.
1075 (pass_asan_O0): Replace declaration with that of...
1076 (make_pass_asan_O0): ...new function.
1077 (pass_tsan): Replace declaration with that of...
1078 (make_pass_tsan): ...new function.
1079 (pass_tsan_O0): Replace declaration with that of...
1080 (make_pass_tsan_O0): ...new function.
1081 (pass_lower_cf): Replace declaration with that of...
1082 (make_pass_lower_cf): ...new function.
1083 (pass_refactor_eh): Replace declaration with that of...
1084 (make_pass_refactor_eh): ...new function.
1085 (pass_lower_eh): Replace declaration with that of...
1086 (make_pass_lower_eh): ...new function.
1087 (pass_lower_eh_dispatch): Replace declaration with that of...
1088 (make_pass_lower_eh_dispatch): ...new function.
1089 (pass_lower_resx): Replace declaration with that of...
1090 (make_pass_lower_resx): ...new function.
1091 (pass_build_cfg): Replace declaration with that of...
1092 (make_pass_build_cfg): ...new function.
1093 (pass_early_tree_profile): Replace declaration with that of...
1094 (make_pass_early_tree_profile): ...new function.
1095 (pass_cleanup_eh): Replace declaration with that of...
1096 (make_pass_cleanup_eh): ...new function.
1097 (pass_sra): Replace declaration with that of...
1098 (make_pass_sra): ...new function.
1099 (pass_sra_early): Replace declaration with that of...
1100 (make_pass_sra_early): ...new function.
1101 (pass_early_ipa_sra): Replace declaration with that of...
1102 (make_pass_early_ipa_sra): ...new function.
1103 (pass_tail_recursion): Replace declaration with that of...
1104 (make_pass_tail_recursion): ...new function.
1105 (pass_tail_calls): Replace declaration with that of...
1106 (make_pass_tail_calls): ...new function.
1107 (pass_tree_loop): Replace declaration with that of...
1108 (make_pass_tree_loop): ...new function.
1109 (pass_tree_loop_init): Replace declaration with that of...
1110 (make_pass_tree_loop_init): ...new function.
1111 (pass_lim): Replace declaration with that of...
1112 (make_pass_lim): ...new function.
1113 (pass_tree_unswitch): Replace declaration with that of...
1114 (make_pass_tree_unswitch): ...new function.
1115 (pass_predcom): Replace declaration with that of...
1116 (make_pass_predcom): ...new function.
1117 (pass_iv_canon): Replace declaration with that of...
1118 (make_pass_iv_canon): ...new function.
1119 (pass_scev_cprop): Replace declaration with that of...
1120 (make_pass_scev_cprop): ...new function.
1121 (pass_empty_loop): Replace declaration with that of...
1122 (make_pass_empty_loop): ...new function.
1123 (pass_record_bounds): Replace declaration with that of...
1124 (make_pass_record_bounds): ...new function.
1125 (pass_graphite): Replace declaration with that of...
1126 (make_pass_graphite): ...new function.
1127 (pass_graphite_transforms): Replace declaration with that of...
1128 (make_pass_graphite_transforms): ...new function.
1129 (pass_if_conversion): Replace declaration with that of...
1130 (make_pass_if_conversion): ...new function.
1131 (pass_loop_distribution): Replace declaration with that of...
1132 (make_pass_loop_distribution): ...new function.
1133 (pass_vectorize): Replace declaration with that of...
1134 (make_pass_vectorize): ...new function.
1135 (pass_slp_vectorize): Replace declaration with that of...
1136 (make_pass_slp_vectorize): ...new function.
1137 (pass_complete_unroll): Replace declaration with that of...
1138 (make_pass_complete_unroll): ...new function.
1139 (pass_complete_unrolli): Replace declaration with that of...
1140 (make_pass_complete_unrolli): ...new function.
1141 (pass_parallelize_loops): Replace declaration with that of...
1142 (make_pass_parallelize_loops): ...new function.
1143 (pass_loop_prefetch): Replace declaration with that of...
1144 (make_pass_loop_prefetch): ...new function.
1145 (pass_iv_optimize): Replace declaration with that of...
1146 (make_pass_iv_optimize): ...new function.
1147 (pass_tree_loop_done): Replace declaration with that of...
1148 (make_pass_tree_loop_done): ...new function.
1149 (pass_ch): Replace declaration with that of...
1150 (make_pass_ch): ...new function.
1151 (pass_ccp): Replace declaration with that of...
1152 (make_pass_ccp): ...new function.
1153 (pass_phi_only_cprop): Replace declaration with that of...
1154 (make_pass_phi_only_cprop): ...new function.
1155 (pass_build_ssa): Replace declaration with that of...
1156 (make_pass_build_ssa): ...new function.
1157 (pass_build_alias): Replace declaration with that of...
1158 (make_pass_build_alias): ...new function.
1159 (pass_build_ealias): Replace declaration with that of...
1160 (make_pass_build_ealias): ...new function.
1161 (pass_dominator): Replace declaration with that of...
1162 (make_pass_dominator): ...new function.
1163 (pass_dce): Replace declaration with that of...
1164 (make_pass_dce): ...new function.
1165 (pass_dce_loop): Replace declaration with that of...
1166 (make_pass_dce_loop): ...new function.
1167 (pass_cd_dce): Replace declaration with that of...
1168 (make_pass_cd_dce): ...new function.
1169 (pass_call_cdce): Replace declaration with that of...
1170 (make_pass_call_cdce): ...new function.
1171 (pass_merge_phi): Replace declaration with that of...
1172 (make_pass_merge_phi): ...new function.
1173 (pass_split_crit_edges): Replace declaration with that of...
1174 (make_pass_split_crit_edges): ...new function.
1175 (pass_pre): Replace declaration with that of...
1176 (make_pass_pre): ...new function.
1177 (pass_profile): Replace declaration with that of...
1178 (make_pass_profile): ...new function.
1179 (pass_strip_predict_hints): Replace declaration with that of...
1180 (make_pass_strip_predict_hints): ...new function.
1181 (pass_lower_complex_O0): Replace declaration with that of...
1182 (make_pass_lower_complex_O0): ...new function.
1183 (pass_lower_complex): Replace declaration with that of...
1184 (make_pass_lower_complex): ...new function.
1185 (pass_lower_vector): Replace declaration with that of...
1186 (make_pass_lower_vector): ...new function.
1187 (pass_lower_vector_ssa): Replace declaration with that of...
1188 (make_pass_lower_vector_ssa): ...new function.
1189 (pass_lower_omp): Replace declaration with that of...
1190 (make_pass_lower_omp): ...new function.
1191 (pass_diagnose_omp_blocks): Replace declaration with that of...
1192 (make_pass_diagnose_omp_blocks): ...new function.
1193 (pass_expand_omp): Replace declaration with that of...
1194 (make_pass_expand_omp): ...new function.
1195 (pass_expand_omp_ssa): Replace declaration with that of...
1196 (make_pass_expand_omp_ssa): ...new function.
1197 (pass_object_sizes): Replace declaration with that of...
1198 (make_pass_object_sizes): ...new function.
1199 (pass_strlen): Replace declaration with that of...
1200 (make_pass_strlen): ...new function.
1201 (pass_fold_builtins): Replace declaration with that of...
1202 (make_pass_fold_builtins): ...new function.
1203 (pass_stdarg): Replace declaration with that of...
1204 (make_pass_stdarg): ...new function.
1205 (pass_early_warn_uninitialized): Replace declaration with that of...
1206 (make_pass_early_warn_uninitialized): ...new function.
1207 (pass_late_warn_uninitialized): Replace declaration with that of...
1208 (make_pass_late_warn_uninitialized): ...new function.
1209 (pass_cse_reciprocals): Replace declaration with that of...
1210 (make_pass_cse_reciprocals): ...new function.
1211 (pass_cse_sincos): Replace declaration with that of...
1212 (make_pass_cse_sincos): ...new function.
1213 (pass_optimize_bswap): Replace declaration with that of...
1214 (make_pass_optimize_bswap): ...new function.
1215 (pass_optimize_widening_mul): Replace declaration with that of...
1216 (make_pass_optimize_widening_mul): ...new function.
1217 (pass_warn_function_return): Replace declaration with that of...
1218 (make_pass_warn_function_return): ...new function.
1219 (pass_warn_function_noreturn): Replace declaration with that of...
1220 (make_pass_warn_function_noreturn): ...new function.
1221 (pass_cselim): Replace declaration with that of...
1222 (make_pass_cselim): ...new function.
1223 (pass_phiopt): Replace declaration with that of...
1224 (make_pass_phiopt): ...new function.
1225 (pass_forwprop): Replace declaration with that of...
1226 (make_pass_forwprop): ...new function.
1227 (pass_phiprop): Replace declaration with that of...
1228 (make_pass_phiprop): ...new function.
1229 (pass_tree_ifcombine): Replace declaration with that of...
1230 (make_pass_tree_ifcombine): ...new function.
1231 (pass_dse): Replace declaration with that of...
1232 (make_pass_dse): ...new function.
1233 (pass_nrv): Replace declaration with that of...
1234 (make_pass_nrv): ...new function.
1235 (pass_rename_ssa_copies): Replace declaration with that of...
1236 (make_pass_rename_ssa_copies): ...new function.
1237 (pass_sink_code): Replace declaration with that of...
1238 (make_pass_sink_code): ...new function.
1239 (pass_fre): Replace declaration with that of...
1240 (make_pass_fre): ...new function.
1241 (pass_check_data_deps): Replace declaration with that of...
1242 (make_pass_check_data_deps): ...new function.
1243 (pass_copy_prop): Replace declaration with that of...
1244 (make_pass_copy_prop): ...new function.
1245 (pass_vrp): Replace declaration with that of...
1246 (make_pass_vrp): ...new function.
1247 (pass_uncprop): Replace declaration with that of...
1248 (make_pass_uncprop): ...new function.
1249 (pass_return_slot): Replace declaration with that of...
1250 (make_pass_return_slot): ...new function.
1251 (pass_reassoc): Replace declaration with that of...
1252 (make_pass_reassoc): ...new function.
1253 (pass_rebuild_cgraph_edges): Replace declaration with that of...
1254 (make_pass_rebuild_cgraph_edges): ...new function.
1255 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
1256 (make_pass_remove_cgraph_callee_edges): ...new function.
1257 (pass_build_cgraph_edges): Replace declaration with that of...
1258 (make_pass_build_cgraph_edges): ...new function.
1259 (pass_local_pure_const): Replace declaration with that of...
1260 (make_pass_local_pure_const): ...new function.
1261 (pass_tracer): Replace declaration with that of...
1262 (make_pass_tracer): ...new function.
1263 (pass_warn_unused_result): Replace declaration with that of...
1264 (make_pass_warn_unused_result): ...new function.
1265 (pass_diagnose_tm_blocks): Replace declaration with that of...
1266 (make_pass_diagnose_tm_blocks): ...new function.
1267 (pass_lower_tm): Replace declaration with that of...
1268 (make_pass_lower_tm): ...new function.
1269 (pass_tm_init): Replace declaration with that of...
1270 (make_pass_tm_init): ...new function.
1271 (pass_tm_mark): Replace declaration with that of...
1272 (make_pass_tm_mark): ...new function.
1273 (pass_tm_memopt): Replace declaration with that of...
1274 (make_pass_tm_memopt): ...new function.
1275 (pass_tm_edges): Replace declaration with that of...
1276 (make_pass_tm_edges): ...new function.
1277 (pass_split_functions): Replace declaration with that of...
1278 (make_pass_split_functions): ...new function.
1279 (pass_feedback_split_functions): Replace declaration with that of...
1280 (make_pass_feedback_split_functions): ...new function.
1281 (pass_strength_reduction): Replace declaration with that of...
1282 (make_pass_strength_reduction): ...new function.
1283 (pass_ipa_lower_emutls): Replace declaration with that of...
1284 (make_pass_ipa_lower_emutls): ...new function.
1285 (pass_ipa_function_and_variable_visibility): Replace declaration with
1287 (make_pass_ipa_function_and_variable_visibility): ...new function.
1288 (pass_ipa_tree_profile): Replace declaration with that of...
1289 (make_pass_ipa_tree_profile): ...new function.
1290 (pass_early_local_passes): Replace declaration with that of...
1291 (make_pass_early_local_passes): ...new function.
1292 (pass_ipa_whole_program_visibility): Replace declaration with that
1294 (make_pass_ipa_whole_program_visibility): ...new function.
1295 (pass_ipa_lto_gimple_out): Replace declaration with that of...
1296 (make_pass_ipa_lto_gimple_out): ...new function.
1297 (pass_ipa_increase_alignment): Replace declaration with that of...
1298 (make_pass_ipa_increase_alignment): ...new function.
1299 (pass_ipa_inline): Replace declaration with that of...
1300 (make_pass_ipa_inline): ...new function.
1301 (pass_ipa_free_lang_data): Replace declaration with that of...
1302 (make_pass_ipa_free_lang_data): ...new function.
1303 (pass_ipa_free_inline_summary): Replace declaration with that of...
1304 (make_pass_ipa_free_inline_summary): ...new function.
1305 (pass_ipa_cp): Replace declaration with that of...
1306 (make_pass_ipa_cp): ...new function.
1307 (pass_ipa_reference): Replace declaration with that of...
1308 (make_pass_ipa_reference): ...new function.
1309 (pass_ipa_pure_const): Replace declaration with that of...
1310 (make_pass_ipa_pure_const): ...new function.
1311 (pass_ipa_pta): Replace declaration with that of...
1312 (make_pass_ipa_pta): ...new function.
1313 (pass_ipa_lto_finish_out): Replace declaration with that of...
1314 (make_pass_ipa_lto_finish_out): ...new function.
1315 (pass_ipa_tm): Replace declaration with that of...
1316 (make_pass_ipa_tm): ...new function.
1317 (pass_ipa_profile): Replace declaration with that of...
1318 (make_pass_ipa_profile): ...new function.
1319 (pass_ipa_cdtor_merge): Replace declaration with that of...
1320 (make_pass_ipa_cdtor_merge): ...new function.
1321 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
1323 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1324 (pass_init_datastructures): Replace declaration with that of...
1325 (make_pass_init_datastructures): ...new function.
1326 (pass_fixup_cfg): Replace declaration with that of...
1327 (make_pass_fixup_cfg): ...new function.
1328 (pass_expand): Replace declaration with that of...
1329 (make_pass_expand): ...new function.
1330 (pass_instantiate_virtual_regs): Replace declaration with that of...
1331 (make_pass_instantiate_virtual_regs): ...new function.
1332 (pass_rtl_fwprop): Replace declaration with that of...
1333 (make_pass_rtl_fwprop): ...new function.
1334 (pass_rtl_fwprop_addr): Replace declaration with that of...
1335 (make_pass_rtl_fwprop_addr): ...new function.
1336 (pass_jump): Replace declaration with that of...
1337 (make_pass_jump): ...new function.
1338 (pass_jump2): Replace declaration with that of...
1339 (make_pass_jump2): ...new function.
1340 (pass_lower_subreg): Replace declaration with that of...
1341 (make_pass_lower_subreg): ...new function.
1342 (pass_cse): Replace declaration with that of...
1343 (make_pass_cse): ...new function.
1344 (pass_fast_rtl_dce): Replace declaration with that of...
1345 (make_pass_fast_rtl_dce): ...new function.
1346 (pass_ud_rtl_dce): Replace declaration with that of...
1347 (make_pass_ud_rtl_dce): ...new function.
1348 (pass_rtl_dce): Replace declaration with that of...
1349 (make_pass_rtl_dce): ...new function.
1350 (pass_rtl_dse1): Replace declaration with that of...
1351 (make_pass_rtl_dse1): ...new function.
1352 (pass_rtl_dse2): Replace declaration with that of...
1353 (make_pass_rtl_dse2): ...new function.
1354 (pass_rtl_dse3): Replace declaration with that of...
1355 (make_pass_rtl_dse3): ...new function.
1356 (pass_rtl_cprop): Replace declaration with that of...
1357 (make_pass_rtl_cprop): ...new function.
1358 (pass_rtl_pre): Replace declaration with that of...
1359 (make_pass_rtl_pre): ...new function.
1360 (pass_rtl_hoist): Replace declaration with that of...
1361 (make_pass_rtl_hoist): ...new function.
1362 (pass_rtl_store_motion): Replace declaration with that of...
1363 (make_pass_rtl_store_motion): ...new function.
1364 (pass_cse_after_global_opts): Replace declaration with that of...
1365 (make_pass_cse_after_global_opts): ...new function.
1366 (pass_rtl_ifcvt): Replace declaration with that of...
1367 (make_pass_rtl_ifcvt): ...new function.
1368 (pass_into_cfg_layout_mode): Replace declaration with that of...
1369 (make_pass_into_cfg_layout_mode): ...new function.
1370 (pass_outof_cfg_layout_mode): Replace declaration with that of...
1371 (make_pass_outof_cfg_layout_mode): ...new function.
1372 (pass_loop2): Replace declaration with that of...
1373 (make_pass_loop2): ...new function.
1374 (pass_rtl_loop_init): Replace declaration with that of...
1375 (make_pass_rtl_loop_init): ...new function.
1376 (pass_rtl_move_loop_invariants): Replace declaration with that of...
1377 (make_pass_rtl_move_loop_invariants): ...new function.
1378 (pass_rtl_unswitch): Replace declaration with that of...
1379 (make_pass_rtl_unswitch): ...new function.
1380 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
1381 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1382 (pass_rtl_doloop): Replace declaration with that of...
1383 (make_pass_rtl_doloop): ...new function.
1384 (pass_rtl_loop_done): Replace declaration with that of...
1385 (make_pass_rtl_loop_done): ...new function.
1386 (pass_web): Replace declaration with that of...
1387 (make_pass_web): ...new function.
1388 (pass_cse2): Replace declaration with that of...
1389 (make_pass_cse2): ...new function.
1390 (pass_df_initialize_opt): Replace declaration with that of...
1391 (make_pass_df_initialize_opt): ...new function.
1392 (pass_df_initialize_no_opt): Replace declaration with that of...
1393 (make_pass_df_initialize_no_opt): ...new function.
1394 (pass_reginfo_init): Replace declaration with that of...
1395 (make_pass_reginfo_init): ...new function.
1396 (pass_inc_dec): Replace declaration with that of...
1397 (make_pass_inc_dec): ...new function.
1398 (pass_stack_ptr_mod): Replace declaration with that of...
1399 (make_pass_stack_ptr_mod): ...new function.
1400 (pass_initialize_regs): Replace declaration with that of...
1401 (make_pass_initialize_regs): ...new function.
1402 (pass_combine): Replace declaration with that of...
1403 (make_pass_combine): ...new function.
1404 (pass_if_after_combine): Replace declaration with that of...
1405 (make_pass_if_after_combine): ...new function.
1406 (pass_ree): Replace declaration with that of...
1407 (make_pass_ree): ...new function.
1408 (pass_partition_blocks): Replace declaration with that of...
1409 (make_pass_partition_blocks): ...new function.
1410 (pass_match_asm_constraints): Replace declaration with that of...
1411 (make_pass_match_asm_constraints): ...new function.
1412 (pass_regmove): Replace declaration with that of...
1413 (make_pass_regmove): ...new function.
1414 (pass_split_all_insns): Replace declaration with that of...
1415 (make_pass_split_all_insns): ...new function.
1416 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1417 (make_pass_fast_rtl_byte_dce): ...new function.
1418 (pass_lower_subreg2): Replace declaration with that of...
1419 (make_pass_lower_subreg2): ...new function.
1420 (pass_mode_switching): Replace declaration with that of...
1421 (make_pass_mode_switching): ...new function.
1422 (pass_sms): Replace declaration with that of...
1423 (make_pass_sms): ...new function.
1424 (pass_sched): Replace declaration with that of...
1425 (make_pass_sched): ...new function.
1426 (pass_ira): Replace declaration with that of...
1427 (make_pass_ira): ...new function.
1428 (pass_reload): Replace declaration with that of...
1429 (make_pass_reload): ...new function.
1430 (pass_clean_state): Replace declaration with that of...
1431 (make_pass_clean_state): ...new function.
1432 (pass_branch_prob): Replace declaration with that of...
1433 (make_pass_branch_prob): ...new function.
1434 (pass_value_profile_transformations): Replace declaration with that
1436 (make_pass_value_profile_transformations): ...new function.
1437 (pass_postreload_cse): Replace declaration with that of...
1438 (make_pass_postreload_cse): ...new function.
1439 (pass_gcse2): Replace declaration with that of...
1440 (make_pass_gcse2): ...new function.
1441 (pass_split_after_reload): Replace declaration with that of...
1442 (make_pass_split_after_reload): ...new function.
1443 (pass_branch_target_load_optimize1): Replace declaration with that
1445 (make_pass_branch_target_load_optimize1): ...new function.
1446 (pass_thread_prologue_and_epilogue): Replace declaration with that
1448 (make_pass_thread_prologue_and_epilogue): ...new function.
1449 (pass_stack_adjustments): Replace declaration with that of...
1450 (make_pass_stack_adjustments): ...new function.
1451 (pass_peephole2): Replace declaration with that of...
1452 (make_pass_peephole2): ...new function.
1453 (pass_if_after_reload): Replace declaration with that of...
1454 (make_pass_if_after_reload): ...new function.
1455 (pass_regrename): Replace declaration with that of...
1456 (make_pass_regrename): ...new function.
1457 (pass_cprop_hardreg): Replace declaration with that of...
1458 (make_pass_cprop_hardreg): ...new function.
1459 (pass_reorder_blocks): Replace declaration with that of...
1460 (make_pass_reorder_blocks): ...new function.
1461 (pass_branch_target_load_optimize2): Replace declaration with that
1463 (make_pass_branch_target_load_optimize2): ...new function.
1464 (pass_leaf_regs): Replace declaration with that of...
1465 (make_pass_leaf_regs): ...new function.
1466 (pass_split_before_sched2): Replace declaration with that of...
1467 (make_pass_split_before_sched2): ...new function.
1468 (pass_compare_elim_after_reload): Replace declaration with that of...
1469 (make_pass_compare_elim_after_reload): ...new function.
1470 (pass_sched2): Replace declaration with that of...
1471 (make_pass_sched2): ...new function.
1472 (pass_stack_regs): Replace declaration with that of...
1473 (make_pass_stack_regs): ...new function.
1474 (pass_stack_regs_run): Replace declaration with that of...
1475 (make_pass_stack_regs_run): ...new function.
1476 (pass_df_finish): Replace declaration with that of...
1477 (make_pass_df_finish): ...new function.
1478 (pass_compute_alignments): Replace declaration with that of...
1479 (make_pass_compute_alignments): ...new function.
1480 (pass_duplicate_computed_gotos): Replace declaration with that of...
1481 (make_pass_duplicate_computed_gotos): ...new function.
1482 (pass_variable_tracking): Replace declaration with that of...
1483 (make_pass_variable_tracking): ...new function.
1484 (pass_free_cfg): Replace declaration with that of...
1485 (make_pass_free_cfg): ...new function.
1486 (pass_machine_reorg): Replace declaration with that of...
1487 (make_pass_machine_reorg): ...new function.
1488 (pass_cleanup_barriers): Replace declaration with that of...
1489 (make_pass_cleanup_barriers): ...new function.
1490 (pass_delay_slots): Replace declaration with that of...
1491 (make_pass_delay_slots): ...new function.
1492 (pass_split_for_shorten_branches): Replace declaration with that of...
1493 (make_pass_split_for_shorten_branches): ...new function.
1494 (pass_split_before_regstack): Replace declaration with that of...
1495 (make_pass_split_before_regstack): ...new function.
1496 (pass_convert_to_eh_region_ranges): Replace declaration with that
1498 (make_pass_convert_to_eh_region_ranges): ...new function.
1499 (pass_shorten_branches): Replace declaration with that of...
1500 (make_pass_shorten_branches): ...new function.
1501 (pass_set_nothrow_function_flags): Replace declaration with that of...
1502 (make_pass_set_nothrow_function_flags): ...new function.
1503 (pass_dwarf2_frame): Replace declaration with that of...
1504 (make_pass_dwarf2_frame): ...new function.
1505 (pass_final): Replace declaration with that of...
1506 (make_pass_final): ...new function.
1507 (pass_rtl_seqabstr): Replace declaration with that of...
1508 (make_pass_rtl_seqabstr): ...new function.
1509 (pass_release_ssa_names): Replace declaration with that of...
1510 (make_pass_release_ssa_names): ...new function.
1511 (pass_early_inline): Replace declaration with that of...
1512 (make_pass_early_inline): ...new function.
1513 (pass_inline_parameters): Replace declaration with that of...
1514 (make_pass_inline_parameters): ...new function.
1515 (pass_update_address_taken): Replace declaration with that of...
1516 (make_pass_update_address_taken): ...new function.
1517 (pass_convert_switch): Replace declaration with that of...
1518 (make_pass_convert_switch): ...new function.
1519 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1520 to a subclass of simple_ipa_opt_pass along with...
1521 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1522 (make_pass_ipa_tree_profile): ...new function.
1523 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1524 subclass of gimple_opt_pass along with...
1525 (pass_data_sra_early): ...new pass_data instance and...
1526 (make_pass_sra_early): ...new function.
1527 (pass_sra): Convert from a global struct to a subclass of
1528 gimple_opt_pass along with...
1529 (pass_data_sra): ...new pass_data instance and...
1530 (make_pass_sra): ...new function.
1531 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1532 gimple_opt_pass along with...
1533 (pass_data_early_ipa_sra): ...new pass_data instance and...
1534 (make_pass_early_ipa_sra): ...new function.
1535 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1536 subclass of gimple_opt_pass along with...
1537 (pass_data_ccp): ...new pass_data instance and...
1538 (make_pass_ccp): ...new function.
1539 (pass_fold_builtins): Convert from a global struct to a subclass of
1540 gimple_opt_pass along with...
1541 (pass_data_fold_builtins): ...new pass_data instance and...
1542 (make_pass_fold_builtins): ...new function.
1543 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1544 subclass of gimple_opt_pass along with...
1545 (pass_data_copy_prop): ...new pass_data instance and...
1546 (make_pass_copy_prop): ...new function.
1547 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1548 global struct to a subclass of gimple_opt_pass along with...
1549 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1550 (make_pass_rename_ssa_copies): ...new function.
1551 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1552 subclass of gimple_opt_pass along with...
1553 (pass_data_dce): ...new pass_data instance and...
1554 (make_pass_dce): ...new function.
1555 (pass_dce_loop): Convert from a global struct to a subclass of
1556 gimple_opt_pass along with...
1557 (pass_data_dce_loop): ...new pass_data instance and...
1558 (make_pass_dce_loop): ...new function.
1559 (pass_cd_dce): Convert from a global struct to a subclass of
1560 gimple_opt_pass along with...
1561 (pass_data_cd_dce): ...new pass_data instance and...
1562 (make_pass_cd_dce): ...new function.
1563 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1564 subclass of gimple_opt_pass along with...
1565 (pass_data_dominator): ...new pass_data instance and...
1566 (make_pass_dominator): ...new function.
1567 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1568 gimple_opt_pass along with...
1569 (pass_data_phi_only_cprop): ...new pass_data instance and...
1570 (make_pass_phi_only_cprop): ...new function.
1571 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1572 subclass of gimple_opt_pass along with...
1573 (pass_data_dse): ...new pass_data instance and...
1574 (make_pass_dse): ...new function.
1575 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1576 a subclass of gimple_opt_pass along with...
1577 (pass_data_forwprop): ...new pass_data instance and...
1578 (make_pass_forwprop): ...new function.
1579 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1580 struct to a subclass of gimple_opt_pass along with...
1581 (pass_data_tree_ifcombine): ...new pass_data instance and...
1582 (make_pass_tree_ifcombine): ...new function.
1583 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1584 subclass of gimple_opt_pass along with...
1585 (pass_data_ch): ...new pass_data instance and...
1586 (make_pass_ch): ...new function.
1587 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1588 subclass of gimple_opt_pass along with...
1589 (pass_data_tree_loop): ...new pass_data instance and...
1590 (make_pass_tree_loop): ...new function.
1591 (pass_tree_loop_init): Convert from a global struct to a subclass of
1592 gimple_opt_pass along with...
1593 (pass_data_tree_loop_init): ...new pass_data instance and...
1594 (make_pass_tree_loop_init): ...new function.
1595 (pass_lim): Convert from a global struct to a subclass of
1596 gimple_opt_pass along with...
1597 (pass_data_lim): ...new pass_data instance and...
1598 (make_pass_lim): ...new function.
1599 (pass_tree_unswitch): Convert from a global struct to a subclass of
1600 gimple_opt_pass along with...
1601 (pass_data_tree_unswitch): ...new pass_data instance and...
1602 (make_pass_tree_unswitch): ...new function.
1603 (pass_predcom): Convert from a global struct to a subclass of
1604 gimple_opt_pass along with...
1605 (pass_data_predcom): ...new pass_data instance and...
1606 (make_pass_predcom): ...new function.
1607 (pass_vectorize): Convert from a global struct to a subclass of
1608 gimple_opt_pass along with...
1609 (pass_data_vectorize): ...new pass_data instance and...
1610 (make_pass_vectorize): ...new function.
1611 (pass_graphite): Convert from a global struct to a subclass of
1612 gimple_opt_pass along with...
1613 (pass_data_graphite): ...new pass_data instance and...
1614 (make_pass_graphite): ...new function.
1615 (pass_graphite_transforms): Convert from a global struct to a subclass
1616 of gimple_opt_pass along with...
1617 (pass_data_graphite_transforms): ...new pass_data instance and...
1618 (make_pass_graphite_transforms): ...new function.
1619 (pass_check_data_deps): Convert from a global struct to a subclass of
1620 gimple_opt_pass along with...
1621 (pass_data_check_data_deps): ...new pass_data instance and...
1622 (make_pass_check_data_deps): ...new function.
1623 (pass_iv_canon): Convert from a global struct to a subclass of
1624 gimple_opt_pass along with...
1625 (pass_data_iv_canon): ...new pass_data instance and...
1626 (make_pass_iv_canon): ...new function.
1627 (pass_scev_cprop): Convert from a global struct to a subclass of
1628 gimple_opt_pass along with...
1629 (pass_data_scev_cprop): ...new pass_data instance and...
1630 (make_pass_scev_cprop): ...new function.
1631 (pass_record_bounds): Convert from a global struct to a subclass of
1632 gimple_opt_pass along with...
1633 (pass_data_record_bounds): ...new pass_data instance and...
1634 (make_pass_record_bounds): ...new function.
1635 (pass_complete_unroll): Convert from a global struct to a subclass of
1636 gimple_opt_pass along with...
1637 (pass_data_complete_unroll): ...new pass_data instance and...
1638 (make_pass_complete_unroll): ...new function.
1639 (pass_complete_unrolli): Convert from a global struct to a subclass of
1640 gimple_opt_pass along with...
1641 (pass_data_complete_unrolli): ...new pass_data instance and...
1642 (make_pass_complete_unrolli): ...new function.
1643 (pass_parallelize_loops): Convert from a global struct to a subclass
1644 of gimple_opt_pass along with...
1645 (pass_data_parallelize_loops): ...new pass_data instance and...
1646 (make_pass_parallelize_loops): ...new function.
1647 (pass_loop_prefetch): Convert from a global struct to a subclass of
1648 gimple_opt_pass along with...
1649 (pass_data_loop_prefetch): ...new pass_data instance and...
1650 (make_pass_loop_prefetch): ...new function.
1651 (pass_iv_optimize): Convert from a global struct to a subclass of
1652 gimple_opt_pass along with...
1653 (pass_data_iv_optimize): ...new pass_data instance and...
1654 (make_pass_iv_optimize): ...new function.
1655 (pass_tree_loop_done): Convert from a global struct to a subclass of
1656 gimple_opt_pass along with...
1657 (pass_data_tree_loop_done): ...new pass_data instance and...
1658 (make_pass_tree_loop_done): ...new function.
1659 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1660 struct to a subclass of gimple_opt_pass along with...
1661 (pass_data_cse_reciprocals): ...new pass_data instance and...
1662 (make_pass_cse_reciprocals): ...new function.
1663 (pass_cse_sincos): Convert from a global struct to a subclass of
1664 gimple_opt_pass along with...
1665 (pass_data_cse_sincos): ...new pass_data instance and...
1666 (make_pass_cse_sincos): ...new function.
1667 (pass_optimize_bswap): Convert from a global struct to a subclass of
1668 gimple_opt_pass along with...
1669 (pass_data_optimize_bswap): ...new pass_data instance and...
1670 (make_pass_optimize_bswap): ...new function.
1671 (pass_optimize_widening_mul): Convert from a global struct to a
1672 subclass of gimple_opt_pass along with...
1673 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1674 (make_pass_optimize_widening_mul): ...new function.
1675 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1676 subclass of gimple_opt_pass along with...
1677 (pass_data_phiopt): ...new pass_data instance and...
1678 (make_pass_phiopt): ...new function.
1679 (pass_cselim): Convert from a global struct to a subclass of
1680 gimple_opt_pass along with...
1681 (pass_data_cselim): ...new pass_data instance and...
1682 (make_pass_cselim): ...new function.
1683 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1684 subclass of gimple_opt_pass along with...
1685 (pass_data_phiprop): ...new pass_data instance and...
1686 (make_pass_phiprop): ...new function.
1687 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1688 subclass of gimple_opt_pass along with...
1689 (pass_data_pre): ...new pass_data instance and...
1690 (make_pass_pre): ...new function.
1691 (pass_fre): Convert from a global struct to a subclass of
1692 gimple_opt_pass along with...
1693 (pass_data_fre): ...new pass_data instance and...
1694 (make_pass_fre): ...new function.
1695 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1696 subclass of gimple_opt_pass along with...
1697 (pass_data_reassoc): ...new pass_data instance and...
1698 (make_pass_reassoc): ...new function.
1699 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1700 subclass of gimple_opt_pass along with...
1701 (pass_data_sink_code): ...new pass_data instance and...
1702 (make_pass_sink_code): ...new function.
1703 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1704 subclass of gimple_opt_pass along with...
1705 (pass_data_strlen): ...new pass_data instance and...
1706 (make_pass_strlen): ...new function.
1707 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1708 struct to a subclass of gimple_opt_pass along with...
1709 (pass_data_build_alias): ...new pass_data instance and...
1710 (make_pass_build_alias): ...new function.
1711 (pass_build_ealias): Convert from a global struct to a subclass of
1712 gimple_opt_pass along with...
1713 (pass_data_build_ealias): ...new pass_data instance and...
1714 (make_pass_build_ealias): ...new function.
1715 (pass_ipa_pta): Convert from a global struct to a subclass of
1716 simple_ipa_opt_pass along with...
1717 (pass_data_ipa_pta): ...new pass_data instance and...
1718 (make_pass_ipa_pta): ...new function.
1719 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1720 subclass of gimple_opt_pass along with...
1721 (pass_data_uncprop): ...new pass_data instance and...
1722 (make_pass_uncprop): ...new function.
1723 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1724 global struct to a subclass of gimple_opt_pass along with...
1725 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1726 (make_pass_late_warn_uninitialized): ...new function.
1727 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1728 to a subclass of gimple_opt_pass along with...
1729 (pass_data_init_datastructures): ...new pass_data instance and...
1730 (make_pass_init_datastructures): ...new function.
1731 (pass_early_warn_uninitialized): Convert from a global struct to a
1732 subclass of gimple_opt_pass along with...
1733 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1734 (make_pass_early_warn_uninitialized): ...new function.
1735 (pass_update_address_taken): Convert from a global struct to a
1736 subclass of gimple_opt_pass along with...
1737 (pass_data_update_address_taken): ...new pass_data instance and...
1738 (make_pass_update_address_taken): ...new function.
1739 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1740 struct to a subclass of gimple_opt_pass along with...
1741 (pass_data_release_ssa_names): ...new pass_data instance and...
1742 (make_pass_release_ssa_names): ...new function.
1743 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1744 subclass of gimple_opt_pass along with...
1745 (pass_data_stdarg): ...new pass_data instance and...
1746 (make_pass_stdarg): ...new function.
1747 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1748 global struct to a subclass of gimple_opt_pass along with...
1749 (pass_data_convert_switch): ...new pass_data instance and...
1750 (make_pass_convert_switch): ...new function.
1751 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1752 to a subclass of gimple_opt_pass along with...
1753 (pass_data_tail_recursion): ...new pass_data instance and...
1754 (make_pass_tail_recursion): ...new function.
1755 (pass_tail_calls): Convert from a global struct to a subclass of
1756 gimple_opt_pass along with...
1757 (pass_data_tail_calls): ...new pass_data instance and...
1758 (make_pass_tail_calls): ...new function.
1759 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1760 struct to a subclass of gimple_opt_pass along with...
1761 (pass_data_lower_vector): ...new pass_data instance and...
1762 (make_pass_lower_vector): ...new function.
1763 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1764 gimple_opt_pass along with...
1765 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1766 (make_pass_lower_vector_ssa): ...new function.
1767 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
1768 to a subclass of gimple_opt_pass along with...
1769 (pass_data_slp_vectorize): ...new pass_data instance and...
1770 (make_pass_slp_vectorize): ...new function.
1771 (pass_ipa_increase_alignment): Convert from a global struct to a
1772 subclass of simple_ipa_opt_pass along with...
1773 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
1774 (make_pass_ipa_increase_alignment): ...new function.
1775 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
1776 gimple_opt_pass along with...
1777 (pass_data_vrp): ...new pass_data instance and...
1778 (make_pass_vrp): ...new function.
1779 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
1780 subclass of simple_ipa_opt_pass along with...
1781 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
1782 (make_pass_ipa_free_lang_data): ...new function.
1783 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
1784 gimple_opt_pass along with...
1785 (pass_data_tsan): ...new pass_data instance and...
1786 (make_pass_tsan): ...new function.
1787 (pass_tsan_O0): Convert from a global struct to a subclass of
1788 gimple_opt_pass along with...
1789 (pass_data_tsan_O0): ...new pass_data instance and...
1790 (make_pass_tsan_O0): ...new function.
1791 * var-tracking.c (pass_variable_tracking): Convert from a global
1792 struct to a subclass of rtl_opt_pass along with...
1793 (pass_data_variable_tracking): ...new pass_data instance and...
1794 (make_pass_variable_tracking): ...new function.
1795 * web.c (pass_web): Convert from a global struct to a subclass of
1796 rtl_opt_pass along with...
1797 (pass_data_web): ...new pass_data instance and...
1798 (make_pass_web): ...new function.
1799 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
1800 declaration with that of...
1801 (make_pass_mode_switch_use): ...new function.
1802 (pass_resolve_sw_modes): Replace declaration with that of...
1803 (make_pass_resolve_sw_modes): ...new function.
1804 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
1805 from a global struct to a subclass of rtl_opt_pass along with...
1806 (pass_data_mode_switch_use): ...new pass_data instance and...
1807 (make_pass_mode_switch_use): ...new function.
1808 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
1809 from a global struct to a subclass of rtl_opt_pass along with...
1810 (pass_data_resolve_sw_modes): ...new pass_data instance and...
1811 (make_pass_resolve_sw_modes): ...new function.
1812 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
1813 struct to a subclass of rtl_opt_pass along with...
1814 (pass_data_insert_vzeroupper): ...new pass_data instance and...
1815 (make_pass_insert_vzeroupper): ...new function.
1816 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
1817 global struct to a subclass of rtl_opt_pass along with...
1818 (pass_data_work_around_errata): ...new pass_data instance and...
1819 (make_pass_work_around_errata): ...new function.
1820 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
1821 struct to a subclass of rtl_opt_pass along with...
1822 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
1823 (make_pass_mips_machine_reorg2): ...new function.
1825 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1827 * passes.c (pass_manager::operator new): New.
1829 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1831 Handwritten part of conversion of passes to C++ classes.
1833 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
1834 (toplev.o): Add dep on PASS_MANAGER_H.
1835 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
1836 of early local pases to reflect this moving from a global to a
1837 member of gcc::pass_manager.
1838 (cgraph_add_new_function): Likewise.
1839 * lto-cgraph.c (lto_output_node): Update for conversion of
1840 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
1841 * passes.c (opt_pass::clone): New.
1842 (opt_pass::gate): New.
1843 (opt_pass::execute): New.
1844 (opt_pass::opt_pass): New.
1845 (pass_manager::execute_early_local_passes): New.
1846 (pass_manager::execute_pass_mode_switching): new.
1847 (finish_optimization_passes): Convert to...
1848 (pass_manager::finish_optimization_passes): ...this.
1849 (finish_optimization_passes): Update for conversion of passes to
1851 (register_dump_files_1): Use has_gate since we cannot portably
1852 check a vtable entry against NULL.
1853 (dump_one_pass): Likewise.
1854 (ipa_write_summaries_2): Likewise.
1855 (ipa_write_optimization_summaries_1): Likewise.
1856 (ipa_read_summaries_1): Likewise.
1857 (ipa_read_optimization_summaries_1): Likewise.
1858 (execute_ipa_stmt_fixups): Likewise.
1859 (pass_manager::pass_manager): Rewrite pass-creation, invoking
1860 pass-creation functions rather than wiring up globals, and
1861 storing the results in fields of pass_manager generated using
1863 (pass_manager::dump_profile_report): Update for conversion of
1864 passes to C++ classes.
1865 (pass_manager::execute_ipa_summary_passes): Likewise.
1866 (execute_one_ipa_transform_pass): Likewise.
1867 (execute_one_pass): Use has_gate and has_execute since we cannot
1868 portably check a vtable entry against NULL.
1869 * pass_manager.h (pass_manager::finish_optimization_passes): New.
1870 (pass_manager): Use pass-instances.def to add fields for the
1871 various pass instances.
1872 * toplev.c (finalize): Update for move of
1873 finish_optimization_passes to a method of gcc::pass_manager.
1874 * toplev.h (finish_optimization_passes): Move to method of class
1876 * tree-pass.h (struct pass_data): New.
1877 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
1878 (opt_pass::gate): Convert to virtual function.
1879 (opt_pass::~opt_pass): New.
1880 (opt_pass::clone): New.
1881 (opt_pass::execute): Convert to virtual function.
1882 (opt_pass::opt_pass): New.
1883 (opt_pass::ctxt_): new.
1884 (gimple_opt_pass): Convert to subclass of opt_pass.
1885 (gimple_opt_pass::gimple_opt_pass): New.
1886 (rtl_opt_pass): Convert to subclass of opt_pass.
1887 (rtl_opt_pass::rtl_opt_pass): New.
1888 (ipa_opt_pass_d): Convert to subclass of opt_pass.
1889 (ipa_opt_pass_d::ipa_opt_pass_d): New.
1890 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
1891 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
1892 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
1893 invocation of pass_mode_switching to reflect this moving from a
1894 global to a member of gcc::pass_manager.
1895 (ix86_option_override): Rework how pass_insert_vzeroupper is
1896 added to the pass_manager to reflect autogenerated changes.
1897 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
1899 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
1901 PR rtl-optimization/57708
1902 * recog.c (peep2_find_free_register): Validate all regs in a
1905 2013-08-05 Jan Hubicka <jh@suse.cz>
1908 * cgraph.c (verify_cgraph_node): Accept local flags from other
1910 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
1911 (function_and_variable_visibility): Likewise.
1912 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
1914 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1916 * graph.c (init_graph_slim_pretty_print): Remove.
1917 (print_graph_cfg): Do not call it. Use local pretty printer.
1918 (start_graph_dump): Likewise.
1920 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1922 * gimple-pretty-print.c (buffer): Remove.
1923 (initialized): Likewise.
1924 (maybe_init_pretty_print): Likewise.
1925 (print_gimple_stmt): Do not call it. Use non-static local
1926 pretty_printer variable.
1927 (print_gimple_expr): Likewise.
1928 (print_gimple_seq): Likewise.
1929 (gimple_dump_bb): Likewise.
1931 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
1933 * asan.c (asan_pp): Remove.
1934 (asan_pp_initialized): Likewise.
1935 (asan_pp_initialize): Likewise.
1936 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
1937 (asan_emit_stack_protection): Tidy. Use local pretty printer.
1938 (asan_add_global): Likewise.
1940 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
1942 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
1943 * diagnostic.h (diagnostic_flush_buffer): Adjust.
1944 * pretty-print.c (pp_formatted_text_data): Likewise.
1945 (pp_indent): Rename from pp_base_indent.
1946 (pp_format): Rename from pp_base_format.
1947 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
1948 (pp_format_verbatim): Rename from pp_base_format_verbatim.
1949 (pp_flush): Rename from pp_base_flush.
1950 (pp_set_line_maximum_length): Rename from
1951 pp_base_set_line_maximum_length.
1952 (pp_clear_output_area): Rename from pp_base_clear_output_area.
1953 (pp_set_prefix): Rename from pp_base_set_prefix.
1954 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
1955 (pp_emit_prefix): Rename from pp_base_emit_prefix.
1956 (pp_append_text): Rename from pp_base_append_text.
1957 (pp_formatted_text): Rename from pp_base_formatted_text.
1958 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
1959 (pp_remaining_character_count_for_line): Rename from
1960 pp_base_remaining_character_count_for_line.
1961 (pp_newline): Rename from pp_base_newline.
1962 (pp_character): Rename from pp_base_character.
1963 (pp_string): Rename from pp_base_string.
1964 (pp_maybe_space): Rename from pp_base_maybe_space.
1965 * asan.c (asan_pp_string): Adjust.
1966 (asan_emit_stack_protection): Likewise.
1967 (asan_add_global): Likewise.
1968 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
1969 * tree-mudflap.c (mf_varname_tree): Likewise.
1970 * tree-pretty-print.c (pp_tree_identifier): Rename from
1971 pp_base_tree_identifier.
1972 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
1973 Declare as function.
1975 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
1977 * pretty-print.h (pp_bar_bar): New.
1978 (pp_ampersand_ampersand): Likewise.
1979 (pp_less_equal): Likewise.
1980 (pp_greater_equal): Likewise.
1981 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
1982 printer functions instead of pp_string or operators and punctuators.
1983 (dump_gimple_call): Likewise.
1984 (dump_gimple_omp_for): Likewise.
1985 (dump_gimple_transaction): Likewise.
1986 (dump_gimple_phi): Likewise.
1987 (pp_gimple_stmt_1): Likewise.
1988 * sched-vis.c (print_insn): Likewise.
1989 * tree-mudflap.c (mf_varname_tree): Likewise.
1990 * tree-pretty-print.c (dump_block_node): Likewise.
1991 (dump_generic_node): Likewise.
1993 2013-08-02 Jan Hubicka <jh@suse.cz>
1995 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
1997 * lto-streamer-out.c (tree_is_indexable): Results decls and
1998 parm decls are not indexable.
1999 (DFS_write_tree_body): Do not follow args and results.
2000 (hash_tree): Likewise.
2001 (output_functions): Rearrange so struct function is needed
2002 only when real body is output; be able to also ouptut abstract
2003 functions; output DECL_ARGUMENTS and DECL_RESULT.
2004 (lto_output): When not in WPA, ale store abstract functions.
2005 (write_symbol): Do not care about RESULT_DECL.
2006 (output_symbol_p): Handle correctly sbtract decls.
2007 * lto-streamer-in.c (input_function): Rearrange so struct
2008 function can be NULL at entry; allow streaming of
2009 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2010 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2011 sanity check during LTO.
2012 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2013 RESULT_DECl and DECL_ARGUMENTS.
2014 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2017 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2019 * pretty-print.h (pp_underscore): New.
2021 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2022 printer functions instead of pp_character.
2023 (dump_binary_rhs): Likewise.
2024 (dump_ternary_rhs): Likewise.
2025 (dump_gimple_call_args): Likewise.
2026 (pp_points_to_solution): Likewise.
2027 (dump_gimple_call): Likewise.
2028 (dump_gimple_switch): Likewise.
2029 (dump_gimple_cond): Likewise.
2030 (dump_gimple_bind): Likewise.
2031 (dump_gimple_try): Likewise.
2032 (dump_gimple_omp_for): Likewise.
2033 (dump_gimple_omp_continue): Likewise.
2034 (dump_gimple_omp_single): Likewise.
2035 (dump_gimple_omp_sections): Likewise.
2036 (dump_gimple_omp_block): Likewise.
2037 (dump_gimple_omp_critical): Likewise.
2038 (dump_gimple_transaction): Likewise.
2039 (dump_gimple_asm): Likewise.
2040 (dump_gimple_phi): Likewise.
2041 (dump_gimple_omp_parallel): Likewise.
2042 (dump_gimple_omp_task): Likewise.
2043 (dump_gimple_omp_atomic_load): Likewise.
2044 (dump_gimple_omp_atomic_store): Likewise.
2045 (dump_gimple_mem_ops): Likewise.
2046 (pp_gimple_stmt_1): Likewise.
2047 (pp_cfg_jump): Likewise.
2048 (dump_implicit_edges): Likewise.
2049 (gimple_dump_bb_for_graph): Likewise.
2050 * graph.c (draw_cfg_node): Likewise.
2051 * langhooks.c (lhd_print_error_function): Likewise.
2052 * sched-vis.c (print_exp): Likewise.
2053 (print_value): Likewise.
2054 (print_pattern): Likewise.
2055 (print_insn): Likewise.
2056 (rtl_dump_bb_for_graph): Likewise.
2057 * tree-pretty-print.c (dump_function_declaration): Likewise.
2058 (dump_array_domain): Likewise.
2059 (dump_omp_clause): Likewise.
2060 (dump_location): Likewise.
2061 (dump_generic_node): Likewise.
2062 (print_struct_decl): Likewise.
2063 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2065 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2067 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2068 candidate table when replacing a candidate statement.
2069 (replace_rhs_if_not_dup): Likewise.
2070 (replace_one_candidate): Likewise.
2072 2013-08-02 Jan Hubicka <jh@suse.cz>
2073 Martin Liska <marxin.liska@gmail.com>
2075 * cgraphunit.c (add_new_function): Fix logic when adding from
2077 (assemble_thunk): Rename to ...
2078 (expand_thunk); .. this one; export; get it working with
2079 general functions; make produced gimple valid.
2080 * cgraph.h (expand_thunk): Declare.
2082 2013-08-02 Jan Hubicka <jh@suse.cz>
2084 * ipa-cp.c (gather_context_independent_values): Use
2085 ipa_get_param_move_cost.
2086 (get_replacement_map): Remove PARAM; move parameter folding
2088 (create_specialized_node): Update.
2089 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2090 assert that we have gimple body; update move_cost.
2091 (count_formal_params): Assert that we have gimple body.
2092 (ipa_dump_param): New function.
2093 (ipa_alloc_node_params): Break out from ...
2094 (ipa_initialize_node_params): ... here.
2095 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2096 (ipa_write_node_info): Stream move costs.
2097 (ipa_read_node_info): Read move costs.
2098 (ipa_update_after_lto_read): Do not recompute node params.
2099 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2100 (ipa_get_param): Check we are not in WPA.
2101 (ipa_get_param_move_cost): New.
2102 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2103 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2104 parm numbers to be present.
2106 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2108 PR rtl-optimization/58048
2109 * lra-constraints.c (process_alt_operands): Don't check asm
2110 operand on register.
2112 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2114 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2115 the implied StoreLoad barrier for atomic operations if before.
2117 2013-08-02 Jan Hubicka <jh@suse.cz>
2118 Martin Liska <marxin.liska@gmail.com>
2120 * cgraph.c (cgraph_function_body_availability): Do not check
2122 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2123 symtab_node_availability): Declare.
2124 * ipa.c (can_replace_by_local_alias): New.
2125 (function_and_variable_visibility): Use it.
2126 * symtab.c (symtab_for_node_and_aliases,
2127 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2129 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2131 PR rtl-optimization/57963
2132 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2133 (lra_constraints): Use them.
2135 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2137 * config/arm/types.md (define_attr "type"): Add "load_acq"
2139 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2141 (cortex_a53_store1): Likewise.
2143 2013-08-01 Jan Hubicka <jh@suse.cz>
2145 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
2146 partitions are not needed.
2148 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
2150 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
2151 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
2152 MAYBE_NON_Q_CLASS_P where appropriate.
2154 2013-08-01 Jan Hubicka <jh@suse.cz>
2156 * cgraph.h (release_function_body): Declare.
2157 * tree.c (free_lang_data_in_decl): Free, parameters and return values
2158 of unused delcarations.
2160 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2162 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
2163 RTL form when subtracting a constant.
2165 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2167 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
2168 Generate canonical plus rtx with negated immediate instead of minus
2170 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
2172 2013-08-01 Jan Hubicka <jh@suse.cz>
2174 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
2175 (cgraph_release_function_body): Likewise.
2176 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2177 * cgraph.h (cgrpah_node): Rename abstract_and_needed
2178 to used_as_abstract_origin.
2179 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
2180 symbols used as abstract origins.
2181 * cgraphunit.c (analyze_functions): Update.
2182 * ipa.c (symtab_remove_unreachable_nodes): Recompute
2183 used_as_abstract_origin.
2184 * tree-inline.c (tree_function_versioning): Update
2185 used_as_abstract_origin; be ready for DECL_RESULT and
2186 DECL_ARGUMENTS to be NULL.
2188 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
2189 for abstract functions.
2190 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
2193 2013-08-01 Jan Hubicka <jh@suse.cz>
2195 * profile.c (compute_value_histograms): Fix thinko.
2197 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2199 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
2200 aarch-common-protos.h to extra_headers.
2201 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2202 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
2203 * config/aarch64/t-aarch64 (aarch-common.o): Define.
2205 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2207 * config/aarch64/aarch64.md (define_attr "type"): Delete.
2208 Include "../arm/types.md". Define "type" attribute for all patterns.
2209 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
2212 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
2214 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
2215 to support power8 load fusion.
2216 (fusion_gpr_mem_load): Likewise.
2218 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
2220 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
2221 declarations for power8 load fusion.
2222 (emit_fusion_gpr_load): Likewise.
2224 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
2225 tuning for power8, turn on fusion mode by default. Turn on sign
2226 extending fusion mode if normal fusion mode is on, and we are at
2228 (fusion_gpr_load_p): New function, return true if we can fuse an
2229 addis instruction with a dependent load to a GPR.
2230 (emit_fusion_gpr_load): Emit the instructions for power8 load
2233 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
2234 (VSX load fusion peepholes): New peepholes to fuse together an
2235 addi instruction with a VSX load instruction.
2237 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
2238 peepholes to fuse an addis instruction with a load to a GPR base
2239 register. If we are supporting sign extending fusions, convert
2240 sign extending loads to zero extending loads and add an explicit
2243 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2245 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
2246 aarch-common-protos.h to extra_headers.
2247 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2248 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
2249 (arm_early_store_addr_dep): Likewise.
2250 (arm_no_early_alu_shift_dep): Likewise.
2251 (arm_no_early_alu_shift_value_dep): Likewise.
2252 (arm_no_early_mul_dep): Likewise.
2253 (arm_no_early_store_addr_dep): Likewise.
2254 (arm_mac_accumulator_is_mul_result): Likewise.
2255 (arm_mac_accumulator_is_result): Likewise.
2256 * config/arm/aarch-common.c: ... here. New file.
2257 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
2259 (arm_early_store_addr_dep): Likewise.
2260 (arm_no_early_alu_shift_dep): Likewise.
2261 (arm_no_early_alu_shift_value_dep): Likewise.
2262 (arm_no_early_mul_dep): Likewise.
2263 (arm_no_early_store_addr_dep): Likewise.
2264 (arm_mac_accumulator_is_mul_result): Likewise.
2265 (arm_mac_accumulator_is_result): Likewise.
2266 * config/arm/aarch-common-protos.h: ... here. New file.
2267 * config/arm/t-arm (aarch-common.o): Define.
2269 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2271 * config/arm/arm.md: Include new file "types.md".
2272 (define_attr "type"): Move from here to ...
2273 (define_attr "mul32"): Likewise.
2274 (define_attr "mul64"): Likewise.
2275 * config/arm/types.md: ... here. New file.
2277 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
2279 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
2280 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
2282 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2284 * gen-pass-instances.awk: Fix offset of substr().
2286 2013-07-31 David Malcolm <dmalcolm@redhat.com>
2288 * Makefile.in (pass-instances.def): New.
2289 (passes.o): Replace dependency on passes.def with one on
2292 * gen-pass-instances.awk: New.
2294 * passes.c (pass_manager::pass_manager): Use pass-instances.def
2295 rather than passes.def, updating local definition of NEXT_PASS
2296 macro to add an extra NUM parameter (currently unused).
2298 2013-07-30 David Malcolm <dmalcolm@redhat.com>
2300 * Makefile.in (PASS_MANAGER_H): New.
2301 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
2302 (passes.o): Likewise.
2303 (statistics.o): Likewise.
2304 (cgraphunit.o): Likewise.
2305 (context.o): Depend on PASS_MANAGER_H.
2307 * pass_manager.h: New.
2309 * cgraphunit.c (cgraph_add_new_function): Update for moves
2310 of globals to fields of pass_manager.
2311 (analyze_function): Likewise.
2312 (expand_function): Likewise.
2313 (ipa_passes): Likewise.
2314 (compile): Likewise.
2316 * context.c (context::context): New.
2317 * context.h (context::context): New.
2318 (context::get_passes): New.
2319 (context::passes_): New.
2321 * lto-cgraph.c (input_node): Update for moves of globals to
2322 fields of pass_manager.
2324 * passes.c (all_passes): Remove, in favor of a field of the
2325 same name within the new class pass_manager.
2326 (all_small_ipa_passes): Likewise.
2327 (all_lowering_passes): Likewise.
2328 (all_regular_ipa_passes): Likewise.
2329 (all_late_ipa_passes): Likewise.
2330 (all_lto_gen_passes): Likewise.
2331 (passes_by_id): Likewise.
2332 (passes_by_id_size): Likewise.
2333 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2334 the new class pass_manager.
2335 (set_pass_for_id): Convert to...
2336 (pass_manager::set_pass_for_id): ...method.
2337 (get_pass_for_id): Convert to...
2338 (pass_manager::get_pass_for_id): ...method.
2339 (register_one_dump_file): Move body of implementation into...
2340 (pass_manager::register_one_dump_file): ...here.
2341 (register_dump_files_1): Convert to...
2342 (pass_manager::register_dump_files_1): ...method.
2343 (register_dump_files): Convert to...
2344 (pass_manager::register_dump_files): ...method.
2345 (create_pass_tab): Update for moves of globals to fields of
2347 (dump_passes): Move body of implementation into...
2348 (pass_manager::dump_passes): ...here.
2349 (register_pass): Move body of implementation into...
2350 (pass_manager::register_pass): ...here.
2351 (init_optimization_passes): Convert into...
2352 (pass_manager::pass_manager): ...constructor for new
2353 pass_manager class, and initialize the pass_lists array.
2354 (check_profile_consistency): Update for moves of globals to
2355 fields of pass_manager.
2356 (dump_profile_report): Move body of implementation into...
2357 (pass_manager::dump_profile_report): ...here.
2358 (ipa_write_summaries_1): Update for moves of pass lists from
2359 being globals to fields of pass_manager.
2360 (ipa_write_optimization_summaries): Likewise.
2361 (ipa_read_summaries): Likewise.
2362 (ipa_read_optimization_summaries): Likewise.
2363 (execute_all_ipa_stmt_fixups): Likewise.
2365 * statistics.c (statistics_fini): Update for moves of globals to
2366 fields of pass_manager.
2368 * toplev.c (general_init): Replace call to
2369 init_optimization_passes with construction of the pass_manager
2372 * tree-pass.h (all_passes): Remove, in favor of a field of the
2373 same name within the new class pass_manager.
2374 (all_small_ipa_passes): Likewise.
2375 (all_lowering_passes): Likewise.
2376 (all_regular_ipa_passes): Likewise.
2377 (all_lto_gen_passes): Likewise.
2378 (all_late_ipa_passes): Likewise.
2379 (passes_by_id): Likewise.
2380 (passes_by_id_size): Likewise.
2381 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2382 the new class pass_manager.
2383 (get_pass_for_id): Remove.
2385 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2387 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
2390 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2392 * arm.md (mulhi3): New expand pattern.
2394 2013-07-30 Jan Hubicka <jh@suse.cz>
2395 Martin Liska <marxin.liska@gmail.com>
2397 * profile.c (compute_value_histograms): Do not ICE when
2398 there is mismatch only on some counters.
2400 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2402 PR rtl-optimization/57637
2403 * function.c (move_insn_for_shrink_wrap): Also check the
2404 GEN set of the LIVE problem for the liveness analysis
2405 if it exists, otherwise give up.
2407 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2409 PR tree-optimization/57993
2410 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2411 replaced statement in the candidate table.
2412 (phi_add_costs): Return infinite cost when the hidden basis does
2413 not dominate all phis on which the candidate is dependent.
2414 (replace_one_candidate): Record replaced statement in the
2417 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2419 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2420 (ashlv2si3): New expander.
2421 (*ashlv2si3_i): New define_insn_and_split.
2422 * predicates.md (float_operation): Allow patterns with three
2425 PR rtl-optimization/58021
2426 * mode-switching.c (create_pre_exit): Always split off preceding
2427 insns if we are not at the basic block head.
2429 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2431 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2432 (UCLIBC_DYNAMIC_LINKER): New macro.
2433 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2435 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2436 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2438 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2439 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2440 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2441 for SF and DF modes. Use ieee_quad_format for TF mode.
2442 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2443 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2444 (mips_option_override): Handle `-mnan=legacy'.
2445 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2446 `-mabs=2008' and `-mnan=2008'.
2447 (OPTION_DEFAULT_SPECS): Add "nan" default.
2448 (ASM_SPEC): Handle `-mnan='.
2449 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2450 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2451 comment accordingly.
2452 (neg<mode>2): Likewise.
2453 * config/mips/mips.opt (mabs, mnan): New options.
2454 * doc/install.texi (Configuration): Document `--with-nan=' option.
2455 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2457 (MIPS Options): Document them.
2458 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2459 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2460 * configure: Regenerate.
2461 * config.in: Regenerate.
2463 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2465 * config/i386/i386.md (float post-reload splitters): Do not check
2466 for subregs of SSE registers.
2468 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2469 H.J. Lu <hongjiu.lu@intel.com>
2473 * config/i386/i386.md (post-reload splitter
2474 to avoid partial SSE reg dependency stalls): New pattern.
2476 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2478 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2480 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2482 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2483 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2485 2013-07-26 Cary Coutant <ccoutant@google.com>
2487 * dwarf2out.c (die_checksum_ordered): Don't include template
2488 instantiations in signature.
2489 (is_template_parameter): New function.
2490 (is_template_instantiation): New function.
2491 (generate_skeleton_bottom_up): Don't include template instantiations
2493 (generate_skeleton): Likewise.
2494 (break_out_comdat_types): Move recursive call to break out nested
2496 (prune_unused_types_mark_generic_parms_dies): Call
2497 is_template_parameter.
2499 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2501 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2502 uses vector registers.
2503 * config/aarch64/iterators.md: Add attributes rtn and vas.
2505 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2506 Richard Earnshaw <richard.earnshaw@arm.com>
2508 * combine.c (simplify_comparison): Re-canonicalize operands
2510 * config/arm/arm.md (movcond_addsi): New splitter.
2512 2013-07-25 Sterling Augustine <saugustine@google.com>
2514 * dwarf2out.c (size_of_pubnames): Move code to...
2515 (include_pubname_in_output): ...here. New.
2516 (want_pubnames): Rearrange.
2517 (output_pubnames): Call include_pubname_in_output. Move assertion.
2519 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2521 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2523 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2526 * doc/extend.texi: Remove obsolete builtins. Fix
2527 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2529 2013-07-25 Jan Hubicka <jh@suse.cz>
2531 * cgraph.c (release_function_body): Break out from ...
2532 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2534 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2535 old_tree in the map.
2536 (create_specialized_node): Update.
2537 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2539 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2540 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2541 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2542 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2545 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2547 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2548 addsi3_carryin_alt2_<optab>): Correct output template.
2550 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2552 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2553 Adjust for arm_restrict_it.
2554 Remove trailing whitespace.
2556 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2558 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2559 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2561 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2563 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2565 PR rtl-optimization/57960
2566 * lra-constraints.c (process_alt_operands): Use the right mode
2567 when checking strict_low.
2569 2013-07-25 Jan Hubicka <jh@suse.cz>
2571 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2572 * cgraph.c (cgraph_remove_node): Do not release function body
2573 when in cgraph streaming.
2574 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
2575 in other partitions are not considered reachable; fix handling of
2578 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2580 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2582 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2587 * config/arm/arm.md ("*sibcall_insn): Replace use of
2588 Ss with US. Adjust output for v5 and v4t.
2589 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
2591 * config/arm/constraints.md ("Ss"): Rename to US.
2593 2013-07-25 Terry Guo <terry.guo@arm.com>
2595 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2596 shift_add/shift_sub0/shift_sub1 RTXs.
2598 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2599 Anton Blanchard <anton@au1.ibm.com>
2601 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2602 (altivec_vpks<VI_char>ss): Likewise.
2603 (altivec_vpks<VI_char>us): Likewise.
2604 (altivec_vpku<VI_char>us): Likewise.
2605 (altivec_vpku<VI_char>um): Likewise.
2607 2013-07-24 David Malcolm <dmalcolm@redhat.com>
2609 Introduce context class.
2611 * Makefile.in (CONTEXT_H): New.
2612 (OBJS): Add context.o.
2613 (toplev.o): Add CONTEXT_H to dependencies.
2616 * toplev.c (general_init): Create the singleton gcc::context instance.
2622 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2624 PR rtl-optimization/57968
2625 * mode-switching.c (create_pre_exit): Allow instructions that
2626 don't set a return register to need a non-exit mode.
2628 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2629 Anton Blanchard <anton@au1.ibm.com>
2631 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2632 operands to vperm for little endian.
2633 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2634 of lvsl to create the control mask for a vperm for little endian.
2636 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2637 Anton Blanchard <anton@au1.ibm.com>
2639 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2640 two operands for little-endian.
2642 2013-07-23 Steve Ellcey <sellcey@mips.com>
2644 * config/mips/mips.c (mips_case_values_threshold): New.
2645 (TARGET_CASE_VALUES_THRESHOLD): Define.
2647 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2648 Anton Blanchard <anton@au1.ibm.com>
2650 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2651 selection of field for vector splat in little endian mode.
2653 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2655 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2656 expanders to rs6000.md.
2657 (ior<mode>3): Likewise.
2658 (and<mode>3): Likewise.
2659 (one_cmpl<mode>2): Likewise.
2660 (nor<mode>3): Likewise.
2661 (andc<mode>3): Likewise.
2662 (eqv<mode>3): Likewise.
2663 (nand<mode>3): Likewise.
2664 (orc<mode>3): Likewise.
2666 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2669 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2670 to split multi-word logical operations.
2671 (rs6000_split_logical_di): Likewise.
2672 (rs6000_split_logical): Likewise.
2674 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2675 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2676 and allow TImode operations in 32-bit.
2677 (vsx_and<mode>3_64bit): Likewise.
2678 (vsx_ior<mode>3_32bit): Likewise.
2679 (vsx_ior<mode>3_64bit): Likewise.
2680 (vsx_xor<mode>3_32bit): Likewise.
2681 (vsx_xor<mode>3_64bit): Likewise.
2682 (vsx_one_cmpl<mode>2_32bit): Likewise.
2683 (vsx_one_cmpl<mode>2_64bit): Likewise.
2684 (vsx_nor<mode>3_32bit): Likewise.
2685 (vsx_nor<mode>3_64bit): Likewise.
2686 (vsx_andc<mode>3_32bit): Likewise.
2687 (vsx_andc<mode>3_64bit): Likewise.
2688 (vsx_eqv<mode>3_32bit): Likewise.
2689 (vsx_eqv<mode>3_64bit): Likewise.
2690 (vsx_nand<mode>3_32bit): Likewise.
2691 (vsx_nand<mode>3_64bit): Likewise.
2692 (vsx_orc<mode>3_32bit): Likewise.
2693 (vsx_orc<mode>3_64bit): Likewise.
2695 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2696 logical types in GPRs.
2698 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2699 logical insns to rs6000.md, and allow TImode operations in
2701 (altivec_ior<mode>3): Likewise.
2702 (altivec_xor<mode>3): Likewise.
2703 (altivec_one_cmpl<mode>2): Likewise.
2704 (altivec_nor<mode>3): Likewise.
2705 (altivec_andc<mode>3): Likewise.
2707 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2708 attributes for moving the 128-bit logical operations into
2710 (BOOL_REGS_OUTPUT): Likewise.
2711 (BOOL_REGS_OP1): Likewise.
2712 (BOOL_REGS_OP2): Likewise.
2713 (BOOL_REGS_UNARY): Likewise.
2714 (BOOL_REGS_AND_CR0): Likewise.
2715 (one_cmpl<mode>2): Add support for DI logical operations on
2716 32-bit, splitting the operations to 32-bit.
2720 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2721 changes to combine the 32/64-bit code, allow logical operations on
2722 TI mode in 32-bit, and to use similar match_operator patterns like
2723 scalar mode uses. Combine the Altivec and VSX code for logical
2724 operations, and move it here.
2725 (ior<mode>3, 128-bit types): Likewise.
2726 (xor<mode>3, 128-bit types): Likewise.
2727 (one_cmpl<mode>3, 128-bit types): Likewise.
2728 (nor<mode>3, 128-bit types): Likewise.
2729 (andc<mode>3, 128-bit types): Likewise.
2730 (eqv<mode>3, 128-bit types): Likewise.
2731 (nand<mode>3, 128-bit types): Likewise.
2732 (orc<mode>3, 128-bit types): Likewise.
2733 (and<mode>3_internal): Likewise.
2734 (bool<mode>3_internal): Likewise.
2735 (boolc<mode>3_internal1): Likewise.
2736 (boolc<mode>3_internal2): Likewise.
2737 (boolcc<mode>3_internal1): Likewise.
2738 (boolcc<mode>3_internal2): Likewise.
2739 (eqv<mode>3_internal1): Likewise.
2740 (eqv<mode>3_internal2): Likewise.
2741 (one_cmpl1<mode>3_internal): Likewise.
2743 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2745 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2746 Rename flag_stack_usage to flag_stack_usage_info.
2748 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2750 * config/microblaze/sync.md: New file.
2751 * config/microblaze/microblaze.md: Include sync.md
2752 * config/microblaze/microblaze.c: Add print_operand 'y'.
2753 * config/microblaze/constraints.md: Add memory_contraint
2754 'Q' which is a single register.
2756 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2758 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2760 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2762 * reload.c (find_reloads): Exit loop once we find this operand
2763 cannot be reloaded somehow for this alternative.
2765 * reload.c (find_reloads): Exit loop once we find a hard register.
2767 * rtlanal.c (computed_jump_p): Exit loop once we find label
2770 * i386.c (ix86_pad_returns): Exit loop after setting replace.
2772 * cfgloopmanip.c (remove_path): Exit loop after setting
2775 * gensupport.c (subst_dup): Avoid loop if code is not
2776 MATCH_DUP nor MATCH_OP_DUP.
2778 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
2780 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
2782 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2784 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
2785 true for SP_REGNUM if mode == ptr_mode.
2786 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
2787 with value R0_REGNUM + 31.
2789 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2791 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
2792 pad pointer-typed argument downward.
2794 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2796 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
2797 and __ILP32__ when the ILP32 model is in use.
2799 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2801 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
2802 (aarch64_load_symref_appropriately): In the case of
2803 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
2804 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
2805 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
2806 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
2807 if 'mode' doesn't equal to 'ptr_mode'.
2808 (aarch64_output_mi_thunk): Add an assertion on the alignment of
2809 'vcall_offset'; change to call aarch64_emit_move differently depending
2810 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
2811 to calculate the upper bound of 'vcall_offset'.
2812 (aarch64_cannot_force_const_mem): Change to also return true if
2814 (aarch64_legitimize_reload_address): In the case of large
2815 displacements, add new local variable 'xmode' and an assertion
2816 based on it; change to use 'xmode' to generate the new rtx and
2818 (aarch64_asm_trampoline_template): Change to generate the template
2819 differently depending on TARGET_ILP32 or not; change to use
2820 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
2821 (aarch64_trampoline_size): Removed.
2822 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
2823 and replace immediate literals with it. Change to use 'ptr_mode'
2824 instead of 'DImode' and call convert_memory_address if the mode
2825 of 'fnaddr' doesn't equal to 'ptr_mode'.
2826 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
2828 (aarch64_elf_asm_destructor): Likewise.
2829 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
2830 on TARGET_ILP32 instead of aarch64_trampoline_size.
2831 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
2832 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
2833 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
2834 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
2835 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
2836 (storewb_pair<GPI:mode>_<P:mode>): ... this.
2837 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
2838 depending on the value of 'mode'.
2839 (add_losym_<mode>): New.
2840 (ldr_got_small_<mode>): New, based on ldr_got_small.
2841 (ldr_got_small): Remove.
2842 (ldr_got_small_sidi): New.
2843 * config/aarch64/iterators.md (P): New.
2844 (PTR): Change to 'ptr_mode' in the condition.
2846 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
2848 * config.gcc (aarch64*-*-*): Support --with-abi.
2849 (aarch64*-*-elf): Support --with-multilib-list.
2850 (aarch64*-*-linux*): Likewise.
2851 (supported_defaults): Add abi to aarch64*-*-*.
2852 * configure.ac: Mention AArch64 for --with-multilib-list.
2853 * configure: Re-generated.
2854 * config/aarch64/biarchilp32.h: New file.
2855 * config/aarch64/biarchlp64.h: New file.
2856 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
2858 (MULTILIB_DEFAULTS): Ditto.
2859 (DRIVER_SELF_SPECS): Ditto.
2860 (ASM_SPEC): Update to also substitute -mabi.
2861 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
2862 file whose name depends on -mabi= and -mbig-endian.
2863 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
2865 (POINTER_SIZE): New define.
2866 (POINTERS_EXTEND_UNSIGNED): Ditto.
2867 (enum aarch64_abi_type): New enumeration tag.
2868 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
2869 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
2870 (TARGET_ILP32): New define.
2871 * config/aarch64/aarch64.opt (mabi): New.
2873 (ilp32, lp64): New values for -mabi.
2874 * config/aarch64/t-aarch64 (comma): New define.
2875 (MULTILIB_OPTIONS): Ditto.
2876 (MULTILIB_DIRNAMES): Ditto.
2877 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
2878 * doc/invoke.texi: Document -mabi for AArch64.
2880 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
2882 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
2884 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2885 Anton Blanchard <anton@au1.ibm.com>
2887 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
2888 endianness when selecting field to splat.
2890 2013-07-22 Eric Christopher <echristo@gmail.com>
2892 * dwarf2out.c (die_odr_checksum): New function to use
2893 CHECKSUM_ macros and ULEB128 for DIE tag.
2894 (generate_type_signature): Use.
2896 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
2898 * config.gcc (sparc*-*-*): Accept leon3 processor.
2899 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
2900 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
2901 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
2902 * config/sparc/sparc.opt (enum processor_type): Add leon3.
2903 (mfix-ut699): Adjust comment.
2904 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
2905 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
2906 (CPP_CPU_SPEC): Likewise.
2907 (ASM_CPU_SPEC): Likewise.
2908 * config/sparc/sparc.c (leon3_cost): New constant.
2909 (sparc_option_override): Add leon3 support.
2910 (mem_ref): New function.
2911 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
2912 (sparc_do_work_around_errata): Look into the instruction in the delay
2913 slot and adjust accordingly. Add fix for the data cache nullify issues
2914 of the UT699. Change insertion position for the NOP.
2915 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
2916 (leon3_load): New reservation.
2917 (leon_store): Bump latency to 2.
2918 (grfpu): New automaton.
2919 (grfpu_alu): New unit.
2920 (grfpu_ds): Likewise.
2921 (leon_fp_alu): Adjust.
2922 (leon_fp_mult): Delete.
2923 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
2924 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
2925 * config/sparc/sparc.md (cpu): Add leon3.
2926 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
2928 (atomic_test_and_set): Likewise.
2931 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
2933 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
2934 default for R5900 targets.
2935 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
2936 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
2937 * config/mips/mips.c (mips_option_override): Report an error for
2938 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
2939 for -march=r5900 -mhard-float.
2941 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2943 * df-problems.c (can_move_insns_across): Exit loop once we
2944 find a non-fixed, non-global register.
2946 * ipa-pure-const.c (propagate_nothrow): Exit loop after
2949 * omega.c (omega_eliminate_red): Break after setting red_found.
2950 (omega_problem_has_red_equations): Similarly after setting found.
2951 (omega_query_variable): Similarly after setting coupled.
2953 2013-07-22 Marek Polacek <polacek@redhat.com>
2955 * gimplify.c: Don't include gimple.h twice.
2957 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2959 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
2960 instead of TARGET_THUMB1.
2961 (Pz): New constraint.
2962 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
2964 (compare_negsi_si): Likewise.
2965 (compare_addsi2_op0): Likewise.
2966 (compare_addsi2_op1): Likewise.
2967 (addsi3_carryin_<optab>): Likewise.
2968 (addsi3_carryin_alt2_<optab>): Likewise.
2969 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
2970 for arm_restrict_it.
2971 (subsi3_carryin): Likewise.
2972 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
2973 (minmax_arithsi): Disable for arm_restrict_it.
2974 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
2975 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
2976 (satsi_<SAT:code>_shift): Likewise.
2977 (arm_shiftsi3): Add alternative for 16-bit encoding.
2978 (arm32_movhf): Disable for arm_restrict_it.
2979 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
2980 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
2982 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
2984 * config/arm/arm.md (attribute "insn"): Delete.
2985 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
2986 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
2987 (not_shiftsi): Update for attribute change.
2988 (not_shiftsi_compare0): Likewise.
2989 (not_shiftsi_compare0_scratch): Likewise.
2990 (arm_one_cmplsi2): Likewise.
2991 (thumb1_one_cmplsi2): Likewise.
2992 (notsi_compare0): Likewise.
2993 (notsi_compare0_scratch): Likewise.
2994 (thumb1_movdi_insn): Likewise.
2995 (arm_movsi_insn): Likewise.
2996 (movhi_insn_arch4): Likewise.
2997 (movhi_bytes): Likewise.
2998 (arm_movqi_insn): Likewise.
2999 (thumb1_movqi_insn): Likewise.
3000 (arm32_movhf): Likewise.
3001 (thumb1_movhf): Likewise.
3002 (arm_movsf_soft_insn): Likewise.
3003 (thumb1_movsf_insn): Likewise.
3004 (thumb_movdf_insn): Likewise.
3005 (movsicc_insn): Likewise.
3006 (movsfcc_soft_insn): Likewise.
3007 (and_scc): Likewise.
3008 (cond_move): Likewise.
3009 (if_move_not): Likewise.
3010 (if_not_move): Likewise.
3011 (if_shift_move): Likewise.
3012 (if_move_shift): Likewise.
3013 (if_shift_shift): Likewise.
3014 (if_not_arith): Likewise.
3015 (if_arith_not): Likewise.
3016 (cond_move_not): Likewise.
3017 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3018 (neon_mov<mode>): Likewise.
3019 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3020 (thumb2_movsi_vfp): Likewise.
3021 (movsf_vfp): Likewise.
3022 (thumb2_movsf_vfp): Likewise.
3023 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3025 (cortexa7_older_only): Likewise.
3026 (cortexa7_younger): Likewise.
3027 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3028 (1020alu_shift_op): Likewise.
3029 (1020alu_shift_reg_op): Likewise.
3030 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3031 (alu_shift_op): Likewise.
3032 (alu_shift_reg_op): Likewise.
3033 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3034 (11_alu_shift_op): Likewise.
3035 (11_alu_shift_reg_op): Likewise.
3036 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3037 (9_alu_shift_reg_op): Likewise.
3038 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3040 (cortex_a15_alu_shift): Likewise.
3041 (cortex_a15_alu_shift_reg): Likewise.
3042 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3043 (cortex_a5_alu_shift): Likewise.
3044 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3046 (cortex_a53_alu_shift): Likewise.
3047 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3049 (cortex_a7_alu_reg): Likewise.
3050 (cortex_a7_alu_shift): Likewise.
3051 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3052 (cortex_a8_alu_shift): Likewise.
3053 (cortex_a8_alu_shift_reg): Likewise.
3054 (cortex_a8_mov): Likewise.
3055 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3056 (cortex_a9_dp_shift): Likewise.
3057 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3058 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3059 (cortex_r4_mov): Likewise.
3060 (cortex_r4_alu_shift): Likewise.
3061 (cortex_r4_alu_shift_reg): Likewise.
3062 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3063 (526_alu_shift_op): Likewise.
3064 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3065 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3066 (626te_alu_shift_op): Likewise.
3067 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3068 (726te_alu_op): Likewise.
3069 (726te_alu_shift_op): Likewise.
3070 (726te_alu_shift_reg_op): Likewise.
3071 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3072 (mp626_alu_shift_op): Likewise.
3073 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3074 (pj4_alu_e1_conds): Likewise.
3075 (pj4_alu): Likewise.
3076 (pj4_alu_conds): Likewise.
3077 (pj4_shift): Likewise.
3078 (pj4_shift_conds): Likewise.
3079 (pj4_alu_shift): Likewise.
3080 (pj4_alu_shift_conds): Likewise.
3082 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3084 * config/arm/predicates.md (shiftable_operator_strict_it):
3086 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3087 Disable cond_exec version for arm_restrict_it.
3088 (thumb2_smaxsi3): Convert to generate cond_exec.
3089 (thumb2_sminsi3): Likewise.
3090 (thumb32_umaxsi3): Likewise.
3091 (thumb2_uminsi3): Likewise.
3092 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3093 (thumb2_neg_abssi2): Likewise.
3094 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3095 (thumb2_movsicc_insn): Adjust alternatives.
3096 (thumb2_mov_negscc): Disable for arm_restrict_it.
3097 (thumb2_mov_negscc_strict_it): New pattern.
3098 (thumb2_mov_notscc_strict_it): New pattern.
3099 (thumb2_mov_notscc): Disable for arm_restrict_it.
3100 (thumb2_ior_scc): Likewise.
3101 (thumb2_ior_scc_strict_it): New pattern.
3102 (thumb2_cond_move): Adjust for arm_restrict_it.
3103 (thumb2_cond_arith): Disable for arm_restrict_it.
3104 (thumb2_cond_arith_strict_it): New pattern.
3105 (thumb2_cond_sub): Adjust for arm_restrict_it.
3106 (thumb2_movcond): Likewise.
3107 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3108 (thumb2_zero_extendhisi2_v6): Likewise.
3109 (thumb2_zero_extendqisi2_v6): Likewise.
3110 (orsi_notsi_si): Likewise.
3111 (orsi_not_shiftsi_si): Likewise.
3113 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3115 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3116 instruction sequence is 1 byte shorter.
3118 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3120 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3121 it is not needed after split.
3123 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3126 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3127 second label for nonlocal goto receivers. Don't output pic base labels
3128 unless we're producing PIC; mark that action unreachable().
3129 (ix86_save_reg): If the function contains a nonlocal label, save the
3131 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3132 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3133 (update_pic_label_number_if_needed): New.
3134 (machopic_output_function_base_name): Adjust for nonlocal receiver
3136 (machopic_should_output_picbase_label): New.
3137 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3138 (nonlocal_goto_receiver): New insn and split.
3140 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3142 * config/aarch64/aarch64-builtins.c
3143 (aarch64_fold_builtin): Fold abs in all modes.
3144 * config/aarch64/aarch64-simd-builtins.def
3145 (abs): Enable for all modes.
3146 * config/aarch64/arm_neon.h
3147 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
3148 (vabs_f64): Add missing intrinsic.
3150 2013-07-19 Ian Bolton <ian.bolton@arm.com>
3152 * config/aarch64/arm_neon.h (vabs_s64): New function
3154 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
3157 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
3158 * config/avr/avr.md (adjust_len): Add `round'.
3159 * config/avr/avr-protos.h (avr_out_round): New prototype.
3160 (avr_out_plus): Add `out_label' argument.
3161 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
3162 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
3163 Handle the case where `insn' is just a pattern.
3164 (avr_out_bitop): Handle the case where `insn' is just a pattern.
3165 (avr_out_round): New function.
3166 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
3168 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
3170 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3171 Add check for flag_stack_usage to handle -fstack-usage support
3173 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
3175 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
3176 interaction for new Power8 flags and VSX.
3178 2013-07-18 Sriraman Tallam <tmsriram@google.com>
3181 * tree-inline.c (expand_call_inline): Emit errors during
3182 early_inlining only if optimization is not turned on.
3184 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3188 * passes.c (init_optimization_passes): Move the construction of
3189 the pass hierarchy into a new passes.def file.
3191 * Makefile.in (passes.o): Add dependency on passes.def.
3193 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3195 * passes.c (init_optimization_passes): Introduce macros for
3196 constructing the tree of passes (INSERT_PASSES_AFTER,
3197 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
3198 TERMINATE_PASS_LIST).
3200 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
3201 Wei Mi <wmi@google.com>
3203 PR rtl-optimization/57878
3204 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
3206 (reload_pseudo_compare_func): Check nregs first for reload
3209 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3211 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
3213 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
3215 * read-rtl.c (validate_const_int): Once an invalid character is
3216 seen, quit the loop.
3218 * gengtype.c (write_roots): Similarly once we find the "deletable"
3219 or "if_marked" option.
3221 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3223 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
3224 "xtab" and "sat". Move value "clz" from here to ...
3225 (attriubte "type"): ... here.
3226 (satsi_<SAT:code>): Delete "insn" attribute.
3227 (satsi_<SAT:code>_shift): Likewise.
3228 (arm_zero_extendqisi2addsi): Likewise.
3229 (arm_extendqisi2addsi): Likewise.
3230 (clzsi2): Update for attribute changes.
3231 (rbitsi2): Likewise.
3232 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
3234 (arm_usatsihi): Likewise.
3235 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3237 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3239 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
3240 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
3241 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
3242 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
3243 in alphabetical order.
3244 (attribute "core_cycles"): Update for attribute changes.
3245 (arm_addsi3): Likewise.
3246 (addsi3_compare0): Likewise.
3247 (addsi3_compare0_scratch): Likewise.
3248 (addsi3_compare_op1): Likewise.
3249 (addsi3_compare_op2): Likewise.
3250 (compare_addsi2_op0): Likewise.
3251 (compare_addsi2_op1): Likewise.
3252 (addsi3_carryin_shift_<optab>): Likewise.
3253 (subsi3_carryin_shift): Likewise.
3254 (rsbsi3_carryin_shift): Likewise.
3255 (arm_subsi3_insn): Likewise.
3256 (subsi3_compare0): Likewise.
3257 (subsi3_compare): Likewise.
3258 (arm_andsi3_insn): Likewise.
3259 (thumb1_andsi3_insn): Likewise.
3260 (andsi3_compare0): Likewise.
3261 (andsi3_compare0_scratch): Likewise.
3262 (zeroextractsi_compare0_scratch
3263 (andsi_not_shiftsi_si): Likewise.
3264 (iorsi3_insn): Likewise.
3265 (iorsi3_compare0): Likewise.
3266 (iorsi3_compare0_scratch): Likewise.
3267 (arm_xorsi3): Likewise.
3268 (thumb1_xorsi3_insn): Likewise.
3269 (xorsi3_compare0): Likewise.
3270 (xorsi3_compare0_scratch): Likewise.
3271 (satsi_<SAT:code>_shift): Likewise.
3273 (arm_shiftsi3): Likewise.
3274 (shiftsi3_compare0): Likewise.
3275 (not_shiftsi): Likewise.
3276 (not_shiftsi_compare0): Likewise.
3277 (not_shiftsi_compare0_scratch): Likewise.
3278 (arm_one_cmplsi2): Likewise.
3279 (thumb_one_complsi2): Likewise.
3280 (notsi_compare0): Likewise.
3281 (notsi_compare0_scratch): Likewise.
3282 (thumb1_zero_extendhisi2): Likewise.
3283 (arm_zero_extendhisi2): Likewise.
3284 (arm_zero_extendhisi2_v6): Likewise.
3285 (arm_zero_extendhisi2addsi): Likewise.
3286 (thumb1_zero_extendqisi2): Likewise.
3287 (thumb1_zero_extendqisi2_v6): Likewise.
3288 (arm_zero_extendqisi2): Likewise.
3289 (arm_zero_extendqisi2_v6): Likewise.
3290 (arm_zero_extendqisi2addsi): Likewise.
3291 (thumb1_extendhisi2): Likewise.
3292 (arm_extendhisi2): Likewise.
3293 (arm_extendhisi2_v6): Likewise.
3294 (arm_extendqisi): Likewise.
3295 (arm_extendqisi_v6): Likewise.
3296 (arm_extendqisi2addsi): Likewise.
3297 (thumb1_extendqisi2): Likewise.
3298 (thumb1_movdi_insn): Likewise.
3299 (arm_movsi_insn): Likewise.
3300 (movsi_compare0): Likewise.
3301 (movhi_insn_arch4): Likewise.
3302 (movhi_bytes): Likewise.
3303 (arm_movqi_insn): Likewise.
3304 (thumb1_movqi_insn): Likewise.
3305 (arm32_movhf): Likewise.
3306 (thumb1_movhf): Likewise.
3307 (arm_movsf_soft_insn): Likewise.
3308 (thumb1_movsf_insn): Likewise.
3309 (movdf_soft_insn): Likewise.
3310 (thumb_movdf_insn): Likewise.
3311 (arm_cmpsi_insn): Likewise.
3312 (cmpsi_shiftsi): Likewise.
3313 (cmpsi_shiftsi_swp): Likewise.
3314 (arm_cmpsi_negshiftsi_si): Likewise.
3315 (movsicc_insn): Likewise.
3316 (movsfcc_soft_insn): Likewise.
3317 (arith_shiftsi): Likewise.
3318 (arith_shiftsi_compare0
3319 (arith_shiftsi_compare0_scratch
3320 (sub_shiftsi): Likewise.
3321 (sub_shiftsi_compare0
3322 (sub_shiftsi_compare0_scratch
3323 (and_scc): Likewise.
3324 (cond_move): Likewise.
3325 (if_plus_move): Likewise.
3326 (if_move_plus): Likewise.
3327 (if_move_not): Likewise.
3328 (if_not_move): Likewise.
3329 (if_shift_move): Likewise.
3330 (if_move_shift): Likewise.
3331 (if_shift_shift): Likewise.
3332 (if_not_arith): Likewise.
3333 (if_arith_not): Likewise.
3334 (cond_move_not): Likewise.
3335 (thumb1_ashlsi3): Set type attribute.
3336 (thumb1_ashrsi3): Likewise.
3337 (thumb1_lshrsi3): Likewise.
3338 (thumb1_rotrsi3): Likewise.
3339 (shiftsi3_compare0_scratch): Likewise.
3340 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
3341 (neon_mov<mode>): Likewise.
3342 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
3344 (thumb2_movsi_insn): Likewise.
3345 (thumb2_cmpsi_neg_shiftsi): Likewise.
3346 (thumb2_extendqisi_v6): Likewise.
3347 (thumb2_zero_extendhisi2_v6): Likewise.
3348 (thumb2_zero_extendqisi2_v6): Likewise.
3349 (thumb2_shiftsi3_short): Likewise.
3350 (thumb2_addsi3_compare0_scratch): Likewise.
3351 (orsi_not_shiftsi_si): Likewise.
3352 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3353 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
3355 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3356 (1020alu_shift_op): Likewise.
3357 (1020alu_shift_reg_op): Likewise.
3358 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3359 (alu_shift_op): Likewise.
3360 (alu_shift_reg_op): Likewise.
3361 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3362 (11_alu_shift_op): Likewise.
3363 (11_alu_shift_reg_op): Likewise.
3364 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3365 (9_alu_shift_reg_op): Likewise.
3366 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3368 (cortex_a15_alu_shift): Likewise.
3369 (cortex_a15_alu_shift_reg): Likewise.
3370 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
3372 (cortex_a5_alu_shift): Likewise.
3373 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
3375 (cortex_a53_alu_shift): Likewise.
3376 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3378 (cortex_a7_alu_reg): Likewise.
3379 (cortex_a7_alu_shift): Likewise.
3380 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
3382 (cortex_a8_alu_shift): Likewise.
3383 (cortex_a8_alu_shift_reg): Likewise.
3384 (cortex_a8_mov): Likewise.
3385 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3386 (cortex_a9_dp_shift): Likewise.
3387 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
3389 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
3391 (cortex_r4_mov): Likewise.
3392 (cortex_r4_alu_shift): Likewise.
3393 (cortex_r4_alu_shift_reg): Likewise.
3394 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3395 (526_alu_shift_op): Likewise.
3396 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3397 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3398 (626te_alu_shift_op): Likewise.
3399 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3400 (726te_alu_op): Likewise.
3401 (726te_alu_shift_op): Likewise.
3402 (726te_alu_shift_reg_op): Likewise.
3403 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3404 (mp626_alu_shift_op): Likewise.
3405 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3406 (pj4_alu_e1_conds): Likewise.
3407 (pj4_alu): Likewise.
3408 (pj4_alu_conds): Likewise.
3409 (pj4_shift): Likewise.
3410 (pj4_shift_conds): Likewise.
3411 (pj4_alu_shift): Likewise.
3412 (pj4_alu_shift_conds): Likewise.
3413 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3415 (cortexa7_older_only): Likewise.
3416 (cortexa7_younger): Likewise.
3418 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3420 * ipa-pure-const.c (generate_summary): Rename to...
3421 (pure_const_generate_summary): ... this.
3423 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3425 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3427 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3430 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3433 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3435 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3436 enabled without -march=zEC12.
3437 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3440 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3442 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3443 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3444 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3445 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3446 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3447 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3448 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3450 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3451 and ISA_HAS_NMADD3_NMSUB3.
3452 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3453 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3454 (nmsub4<mode>, nmsub3<mode>): Likewise.
3455 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3457 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3459 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3460 TARGET_MIPS5400 checking.
3462 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3463 Peter Bergner <bergner@vnet.ibm.com>
3465 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3466 registers in the comment.
3467 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3468 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3469 rather than FIRST_PSEUDO_REGISTERS.
3471 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3473 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3474 enable extra ISA flags with TARGET_HTM.
3476 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3478 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3481 2013-07-15 Cong Hou <congh@google.com>
3483 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3484 in compare function for sorting.
3486 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3488 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3489 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3490 * config/rs6000/rs6000.opt: Add -mhtm option.
3491 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3492 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3493 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3494 __HTM__ if the HTM instructions are available.
3495 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3496 htm_spr_reg_operand): New define_predicates.
3497 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3498 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3500 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3501 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3502 HTM builtin functions.
3503 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3504 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3505 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3506 (rs6000_builtin_mask_calculate): Likewise.
3507 (rs6000_option_override_internal): Likewise.
3508 (bdesc_htm): Add new HTM builtin support.
3509 (htm_spr_num): New function.
3510 (htm_spr_regno): Likewise.
3511 (rs6000_htm_spr_icode): Likewise.
3512 (htm_expand_builtin): Likewise.
3513 (htm_init_builtins): Likewise.
3514 (rs6000_expand_builtin): Add support for HTM builtin functions.
3515 (rs6000_init_builtins): Likewise.
3516 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3518 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3519 (TARGET_HTM, MASK_HTM): Define macros.
3520 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3521 (FIXED_REGISTERS): Likewise.
3522 (CALL_USED_REGISTERS): Likewise.
3523 (CALL_REALLY_USED_REGISTERS): Likewise.
3524 (REG_ALLOC_ORDER): Likewise.
3525 (enum reg_class): Likewise.
3526 (REG_CLASS_NAMES): Likewise.
3527 (REG_CLASS_CONTENTS): Likewise.
3528 (REGISTER_NAMES): Likewise.
3529 (ADDITIONAL_REGISTER_NAMES): Likewise.
3530 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3531 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3532 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3533 * config/rs6000/htm.md: New file.
3534 * config/rs6000/htmintrin.h: New file.
3535 * config/rs6000/htmxlintrin.h: New file.
3537 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3539 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3540 Define SYMBOL_TINY_GOT, update comment.
3541 * config/aarch64/aarch64.c
3542 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3543 (aarch64_expand_mov_immediate): Likewise.
3544 (aarch64_print_operand): Likewise.
3545 (aarch64_classify_symbol): Likewise.
3546 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3547 (ldr_got_tiny): Define.
3549 2013-07-13 Tobias Grosser <tobias@grosser.es>
3551 PR tree-optimization/54094
3552 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3553 scheduling dimension for the parallelism check from the polyhedral
3554 information in the AST.
3555 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3558 2013-07-13 Jason Merrill <jason@redhat.com>
3560 * print-tree.c (debug_vec_tree): Use debug_raw.
3561 (debug_raw (vec<tree, va_gc> &)): New.
3562 (debug_raw (vec<tree, va_gc> *)): New.
3563 * tree.h: Declare them.
3565 2013-07-13 Bin Cheng <bin.cheng@arm.com>
3567 * ifcvt.c (ifcvt_after_combine): New static variable.
3568 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3570 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3571 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3572 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3574 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3576 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3578 2013-07-12 Michael Matz <matz@suse.de>
3581 * convert.c (convert_to_real): Reject non-float inner types.
3583 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3585 * config/aarch64/aarch64-protos.h
3586 (aarch64_simd_immediate_valid_for_move): Remove.
3587 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3588 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3590 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3592 2013-07-11 Steve Ellcey <sellcey@mips.com>
3594 * config/mips/mips.c (mips_conditional_register_usage): Do not
3595 use t[0-7] registers in MIPS16 mode when optimizing for size.
3597 2013-07-11 Sriraman Tallam <tmsriram@google.com>
3599 * config/i386/i386.c (dispatch_function_versions): Fix array
3600 indexing of function_version_info to match actual_versions.
3602 2013-07-11 Teresa Johnson <tejohnson@google.com>
3604 * vec.h (struct va_gc): Move release out-of-line.
3605 (va_gc::release): Call ggc_free on released vec.
3607 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3609 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3610 Require GOT register as additional operand in UNSPEC.
3611 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3612 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3613 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3614 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3615 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3616 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3617 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3619 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
3622 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3623 name seen by assembler/linker rather if available.
3625 2013-07-11 Andreas Schwab <schwab@suse.de>
3627 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3629 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3631 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
3633 2013-07-10 Joseph Myers <joseph@codesourcery.com>
3635 * doc/tm.texi.in: Move hook documentation to ....
3636 * target.def: ... here.
3638 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3640 * doc/tm.texi: Regenerate.
3642 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3645 * doc/invoke.texi: Document -Wconditionally-supported.
3647 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3650 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3653 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3656 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3657 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3658 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3659 Remove duplicate devices.
3660 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3661 * config/avr/t-multilib: Regenerate.
3662 * config/avr/avr-tables.opt: Regenerate.
3663 * doc/avr-mmcu.texi: Regenerate.
3665 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3668 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3670 2013-07-10 Graham Stott <graham.stott@btinternet.com>
3672 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3673 the cost of MULT when optimizing for size.
3675 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3677 * config/cr16/cr16-protos.h: Don't include target.h.
3679 2013-07-09 Joseph Myers <joseph@codesourcery.com>
3681 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3682 adjust register size for TDmode and TFmode for VSX registers.
3684 2013-07-08 Kai Tietz <ktietz@redhat.com>
3687 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3688 hook_bool_const_tree_true.
3690 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3692 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3693 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3694 * config/s390/s390.md: Define FPR*_REGNUM constants.
3695 Fix FPR2_REGNUM constant (18 -> 17).
3696 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3697 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3698 ("trunc<BFP:mode><DFP_ALL:mode>2")
3699 ("trunc<DFP_ALL:mode><BFP:mode>2")
3700 ("*extend<BFP:mode><DFP_ALL:mode>2")
3701 ("*extend<DFP_ALL:mode><BFP:mode>2")
3702 ("extend<BFP:mode><DFP_ALL:mode>2")
3703 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3706 2013-07-08 Graham Stott <graham.stott@btinternet.com>
3708 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
3710 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3712 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3713 and cfun_fpr_bit_p to cfun_fpr_save_p.
3714 (s390_frame_area, s390_register_info, s390_frame_info)
3715 (s390_emit_prologue, s390_emit_epilogue)
3716 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3718 * config/s390/s390.h: Define *_REGNUM macros for floating point
3721 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3723 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3725 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3727 PR rtl-optimization/57786
3728 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3729 and break out of the loop when it is set to false.
3731 2013-07-08 Jakub Jelinek <jakub@redhat.com>
3734 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3735 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3736 (const_int 63)) 0)).
3737 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3738 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3739 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3741 PR rtl-optimization/57829
3742 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3743 mask bits outside of mode are just sign-extension from mode to HWI.
3745 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3747 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3748 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3749 adjust_address instead of change_address to keep info about alignment.
3750 (emit_strmov): Remove.
3751 (emit_memmov): New function.
3752 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3753 (expand_movmem_epilogue): Likewise and return updated rtx for
3755 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3756 destination and source.
3757 (decide_alignment): Refactor, handle vector_loop.
3758 (ix86_expand_movmem): Likewise.
3759 (ix86_expand_setmem): Likewise.
3760 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3762 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
3764 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3765 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
3767 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3769 * config/i386/sse.md (sse_movlhps): Change alternative 3
3770 of operand 2 to "m".
3772 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
3775 * config/i386/sse.md (iptr): New mode attribute.
3776 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
3777 (<sse>_vm<plusminus_insn><mode>3): Ditto.
3778 (<sse>_vmmul<mode>3): Ditto.
3779 (<sse>_vmdiv<mode>3): Ditto.
3780 (sse_vmrcpv4sf2): Ditto.
3781 (<sse>_vmsqrt<mode>2): Ditto.
3782 (sse_vmrsqrtv4sf2): Ditto.
3783 (<sse>_vm<code><mode>3): Ditto.
3784 (avx_vmcmp<mode>3): Ditto.
3785 (<sse>_vmmaskcmp<mode>3): Ditto.
3786 (<sse>_comi): Ditto.
3787 (<sse>_ucomi): Ditto.
3788 (*xop_vmfrcz_<mode>): Ditto.
3789 (*fmai_fmadd_<mode>): Ditto.
3790 (*fmai_fmsub_<mode>): Ditto.
3791 (*fmai_fnmadd_<mode>): Ditto.
3792 (*fmai_fnmsub_<mode>): Ditto.
3793 (*fma4i_vmfmadd_<mode>): Ditto.
3794 (*fma4i_vmfmsub_<mode>): Ditto.
3795 (*fma4i_vmfnmadd_<mode>): Ditto.
3796 (*fma4i_vmfnmsub_<mode>): Ditto.
3797 (*xop_vmfrcz_<mode>): Ditto.
3798 (sse_cvtps2pi): Ditto.
3799 (sse_cvttps2pi): Ditto.
3800 (sse_cvtss2si): Ditto.
3801 (sse_cvtss2si_2): Ditto.
3802 (sse_cvtss2siq_2): Ditto.
3803 (sse_cvttss2si): Ditto.
3804 (sse_cvttss2siq): Ditto.
3805 (sse_cvtsd2si): Ditto.
3806 (sse_cvtsd2si_2): Ditto.
3807 (sse_cvtsd2siq_2): Ditto.
3808 (sse_cvttsd2si): Ditto.
3809 (sse_cvttsd2siq): Ditto.
3810 (sse_cvtsd2ss): Ditto.
3811 (sse_cvtss2sd): Ditto.
3812 (avx2_pbroadcast<mode>): Ditto.
3813 (avx2_pbroadcast<mode>_1): Ditto.
3814 (*avx_vperm_broadcast_v4sf): Ditto.
3816 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
3817 (sse_movlhps): Ditto.
3818 (sse_storehps): Ditto.
3819 (sse_loadhps): Ditto.
3820 (sse_storelps): Ditto.
3821 (sse_loadlps): Ditto.
3822 (*vec_concatv4sf): Ditto.
3823 (*vec_interleave_highv2df): Ditto.
3824 (*vec_interleave_lowv2df): Ditto.
3825 (*vec_extractv2df_1_sse): Ditto.
3826 (*vec_extractv2df_0_sse): Ditto.
3827 (sse2_storelpd): Ditto.
3828 (sse2_loadlpd): Ditto.
3829 (sse2_movsd): Ditto.
3830 (*vec_concatv4si): Ditto.
3831 (vec_concatv2di): Ditto.
3833 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
3834 for Intel asm dialect.
3835 (mmx_punpcklwd): Ditto.
3836 (mmx_punpckldq): Ditto.
3838 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
3839 for intel assembler dialect.
3841 2013-07-06 Jakub Jelinek <jakub@redhat.com>
3844 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
3845 for BUILT_IN_C{LZ,LRSB}*.
3846 * tree.h (CASE_INT_FN): Add FN##IMAX case.
3847 * tree-vrp.c (extract_range_basic): Handle
3848 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
3849 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
3850 fall thru to code calling set_value*.
3851 * builtins.c (expand_builtin): Remove *IMAX cases.
3852 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
3853 if width is bigger than 2*HWI.
3855 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
3857 PR rtl-optimization/55342
3858 * lra-int.h (lra_subreg_reload_pseudos): New.
3859 * lra.c: Add undoing optional reloads to the block diagram.
3860 (lra_subreg_reload_pseudos): New.
3861 (lra_optional_reload_pseudos): Change comments.
3862 (lra): Init and clear lra_subreg_reload_pseudos. Clear
3863 lra_optional_reload_pseudos after undo transformations.
3864 * lra-assigns.c (pseudo_prefix_title): New.
3865 (lra_setup_reg_renumber): Use it.
3866 (spill_for): Ditto. Check subreg reload pseudos too.
3867 (assign_by_spills): Consider subreg reload pseudos too.
3868 * lra-constraints.c (simplify_operand_subreg): Use
3869 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
3870 (curr_insn_transform): Recognize and do optional reloads.
3871 (undo_optional_reloads): New.
3872 (lra_undo_inheritance): Call undo_optional_reloads.
3874 2013-07-05 Thomas Quinot <quinot@adacore.com>
3876 * tree-complex.c (expand_complex_operations_1): Fix typo.
3878 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
3880 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
3881 (tune_params): New member 'const vec_costs'.
3882 * config/aarch64/aarch64.c (generic_vector_cost): New.
3883 (generic_tunings): New member 'generic_vector_cost'.
3884 (aarch64_builtin_vectorization_cost): New.
3885 (aarch64_add_stmt_cost): New.
3886 (TARGET_VECTORIZE_ADD_STMT_COST): New.
3887 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
3889 2013-07-03 Jakub Jelinek <jakub@redhat.com>
3892 * config/i386/predicates.md (vsib_address_operand): Disallow
3893 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
3895 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
3898 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
3899 expand_builtin_setjmp_receiver.
3900 (expand_label): Adjust, call expand_builtin_setjmp_receiver
3901 with NULL for the label parameter.
3902 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
3903 the frame-pointer. Adjust comments.
3904 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
3905 only if LABEL is non-NULL.
3907 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
3909 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
3910 (ARM_ABI_AAPCS64): Ditto.
3912 (ARM_DEFAULT_ABI): Ditto.
3914 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
3916 * config/aarch64/aarch64-builtins.c
3917 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
3918 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
3920 * config/aarch64/aarch64-simd.md
3921 (aarch64_ld1<VALL:mode>): New.
3922 (aarch64_st1<VALL:mode>): Likewise.
3923 * config/aarch64/arm_neon.h
3924 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
3926 2013-07-02 Sriraman Tallam <tmsriram@google.com>
3928 * config/i386/i386.c (gate_insert_vzeroupper): Check if
3930 (ix86_option_override_internal):Turn on all -mavx target flags by
3931 default as they are dependent on AVX anyway.
3933 2013-07-02 Cary Coutant <ccoutant@google.com>
3935 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
3937 (loc_checksum_ordered): Likewise.
3938 (hash_loc_operands): Remove inline keyword.
3940 2013-07-02 Jakub Jelinek <jakub@redhat.com>
3942 PR tree-optimization/57741
3943 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
3944 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
3945 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
3946 Allow REAL_CST step_exprs if flag_associative_math.
3947 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
3949 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3951 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
3953 2013-07-02 Ian Bolton <ian.bolton@arm.com>
3955 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
3957 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3959 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
3961 (iorsi3_insn): Likewise.
3962 (arm_xorsi3): Likewise.
3964 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
3966 * arm.md (attribute "wtype"): Delete. Move attribute values from here
3968 (attribute "type"): ... here, and prefix with "wmmx_".
3969 (attribute "core_cycles"): Update for attribute changes.
3970 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
3971 (tbcstv4hi): Likewise.
3972 (tbcstv2si): Likewise.
3973 (iwmmxt_iordi3): Likewise.
3974 (iwmmxt_xordi3): Likewise.
3975 (iwmmxt_anddi3): Likewise.
3976 (iwmmxt_nanddi3): Likewise.
3977 (iwmmxt_arm_movdi): Likewise.
3978 (iwmmxt_movsi_insn): Likewise.
3979 (mov<mode>_internal): Likewise.
3980 (and<mode>3_iwmmxt): Likewise.
3981 (ior<mode>3_iwmmxt): Likewise.
3982 (xor<mode>3_iwmmxt): Likewise.
3983 (add<mode>3_iwmmxt): Likewise.
3984 (ssaddv8qi3): Likewise.
3985 (ssaddv4hi3): Likewise.
3986 (ssaddv2si3): Likewise.
3987 (usaddv8qi3): Likewise.
3988 (usaddv4hi3): Likewise.
3989 (usaddv2si3): Likewise.
3990 (sub<mode>3_iwmmxt): Likewise.
3991 (sssubv8qi3): Likewise.
3992 (sssubv4hi3): Likewise.
3993 (sssubv2si3): Likewise.
3994 (ussubv8qi3): Likewise.
3995 (ussubv4hi3): Likewise.
3996 (ussubv2si3): Likewise.
3997 (mulv4hi3_iwmmxt): Likewise.
3998 (smulv4hi3_highpart): Likewise.
3999 (umulv4hi3_highpart): Likewise.
4000 (iwmmxt_wmacs): Likewise.
4001 (iwmmxt_wmacsz): Likewise.
4002 (iwmmxt_wmacu): Likewise.
4003 (iwmmxt_wmacuz): Likewise.
4004 (iwmmxt_clrdi): Likewise.
4005 (iwmmxt_clrv8qi): Likewise.
4006 (iwmmxt_clr4hi): Likewise.
4007 (iwmmxt_clr2si): Likewise.
4008 (iwmmxt_uavgrndv8qi3): Likewise.
4009 (iwmmxt_uavgrndv4hi3): Likewise.
4010 (iwmmxt_uavgv8qi3): Likewise.
4011 (iwmmxt_uavgv4hi3): Likewise.
4012 (iwmmxt_tinsrb): Likewise.
4013 (iwmmxt_tinsrh): Likewise.
4014 (iwmmxt_tinsrw): Likewise.
4015 (iwmmxt_textrmub): Likewise.
4016 (iwmmxt_textrmsb): Likewise.
4017 (iwmmxt_textrmuh): Likewise.
4018 (iwmmxt_textrmsh): Likewise.
4019 (iwmmxt_textrmw): Likewise.
4020 (iwmxxt_wshufh): Likewise.
4021 (eqv8qi3): Likewise.
4022 (eqv4hi3): Likewise.
4023 (eqv2si3): Likewise.
4024 (gtuv8qi3): Likewise.
4025 (gtuv4hi3): Likewise.
4026 (gtuv2si3): Likewise.
4027 (gtv8qi3): Likewise.
4028 (gtv4hi3): Likewise.
4029 (gtv2si3): Likewise.
4030 (smax<mode>3_iwmmxt): Likewise.
4031 (umax<mode>3_iwmmxt): Likewise.
4032 (smin<mode>3_iwmmxt): Likewise.
4033 (umin<mode>3_iwmmxt): Likewise.
4034 (iwmmxt_wpackhss): Likewise.
4035 (iwmmxt_wpackwss): Likewise.
4036 (iwmmxt_wpackdss): Likewise.
4037 (iwmmxt_wpackhus): Likewise.
4038 (iwmmxt_wpackwus): Likewise.
4039 (iwmmxt_wpackdus): Likewise.
4040 (iwmmxt_wunpckihb): Likewise.
4041 (iwmmxt_wunpckihh): Likewise.
4042 (iwmmxt_wunpckihw): Likewise.
4043 (iwmmxt_wunpckilb): Likewise.
4044 (iwmmxt_wunpckilh): Likewise.
4045 (iwmmxt_wunpckilw): Likewise.
4046 (iwmmxt_wunpckehub): Likewise.
4047 (iwmmxt_wunpckehuh): Likewise.
4048 (iwmmxt_wunpckehuw): Likewise.
4049 (iwmmxt_wunpckehsb): Likewise.
4050 (iwmmxt_wunpckehsh): Likewise.
4051 (iwmmxt_wunpckehsw): Likewise.
4052 (iwmmxt_wunpckelub): Likewise.
4053 (iwmmxt_wunpckeluh): Likewise.
4054 (iwmmxt_wunpckeluw): Likewise.
4055 (iwmmxt_wunpckelsb): Likewise.
4056 (iwmmxt_wunpckelsh): Likewise.
4057 (iwmmxt_wunpckelsw): Likewise.
4058 (ror<mode>3): Likewise.
4059 (ashr<mode>3_iwmmxt): Likewise.
4060 (lshr<mode>3_iwmmxt): Likewise.
4061 (ashl<mode>3_iwmmxt): Likewise.
4062 (ror<mode>3_di): Likewise.
4063 (ashr<mode>3_di): Likewise.
4064 (lshr<mode>3_di): Likewise.
4065 (ashl<mode>3_di): Likewise.
4066 (iwmmxt_wmadds): Likewise.
4067 (iwmmxt_wmaddu): Likewise.
4068 (iwmmxt_tmia): Likewise.
4069 (iwmmxt_tmiaph): Likewise.
4070 (iwmmxt_tmiabb): Likewise.
4071 (iwmmxt_tmiatb): Likewise.
4072 (iwmmxt_tmiabt): Likewise.
4073 (iwmmxt_tmiatt): Likewise.
4074 (iwmmxt_tmovmskb): Likewise.
4075 (iwmmxt_tmovmskh): Likewise.
4076 (iwmmxt_tmovmskw): Likewise.
4077 (iwmmxt_waccb): Likewise.
4078 (iwmmxt_wacch): Likewise.
4079 (iwmmxt_waccw): Likewise.
4080 (iwmmxt_waligni): Likewise.
4081 (iwmmxt_walignr): Likewise.
4082 (iwmmxt_walignr0): Likewise.
4083 (iwmmxt_walignr1): Likewise.
4084 (iwmmxt_walignr2): Likewise.
4085 (iwmmxt_walignr3): Likewise.
4086 (iwmmxt_wsadb): Likewise.
4087 (iwmmxt_wsadh): Likewise.
4088 (iwmmxt_wsadbz): Likewise.
4089 (iwmmxt_wsadhz): Likewise.
4090 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4091 (iwmmxt_wabsdiffb): Likewise.
4092 (iwmmxt_wabsdiffh): Likewise.
4093 (iwmmxt_wabsdiffw): Likewise.
4094 (iwmmxt_waddsubhx): Likewise
4095 (iwmmxt_wsubaddhx): Likewise.
4096 (addc<mode>3): Likewise.
4097 (iwmmxt_avg4): Likewise.
4098 (iwmmxt_avg4r): Likewise.
4099 (iwmmxt_wmaddsx): Likewise.
4100 (iwmmxt_wmaddux): Likewise.
4101 (iwmmxt_wmaddsn): Likewise.
4102 (iwmmxt_wmaddun): Likewise.
4103 (iwmmxt_wmulwsm): Likewise.
4104 (iwmmxt_wmulwum): Likewise.
4105 (iwmmxt_wmulsmr): Likewise.
4106 (iwmmxt_wmulumr): Likewise.
4107 (iwmmxt_wmulwsmr): Likewise.
4108 (iwmmxt_wmulwumr): Likewise.
4109 (iwmmxt_wmulwl): Likewise.
4110 (iwmmxt_wqmulm): Likewise.
4111 (iwmmxt_wqmulwm): Likewise.
4112 (iwmmxt_wqmulmr): Likewise.
4113 (iwmmxt_wqmulwmr): Likewise.
4114 (iwmmxt_waddbhusm): Likewise.
4115 (iwmmxt_waddbhusl): Likewise.
4116 (iwmmxt_wqmiabb): Likewise.
4117 (iwmmxt_wqmiabt): Likewise.
4118 (iwmmxt_wqmiatb): Likewise.
4119 (iwmmxt_wqmiatt): Likewise.
4120 (iwmmxt_wqmiabbn): Likewise.
4121 (iwmmxt_wqmiabtn): Likewise.
4122 (iwmmxt_wqmiatbn): Likewise.
4123 (iwmmxt_wqmiattn): Likewise.
4124 (iwmmxt_wmiabb): Likewise.
4125 (iwmmxt_wmiabt): Likewise.
4126 (iwmmxt_wmiatb): Likewise.
4127 (iwmmxt_wmiatt): Likewise.
4128 (iwmmxt_wmiabbn): Likewise.
4129 (iwmmxt_wmiabtn): Likewise.
4130 (iwmmxt_wmiatbn): Likewise.
4131 (iwmmxt_wmiattn): Likewise.
4132 (iwmmxt_wmiawbb): Likewise.
4133 (iwmmxt_wmiawbt): Likewise.
4134 (iwmmxt_wmiawtb): Likewise.
4135 (iwmmxt_wmiawtt): Likewise.
4136 (iwmmxt_wmiawbbn): Likewise.
4137 (iwmmxt_wmiawbtn): Likewise.
4138 (iwmmxt_wmiawtbn): Likewise.
4139 (iwmmxt_wmiawttn): Likewise.
4140 (iwmmxt_wmerge): Likewise.
4141 (iwmmxt_tandc<mode>3): Likewise.
4142 (iwmmxt_torc<mode>3): Likewise.
4143 (iwmmxt_torvsc<mode>3): Likewise.
4144 (iwmmxt_textrc<mode>3): Likewise.
4145 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
4146 (wmmxt_pack): Likewise.
4147 (wmmxt_mult_c1): Likewise.
4148 (wmmxt_mult_c2): Likewise.
4149 (wmmxt_alu_c1): Likewise.
4150 (wmmxt_alu_c2): Likewise.
4151 (wmmxt_alu_c3): Likewise.
4152 (wmmxt_transfer_c1): Likewise.
4153 (wmmxt_transfer_c2): Likewise.
4154 (wmmxt_transfer_c3): Likewise.
4155 (marvell_f_iwmmxt_wstr): Likewise.
4156 (marvell_f_iwmmxt_wldr): Likewise.
4158 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
4160 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
4162 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4165 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4166 * lra-constraints.c (need_for_split_p): Check call used hard regs
4167 living through calls.
4169 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
4170 call used regs for call insn.
4172 2013-06-28 Jakub Jelinek <jakub@redhat.com>
4175 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
4176 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
4178 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
4180 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
4182 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4184 * lra-constraints.c (need_for_split_p): Check call used hard regs
4185 living through calls.
4187 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
4190 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
4191 to tie with any other modes. Eliminate Altivec vector mode tests,
4192 since these are a subset of ALTIVEC or VSX vector modes. Simplify
4193 code, to return 0 if testing MODE2 for a condition, if we've
4194 already tested MODE1 for the same condition.
4196 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4198 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
4201 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4203 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4204 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
4206 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4208 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
4210 (aarch64_symbolic_constant_p): Remove.
4211 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
4212 static. Fix line length and white space.
4213 (aarch64_symbolic_constant_p): Remove.
4214 * config/aarch64/predicates.md (aarch64_valid_symref):
4215 Use aarch64_classify_symbol_expression.
4217 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4219 * config/arm/constraints.md (Ts): New constraint.
4220 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
4222 (compare_scc): Use "Ts" constraint for operand 0.
4223 (ior_scc_scc): Likewise.
4224 (and_scc_scc): Likewise.
4225 (and_scc_scc_nodom): Likewise.
4226 (ior_scc_scc_cmp): Likewise for operand 7.
4227 (and_scc_scc_cmp): Likewise.
4228 * config/arm/thumb2.md (thumb2_movsi_insn):
4229 Add alternatives for 16-bit encodings.
4230 (thumb2_movhi_insn): Likewise.
4231 (thumb2_movsicc_insn): Likewise.
4232 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
4233 (thumb2_negscc): Use "Ts" constraint.
4234 Move mvn instruction outside cond_exec block.
4235 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
4236 for 16-bit encodings.
4238 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4240 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
4242 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
4243 (mulsi3subsi): Likewise.
4244 (mulsidi3adddi): Likewise.
4245 (mulsidi3_v6): Likewise.
4246 (umulsidi3_v6): Likewise.
4247 (umulsidi3adddi_v6): Likewise.
4248 (smulsi3_highpart_v6): Likewise.
4249 (umulsi3_highpart_v6): Likewise.
4250 (mulhisi3tb): Likewise.
4251 (mulhisi3bt): Likewise.
4252 (mulhisi3tt): Likewise.
4253 (maddhisi4): Likewise.
4254 (maddhisi4tb): Likewise.
4255 (maddhisi4tt): Likewise.
4256 (maddhidi4): Likewise.
4257 (maddhidi4tb): Likewise.
4258 (maddhidi4tt): Likewise.
4259 (zeroextractsi_compare0_scratch): Likewise.
4260 (insv_zero): Likewise.
4261 (insv_t2): Likewise.
4262 (anddi_notzesidi_di): Likewise.
4263 (anddi_notsesidi_di): Likewise.
4264 (andsi_notsi_si): Likewise.
4265 (iordi_zesidi_di): Likewise.
4266 (xordi_zesidi_di): Likewise.
4267 (andsi_iorsi3_notsi): Likewise.
4269 (smax_m1): Likewise.
4271 (not_shiftsi): Likewise.
4272 (unaligned_loadsi): Likewise.
4273 (unaligned_loadhis): Likewise.
4274 (unaligned_loadhiu): Likewise.
4275 (unaligned_storesi): Likewise.
4276 (unaligned_storehi): Likewise.
4277 (extv_reg): Likewise.
4278 (extzv_t2): Likewise.
4280 (udivsi3): Likewise.
4281 (arm_zero_extendhisi2addsi): Likewise.
4282 (arm_zero_extendqisi2addsi): Likewise.
4283 (compareqi_eq0): Likewise.
4284 (arm_extendhisi2_v6): Likewise.
4285 (arm_extendqisi2addsi): Likewise.
4286 (arm_movt): Likewise.
4287 (thumb2_ldrd): Likewise.
4288 (thumb2_ldrd_base): Likewise.
4289 (thumb2_ldrd_base_neg): Likewise.
4290 (thumb2_strd): Likewise.
4291 (thumb2_strd_base): Likewise.
4292 (thumb2_strd_base_neg): Likewise.
4293 (arm_negsi2): Add alternative for 16-bit encoding.
4294 (arm_one_cmplsi2): Likewise.
4296 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4298 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
4299 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
4300 (movdfcc): Likewise.
4301 * config/arm/vfp.md (*thumb2_movsf_vfp):
4302 Disable predication for arm_restrict_it.
4303 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
4304 (*thumb2_movdfcc_vfp): Likewise.
4305 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
4306 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
4307 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
4308 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
4309 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
4310 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
4311 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
4312 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
4313 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
4314 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
4315 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
4316 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
4317 Disable predication for arm_restrict_it.
4319 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
4321 * config/i386/bmiintrin.h (_bextr_u32): New.
4322 (_bextr_u64): Ditto.
4324 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
4326 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
4327 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
4328 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
4329 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
4330 * config/mips/n32-elf.h: ...this new file.
4332 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
4335 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
4336 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
4338 2013-06-27 Catherine Moore <clm@codesourcery.com>
4340 * config/mips/mips-tables.opt: Regenerate.
4341 * config/mips/mips-cpus.def: Add m14ke and m14kec.
4342 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
4343 * doc/invoke.texi: Add -m14kc.
4345 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4348 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
4349 constraints of operand 1 and 2.
4352 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
4353 to match RTL canonicalization. Swap predicates and
4354 constraints of operand 1 and 2.
4356 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
4358 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
4359 Process OP_INOUT regs for splitting too.
4361 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4363 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
4364 decl before the loop, initialize to NULL.
4365 (vectorizable_load): Initialize ptr_incr to NULL.
4367 2013-06-27 Martin Jambor <mjambor@suse.cz>
4370 * ipa-ref.h (ipa_maybe_record_reference): Declare.
4371 * ipa-ref.c (ipa_maybe_record_reference): New function.
4372 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
4373 * ipa-cp.c (create_specialized_node): Record potential references from
4375 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
4377 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4379 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
4380 parameter 'mode' of type 'enum machine_mode mode'; change to pass
4381 'mode' to force_reg.
4382 (aarch64_add_offset): Update calls to aarch64_force_temporary.
4383 (aarch64_expand_mov_immediate): Likewise.
4385 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4387 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
4388 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
4390 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4392 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
4393 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4394 (struct machine_function): Add tbegin_p.
4395 (s390_canonicalize_comparison): Fold CC mode compares to
4396 conditional jump if possible.
4397 (s390_emit_jump): Return the emitted jump.
4398 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4399 Handle CCRAWmode compares.
4400 (s390_option_override): Default to -mhtm if available.
4401 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4402 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4403 FPRs instead of df_regs_ever_live_p.
4404 (s390_optimize_nonescaping_tx): New function.
4405 (s390_init_frame_layout): Extend clobbered_regs array to cover
4407 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4408 (s390_expand_tbegin): New function.
4409 (enum s390_builtin): New enum definition.
4410 (code_for_builtin): New array definition.
4411 (s390_init_builtins): New function.
4412 (s390_expand_builtin): New function.
4413 (TARGET_INIT_BUILTINS): Define.
4414 (TARGET_EXPAND_BUILTIN): Define.
4415 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4416 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4417 (s390_alc_comparison): Likewise.
4418 * config/s390/s390-modes.def: Add CCRAWmode.
4419 * config/s390/s390.h (processor_flags): Add PF_TX.
4420 (TARGET_CPU_HTM): Define macro.
4421 (TARGET_HTM): Define macro.
4422 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4423 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4424 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4425 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4426 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4427 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4428 ("*cc_to_int"): Move up.
4429 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4430 constants other than 0.
4431 ("*ccraw_to_int"): New insn and splitter definition.
4432 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4433 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4434 ("tx_assist"): New expander.
4435 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4436 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4437 * config/s390/s390.opt: Add -mhtm option.
4438 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4439 * config/s390/htmxlintrin.h: New file.
4440 * config/s390/htmintrin.h: New file.
4441 * config/s390/s390intrin.h: New file.
4442 * doc/extend.texi: Document htm builtins.
4443 * config.gcc: Add the new header files to extra_headers.
4445 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4447 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4448 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4450 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4451 Pat Haugen <pthaugen@us.ibm.com>
4452 Peter Bergner <bergner@vnet.ibm.com>
4454 * config/rs6000/power8.md: New.
4455 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4456 setting for power8 entry.
4457 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4458 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4459 test for Power4/Power5 only.
4460 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4462 (force_new_group): Adjust comment.
4463 * config/rs6000/rs6000.md: Include power8.md.
4465 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4467 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4468 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4470 (tune_params): Update comment.
4471 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4472 (arm_max_conditional_execute): New function.
4473 (thumb2_final_prescan_insn): Use max_insn_skipped and
4474 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4476 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4478 PR tree-optimization/57705
4479 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4480 SSA_NAME step, provided that it is not defined inside the loop.
4481 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4482 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4484 2013-06-25 Martin Jambor <mjambor@suse.cz>
4487 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4488 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4490 (ipa_note_param_call): Initialize member_ptr flag.
4491 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4492 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4493 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4494 (ipa_read_indirect_edge_info): Likewise.
4496 2013-06-25 Richard Biener <rguenther@suse.de>
4499 * passes.c (init_optimization_passes): Move pass_fold_builtins
4500 and pass_dce earlier with -Og.
4502 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4504 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4505 <BIT_FIELD_REF>: Remove trailing TAB.
4506 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4509 2013-06-24 Martin Jambor <mjambor@suse.cz>
4511 PR tree-optimization/57358
4512 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4513 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4514 (ipa_analyze_params_uses): Generate pessimistic info when true.
4516 2013-06-24 Martin Jambor <mjambor@suse.cz>
4518 PR tree-optimization/57539
4519 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4520 global.inlined_to of the new node to it. All callers changed.
4521 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4522 inlining_into, pass it to cgraph_clone_node.
4523 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4524 ipa_free_edge_args_substructures.
4525 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4526 rdesc linked list. Do not assert rdesc edges have inlined caller.
4527 Assert we have found an rdesc in the rdesc list.
4529 2013-06-24 Richard Biener <rguenther@suse.de>
4531 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4532 (pointer_set_lookup): Declare.
4533 (class pointer_map): New template class implementing a
4534 generic pointer to T map.
4535 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4536 pointer_map<T>::contains, pointer_map<T>::insert,
4537 pointer_map<T>::traverse): New functions.
4538 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4539 (pointer_set_lookup): New function.
4540 (pointer_set_contains): Use pointer_set_lookup.
4541 (pointer_set_insert): Likewise.
4542 (insert_aux): Remove.
4543 (struct pointer_map_t): Embed a pointer_set_t.
4544 (pointer_map_create): Adjust.
4545 (pointer_map_destroy): Likewise.
4546 (pointer_map_contains): Likewise.
4547 (pointer_map_insert): Likewise.
4548 (pointer_map_traverse): Likewise.
4549 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4550 pointer_map<unsigned> instead of a pointer_map_t.
4551 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4552 (streamer_tree_cache_lookup): Likewise.
4553 (streamer_tree_cache_create): Likewise.
4554 (streamer_tree_cache_delete): Likewise.
4555 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4556 pointer_map<unsigned> instead of a pointer_map_t.
4557 (lto_init_tree_ref_encoder): Adjust.
4558 (lto_destroy_tree_ref_encoder): Likewise.
4559 * lto-section-out.c (lto_output_decl_index): Likewise.
4560 (lto_record_function_out_decl_state): Likewise.
4561 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4563 2013-06-24 Richard Biener <rguenther@suse.de>
4565 PR tree-optimization/57488
4566 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4568 2013-06-24 Alan Modra <amodra@gmail.com>
4570 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4571 (gen_easy_altivec_constant): Likewise.
4572 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4573 easy_vector_constant_msb): Likewise.
4575 2013-06-23 Jakub Jelinek <jakub@redhat.com>
4578 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4579 add missing return true.
4581 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4584 * config/sh/predicates.md (general_extend_operand): Invoke
4585 general_movsrc_operand for memory operands.
4586 (general_movsrc_operand): Allow reg+reg addressing, do not use
4587 general_operand for memory operands.
4589 2013-06-23 Sriraman Tallam <tmsriram@google.com>
4591 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4592 when current target options does not apply.
4593 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4594 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
4595 * config/i386/bmiintrin.h: Pass appropriate target
4596 attributes to header.
4597 * config/i386/mmintrin.h: Ditto.
4598 * config/i386/nmmintrin.h: Ditto.
4599 * config/i386/avx2intrin.h: Ditto.
4600 * config/i386/fxsrintrin.h: Ditto.
4601 * config/i386/tbmintrin.h: Ditto.
4602 * config/i386/xsaveintrin.h: Ditto.
4603 * config/i386/f16cintrin.h: Ditto.
4604 * config/i386/xtestintrin.h: Ditto.
4605 * config/i386/xsaveoptintrin.h: Ditto.
4606 * config/i386/bmi2intrin.h: Ditto.
4607 * config/i386/lzcntintrin.h: Ditto.
4608 * config/i386/smmintrin.h: Ditto.
4609 * config/i386/wmmintrin.h: Ditto.
4610 * config/i386/x86intrin.h: Remove all header include guards.
4611 * config/i386/prfchwintrin.h: Ditto.
4612 * config/i386/pmmintrin.h: Ditto.
4613 * config/i386/tmmintrin.h: Ditto.
4614 * config/i386/xmmintrin.h: Ditto.
4615 * config/i386/popcntintrin.h: Ditto.
4616 * config/i386/rdseedintrin.h: Ditto.
4617 * config/i386/ammintrin.h: Ditto.
4618 * config/i386/emmintrin.h: Ditto.
4619 * config/i386/immintrin.h: Remove all header include guards.
4620 * config/i386/fma4intrin.h: Ditto.
4621 * config/i386/lwpintrin.h: Ditto.
4622 * config/i386/xopintrin.h: Ditto.
4623 * config/i386/ia32intrin.h: Ditto.
4624 * config/i386/avxintrin.h: Ditto.
4625 * config/i386/rtmintrin.h: Ditto.
4626 * config/i386/fmaintrin.h: Ditto.
4627 * config/i386/mm3dnow.h: Ditto.
4629 2013-06-22 Sriraman Tallam <tmsriram@google.com>
4631 * common/config/i386/i386-common.c: Handle LZCNT.
4633 2013-06-22 Andi Kleen <ak@linux.intel.com>
4635 * doc/extend.texi: Use __atomic_store_n instead of
4636 __atomic_store in HLE example.
4638 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4640 * config/sh/sh.c: Remove <cstdlib> workaround.
4642 2013-06-21 Andi Kleen <ak@linux.intel.com>
4644 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
4646 2013-06-21 Andi Kleen <ak@linux.intel.com>
4648 * doc/extend.texi: Document that __atomic_clear and
4649 __atomic_test_and_set should only be used with bool.
4651 2013-06-20 Jan Hubicka <jh@suse.cz>
4653 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4655 * tree.c (decls_same_for_odr): New function.
4656 (same_for_edr): New function.
4657 (types_same_for_odr): New function.
4658 (get_binfo_at_offset): Use it.
4659 * tree.h (types_same_for_odr): Declare.
4661 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
4662 Jason Merrill <jason@redhat.com>
4664 * system.h: Include <cstdlib> as well as <stdlib.h>.
4666 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
4669 * config/i386/i386.c (construct_container): Report error if
4670 long double is used with disabled x87 float returns.
4672 2013-06-20 Jan Hubicka <jh@suse.cz>
4674 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4676 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4678 PR rtl-optimization/57425
4679 PR rtl-optimization/57569
4680 * alias.c (write_dependence_p): Remove parameters mem_mode and
4681 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4682 Changed all callers.
4683 (canon_anti_dependence): Get comments and semantics in sync.
4684 Add parameter mem_canonicalized. Changed all callers.
4685 * rtl.h (canon_anti_dependence): Update prototype.
4687 2013-06-20 Richard Biener <rguenther@suse.de>
4689 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4690 byte case, inline streamer_read_uchar and defer section
4693 2013-06-20 Richard Biener <rguenther@suse.de>
4695 PR tree-optimization/57584
4696 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4697 SSA names into the expanded expression that take part in
4698 abnormal coalescing.
4700 2013-06-19 Sharad Singhai <singhai@google.com>
4702 * gcov.c (print_usage): Handle new option.
4703 (process_args): Ditto.
4704 (get_gcov_intermediate_filename): New function.
4705 (output_intermediate_file): New function.
4706 (output_gcov_file): New function
4707 (generate_results): Handle new option.
4708 (release_function): Relase demangled name.
4709 (read_graph_file): Handle demangled name.
4710 (output_lines): Ditto.
4711 * doc/gcov.texi: Document gcov intermediate format.
4713 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4716 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4717 (lra_emit_add): Use the functions. Add comment about Y as an
4720 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
4723 * collect2.c (collect_atexit): New.
4724 (collect_exit): Delete.
4725 (main): Register collect_atexit with atexit.
4726 (collect_wait): Change collect_exit to exit.
4728 * collect2.h (collect_exit): Delete.
4729 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4731 2013-06-19 Wei Mi <wmi@google.com>
4733 PR rtl-optimization/57518
4734 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4735 if regno is used in paradoxical subreg.
4736 (update_equiv_regs): Check pdx_subregs[regno] before
4737 set a reg to be equivalent with a mem.
4739 2013-06-19 Matthias Klose <doko@ubuntu.com>
4742 * file-find.h (find_a_file): Add a mode parameter.
4743 * file-find.c (find_a_file): Likewise.
4744 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4745 with X_OK for the executables.
4746 * collect2.c (main): Call find_a_file with X_OK.
4748 2013-06-19 Steve Ellcey <sellcey@mips.com>
4751 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4752 Use NEXT_INSN instead of next_real_insn.
4754 2013-06-19 Jan Hubicka <jh@suse.cz>
4756 * cgraph.h (const_value_known_p): Replace by ...
4757 (ctor_for_folding): .. this one.
4758 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4759 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4760 * expr.c (expand_expr_real_1): Likewise.
4761 (string_constant): Likewise.
4762 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4763 * ipa.c (process_references): Likewise.
4764 (symtab_remove_unreachable_nodes): Likewise.
4765 * ipa-inline-analysis.c (param_change_prob): Likewise.
4766 * gimple-fold.c (canonicalize_constructor_val): Likewise.
4767 (get_base_constructor): Likwise.
4768 * varpool.c (varpool_remove_node): Likewise.
4769 (varpool_remove_initializer): LIkewise.
4770 (dump_varpool_node): LIkwise.
4771 (const_value_known_p): Rewrite to ...
4772 (ctor_for_folding): ... this one.
4774 2013-06-19 Jakub Jelinek <jakub@redhat.com>
4777 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
4778 PERSONALITY in $PATH derived prefixes.
4780 2013-06-19 Jeff Law <law@redhat.com>
4782 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
4785 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
4786 (simplify_bitwise_binary): Use it to simpify certain binary ops on
4789 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
4791 * config/arm/vfp.md: Move VFP instruction classification documentation
4793 * config/arm/arm.md: ... here. Update instruction classification
4796 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
4798 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
4799 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
4800 pattern. Use more efficient sequences on ARMv5 and Thumb2.
4802 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
4805 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4806 with NEXT_INSN. Use tablejump_p to check for jump table data
4809 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
4812 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
4813 that now in C++ the value is correct per the C++ standards.
4815 2013-06-19 Richard Biener <rguenther@suse.de>
4817 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
4820 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4823 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4826 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
4827 with next_active_insn.
4829 2013-06-18 Sriraman Tallam <tmsriram@google.com>
4831 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
4832 functions are inlined during failures to flag an error.
4833 * tree-inline.c (expand_call_inline): Allow the error to be flagged
4834 in early inline pass.
4836 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
4838 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
4841 2013-06-18 Julian Brown <julian@codesourcery.com>
4843 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
4844 Permit virtual register pre-reload if !strict.
4845 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
4847 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
4849 * config/arm/neon.md (movmisalign<mode>): Use
4850 neon_perm_struct_or_reg_operand instead of
4851 neon_struct_or_register_operand.
4852 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
4853 neon_permissive_struct_operand instead of neon_struct_operand.
4854 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
4855 neon_vector_mem_operand.
4856 * config/arm/predicates.md (neon_struct_operand): Adjust call to
4857 neon_vector_mem_operand.
4858 (neon_permissive_struct_operand): New.
4859 (neon_struct_or_register_operand): Rename to...
4860 (neon_perm_struct_or_reg_operand): This. Adjust call to
4861 neon_vector_mem_operand.
4863 2013-06-18 Richard Biener <rguenther@suse.de>
4865 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
4866 * lto-streamer.h: Include pointer-set.h.
4867 (struct lto_decl_slot): Remove.
4868 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
4869 Remove next_index entry.
4870 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4871 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4872 (lto_init_tree_ref_encoder): Adjust.
4873 (lto_destroy_tree_ref_encoder): Likewise.
4874 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
4875 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
4876 (lto_output_decl_index): Adjust.
4877 (lto_new_out_decl_state): Likewise.
4878 (lto_record_function_out_decl_state): Likewise.
4879 * lto-streamer-out.c (copy_function): Likewise.
4881 2013-06-18 Richard Biener <rguenther@suse.de>
4883 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
4884 * cgraphunit.c: Include cfgloop.h.
4885 (init_lowered_empty_function): Initialize the loop tree.
4886 (assemble_thunk): Insert new BBs into loops.
4888 2013-06-18 Richard Biener <rguenther@suse.de>
4890 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
4891 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
4892 the map from cache entry to cache index optional.
4893 (streamer_tree_cache_replace_tree): Adjust accordingly.
4894 (streamer_tree_cache_append): Likewise.
4895 (streamer_tree_cache_delete): Likewise.
4896 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
4897 streamer cache map from cache entry to cache index.
4898 * lto-streamer-out.c (create_output_block): Adjust.
4900 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
4902 * config/arm/arm.md (attribute "insn"): Move multiplication and
4903 division attributes to...
4904 (attribute "type"): ... here. Remove mult.
4905 (attribute "mul32"): New attribute.
4906 (attribute "mul64"): Add umaal.
4907 (*arm_mulsi3): Update attributes.
4908 (*arm_mulsi3_v6): Likewise.
4909 (*thumb_mulsi3): Likewise.
4910 (*thumb_mulsi3_v6): Likewise.
4911 (*mulsi3_compare0): Likewise.
4912 (*mulsi3_compare0_v6): Likewise.
4913 (*mulsi_compare0_scratch): Likewise.
4914 (*mulsi_compare0_scratch_v6): Likewise.
4915 (*mulsi3addsi): Likewise.
4916 (*mulsi3addsi_v6): Likewise.
4917 (*mulsi3addsi_compare0): Likewise.
4918 (*mulsi3addsi_compare0_v6): Likewise.
4919 (*mulsi3addsi_compare0_scratch): Likewise.
4920 (*mulsi3addsi_compare0_scratch_v6): Likewise.
4921 (*mulsi3subsi): Likewise.
4922 (*mulsidi3adddi): Likewise.
4923 (*mulsi3addsi_v6): Likewise.
4924 (*mulsidi3adddi_v6): Likewise.
4925 (*mulsidi3_nov6): Likewise.
4926 (*mulsidi3_v6): Likewise.
4927 (*umulsidi3_nov6): Likewise.
4928 (*umulsidi3_v6): Likewise.
4929 (*umulsidi3adddi): Likewise.
4930 (*umulsidi3adddi_v6): Likewise.
4931 (*smulsi3_highpart_nov6): Likewise.
4932 (*smulsi3_highpart_v6): Likewise.
4933 (*umulsi3_highpart_nov6): Likewise.
4934 (*umulsi3_highpart_v6): Likewise.
4935 (mulhisi3): Likewise.
4936 (*mulhisi3tb): Likewise.
4937 (*mulhisi3bt): Likewise.
4938 (*mulhisi3tt): Likewise.
4939 (maddhisi4): Likewise.
4940 (*maddhisi4tb): Likewise.
4941 (*maddhisi4tt): Likewise.
4942 (maddhidi4): Likewise.
4943 (*maddhidi4tb): Likewise.
4944 (*maddhidi4tt): Likewise.
4946 (udivsi3): Likewise.
4947 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
4948 (thumb2_mulsi_short_compare0): Likewise.
4949 (thumb2_mulsi_short_compare0_scratch): Likewise.
4950 * config/arm/arm1020e.md (1020mult1): Update attribute change.
4951 (1020mult2): Likewise.
4952 (1020mult3): Likewise.
4953 (1020mult4): Likewise.
4954 (1020mult5): Likewise.
4955 (1020mult6): Likewise.
4956 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
4958 (cortex_a15_mult64): Likewise.
4959 (cortex_a15_sdiv): Likewise.
4960 (cortex_a15_udiv): Likewise.
4961 * config/arm/arm1026ejs.md (mult1): Update attribute change.
4967 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
4968 (pj4_ir_div): Likewise.
4969 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
4970 (11_mult2): Likewise.
4971 (11_mult3): Likewise.
4972 (11_mult4): Likewise.
4973 (11_mult5): Likewise.
4974 (11_mult6): Likewise.
4975 (11_mult7): Likewise.
4976 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
4977 (cortex_a8_mla): Likewise.
4978 (cortex_a8_mull): Likewise.
4979 (cortex_a8_smulwy): Likewise.
4980 (cortex_a8_smlald): Likewise.
4981 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
4982 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
4983 (cortex_r4_mul_3): Likewise.
4984 (cortex_r4_mla_4): Likewise.
4985 (cortex_r4_mla_3): Likewise.
4986 (cortex_r4_smlald): Likewise.
4987 (cortex_r4_mull): Likewise.
4988 (cortex_r4_sdiv): Likewise.
4989 (cortex_r4_udiv): Likewise.
4990 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
4991 (cortex_a7_idiv): Likewise.
4992 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
4993 (9_mult2): Likewise.
4994 (9_mult3): Likewise.
4995 (9_mult4): Likewise.
4996 (9_mult5): Likewise.
4997 (9_mult6): Likewise.
4998 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
4999 (cortex_a53_sdiv): Likewise.
5000 (cortex_a53_udiv): Likewise.
5001 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5002 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5003 (mp626_mult2): Likewise.
5004 (mp626_mult3): Likewise.
5005 (mp626_mult4): Likewise.
5006 * config/arm/fa526.md (526_mult1): Update attribute change.
5007 (526_mult2): Likewise.
5008 * config/arm/arm-generic.md (mult): Update attribute change.
5009 (mult_ldsched_strongarm): Likewise.
5010 (mult_ldsched): Likewise.
5011 (multi_cycle): Likewise.
5012 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5013 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5014 (606te_mult2): Likewise.
5015 (606te_mult3): Likewise.
5016 (606te_mult4): Likewise.
5017 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5018 (cortex_a9_mac16): Likewise.
5019 (cortex_a9_multiply): Likewise.
5020 (cortex_a9_mac): Likewise.
5021 (cortex_a9_multiply_long): Likewise.
5022 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5023 (626te_mult2): Likewise.
5024 (626te_mult3): Likewise.
5025 (626te_mult4): Likewise.
5027 2013-06-18 Richard Biener <rguenther@suse.de>
5030 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5032 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5035 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5036 with next_active_insn.
5038 2013-06-18 Alan Modra <amodra@gmail.com>
5040 * config/rs6000/rs6000.h (enum data_align): New.
5041 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5042 (DATA_ABI_ALIGNMENT): Define.
5043 (CONSTANT_ALIGNMENT): Correct comment.
5044 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5045 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5047 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5049 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5050 ATTRIBUTE_UNUSED marking.
5052 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5054 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5055 alternative and update.
5056 (aarch64_dup_lanedi): Delete.
5057 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5058 * config/aarch64/aarch64-simd-builtins.def: Update.
5060 2013-06-17 Richard Biener <rguenther@suse.de>
5062 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5063 (lto_input_scc): Declare.
5064 (lto_input_tree_1): Likewise.
5065 (struct lto_stats_d): Add num_tree_bodies_output and
5066 num_pickle_refs_output.
5067 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5068 (lto_read_tree_1): Split out from ...
5069 (lto_read_tree): ... this.
5070 (lto_input_scc): New function.
5071 (lto_input_tree_1): Split out from ...
5072 (lto_input_tree): ... this. Handle LTO_tree_scc.
5073 (lto_data_in_create): Create the streamer cache without hashes.
5074 * lto-streamer-out.c (create_output_block): Create the streamer
5075 cache with hashes when not doing WPA.
5076 (lto_write_tree_1): Split out from ...
5077 (lto_write_tree): ... this.
5078 (get_symbol_initial_value): New function.
5079 (lto_output_tree_1): Split out from ...
5080 (lto_output_tree): ... this. Write trees as series of SCCs
5081 using a DFS walk via DFS_write_tree.
5082 (struct sccs, struct scc_entry): New types.
5083 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5084 (DFS_write_tree_body): New function.
5085 (DFS_write_tree): Likewise.
5086 (hash_tree): Likewise.
5087 (scc_entry_compare): Likewise.
5088 (hash_scc): Likewise.
5089 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5090 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5091 TREE_CHAIN as regular reference.
5092 (streamer_read_integer_cst): Remove.
5093 (streamer_get_pickled_tree): Adjust.
5094 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5095 of DECL_EXTERNALs in BLOCK_VARS for now.
5096 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5098 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5099 Add hash value argument and record that if hashes are recorded
5101 (streamer_tree_cache_insert_1): Adjust.
5102 (streamer_tree_cache_insert): Likewise.
5103 (streamer_tree_cache_insert_at): Rename to ...
5104 (streamer_tree_cache_replace_tree): ... this and adjust.
5105 (streamer_tree_cache_append): Adjust.
5106 (record_common_node): Likewise.
5107 (streamer_tree_cache_create): Add argument whether to
5108 record hash values together with trees.
5109 (streamer_tree_cache_delete): Adjust.
5110 * tree-streamer.h (struct streamer_tree_cache_d): Add
5112 (streamer_read_integer_cst): Remove.
5113 (streamer_tree_cache_insert): Adjust.
5114 (streamer_tree_cache_append): Likewise.
5115 (streamer_tree_cache_insert_at): Rename to ...
5116 (streamer_tree_cache_replace_tree): ... this and adjust.
5117 (streamer_tree_cache_create): Add argument whether to record hashes.
5118 (streamer_tree_cache_get): Rename to ...
5119 (streamer_tree_cache_get_tree): ... this.
5120 (streamer_tree_cache_get_hash): New function.
5121 * tree.c (cache_integer_cst): New function.
5122 * tree.h (cache_integer_cst): Declare.
5123 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5124 * lto-symtab.c (lto_varpool_replace_node): Only release
5125 DECL_INITIAL of non-prevailing decls.
5126 * varpool.c (varpool_remove_initializer): Do not release
5127 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5129 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5131 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5132 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5133 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5134 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5135 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5136 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5137 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5138 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5139 instead of TARGET_64BIT.
5140 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5141 Require ISA_HAS_<D>DIV.
5143 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
5145 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
5146 (mips*-*-linux*): Move default with_llsc setting to where other
5148 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
5150 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
5151 Likewise. Remove default with_tune setting. Move default float
5152 setting to its own block. Handle with_llsc in the same block as above.
5154 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
5156 PR rtl-optimization/57425
5157 PR rtl-optimization/57569
5158 * alias.c (write_dependence_p): Add new parameters mem_mode,
5159 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
5160 Changed all callers.
5161 (canon_anti_dependence): New function.
5162 * cse.c (check_dependence): Use canon_anti_dependence.
5163 * cselib.c (cselib_invalidate_mem): Likewise.
5164 * rtl.h (canon_anti_dependence): Declare.
5166 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5168 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
5169 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
5170 ".set mips3" for 64-bit targets.
5172 2013-06-15 Dehao Chen <dehao@google.com>
5174 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
5175 * gimple-low.c (gimple_check_call_matching_types): Likewise.
5176 (gimple_check_call_args): Likewise.
5177 * value-prof.c (check_ic_target): Likewise.
5178 * ipa-inline.c (early_inliner): Likewise.
5179 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
5180 * cgraph.c (cgraph_create_edge_1): Likewise.
5181 (cgraph_make_edge_direct): Likewise.
5183 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
5186 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
5187 rs6000_output_move_128bit to handle emitting quad memory
5188 operations. Set attribute length to 8 bytes.
5190 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
5192 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
5194 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
5195 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
5196 (aarch64_<su>mlsl<mode>): Likewise.
5198 2013-06-14 Mike Stump <mikestump@comcast.net>
5200 * Makefile.in (TARGET_H): Add insn-codes.h.
5202 2013-06-14 Alan Modra <amodra@gmail.com>
5206 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
5207 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
5208 bitfield expansion when EXPAND_MEMORY.
5209 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
5211 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
5213 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
5214 test for clearing quad memory on 32-bit later.
5216 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5218 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
5219 (fold_negate_expr): Likewise.
5220 (fold_real_zero_addition_p): Handle vectors.
5221 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
5223 2013-06-14 Alan Modra <amodra@gmail.com>
5225 * varasm.c (force_const_mem): Revert 2013-06-07 change.
5227 2013-06-13 Jan Hubicka <jh@suse.cz>
5229 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
5230 Local comdats are not externally visible.
5231 * symtab.c (dump_symtab_base): Dump externally visible.
5232 (verify_symtab_base): Verify back links in the symtab hash.
5234 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5236 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
5237 CONVERT_EXPR as equal nodes.
5239 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5241 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
5243 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5245 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
5246 Generalize to complex and vector.
5247 * tree.c (build_all_ones_cst): New function.
5248 * tree.h (build_all_ones_cst): Declare it.
5250 2013-06-13 Alan Modra <amodra@gmail.com>
5252 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
5253 * config/rs6000/rs6000.md (signbittf2): New insn.
5254 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
5255 (abstf2_internal, cmptf_internal2): Likewise.
5256 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
5258 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5259 Pat Haugen <pthaugen@us.ibm.com>
5260 Peter Bergner <bergner@vnet.ibm.com>
5262 * config/rs6000/rs6000.c (emit_load_locked): Add support for
5263 power8 byte, half-word, and quad-word atomic instructions.
5264 (emit_store_conditional): Likewise.
5265 (rs6000_expand_atomic_compare_and_swap): Likewise.
5266 (rs6000_expand_atomic_op): Likewise.
5268 * config/rs6000/sync.md (larx): Add new modes for power8.
5270 (AINT): New mode iterator to include TImode as well as normal
5271 integer modes on power8.
5272 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
5273 that VSX registers are not considered. Use AINT mode iterator
5274 instead of INT1 to allow inclusion of quad word atomic operations
5276 (load_locked<mode>): Likewise.
5277 (store_conditional<mode>): Likewise.
5278 (atomic_compare_and_swap<mode>): Likewise.
5279 (atomic_exchange<mode>): Likewise.
5280 (atomic_nand<mode>): Likewise.
5281 (atomic_fetch_<fetchop_name><mode>): Likewise.
5282 (atomic_nand_fetch<mode>): Likewise.
5283 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
5285 (ATOMIC): On power8, add QImode, HImode modes.
5286 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
5287 modes that promote to SImode.
5288 (load_lockedti): Convert TImode arguments to PTImode, so that we
5289 get a guaranteed even/odd register pair.
5290 (load_lockedpti): Likewise.
5291 (store_conditionalti): Likewise.
5292 (store_conditionalpti): Likewise.
5294 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
5295 atomic load/store instructions.
5298 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
5300 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
5302 (insn_count): New attribute, with most cases extracted from...
5303 (length): ...here. Redefine most cases in terms of insn_count.
5304 (single_insn): Delete.
5305 (can_delay): Use insn_count to check for single instructions.
5306 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
5307 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
5308 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
5309 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
5310 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
5311 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
5312 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
5313 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
5314 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
5315 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
5316 rather than "length".
5317 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
5318 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
5319 Use "insn_count" rather than "length".
5320 * config/mips/mips-dsp.md
5321 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
5322 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
5325 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
5327 PR tree-optimization/57361
5328 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
5330 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
5332 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
5334 (aarch64_simd_combine<mode>): New instruction expansion.
5335 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
5337 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
5338 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
5340 2013-06-12 Jan Hubicka <jh@suse.cz>
5342 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
5343 decl has when in streaming stage.
5344 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
5345 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
5347 2013-06-12 Roland Stigge <stigge@antcom.de>
5350 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
5352 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5354 PR tree-optimization/57537
5355 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
5356 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
5358 2013-06-12 Richard Biener <rguenther@suse.de>
5360 * data-streamer.h (streamer_write_char_stream): CSE
5361 obs->current_pointer.
5362 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
5363 streamer_write_char_stream manually and optimize the resulting loop.
5364 (streamer_write_hwi_stream): Likewise.
5366 2013-06-12 Jan Hubicka <jh@suse.cz>
5368 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
5369 * cgraph.h (varpool_create_empty_node): Declare.
5370 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
5372 * symtab.c (symtab_unregister_node): Be lax about missin entries
5374 (symtab_get_node): Update comment.
5375 * varpool.c (varpool_create_empty_node): Break out from ...
5376 (varpool_node_for_decl): ... here.
5377 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
5379 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
5381 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
5382 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
5383 part. Use straight-line flow at the end.
5384 <COMPONENT_REF>: Remove superfluous else.
5385 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
5387 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5390 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
5391 target hook even for !TREE_PUBLIC decls. If no resolution info
5392 is available, return false for common and external decls.
5394 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5396 * config/rl78/constraints.md (U): New constraint.
5397 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5400 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5403 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5404 to allow returning address to AT_PLATFORM name.
5406 2013-06-11 Jan Hubicka <jh@suse.cz>
5408 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5409 * cgraph.h (symtab_node_base): Add weakref flag.
5410 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5411 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5412 (output_weakrefs): Use weakref flag.
5413 * fold-const.c (simple_operand_p): Handle WEAK.
5414 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5415 * ipa.c (varpool_externally_visible_p): Drop weakref.
5416 (function_and_variable_visibility): Update comment; fix weakref
5417 sanity checks; do not clear DECL_WEAK on them.
5418 * lto-cgraph.c (lto_output_node): update.
5419 (lto_output_varpool_node): Update.
5420 (input_overwrite_node): Update.
5421 (input_node): Update.
5422 (input_varpool_node): Update.
5423 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5424 (lto_symtab_merge_symbols): Add sanity check.
5425 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5426 * passes.c (rest_of_decl_compilation): Set static flag, too.
5427 * symtab.c (dump_symtab_base): Dump weakref.
5428 (verify_symtab_base): Sanity check weakrefs.
5429 (symtab_make_decl_local): Remove duplicated code.
5430 (symtab_alias_ultimate_target): Simplify.
5431 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5433 2013-06-11 DJ Delorie <dj@redhat.com>
5435 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5436 (rl78_unwind_word_mode): New.
5438 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5440 * final.c (debug_prefix_maps): Make static.
5442 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5444 * function.c (initial_trampoline): Remove stray copy.
5446 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5448 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5450 2013-06-11 Martin Jambor <mjambor@suse.cz>
5452 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5453 within bounds at the beginning of the function.
5455 2013-06-11 Alan Modra <amodra@gmail.com>
5457 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5459 (get_named_section): Don't NULL !DECL_P decl.
5461 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5463 * doc/invoke.texi (core-avx2): Document.
5465 (atom): Updated with MOVBE.
5467 2013-06-11 Richard Biener <rguenther@suse.de>
5469 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5471 2013-06-11 Anton Blanchard <anton@samba.org>
5473 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5474 correct shift value in little-endian mode.
5476 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5479 * varasm.c (get_variable_align): Move #endif to the right place.
5481 2013-06-10 Cary Coutant <ccoutant@google.com>
5483 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5484 for hash so that hash table traversal order is deterministic.
5486 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5487 Pat Haugen <pthaugen@us.ibm.com>
5488 Peter Bergner <bergner@vnet.ibm.com>
5490 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5491 of vectors in GPRS if they are direct moves or quad word load or
5494 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5496 (direct_move_p): Likewise.
5497 (quad_load_store_p): Likewise.
5499 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5500 classes into bins based on the physical register type.
5501 (reg_class_to_reg_type): Likewise.
5502 (IS_STD_REG_TYPE): Likewise.
5503 (IS_FP_VECT_REG_TYPE): Likewise.
5504 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5505 use direct move instructions.
5506 (reload_gpr_vsx): Likewise.
5507 (reload_vsx_gpr): Likewise.
5508 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5509 information that is a simplification of register classes. Also
5510 precalculate direct move reload helpers.
5511 (direct_move_p): New function to return true if the operation can
5512 be done as a direct move instruciton.
5513 (quad_load_store_p): New function to return true if the operation
5514 is a quad memory operation.
5515 (rs6000_legitimize_address): If quad memory, only allow register
5516 indirect for TImode addresses.
5517 (rs6000_legitimate_address_p): Likewise.
5518 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5519 (rs6000_reload_register_type): Likewise.
5520 (register_to_reg_type): Return register type.
5521 (rs6000_secondary_reload_simple_move): New helper function for
5522 secondary reload and secondary memory needed to identify anything
5523 that is a simple move, and does not need reloading.
5524 (rs6000_secondary_reload_direct_move): New helper function for
5525 secondary reload to identify cases that can be done with several
5526 instructions via the direct move instructions.
5527 (rs6000_secondary_reload_move): New helper function for secondary
5528 reload to identify moves between register types that can be done.
5529 (rs6000_secondary_reload): Add support for quad memory operations
5530 and for direct move.
5531 (rs6000_secondary_memory_needed): Likewise.
5532 (rs6000_debug_secondary_memory_needed): Change argument names.
5533 (rs6000_output_move_128bit): New function to return the move to
5534 use for 128-bit moves, including knowing about the various
5535 limitations of quad memory operations.
5537 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5538 memory operations. call rs6000_output_move_128bit for the actual
5539 instruciton(s) to generate.
5540 (vsx_movti_64bit): Likewise.
5542 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5543 (UNSPEC_P8V_MTVSRWZ): Likewise.
5544 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5545 (UNSPEC_P8V_MTVSRD): Likewise.
5546 (UNSPEC_P8V_XXPERMDI): Likewise.
5547 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5548 (UNSPEC_FUSION_GPR): Likewise.
5549 (FMOVE128_GPR): New iterator for direct move.
5550 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
5553 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5554 loads and direct move instructions.
5555 (zero_extendsidi2_lfiwzx): Likewise.
5556 (extendsidi2_lfiwax): Likewise.
5557 (extendsidi2_nocell): Likewise.
5558 (floatsi<mode>2_lfiwax): Likewise.
5560 (floatunssi<mode>2_lfiwzx): Likewise.
5562 (fix_trunc<mode>_stfiwx): Likewise.
5563 (fixuns_trunc<mode>_stfiwx): Likewise.
5564 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5565 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5566 (parity<mode>2_cmpb): Set length/type attr.
5567 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5568 for 'mr.' to fast_compare.
5569 (bpermd_<mode>): Change type attr to popcnt.
5570 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5571 (p8_mtvsrwz_1): Likewise.
5572 (p8_mtvsrwz_2): Likewise.
5573 (reload_fpr_from_gpr<mode>): Likewise.
5574 (p8_mtvsrd_1): Likewise.
5575 (p8_mtvsrd_2): Likewise.
5576 (p8_xxpermdi_<mode>): Likewise.
5577 (reload_vsx_from_gpr<mode>): Likewise.
5578 (reload_vsx_from_gprsf): Likewise.
5579 (p8_mfvsrd_3_<mode>): LIkewise.
5580 (reload_gpr_from_vsx<mode>): Likewise.
5581 (reload_gpr_from_vsxsf): Likewise.
5582 (p8_mfvsrd_4_disf): Likewise.
5583 (multi-word GPR splits): Do not split direct moves or quad memory
5586 2013-06-10 David Malcolm <dmalcolm@redhat.com>
5588 * tree-into-ssa.c (interesting_blocks): Make static.
5590 2013-06-10 Jakub Jelinek <jakub@redhat.com>
5593 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5594 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5595 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5596 (get_variable_align): New function.
5597 (get_variable_section, emit_bss, emit_common,
5598 assemble_variable_contents, place_block_symbol): Use
5599 get_variable_align instead of DECL_ALIGN.
5600 (assemble_noswitch_variable): Add align argument, use it
5601 instead of DECL_ALIGN.
5602 (assemble_variable): Adjust caller. Use get_variable_align
5603 instead of DECL_ALIGN.
5604 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5606 (DATA_ABI_ALIGNMENT): Define.
5607 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5608 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5609 opt is false, only return the psABI mandated alignment increase.
5610 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5611 (DATA_ABI_ALIGNMENT): ... this.
5612 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5613 (DATA_ABI_ALIGNMENT): ... this.
5614 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5615 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5616 (DATA_ABI_ALIGNMENT): ... this.
5617 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5618 * doc/tm.texi: Regenerated.
5620 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
5622 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5623 cmp_code to construct REG_EQUAL note.
5625 2013-06-09 Jakub Jelinek <jakub@redhat.com>
5628 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5629 that operands[2] doesn't overlap with operands[0].
5631 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
5632 Jan Hubicka <jh@suse.cz>
5634 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5635 hack to mark symbols as used.
5637 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5639 PR rtl-optimization/57559
5640 * lra-constraints.c (process_alt_operands): Don't discourage
5641 memory with known offset for offsetable memory constraint.
5642 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5644 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5646 * varasm.c (struct oc_local_state): Reorder fields.
5647 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5648 and adjust accordingly.
5649 (output_constructor): Reorder initialization code and adjust call to
5650 output_constructor_bitfield.
5652 2013-06-07 Jan Hubicka <jh@suse.cz>
5654 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5656 2013-06-07 David Malcolm <dmalcolm@redhat.com>
5658 * tree-object-size.c (unknown): Make const.
5660 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5662 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5663 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5664 for last alternative in the cpu_facility attribute.
5666 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5669 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5670 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5671 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5672 * config/arm/constraints.md (Dg): New constraint.
5673 * config/arm/neon.md (xordi3_neon): Remove.
5674 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5675 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5677 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5679 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5680 Clean up alternatives.
5682 2013-06-07 Alan Modra <amodra@gmail.com>
5684 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5687 2013-06-07 Alan Modra <amodra@gmail.com>
5689 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5691 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5693 * config/arm/constraints.md (Df): New constraint.
5694 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5695 Correct length attribute for last two alternatives.
5697 2013-06-07 Alan Modra <amodra@gmail.com>
5699 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5700 override user -mfp-in-toc.
5701 (offsettable_ok_by_alignment): Consider just the current access
5702 rather than the whole object, unless BLKmode. Handle
5703 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5704 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5705 for -mcmodel=medium.
5706 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5707 override user -mfp-in-toc or -msum-in-toc. Default to
5708 -mno-fp-in-toc for -mcmodel=medium.
5710 2013-06-06 DJ Delorie <dj@redhat.com>
5712 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5713 TARGET_VALID_POINTER_MODE.
5715 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5716 Pat Haugen <pthaugen@us.ibm.com>
5717 Peter Bergner <bergner@vnet.ibm.com>
5719 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5720 Document new power8 builtins.
5722 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5723 condition code register, to allow 128-bit logical operations to be
5724 done in the VSX or GPR registers.
5725 (nor<mode>3): Use the canonical form for nor.
5726 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5727 vclz*, and vpopcnt* vector instructions.
5728 (nand<mode>3): Likewise.
5729 (orc<mode>3): Likewise.
5730 (clz<mode>2): LIkewise.
5731 (popcount<mode>2): Likewise.
5733 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5734 that only the GPRs are recognized.
5736 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5737 support for new power8 builtins.
5739 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5741 (xscvdpspn): Likewise.
5747 (vpopcnt): Likewise.
5748 (vpopcntb): Likewise.
5749 (vpopcnth): Likewise.
5750 (vpopcntw): Likewise.
5751 (vpopcntd): Likewise.
5756 (eqv_v16qi3): Likewise.
5757 (eqv_v8hi3): Likewise.
5758 (eqv_v4si3): Likewise.
5759 (eqv_v2di3): Likewise.
5760 (eqv_v4sf3): Likewise.
5761 (eqv_v2df3): Likewise.
5763 (nand_v16qi3): Likewise.
5764 (nand_v8hi3): Likewise.
5765 (nand_v4si3): Likewise.
5766 (nand_v2di3): Likewise.
5767 (nand_v4sf3): Likewise.
5768 (nand_v2df3): Likewise.
5770 (orc_v16qi3): Likewise.
5771 (orc_v8hi3): Likewise.
5772 (orc_v4si3): Likewise.
5773 (orc_v2di3): Likewise.
5774 (orc_v4sf3): Likewise.
5775 (orc_v2df3): Likewise.
5777 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
5778 allow power8 quad mode in 64-bit.
5779 (rs6000_builtin_vectorized_function): Add support to vectorize
5780 ISA 2.07 count leading zeros, population count builtins.
5781 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
5782 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
5783 (builtin_function_type): Add vgbbd builtin function which takes an
5785 (altivec_expand_vec_perm_const): Add support for new power8 merge
5788 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
5789 that does not include TImdoe for use with 32-bit.
5790 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
5792 (UNSPEC_VSX_CVDPSPN): Likewise.
5793 (vsx_xscvdpspn): Likewise.
5794 (vsx_xscvspdpn): Likewise.
5795 (vsx_xscvdpspn_scalar): Likewise.
5796 (vsx_xscvspdpn_directmove): Likewise.
5797 (vsx_and<mode>3): Split logical operations into 32-bit and
5798 64-bit. Add support to do logical operations on TImode as well as
5799 VSX vector types. Allow logical operations to be done in either
5800 VSX registers or in general purpose registers in 64-bit mode. Add
5801 splitters if GPRs were used. For AND, add clobber of CCmode to
5802 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
5804 (vsx_and<mode>3_32bit): Likewise.
5805 (vsx_and<mode>3_64bit): Likewise.
5806 (vsx_ior<mode>3): Likewise.
5807 (vsx_ior<mode>3_32bit): Likewise.
5808 (vsx_ior<mode>3_64bit): Likewise.
5809 (vsx_xor<mode>3): Likewise.
5810 (vsx_xor<mode>3_32bit): Likewise.
5811 (vsx_xor<mode>3_64bit): Likewise.
5812 (vsx_one_cmpl<mode>2): Likewise.
5813 (vsx_one_cmpl<mode>2_32bit): Likewise.
5814 (vsx_one_cmpl<mode>2_64bit): Likewise.
5815 (vsx_nor<mode>3): Likewise.
5816 (vsx_nor<mode>3_32bit): Likewise.
5817 (vsx_nor<mode>3_64bit): Likewise.
5818 (vsx_andc<mode>3): Likewise.
5819 (vsx_andc<mode>3_32bit): Likewise.
5820 (vsx_andc<mode>3_64bit): Likewise.
5821 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
5822 and xxlorc instructions.
5823 (vsx_eqv<mode>3_64bit): Likewise.
5824 (vsx_nand<mode>3_32bit): Likewise.
5825 (vsx_nand<mode>3_64bit): Likewise.
5826 (vsx_orc<mode>3_32bit): Likewise.
5827 (vsx_orc<mode>3_64bit): Likewise.
5829 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
5831 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
5833 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
5834 (p8_vmrgow): Likewise.
5835 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
5836 GPRs to be split under VSX.
5837 (p8v_clz<mode>2): Add power8 count leading zero support.
5838 (p8v_popcount<mode>2): Add power8 population count support.
5839 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
5842 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
5845 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
5847 (vec_nand): Likewise.
5848 (vec_vclz): Likewise.
5849 (vec_vclzb): Likewise.
5850 (vec_vclzd): Likewise.
5851 (vec_vclzh): Likewise.
5852 (vec_vclzw): Likewise.
5853 (vec_vgbbd): Likewise.
5854 (vec_vmrgew): Likewise.
5855 (vec_vmrgow): Likewise.
5856 (vec_vpopcnt): Likewise.
5857 (vec_vpopcntb): Likewise.
5858 (vec_vpopcntd): Likewise.
5859 (vec_vpopcnth): Likewise.
5860 (vec_vpopcntw): Likewise.
5862 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5864 PR rtl-optimization/57468
5865 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
5868 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5870 PR rtl-optimization/57459
5871 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
5872 type when setting live regs.
5874 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
5876 * config/s390/s390.opt (mlra): New option.
5877 * config/s390/s390.c (s390_decompose_address): Check displacement
5878 for all registers for LRA.
5879 (s390_secondary_reload): Don't used secondary reloads for LRA.
5880 (s390_lra_p): New function.
5881 (TARGET_LRA_P): Define.
5882 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
5883 of attribute cpu_facility to zarch for the last alternative.
5884 (*cmpmem_short): Ditto.
5886 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
5888 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
5889 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
5890 (arm_expand_prologue): Likewise.
5892 2013-06-06 Teresa Johnson <tejohnson@google.com>
5895 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
5896 as this is now done by redirect_edge_and_branch_force.
5897 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
5898 barriers, and fix interaction with splitting.
5899 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
5900 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
5901 reflect changes made in the routine.
5902 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
5903 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
5904 since this is called in cfglayout mode, and replace partition fixup
5905 with assert as that is now done by force_nonfallthru_and_redirect.
5906 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
5907 already be marked with region crossing note.
5908 (insert_section_boundary_note): Make non-static, gate on flag
5909 has_bb_partition, rewrite to also check for multiple partitions.
5910 (rest_of_handle_reorder_blocks): Remove call to
5911 insert_section_boundary_note, now done later during free_cfg.
5912 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
5913 * bb-reorder.h (insert_section_boundary_note): Declare.
5914 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
5915 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
5916 invoke insert_section_boundary_note.
5917 (try_redirect_by_replacing_jump): Remove unnecessary
5918 check for region crossing note.
5919 (fixup_partition_crossing): New function.
5920 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
5921 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
5922 in non-cfglayout mode.
5923 (force_nonfallthru_and_redirect): Fixup partition boundaries,
5924 remove old code that tried to do this. Emit barrier correctly
5925 when we are in cfglayout mode.
5926 (last_bb_in_partition): New function.
5927 (rtl_split_edge): Correctly fixup partition boundaries.
5928 (commit_one_edge_insertion): Remove old code that tried to
5929 fixup region crossing edge since this is now handled in
5930 split_block, and set up insertion point correctly since
5931 block may now end in a jump.
5932 (verify_hot_cold_block_grouping): Guard against checking when not in
5933 linearized RTL mode.
5934 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
5936 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
5937 rtl_verify_flow_info, so not called in cfglayout mode.
5938 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
5939 (fixup_reorder_chain): Remove old code that attempted to fixup region
5940 crossing note as this is now handled in force_nonfallthru_and_redirect.
5941 (duplicate_insn_chain): Don't duplicate switch section notes.
5942 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
5944 * basic-block.h (emit_barrier_after_bb): Declare.
5946 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5948 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
5949 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
5950 arm_usatsihi): Adjust alternatives for arm_restrict_it.
5952 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5954 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
5956 * config/arm/ldmstm.md: Regenerate.
5958 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5960 * config/arm/sync.md (atomic_loaddi_1):
5961 Disable predication for arm_restrict_it.
5962 (arm_load_exclusive<mode>): Likewise.
5963 (arm_load_exclusivesi): Likewise.
5964 (arm_load_exclusivedi): Likewise.
5965 (arm_load_acquire_exclusive<mode>): Likewise.
5966 (arm_load_acquire_exclusivesi): Likewise.
5967 (arm_load_acquire_exclusivedi): Likewise.
5968 (arm_store_exclusive<mode>): Likewise.
5969 (arm_store_exclusive<mode>): Likewise.
5970 (arm_store_release_exclusivedi): Likewise.
5971 (arm_store_release_exclusive<mode>): Likewise.
5973 2013-06-06 Richard Biener <rguenther@suse.de>
5975 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
5977 (lto_tag_is_tree_code_p): Adjust.
5978 (lto_tag_is_gimple_code_p): Likewise.
5979 (lto_gimple_code_to_tag): Likewise.
5980 (lto_tag_to_gimple_code): Likewise.
5981 (lto_tree_code_to_tag): Likewise.
5982 (lto_tag_to_tree_code): Likewise.
5983 * data-streamer.h (streamer_write_hwi_in_range): Use
5984 uhwi streaming to stream the normalized range.
5985 (streamer_read_hwi_in_range): Likewise.
5987 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5989 * config/arm/arm.md (enabled_for_depr_it): New attribute.
5990 (predicable_short_it): Likewise.
5991 (predicated): Likewise.
5992 (enabled): Handle above.
5993 (define_cond_exec): Set predicated attribute to yes.
5995 2013-06-05 Mike Stump <mikestump@comcast.net>
5997 * gdbinit.in (__FUNCTION__): Add.
5999 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6001 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6002 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6004 2013-06-05 Jan Hubicka <jh@suse.cz>
6006 * varasm.c (mark_decl_referenced): Revert the removal until targets
6009 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6011 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6012 instead of mark_decl_referenced.
6014 2013-06-05 Jan Hubicka <jh@suse.cz>
6016 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6017 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6018 and symtab_used_from_object_file_p.
6019 (cgraph_make_node_local_1): Clear forced_by_abi.
6020 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6021 * cgraph.h (symtab_node_base): Add forced_by_abi.
6022 (decide_is_variable_needed): Remove.
6023 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6024 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6025 (decide_is_symbol_needed): ... this one; handle symbols in general;
6026 always analyze virtuals; honnor forced_by_abi.
6027 (cgraph_finalize_function): Update.
6028 (varpool_finalize_decl): Update.
6029 (symbol_defined_and_needed): Remove.
6030 (analyze_functions): Update.
6031 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6032 output_refs, input_overwrite_node): Handle forced_by_abi.
6033 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6034 (address_taken_from_non_vtable_p): ... this one.
6035 (comdat_can_be_unshared_p_1): New function.
6036 (cgraph_comdat_can_be_unshared_p): Rename to ...
6037 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6038 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6039 (function_and_variable_visibility): Clear forced_by_abi as needed.
6040 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6041 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6042 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6043 * varpool.c (decide_is_variable_needed): Remove.
6045 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6047 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6048 (arm_option_override): Override arm_restrict_it where appropriate.
6049 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6050 * config/arm/arm.opt (mrestrict-it): New command-line option.
6051 * doc/invoke.texi: Document -mrestrict-it.
6053 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6055 * tsan.c (tsan_atomic_table): Make const.
6057 2013-06-05 Richard Biener <rguenther@suse.de>
6059 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6060 index associated with the tree we are supposed to replace.
6061 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6062 TREE_ASM_WRITTEN as zero for everything but SSA names.
6064 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6066 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6068 2013-06-04 Jan Hubicka <jh@suse.cz>
6070 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6071 (input_node, input_varpool_node): Handle correctly external same
6073 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6074 nodes at ltrans stage.
6076 2013-06-04 Jan Hubicka <jh@suse.cz>
6078 * ipa-inline.c (update_caller_keys): Fix availability test.
6079 (update_callee_keys): Likewise.
6080 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6081 to follow ELF standard.
6083 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6085 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6086 (mips64r5900el-*-elf*): New configurations.
6087 * config/mips/mips-cpus.def (r5900): New processor.
6088 * config/mips/mips-tables.opt: Regenerate.
6089 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6090 (mips_issue_rate): Handle PROCESSOR_R5900.
6091 (mips_reorg_process_insns): Force reorder mode for the R5900.
6092 * config/mips/mips.h (TARGET_MIPS5900): Define.
6093 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6095 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6097 * config/mips/mips.md (processor): Add r5900.
6098 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6100 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6102 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6103 into function to generate MOVI instruction.
6104 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6105 (aarch64_preferred_simd_mode): Turn into wrapper.
6106 (aarch64_output_scalar_simd_mov_immediate): New function.
6107 * config/aarch64/aarch64-protos.h: Add prototype for above.
6109 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6111 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6112 element_char member.
6113 (sizetochar): Return signed char.
6114 (aarch64_simd_valid_immediate): Remove elchar and other
6115 unnecessary variables.
6116 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6117 Calculate element_char as required.
6118 * config/aarch64/aarch64-protos.h: Update and move prototype
6119 for aarch64_output_simd_mov_immediate.
6120 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6123 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6125 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6126 information completed by aarch64_simd_valid_immediate.
6127 (aarch64_legitimate_constant_p): Update arguments.
6128 (aarch64_simd_valid_immediate): Work with struct rather than many
6130 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6131 (aarch64_simd_make_constant): Update arguments.
6132 (aarch64_output_simd_mov_immediate): Work with struct rather than
6133 many pointers. Output immediate directly rather than as operand.
6134 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6136 * config/aarch64/constraints.md (Dn): Update arguments.
6138 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6140 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6142 (aarch64_simd_immediate_valid_for_move): Remove.
6143 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
6144 (aarch64_simd_make_constant): Update call.
6145 (aarch64_output_simd_mov_immediate): Update call.
6146 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6148 * config/aarch64/constraints.md (Dn): Update call.
6150 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6152 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
6153 return type to bool for prototype.
6154 (aarch64_legitimate_constant_p): Check for true instead of not -1.
6155 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
6156 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
6158 2013-06-04 Catherine Moore <clm@codesourcery.com>
6160 * config/mips/mips.opt (meva): New.
6161 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
6162 (ASM_SPEC): Handle -meva.
6163 * doc/invoke.texi (meva): Document.
6165 2013-06-04 Alan Modra <amodra@gmail.com>
6167 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
6170 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6172 * rtl.def: Add extra fourth optional field to define_cond_exec.
6173 * gensupport.c (process_one_cond_exec): Process attributes from
6175 * doc/md.texi: Document fourth field in define_cond_exec.
6177 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
6179 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
6180 out the processing order as in store_bit_field_1.
6182 2013-06-04 Jan Hubicka <jh@suse.cz>
6185 * cgraphunit.c (cgraph_process_same_body_aliases): Create
6186 non-VAR_DECL node if it does not exist yet.
6188 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
6190 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
6191 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
6192 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
6193 target_cpu_default setting.
6195 2013-06-03 Teresa Johnson <tejohnson@google.com>
6197 * dumpfile.c (opt_info_switch_p): Change -fopt-info
6198 default to -fopt-info=optimized instead of all.
6199 * doc/invoke.texi: Ditto.
6200 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
6201 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
6202 (execute_vect_slp): Emit BB vectorization success under
6203 MSG_OPTIMIZED_LOCATIONS.
6204 * tree-vect-slp.c (vect_slp_transform_bb): Change
6205 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
6206 * tree-vect-loop.c (vect_transform_loop): Ditto.
6208 2013-06-03 Jason Merrill <jason@redhat.com>
6211 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
6212 Use TARGET_EXPR for C++.
6214 2013-06-03 Jakub Jelinek <jakub@redhat.com>
6216 PR rtl-optimization/57268
6217 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
6218 if DEBUG_INSN_P (insn).
6221 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6223 PR rtl-optimization/57268
6224 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6225 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6227 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6229 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
6230 (ix86_avoid_lea_for_addr): Likewise.
6231 (exact_dependency_1): Likewise.
6232 (ix86_adjust_cost): Likewise.
6233 (swap_top_of_ready_list): Fix formatting and !reload_completed check
6235 (do_reorder_for_imul): Fix typo, formatting and
6236 !reload_completed check removed.
6237 (ix86_sched_reorder): Fix typo and formatting.
6238 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
6241 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
6243 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
6245 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
6247 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
6248 <STRING_CST>: Likewise.
6249 <VECTOR_CST>: Likewise.
6251 2013-06-01 Janus Weil <janus@gcc.gnu.org>
6252 Mikael Morin <mikael@gcc.gnu.org>
6254 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
6255 * config.in: Regenerated.
6256 * configure: Regenerated.
6258 2013-06-01 Jan Hubicka <jh@suse.cz>
6261 * cgraphunit.c (compile): When weakref is not supported,
6262 set up transparent aliases before final output pass.
6263 * varasm.c (assemble_alias): Do not try to do it here.
6265 2013-06-01 Jan Hubicka <jh@suse.cz>
6268 * passes.c (for_per_function): Skip unanalyzed functions.
6270 2013-06-01 Jan Hubicka <jh@suse.cz>
6272 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
6273 (lto_symtab_merge_symbols_1): ... this one.
6274 (lto_symtab_merge_cgraph_nodes): Rename to ...
6275 (lto_symtab_merge_symbols): ... this one; simplify.
6276 * cgraph.c (same_body_aliases_done): Rename to ...
6277 (cpp_implicit_aliases_done): ... this one.
6278 (cgraph_create_function_alias): Update.
6279 (cgraph_same_body_alias): Update.
6280 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
6281 (verify_edge_corresponds_to_fndecl): Simplify.
6282 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
6283 (cgraph_node): Remove same_body_alias.
6284 (varpool_node): Remove alias_of and extra_name_alias.
6285 (same_body_aliases_done): Rename to ..
6286 (cpp_implicit_aliases_done): ... this one.
6287 (symtab_alias_ultimate_target): Add default parameter.
6288 (symtab_resolve_alias): New function.
6289 (fixup_same_cpp_alias_visibility): Declare.
6290 (cgraph_function_node): Add default parameter.
6291 (cgraph_node_asm_name): Likewise.
6292 (cgraph_function_or_thunk_node): Add default parameter; do
6293 not ICE when it is NULL.
6294 (varpool_variable_node): Likewise.
6295 * tree-emutls.c (create_emultls_var): Update.
6296 (ipa_lower_emutls): Update.
6297 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
6298 (cgraph_reset_node): Reset alias info.
6299 (cgraph_finalize_function): Update.
6300 (fixup_same_cpp_alias_visibility): Move to symtab.c.
6301 (analyze_function): Simplify.
6302 (cgraph_process_same_body_aliases): Simplify.
6303 (analyze_functions): Fixup same body aliases.
6304 (handle_alias_pairs): Simplify.
6305 (assemble_thunk): Update.
6306 (assemble_thunks_and_aliases): Update.
6307 (output_weakrefs): Rewrite.
6308 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
6309 (lto_output_varpool_node): Likewise.
6310 (compute_ltrans_boundary): Remve assert.
6311 (get_alias_symbol): New functoin.
6312 (input_node): Rewrite alias handling.
6313 (input_varpool_node): Likewise.
6314 * ipa-pure-const.c (propagate_pure_const): Fix formating.
6315 * ipa.c (process_references): Handle weakrefs correctly.
6316 (symtab_remove_unreachable_nodes): Likewise.
6317 * trans-mem.c (get_cg_data): Update.
6318 (ipa_tm_create_version_alias): Update.
6319 (ipa_tm_execute): Update.
6320 * symtab.c (dump_symtab_base): Dump aliases.
6321 (verify_symtab_base): Verify aliases.
6322 (symtab_node_availability): New function.
6323 (symtab_alias_ultimate_target): Simplify.
6324 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
6325 handle all the fixup cases.
6326 (symtab_resolve_alias): New function.
6327 * passes.c (ipa_write_summaries): Handle weakrefs.
6328 * varpool.c (varpool_analyze_node): Simplify.
6329 (assemble_aliases): Update.
6330 (varpool_create_variable_alias): Simplify.
6331 (varpool_extra_name_alias): Simplify.
6332 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
6333 (lto_symtab_merge_symbols): ... this one.
6335 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
6338 PR rtl-optimization/57268
6339 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6340 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6342 2013-06-01 Tobias Burnus <burnus@net-b.de>
6345 2013-05-31 Tobias Burnus <burnus@net-b.de>
6348 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
6351 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6353 PR rtl-optimization/57268
6354 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6355 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6357 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6359 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
6360 unordered comparison operators when -fno-trapping-math is in effect
6362 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
6363 and implement unordered comparison operators properly on the e500.
6365 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6367 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
6368 for constant scalar integers.
6369 (simplify_relational_operation_1): Likewise.
6371 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
6373 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
6374 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
6377 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6378 Igor Zamyatin <igor.zamyatin@intel.com>
6380 Silvermont (SLM) architecture performance tuning.
6381 * config/i386/i386.h (enum ix86_tune_indices): Add
6382 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
6383 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
6385 * config/i386/i386.c (initial_ix86_tune_features)
6386 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
6387 (ix86_lea_outperforms): Handle Silvermont tuning.
6388 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
6390 (ix86_use_lea_for_mov): Likewise.
6391 (ix86_avoid_lea_for_addr): Likewise.
6392 (ix86_lea_for_add_ok): Likewise.
6393 (exact_dependency_1): New function.
6394 (exact_store_load_dependency): Likewise.
6395 (ix86_adjust_cost): Handle Silvermont tuning.
6396 (do_reoder_for_imul): Likewise.
6397 (swap_top_of_ready_list): New function.
6398 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6400 * config/i386/i386.md (peepholes that split memory operand in fp
6403 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6405 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6406 Remove un-necessary braces.
6408 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6410 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6411 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6413 2013-05-31 Tobias Burnus <burnus@net-b.de>
6416 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6418 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6421 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6422 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6423 * config/arm/neon.md (iordi3_neon): Remove.
6424 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6425 * config/arm/predicates.md (imm_for_neon_logic_operand):
6426 Move to earlier in the file.
6427 (neon_logic_op2): Likewise.
6428 (arm_iordi_operand_neon): New predicate.
6430 2013-05-31 Richard Biener <rguenther@suse.de>
6432 PR tree-optimization/57478
6433 PR tree-optimization/57453
6434 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6437 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6439 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6440 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6442 2013-05-30 Tobias Burnus <burnus@net-b.de>
6443 Thomas Koenig <tkoenig@gcc.gnu.org>
6446 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6447 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6449 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6451 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6453 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6455 * target.def (register_usage_leveling_p): New hook.
6456 * targhooks.c (default_register_usage_leveling_p): New.
6457 * targhooks.h (default_register_usage_leveling_p): New prototype.
6458 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6459 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6460 * doc/tm.texi: Update.
6461 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6463 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6465 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6466 (*insv_reg<mode>): New define_insn.
6468 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6470 PR rtl-optimization/57439
6471 * postreload.c (move2add_valid_value_p): Check that we have
6472 a zero subreg_regno_offset when accessing the register in
6475 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6476 Igor Zamyatin <igor.zamyatin@intel.com>
6478 Silvermont (SLM) architecture pipeline model, tuning and
6480 * config.gcc: Add slm config options and target.
6482 * config/i386/slm.md: New.
6484 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6486 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6488 (ix86_target_macros_internal): Likewise.
6490 * gcc/config/i386/i386.c (slm_cost): New cost.
6491 (m_SLM): New macro flag.
6492 (initial_ix86_tune_features): Set m_SLM.
6493 (x86_accumulate_outgoing_args): Likewise.
6494 (x86_arch_always_fancy_math_387): Likewise.
6495 (processor_target_table): Add slm cost.
6496 (cpu_names): Add slm cpu name.
6497 (x86_option_override_internal): Set SLM ISA.
6498 (ix86_issue_rate): New case PROCESSOR_SLM.
6499 (ia32_multipass_dfa_lookahead): Likewise.
6500 (fold_builtin_cpu): Add slm.
6502 * config/i386/i386.h (TARGET_SLM): New target macro.
6503 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6504 (processor_type): Add PROCESSOR_SLM.
6506 * config/i386/i386.md (cpu): Add new value "slm".
6507 (slm.md): Include slm.md.
6509 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6510 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6512 * config/arm/arm-protos.h: Add and update function protos.
6513 * config/arm/arm.c (use_simple_return_p): New added.
6514 (thumb2_expand_return): Check simple_return flag.
6515 * config/arm/arm.md: Add simple_return and conditional simple_return.
6516 * config/arm/iterators.md: Add iterator for return and simple_return.
6518 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6520 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6521 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6522 (arm_emit_vfp_multi_reg_pop): Likewise.
6523 (thumb2_emit_ldrd_pop): Likewise.
6524 (arm_expand_epilogue): Add misc REG_CFA notes.
6525 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6527 2013-05-29 Lawrence Crowl <crowl@google.com>
6529 * config/arm/t-arm: Update for below.
6531 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6532 Change type to hash_table. Update dependent calls and types.
6534 * config/i386/t-cygming: Update for below.
6536 * config/i386/t-interix: Update for below.
6538 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6539 Change type to hash_table. Update dependent calls and types.
6540 (i386_find_on_wrapper_list::wrappers): Likewise.
6542 * config/ia64/t-ia64: Update for below.
6544 * config/ia64/ia64.c (bundle_state_table):
6545 Change type to hash_table. Update dependent calls and types.
6547 * config/mips/mips.c (mips_reorg_process_insns::htab):
6548 Change type to hash_table. Update dependent calls and types.
6550 * config/sol2.c (solaris_comdat_htab):
6551 Change type to hash_table. Update dependent calls and types.
6553 * config/t-sol2: Update for above.
6555 2013-05-29 Teresa Johnson <tejohnson@google.com>
6557 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6558 functions are not yet marked as defined.
6560 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6561 Pat Haugen <pthaugen@us.ibm.com>
6562 Peter Bergner <bergner@vnet.ibm.com>
6564 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6568 (vrotl<mode>3): Likewise.
6569 (vashl<mode>3): Likewise.
6570 (vlshr<mode>3): Likewise.
6571 (vashr<mode>3): Likewise.
6573 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6574 support for power8 V2DI builtins.
6576 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6577 power8 V2DI builtins.
6578 (vupkhsw): Likewise.
6579 (vupklsw): Likewise.
6580 (vaddudm): Likewise.
6585 (vpkudum): Likewise.
6586 (vpksdss): Likewise.
6587 (vpkudus): Likewise.
6588 (vpksdus): Likewise.
6593 (vsubudm): Likewise.
6594 (vcmpequd): Likewise.
6595 (vcmpgtsd): Likewise.
6596 (vcmpgtud): Likewise.
6597 (vcmpequd_p): Likewise.
6598 (vcmpgtsd_p): Likewise.
6599 (vcmpgtud_p): Likewise.
6600 (vupkhsw): Likewise.
6601 (vupklsw): Likewise.
6602 (vaddudm): Likewise.
6607 (vpksdss): Likewise.
6608 (vpksdus): Likewise.
6609 (vpkudum): Likewise.
6610 (vpkudus): Likewise.
6615 (vsubudm): Likewise.
6617 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6618 support for power8 V2DI instructions.
6620 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6621 power8 V2DI instructions. Combine pack and unpack insns to use an
6622 iterator for each mode. Check whether a particular mode supports
6623 Altivec instructions instead of just checking TARGET_ALTIVEC.
6624 (UNSPEC_VPKUWUM): Likewise.
6625 (UNSPEC_VPKSHSS): Likewise.
6626 (UNSPEC_VPKSWSS): Likewise.
6627 (UNSPEC_VPKUHUS): Likewise.
6628 (UNSPEC_VPKSHUS): Likewise.
6629 (UNSPEC_VPKUWUS): Likewise.
6630 (UNSPEC_VPKSWUS): Likewise.
6631 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6632 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6633 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6634 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6635 (UNSPEC_VUPKHSB): Likewise.
6636 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6637 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6638 (UNSPEC_VUPKHSH): Likewise.
6639 (UNSPEC_VUPKLSB): Likewise.
6640 (UNSPEC_VUPKLSH): Likewise.
6642 (VI_char): Likewise.
6643 (VI_scalar): Likewise.
6644 (VI_unit): Likewise.
6646 (VP_small): Likewise.
6647 (VP_small_lc): Likewise.
6648 (VU_char): Likewise.
6649 (add<mode>3): Likewise.
6650 (altivec_vaddcuw): Likewise.
6651 (altivec_vaddu<VI_char>s): Likewise.
6652 (altivec_vadds<VI_char>s): Likewise.
6653 (sub<mode>3): Likewise.
6654 (altivec_vsubcuw): Likewise.
6655 (altivec_vsubu<VI_char>s): Likewise.
6656 (altivec_vsubs<VI_char>s): Likewise.
6657 (altivec_vavgs<VI_char>): Likewise.
6658 (altivec_vcmpbfp): Likewise.
6659 (altivec_eq<mode>): Likewise.
6660 (altivec_gt<mode>): Likewise.
6661 (altivec_gtu<mode>): Likewise.
6662 (umax<mode>3): Likewise.
6663 (smax<mode>3): Likewise.
6664 (umin<mode>3): Likewise.
6665 (smin<mode>3): Likewise.
6666 (altivec_vpkuhum): Likewise.
6667 (altivec_vpkuwum): Likewise.
6668 (altivec_vpkshss): Likewise.
6669 (altivec_vpkswss): Likewise.
6670 (altivec_vpkuhus): Likewise.
6671 (altivec_vpkshus): Likewise.
6672 (altivec_vpkuwus): Likewise.
6673 (altivec_vpkswus): Likewise.
6674 (altivec_vpks<VI_char>ss): Likewise.
6675 (altivec_vpks<VI_char>us): Likewise.
6676 (altivec_vpku<VI_char>us): Likewise.
6677 (altivec_vpku<VI_char>um): Likewise.
6678 (altivec_vrl<VI_char>): Likewise.
6679 (altivec_vsl<VI_char>): Likewise.
6680 (altivec_vsr<VI_char>): Likewise.
6681 (altivec_vsra<VI_char>): Likewise.
6682 (altivec_vsldoi_<mode>): Likewise.
6683 (altivec_vupkhsb): Likewise.
6684 (altivec_vupkhs<VU_char>): Likewise.
6685 (altivec_vupkls<VU_char>): Likewise.
6686 (altivec_vupkhsh): Likewise.
6687 (altivec_vupklsb): Likewise.
6688 (altivec_vupklsh): Likewise.
6689 (altivec_vcmpequ<VI_char>_p): Likewise.
6690 (altivec_vcmpgts<VI_char>_p): Likewise.
6691 (altivec_vcmpgtu<VI_char>_p): Likewise.
6692 (abs<mode>2): Likewise.
6693 (vec_unpacks_hi_v16qi): Likewise.
6694 (vec_unpacks_hi_v8hi): Likewise.
6695 (vec_unpacks_lo_v16qi): Likewise.
6696 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6697 (vec_unpacks_lo_v8hi): Likewise.
6698 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6699 (vec_pack_trunc_v8h): Likewise.
6700 (vec_pack_trunc_v4si): Likewise.
6701 (vec_pack_trunc_<mode>): Likewise.
6703 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6705 (vec_vmaxsd): Likewise.
6706 (vec_vmaxud): Likewise.
6707 (vec_vminsd): Likewise.
6708 (vec_vminud): Likewise.
6709 (vec_vpksdss): Likewise.
6710 (vec_vpksdus): Likewise.
6711 (vec_vpkudum): Likewise.
6712 (vec_vpkudus): Likewise.
6713 (vec_vrld): Likewise.
6714 (vec_vsld): Likewise.
6715 (vec_vsrad): Likewise.
6716 (vec_vsrd): Likewise.
6717 (vec_vsubudm): Likewise.
6718 (vec_vupkhsw): Likewise.
6719 (vec_vupklsw): Likewise.
6721 2013-05-29 Jan Hubicka <jh@suse.cz>
6723 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6724 flags; reorder rest of fields in more consistent way.
6725 (varpool_node): Remove analyzed, finalized and alias.
6726 (cgraph_ndoe): Likewise.
6727 (symtab_alias_ultimate_target): New function.
6728 (cgraph_function_node): Move offline.
6729 (cgraph_reset_node): Declare.
6730 (cgraph_comdat_can_be_unshared_p): Remove.
6731 (varpool_remove_initializer): Declare.
6732 (varpool_first_defined_variable, varpool_next_defined_variable
6733 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6734 (cgraph_function_with_gimple_body_p): Update.
6735 (varpool_all_refs_explicit_p): Update.
6736 (symtab_alias_target): New function.
6737 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6738 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
6739 (cgraph_function_or_thunk_node): Simplify using
6740 symtab_alias_ultimate_target.
6741 (varpool_variable_node): Likewise.
6742 * cgraph.c (cgraph_create_function_alias): Update.
6743 (cgraph_add_thunk): Update.
6744 (cgraph_remove_node): Update.
6745 (dump_cgraph_node): Do not dump removed flags.
6746 (cgraph_function_body_availability): Update.
6747 (cgraph_propagate_frequency): Update.
6748 (verify_cgraph_node): Check sanity of local flag.
6749 (cgraph_function_node): Move here from cgraph.h; revamp for
6750 cgraph_function_or_thunk_node.
6751 * lto-symtab.c (lto_varpool_replace_node): Update.
6752 (lto_symtab_resolve_can_prevail_p): Update.
6753 (lto_symtab_merge_cgraph_nodes): Update.
6754 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6755 propagate_constants_accross_call, devirtualization_time_bonus,
6756 ipcp_propagate_stage): Update.
6757 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
6758 * ipa-inline-transform.c (clone_inlined_nodes,
6759 preserve_function_body_p): Update.
6760 * ipa-reference.c (propagate): Update.
6761 (write_node_summary_p): Update.
6762 * toplev.c (wrapup_global_declaration_2): Update.
6763 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6764 (analyze_function) ... this one.
6765 (cgraph_process_new_functions): Update.
6766 (cgraph_reset_node): Export.
6767 (cgraph_finalize_function): Update.
6768 (cgraph_add_new_function): Update.
6769 (process_function_and_variable_attributes): Update.
6770 (varpool_finalize_decl): Update.
6771 (symbol_finalized): Remove.
6772 (symbol_finalized_and_needed): Rename to ...
6773 (symbol_defined_and_needed): ... update.
6774 (cgraph_analyze_functions): Update.
6775 (handle_alias_pairs): Update.
6776 (mark_functions_to_output): Update.
6777 (assemble_thunk): Update.
6778 (output_in_order): Update.
6779 (output_weakrefs): Update.
6780 (finalize_compilation_unit): Update.
6781 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
6782 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
6783 input_node, input_varpool_node): Update.
6784 * dbxout.c (dbxout_expand_expr): Update.
6785 * cgraphclones.c (cgraph_clone_node): Update.
6786 (cgraph_copy_node_for_versioning): Update.
6787 (cgraph_materialize_clone): Update.
6788 (cgraph_materialize_all_clones): Update.
6789 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
6790 propagate_pure_const, propagate_nothrow): Update.
6791 * lto-streamer-out.c (lto_output, write_symbol): Update.
6792 * ipa-utils.c (ipa_reverse_postorder): Update.
6793 * ipa-inline.c (can_inline_edge_p): Update.
6794 (update_caller_keys, ipa_inline): Update.
6795 * dwarf2out.c (reference_to_unused,
6796 premark_types_used_by_global_vars_helper): Update.
6797 * tree-eh.c (tree_could_trap_p): Update.
6798 * ipa-split.c (consider_split, execute_split_functions): Update.
6799 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
6800 has_addr_references_p): Update; move ahead in file for better
6802 (process_references): Simplify.
6803 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
6805 (cgraph_comdat_can_be_unshared_p): Make static.
6806 (cgraph_externally_visible_p): Update.
6807 (varpool_externally_visible_p): Update.
6808 (function_and_variable_visibility): Update.
6809 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
6810 ipa_tm_mark_force_output_node): Update.
6811 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
6812 estimate_edge_devirt_benefit, inline_generate_summary,
6813 inline_write_summary): Update.
6814 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
6815 * ipa-prop.c (ipa_compute_jump_functions): Update.
6816 (ipa_print_node_params, ipa_prop_read_section,
6817 ipa_update_after_lto_read, read_replacements_section): Update.
6818 * varasm.c (mark_decl_referenced): Update.
6819 (assemble_alias, dump_tm_clone_pairs): Update.
6820 * tree-inline.c (copy_bb): Update.
6821 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
6823 * symtab.c (dump_symtab_base): Print new flags.
6824 (verify_symtab_base): Verify new flags.
6825 (symtab_alias_ultimate_target): New function.
6826 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
6827 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
6829 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
6831 * i386.c (ix86_get_function_versions_dispatcher,
6832 ix86_generate_version_dispatcher_body): Update.
6833 (fold_builtin_cpu): Use varpool_add_new_variable.
6834 * varpool.c (varpool_remove_initializer): Break out from ...
6835 (varpool_remove_node): ... this one.
6836 (dump_varpool_node, varpool_node_for_asm,
6837 cgraph_variable_initializer_availability, varpool_analyze_node,
6838 varpool_assemble_decl, varpool_remove_unreferenced_decls,
6839 varpool_finalize_named_section_flags, varpool_create_variable_alias):
6842 2013-05-29 Jan Hubicka <jh@suse.cz>
6844 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
6846 2013-05-29 Easwaran Raman <eraman@google.com>
6848 PR tree-optimization/57442
6849 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
6850 when control exits the main loop.
6852 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
6854 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
6856 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
6857 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
6858 * rx/t-rx: Add rx100 under multi library matches option for nofpu
6861 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6863 PR tree-optimization/57441
6864 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
6865 Don't limit size of incr_vec to number of candidates.
6867 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
6869 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
6870 and mips16 directories.
6871 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
6872 (MULTILIB_DIRNAMES): Ditto.
6873 (MULTILIB_EXCEPTIONS): Add new exceptions.
6874 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
6875 (MULTILIB_DIRNAMES): Ditto.
6876 (MULTILIB_EXCEPTIONS): Add new exceptions.
6878 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6879 Marcus Shawcroft <marcus.shawcroft@arm.com>
6881 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
6882 SYMBOL_TINY_ABSOLUTE.
6883 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
6884 SYMBOL_TINY_ABSOLUTE.
6885 (aarch64_expand_mov_immediate): Likewise.
6886 (aarch64_classify_symbol): Likewise.
6887 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
6888 Permit SYMBOL_TINY_ABSOLUTE.
6889 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
6891 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
6892 Marcus Shawcroft <marcus.shawcroft@arm.com>
6894 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
6895 Refactor if/switch. Replace gcc_assert with if.
6897 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
6899 * config/i386/i386.c (initial_ix86_tune_features): Enable
6900 FP Reassociation for AMD bdver1 and bdver2.
6902 2013-05-29 Martin Jambor <mjambor@suse.cz>
6904 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
6905 and IMAGPART_EXPR do not occur within other handled_components.
6907 2013-05-29 Richard Biener <rguenther@suse.de>
6909 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
6910 access on whether the use is in the BB we currently try to
6912 (vect_bb_vectorization_profitable_p): Pass the BB we currently
6913 vectorize to vect_bb_slp_scalar_cost.
6915 2013-05-29 Richard Biener <rguenther@suse.de>
6917 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
6918 computing scalar cost offsetted by stmts that are kept live
6920 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
6921 for computation of scalar cost.
6923 2013-05-28 Steve Ellcey <sellcey@mips.com>
6925 * config/mips/mips-cpus.def (mips32r2): Change processor type.
6927 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6929 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
6930 array notation built-in reduction functions.
6931 * doc/passes.texi (Passes): Added documentation about changes done
6933 * doc/invoke.texi (C Dialect Options): Added documentation about
6934 the -fcilkplus flag.
6935 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
6936 (BUILTINS_DEF): Depend on cilkplus.def.
6937 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
6938 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
6939 * cilkplus.def: New file.
6941 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
6943 PR rtl-optimization/57439
6944 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
6946 2013-05-28 Easwaran Raman <eraman@google.com>
6948 PR tree-optimization/57337
6949 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
6950 (find_insert_point): Correctly identify the insertion point
6951 when two statements with the same UID is compared.
6953 2013-05-28 Richard Biener <rguenther@suse.de>
6955 PR tree-optimization/56787
6956 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
6957 from the list of data references.
6958 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
6960 (vect_analyze_loop_operations): Likewise.
6961 (vect_transform_loop): Remove clobbers.
6963 2013-05-28 Martin Jambor <mjambor@suse.cz>
6965 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
6966 and REALPART_EXPRs have scalar type.
6968 2013-05-28 Richard Biener <rguenther@suse.de>
6970 PR tree-optimization/57411
6971 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
6973 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
6974 virtual operand propagation.
6976 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6978 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
6979 destination register for bmasksi_vis.
6980 (vector_init_bshuffle): Likewise.
6981 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
6983 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
6985 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
6986 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
6987 mode if the instruction isn't available in the original mode.
6988 * config/sparc/sparc.opt (mfix-ut699): New option.
6989 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
6990 (divdf3): Turn into expander.
6991 (divdf3_nofix): New insn.
6992 (divdf3_fix): Likewise.
6993 (divsf3): Disable if -mfix-ut699.
6994 (sqrtdf2): Turn into expander.
6995 (sqrtdf2_nofix): New insn.
6996 (sqrtdf2_fix): Likewise.
6997 (sqrtsf2): Disable if -mfix-ut699.
6999 2013-05-27 Richard Biener <rguenther@suse.de>
7002 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7003 block for the new loop.
7005 2013-05-27 Richard Biener <rguenther@suse.de>
7007 PR tree-optimization/57343
7008 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7009 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7010 (number_of_iterations_cond): Do not build the folded tree.
7012 2013-05-27 Richard Biener <rguenther@suse.de>
7016 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7017 OEP_CONSTANT_ADDRESS_OF retained.
7019 PR tree-optimization/57417
7020 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7022 (set_ssa_val_to): Compare addresses using
7023 get_addr_base_and_unit_offset.
7025 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7027 PR rtl-optimization/56833
7028 * postreload.c (move2add_record_mode): New function.
7029 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7030 (move2add_use_add2_insn): Use move2add_record_sym_value.
7031 (move2add_use_add3_insn): Likewise.
7032 (reload_cse_move2add): Use move2add_valid_value_p and
7033 move2add_record_mode. Invalidate call-clobbered and REG_INC
7034 affected regs by setting reg_mode to VOIDmode.
7035 (move2add_note_store): Don't pretend the inside of a SUBREG is
7036 the actual destination. Invalidate single/leading registers by
7037 setting reg_mode to VOIDmode.
7038 Use move2add_record_sym_value, move2add_valid_value_p and
7039 move2add_record_mode.
7041 2013-05-27 Richard Biener <rguenther@suse.de>
7043 PR tree-optimization/57396
7044 * tree-affine.c (double_int_constant_multiple_p): Properly
7045 return false for val == 0 and div != 0.
7047 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7049 * config/mips/mips.h: Use #elif in preprocessor conditions.
7051 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7054 * config/mips/constraints.md (kl): New constraint.
7055 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7056 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7057 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7059 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7061 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7064 * config/mips/mips.c (mips_can_inline_p): New function.
7065 (TARGET_CAN_INLINE_P): Define.
7067 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7069 * sched-int.h (ds_t, dw_t): Make unsigned int.
7070 Fix documentation that describes how all the ds_t bits are used.
7071 Reserve the last bit for delayed-branch scheduling.
7072 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7073 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7074 (gen_dep_weak_1): Remove prototype.
7075 * sched-deps.c (get_dep_weak_1): Make static.
7076 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7077 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7078 * doc/tm.texi: Regenerate.
7079 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7081 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7084 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7086 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7087 Sandra Loosemore <sandra@codesourcery.com>
7089 * config.gcc (powerpc-*): Allow native for with-cpu.
7091 2013-05-24 Jeff Law <law@redhat.com>
7093 PR tree-optimization/57124
7094 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7095 conversion feeding a condition if the range has an overflow
7096 if -fstrict-overflow. Add warnings for when we do make the
7099 2013-05-24 Dehao Chen <dehao@google.com>
7101 * tree-cfg.c (locus_discrim_map): Fix the typo.
7102 (locus_discrim_hasher): Likewise.
7103 (locus_discrim_hasher::hash): Likewise.
7104 (locus_discrim_hasher::equal): Likewise.
7106 2013-05-24 Martin Jambor <mjambor@suse.cz>
7108 PR tree-optimization/57294
7109 * cgraph.h (ipa_record_stmt_references): Declare.
7110 * cgraphbuild.c (ipa_record_stmt_references): New function.
7111 (build_cgraph_edges): Use ipa_record_stmt_references.
7112 (rebuild_cgraph_edges): Likewise.
7113 (cgraph_rebuild_references): Likewise.
7114 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7115 associated with the old statement and build references from the
7116 newly built statements.
7117 * ipa-ref.c (ipa_remove_stmt_references): New function.
7118 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7120 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7122 * lra-constraints.c (emit_spill_move): Use smaller mode for
7124 (check_and_process_move): Consider mem-reg moves for secondary
7126 (curr_insn_transform): Don't lose insns emitted before for
7127 secondary memory moves.
7128 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7129 reg set up in the current insn.
7131 2013-05-24 Dehao Chen <dehao@google.com>
7133 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7135 (locus_descrim_hasher::equal): Likewise.
7136 (build_gimple_cfg): New discriminator assignment algorithm.
7137 (make_edges): Likewise.
7138 (next_discriminator_for_locus): Likewise.
7139 (same_line_p): Likewise.
7140 (assign_discriminators): Likewise.
7141 (make_cond_expr_edges): Likewise.
7142 (make_gimple_switch_edges): Likewise.
7143 (make_goto_expr_edges): Likewise.
7144 (make_gimple_asm_edges): Likewise.
7146 2013-05-24 Ian Bolton <ian.bolton@arm.com>
7148 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
7149 X format specifier to only display bottom 16 bits.
7150 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
7151 immediate to match for operand 2, since it will be masked.
7153 2013-05-24 Richard Biener <rguenther@suse.de>
7155 PR tree-optimization/57287
7156 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
7157 all SSA names that occur in abnormal PHIs.
7159 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
7162 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
7163 that index is not negative.
7165 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
7167 PR rtl-optimization/55177
7168 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
7169 (simplify_byte_swapping_operation): New.
7170 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
7171 (simplify_relational_operation_1): Deal with BSWAP.
7173 2013-05-23 Richard Henderson <rth@redhat.com>
7176 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
7177 (ix86_reorg): Call it.
7179 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
7182 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
7183 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
7184 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
7186 2013-05-23 Christian Bruel <christian.bruel@st.com>
7189 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
7191 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7192 Marcus Shawcroft <marcus.shawcroft@arm.com>
7194 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
7195 * config/aarch64/constraints.md (Usa): Remove.
7196 * doc/md.texi (AArch64 Usa): Remove.
7198 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7199 Marcus Shawcroft <marcus.shawcroft@arm.com>
7201 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
7202 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
7203 * config/aarch64/predicates.md (aarch64_const_address): Remove.
7204 (aarch64_mov_operand): Use aarch64_mov_operand_p.
7206 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
7208 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
7209 instruction (AdvSIMD).
7210 * config/aarch64/aarch64-builtins.c
7211 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
7212 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
7214 2013-05-23 Martin Jambor <mjambor@suse.cz>
7217 * tree.h (contains_bitfld_component_ref_p): Declare.
7218 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
7219 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
7221 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
7222 not access a bit-field. Assert all final offsets are byte-aligned.
7224 2013-05-23 Richard Biener <rguenther@suse.de>
7226 PR tree-optimization/57380
7227 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
7228 least one invariant or re-used load.
7229 * passes.c (init_optimization_passes): Move pass_phiprop before
7232 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
7234 * config/aarch64/aarch64-simd.md
7235 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
7237 2013-05-23 Richard Biener <rguenther@suse.de>
7240 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7241 OEP_CONSTANT_ADDRESS_OF retained.
7243 2013-05-23 Jakub Jelinek <jakub@redhat.com>
7246 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
7247 don't lower unit. Handle unit not being always BITS_PER_WORD.
7249 2013-05-23 Richard Biener <rguenther@suse.de>
7251 PR rtl-optimization/57341
7252 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
7253 instead of true_dependence.
7255 2013-05-22 David Malcolm <dmalcolm@redhat.com>
7257 * bb-reorder.c (branch_threshold): Make const.
7258 (exec_threshold): Ditto.
7260 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7261 Pat Haugen <pthaugen@us.ibm.com>
7262 Peter Bergner <bergner@vnet.ibm.com>
7264 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
7265 documentation for the power8 crypto builtins.
7267 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
7269 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
7270 macros for defining power8 builtin functions.
7271 (BU_P8V_AV_2): Likewise.
7272 (BU_P8V_AV_P): Likewise.
7273 (BU_P8V_VSX_1): Likewise.
7274 (BU_P8V_OVERLOAD_1): Likewise.
7275 (BU_P8V_OVERLOAD_2): Likewise.
7276 (BU_CRYPTO_1): Likewise.
7277 (BU_CRYPTO_2): Likewise.
7278 (BU_CRYPTO_3): Likewise.
7279 (BU_CRYPTO_OVERLOAD_1): Likewise.
7280 (BU_CRYPTO_OVERLOAD_2): Likewise.
7281 (XSCVSPDP): Fix typo, point to the correct instruction.
7282 (VCIPHER): Add power8 crypto builtins.
7283 (VCIPHERLAST): Likewise.
7284 (VNCIPHER): Likewise.
7285 (VNCIPHERLAST): Likewise.
7286 (VPMSUMB): Likewise.
7287 (VPMSUMH): Likewise.
7288 (VPMSUMW): Likewise.
7289 (VPERMXOR_V2DI): Likewise.
7290 (VPERMXOR_V4SI: Likewise.
7291 (VPERMXOR_V8HI: Likewise.
7292 (VPERMXOR_V16QI: Likewise.
7293 (VSHASIGMAW): Likewise.
7294 (VSHASIGMAD): Likewise.
7296 (VPERMXOR): Likewise.
7297 (VSHASIGMA): Likewise.
7299 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7300 __CRYPTO__ if the crypto instructions are available.
7301 (altivec_overloaded_builtins): Add support for overloaded power8
7304 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
7305 support for power8 crypto builtins.
7306 (builtin_function_type): Likewise.
7307 (altivec_init_builtins): Add support for builtins that take vector
7308 long long (V2DI) arguments.
7310 * config/rs6000/crypto.md: New file, define power8 crypto
7313 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7314 Pat Haugen <pthaugen@us.ibm.com>
7315 Peter Bergner <bergner@vnet.ibm.com>
7317 * doc/invoke.texi (Option Summary): Add power8 options.
7318 (RS/6000 and PowerPC Options): Likewise.
7320 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
7321 constraints.md instead of rs6000.h. Reorder w* constraints. Add
7322 wm, wn, wr documentation.
7324 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
7325 registers if direct move instructions are enabled.
7326 (wn): New constraint for no registers.
7327 (wq): New constraint for quad word even GPR registers.
7328 (wr): New constraint if 64-bit instructions are enabled.
7329 (wv): New constraint if power8 vector instructions are enabled.
7330 (wQ): New constraint for quad word memory locations.
7332 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
7333 constraint for 0..15 for crypto instructions.
7334 (gpc_reg_operand): If VSX allow registers in VSX registers as well
7335 as GPR and floating point registers.
7336 (int_reg_operand): New predicate to match only GPR registers.
7337 (base_reg_operand): New predicate to match base registers.
7338 (quad_int_reg_operand): New predicate to match even GPR registers
7339 for quad memory operations.
7340 (vsx_reg_or_cint_operand): New predicate to allow vector logical
7341 operations in both GPR and VSX registers.
7342 (quad_memory_operand): New predicate for quad memory operations.
7343 (reg_or_indexed_operand): New predicate for direct move support.
7345 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
7346 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
7347 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
7348 (POWERPC_MASKS): Add power8 options.
7349 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
7352 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7353 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
7355 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
7356 (-mpower8-fusion): New power8 options.
7357 (-mpower8-fusion-sign): Likewise.
7358 (-mpower8-vector): Likewise.
7359 (-mcrypto): Likewise.
7360 (-mdirect-move): Likewise.
7361 (-mquad-memory): Likewise.
7363 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
7365 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
7367 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
7368 (rs6000_debug_vector_unit): Add p8_vector.
7369 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
7370 definitions. Also print fusion state.
7371 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
7372 (rs6000_builtin_mask_calculate): Add power8 builtin support.
7373 (rs6000_option_override_internal): Add support for power8.
7374 (rs6000_common_init_builtins): Add debugging for skipped builtins
7376 (rs6000_adjust_cost): Add power8 support.
7377 (rs6000_issue_rate): Likewise.
7378 (insn_must_be_first_in_group): Likewise.
7379 (insn_must_be_last_in_group): Likewise.
7380 (force_new_group): Likewise.
7381 (rs6000_register_move_cost): Likewise.
7382 (rs6000_opt_masks): Likewise.
7384 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
7385 power8 capable assembler, default to power7 options.
7386 (TARGET_DIRECT_MOVE): Likewise.
7387 (TARGET_CRYPTO): Likewise.
7388 (TARGET_P8_VECTOR): Likewise.
7389 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
7390 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
7391 (VECTOR_MEM_P8_VECTOR_P): Likewise.
7392 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
7393 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
7394 (TARGET_XSCVDPSPN): Likewise.
7395 (TARGET_XSCVSPDPN): Likewsie.
7396 (TARGET_SYNC_HI_QI): Likewise.
7397 (TARGET_SYNC_TI): Likewise.
7398 (MASK_CRYPTO): Likewise.
7399 (MASK_DIRECT_MOVE): Likewise.
7400 (MASK_P8_FUSION): Likewise.
7401 (MASK_P8_VECTOR): Likewise.
7402 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7403 temporary used by some of the direct move instructions to get two FP
7404 temporary registers does not force creation of a stack frame.
7405 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7406 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7407 that any VSX registers are tieable, even if they are also an
7408 Altivec vector mode.
7409 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7410 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7411 (RS6000_BTM_CRYPTO): Likewise.
7412 (RS6000_BTM_COMMON): Likewise.
7414 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7415 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7416 (enum rs6000_vector): Add power8 vector support.
7418 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7422 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7423 (any_sibcall_could_use_r3): this and handle indirect calls.
7424 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7426 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7428 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7430 2013-05-22 Richard Biener <rguenther@suse.de>
7433 * profile.c (branch_prob): Do not split blocks that are
7434 abnormally receiving from ECF_RETURNS_TWICE functions.
7436 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7438 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7439 address mode. Move pointer mode initialization to the same place.
7441 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7443 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7444 while it has any effect.
7446 2013-05-21 Easwaran Raman <eraman@google.com>
7448 PR tree-optimization/57322
7449 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7450 UID of the statement added to the BB to be 1.
7452 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7454 PR tree-optimization/57331
7455 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7456 of conversion from pointer type to integral type with integer.
7458 2013-05-21 Martin Jambor <mjambor@suse.cz>
7461 * ipa-prop.c (ipa_read_node_info): Process param_used and
7462 controlled_uses in the same order as when writing.
7464 2013-05-21 Magnus Granberg <baldrick@free.fr>
7467 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7469 2013-05-21 Richard Biener <rguenther@suse.de>
7471 PR tree-optimization/57318
7472 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7473 estimate stmts with side-effects as likely eliminated.
7475 2013-05-21 Richard Biener <rguenther@suse.de>
7477 PR tree-optimization/57330
7478 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7479 preserve the call stmts fntype.
7481 2013-05-21 Richard Biener <rguenther@suse.de>
7483 PR tree-optimization/57303
7484 * tree-ssa-sink.c (statement_sink_location): Improve killing
7485 stmt detection and properly handle self-assignments.
7487 2013-05-21 Christian Bruel <christian.bruel@st.com>
7489 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7490 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7491 registers. Set register size out of the PARALLEL loop.
7493 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7496 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7497 (*fmasf4, *fmasf4_media): New insns.
7499 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7501 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7502 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7503 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7504 (mips_idiv_insns): Update the comments to say that the returned
7505 instruction counts are in units of BASE_INSN_LENGTH.
7506 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7507 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7508 using 2 rather than 4 as the length of indirect MIPS16 and
7509 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7510 length of a NOP. Don't divide MIPS16 lengths by 2.
7511 (mips16_split_long_branches): Assume a branch is long if the
7512 length is greater than 4 rather than 8.
7513 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7514 rather than multiplying them by 2. Multiply instruction counts
7515 by BASE_INSN_LENGTH rather than 4.
7516 (*jump_mips16, tls_get_tp_mips16_<mode>)
7517 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7519 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7521 * config/mips/mips.md (extended_mips16): Remove branch case.
7522 (length): Remove duplicated extended_mips16 test.
7524 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7526 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7528 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7530 * recog.h (Recog_data): Rename to...
7531 (recog_data_d): ...this.
7532 (recog_data): Update accordingly.
7533 * recog.c (recog_data): Likewise.
7534 * reload.c (save_recog_data): Likewise.
7535 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7536 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7538 2013-05-17 Julian Brown <julian@codesourcery.com>
7540 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7541 found in a REG_EQUAL note, invalidate it.
7543 2013-05-17 Easwaran Raman <eraman@google.com>
7545 * tree-ssa-reassoc.c (find_insert_point): New function.
7546 (insert_stmt_after): Likewise.
7547 (get_def_stmt): Likewise.
7548 (ensure_ops_are_available): Likewise.
7549 (not_dominated_by): Likewise.
7550 (rewrite_expr_tree): Do not move statements beyond what is
7551 necessary. Remove call to swap_ops_for_binary_stmt...
7552 (reassociate_bb): ... and move it here.
7553 (build_and_add_sum): Assign UIDs for new statements.
7554 (linearize_expr): Likewise.
7555 (do_reassoc): Renumber gimple statement UIDs.
7557 2013-05-17 Jan Hubicka <jh@suse.cz>
7559 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7561 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
7562 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7564 (output_weakrefs): Update.
7566 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7567 Martin Jambor <mjambor@suse.cz>
7570 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7571 value that corresponds to the given aggval is found in values vector.
7573 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
7575 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7576 sse, sse2, sse3, ssse3 and sse4a flags to options.
7578 2013-05-17 David Malcolm <dmalcolm@redhat.com>
7580 * gengtype-state.c: (s_expr_writer): New class, to handle
7581 prettifying of output layout of s-expressions.
7582 (state_writer): New class, to write out gtype.state.
7583 (state_written_type_count): Move this variable into member data of
7585 (s_expr_writer::s_expr_writer): New code: constructor for new class
7586 (state_writer::state_writer(): ditto
7587 (s_expr_writer::write_new_line): New function
7588 (s_expr_writer::write_any_indent): ditto
7589 (s_expr_writer::begin_s_expr): ditto
7590 (s_expr_writer::end_s_expr): ditto
7591 (write_state_fileloc): convert to method of state_writer...
7592 (state_writer:: write_state_fileloc): ...and use methods of
7593 s_expr_writer to write indentation into the gtype.state output file
7594 to visually represent the hierarchical structure of the list
7596 (write_state_fields): ditto, renaming to...
7597 (state_writer::write_state_fields)
7598 (write_state_a_string): ditto, renaming to...
7599 (state_writer::write_state_a_string)
7600 (write_state_string_option): ditto, renaming to...
7601 (state_writer::write_state_string_option)
7602 (write_state_type_option): ditto, renaming to...
7603 (state_writer::write_state_type_option)
7604 (write_state_nested_option): ditto, renaming to...
7605 (state_writer::write_state_nested_option)
7606 (write_state_option): ditto, renaming to...
7607 (state_writer::write_state_option)
7608 (write_state_options): ditto, renaming to...
7609 (state_writer::write_state_options)
7610 (write_state_lang_bitmap): ditto, renaming to...
7611 (state_writer::write_state_lang_bitmap)
7612 (write_state_version): ditto, renaming to...
7613 (state_writer::write_state_version)
7614 (write_state_scalar_type): ditto, renaming to...
7615 (state_writer::write_state_scalar_type)
7616 (write_state_string_type): ditto, renaming to...
7617 (state_writer::write_state_string_type)
7618 (write_state_undefined_type): ditto, renaming to...
7619 (state_writer::write_state_undefined_type)
7620 (write_state_struct_union_type): ditto, renaming to...
7621 (state_writer::write_state_struct_union_type)
7622 (write_state_struct_type): ditto, renaming to...
7623 (state_writer::write_state_struct_type)
7624 (write_state_user_struct_type): ditto, renaming to...
7625 (state_writer::write_state_user_struct_type)
7626 (write_state_lang_struct_type): ditto, renaming to...
7627 (state_writer::write_state_lang_struct_type)
7628 (write_state_param_struct_type): ditto, renaming to...
7629 (state_writer::write_state_param_struct_type)
7630 (write_state_pointer_type): ditto, renaming to...
7631 (state_writer::write_state_pointer_type)
7632 (write_state_array_type): ditto, renaming to...
7633 (state_writer::write_state_array_type)
7634 (write_state_gc_used): ditto, renaming to...
7635 (state_writer::write_state_gc_used)
7636 (write_state_common_type_content): ditto, renaming to...
7637 (state_writer::write_state_common_type_content)
7638 (write_state_type): ditto, renaming to...
7639 (state_writer::write_state_type)
7640 (write_state_pair_list): ditto, renaming to...
7641 (state_writer::write_state_pair_list)
7642 (write_state_pair): ditto, renaming to...
7643 (state_writer::write_state_pair)
7644 (write_state_typedefs): ditto, renaming to...
7645 (state_writer::write_state_typedefs)
7646 (write_state_structures): ditto, renaming to...
7647 (state_writer::write_state_structures)
7648 (write_state_param_structs): ditto, renaming to...
7649 (state_writer::write_state_param_structs)
7650 (write_state_variables): ditto, renaming to...
7651 (state_writer::write_state_variables)
7652 (write_state_srcdir): ditto, renaming to...
7653 (state_writer::write_state_srcdir)
7654 (write_state_files_list): ditto, renaming to...
7655 (state_writer::write_state_files_list)
7656 (write_state_languages): ditto, renaming to...
7657 (state_writer::write_state_languages)
7658 (write_state): create a state_writer instance and use it when
7659 writing out the state file
7661 2013-05-17 Mike Stump <mikestump@comcast.net>
7663 PR rtl-optimization/57304
7664 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7665 accessing DF_REF_REAL_LOC.
7667 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7669 PR rtl-optimization/57281
7670 PR rtl-optimization/57300
7671 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7672 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7673 what the other splitter did if the registers are dead.
7675 2013-05-17 Richard Biener <rguenther@suse.de>
7677 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7680 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7682 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7685 2013-05-17 Marek Polacek <polacek@redhat.com>
7687 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7688 length when doing non-zero store of storing '\0' to '\0'.
7690 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7692 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7693 vect_external_def oprnd1 with loop_vinfo, try to emit
7694 optional cast, negation and and stmts on the loop preheader
7695 edge instead of into the pattern def seq.
7697 PR tree-optimization/57051
7698 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7699 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7701 2013-05-16 Nick Clifton <nickc@redhat.com>
7703 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7704 (rl78_is_naked_func): New function.
7705 (rl78_expand_prologue): Skip prologue generation for naked functions.
7706 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
7707 * doc/extend.texi (naked): Add RL78 to the list of processors
7708 that supports this attribute.
7710 2013-05-16 Jeff Law <law@redhat.com>
7712 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7714 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7716 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7717 cache parameters using detect_caches_amd also for CYRIX,
7718 NSC and TM2 signatures.
7720 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7721 Dzianis Kahanovich <mahatma@eu.by>
7725 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7726 VIA/Centaur processors and determine their cache parameters
7727 using detect_caches_amd.
7729 2013-05-16 Teresa Johnson <tejohnson@google.com>
7731 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7732 (rtl_verify_edges): New function.
7733 (rtl_verify_bb_insns): Ditto.
7734 (rtl_verify_bb_pointers): Ditto.
7735 (rtl_verify_bb_insn_chain): Ditto.
7736 (rtl_verify_fallthru): Ditto.
7737 (rtl_verify_bb_layout): Ditto.
7738 (rtl_verify_flow_info_1): Outline checks into new functions.
7739 (rtl_verify_flow_info): Ditto.
7741 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
7743 * cfghooks.c (copy_bbs): Add update_dominance argument.
7744 * cfghooks.h (copy_bbs): Update prototype.
7745 * tree-cfg.c (gimple_duplicate_sese_region):
7746 Add update_dominance argument.
7747 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7748 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7749 gimple_duplicate_sese_region call.
7750 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7751 Update copy_bbs call.
7752 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7753 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7755 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7757 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7758 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7759 vect_recog_rotate_pattern.
7760 (vect_recog_rotate_pattern): New function.
7762 2013-05-16 Jason Merrill <jason@redhat.com>
7764 * Makefile.in (LLINKER): New variable.
7765 (mostlyclean): Remove link mutex.
7766 * configure.ac: Handle --enable-link-mutex.
7767 * lock-and-run.sh: New script.
7769 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7772 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
7775 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7777 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
7779 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
7781 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
7782 * config/arm/arm.c (next_consecutive_mem): New function.
7783 (gen_movmem_ldrd_strd): Likewise.
7784 * config/arm/arm.md (movmemqi): Update condition and code.
7785 (unaligned_loaddi, unaligned_storedi): New patterns.
7787 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7789 * config.gcc: Obsolete *-*-solaris2.9*.
7790 * doc/install.texi (Specific, *-*-solaris2*): Document it.
7792 2013-05-16 Richard Biener <rguenther@suse.de>
7794 * passes.c (init_optimization_passes): Move pass_parallelize_loops
7795 earlier, after GRAPHITE transforms and IV canonicalization.
7797 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7799 * omp-low.c (extract_omp_for_data): For collapsed loops,
7800 if at least one of the loops is known at compile time to
7801 iterate zero times, set count to 0.
7802 (expand_omp_regimplify_p): New function.
7803 (expand_omp_for_generic): For collapsed loops, if at least
7804 one of the loops isn't known to iterate at least once,
7805 add runtime check with setting count to 0.
7806 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
7807 For unsigned types if it isn't known at compile time that
7808 the loop will iterate at least once, add runtime check to bypass
7809 the whole loop if initial condition isn't true.
7811 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
7813 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
7815 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
7818 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
7819 transformations to avoid an infinite loop.
7821 2013-05-16 Marek Polacek <polacek@redhat.com>
7823 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
7825 2013-05-15 Leif Ekblad <leif@rdos.net>
7827 * config/i386/i386.c (ix86_decompose_address): Use
7828 DEFAULT_TLS_SEG_REG to access TLS segment register.
7829 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
7830 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
7831 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
7833 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
7836 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
7837 sibling calls to functions that would normally be lazily bound,
7838 unless $gp is call-clobbered.
7840 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
7842 * config/i386/i386.c (ix86_option_override_internal): Update
7843 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
7844 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
7845 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
7846 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
7848 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
7850 2013-05-15 Andreas Schwab <schwab@suse.de>
7852 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
7853 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
7856 2013-05-15 Teresa Johnson <tejohnson@google.com>
7858 * loop-unroll.c (report_unroll_peel): Check decision before
7859 emitting unroll/peel message.
7861 2013-05-15 Teresa Johnson <tejohnson@google.com>
7863 * function.h (has_bb_partition): New rtl_data flag.
7864 (bb_reorder_complete): Ditto.
7865 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
7866 instead of flag_reorder_blocks_and_partition.
7867 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
7868 with some enhancements.
7869 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
7870 * bb-reorder.c (connect_traces): Check for has_bb_partition
7871 instead of flag_reorder_blocks_and_partition.
7872 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
7873 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
7874 verify_hot_cold_block_grouping.
7875 (partition_hot_cold_basic_blocks): Set has_bb_partition.
7877 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7880 * config/arm/predicates.md (call_insn_operand): New predicate.
7881 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
7882 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
7883 if insn is not a tail call.
7884 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
7886 * config/arm/arm.h (enum reg_class): New caller save register class.
7887 (REG_CLASS_NAMES): Likewise.
7888 (REG_CLASS_CONTENTS): Likewise.
7889 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
7892 2013-05-15 Richard Biener <rguenther@suse.de>
7894 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
7895 of MSG_OPTIMIZED_LOCATIONS.
7896 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
7897 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
7899 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
7900 of MSG_OPTIMIZED_LOCATIONS.
7901 (execute_vect_slp): Likewise.
7902 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
7903 (vect_create_cond_for_alias_checks): Likewise.
7904 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
7905 (vect_recog_widen_mult_pattern): Likewise.
7906 (vect_recog_widen_sum_pattern): Likewise.
7907 (vect_recog_over_widening_pattern): Likewise.
7908 (vect_recog_widen_shift_pattern): Likewise.
7909 (vect_recog_vector_vector_shift_pattern): Likewise.
7910 (vect_recog_divmod_pattern): Likewise.
7911 (vect_recog_mixed_size_cond_pattern): Likewise.
7912 (vect_recog_bool_pattern): Likewise.
7913 (vect_pattern_recog_1): Likewise.
7915 2013-05-15 Martin Jambor <mjambor@suse.cz>
7917 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
7918 non-functions to builtin_unreachable.
7919 * ipa-inline-transform.c (inline_call): Do not assert estimates were
7920 correct when new direct edges were discovered.
7922 2013-05-15 Martin Jambor <mjambor@suse.cz>
7924 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
7925 header, print symbol order instead of node uid, print more information
7926 about indirect edge targets.
7927 (ipa_make_edge_direct_to_target): Print symbol order instead of node
7929 (ipa_make_edge_direct_to_target): Likewise.
7930 (remove_described_reference): Likewise.
7931 (propagate_controlled_uses): Likewise.
7932 (ipa_print_node_params): Also print symbol order.
7933 (ipcp_transform_function): Print symbol order instead of node uids.
7934 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
7935 (cgraph_get_create_real_symbol_node): Likewise.
7936 * ipa-cp.c (print_lattice): Likewise.
7937 (print_all_lattices): Likewise.
7938 (determine_versionability): Likewise.
7939 (initialize_node_lattices): Likewise.
7940 (estimate_local_effects): Likewise.
7941 (update_profiling_info): Likewise.
7942 (create_specialized_node): Likewise.
7943 (perhaps_add_new_callers): Likewise.
7944 (decide_about_value): Likewise.
7945 (decide_whether_version_node): Likewise.
7946 (identify_dead_nodes): Likewise.
7947 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
7948 (dump_inline_summary): Likewise.
7949 (estimate_node_size_and_time): Likewise.
7950 (inline_analyze_function): Likewise.
7951 * ipa-inline.c (report_inline_failed_reason): Likewise.
7952 (want_early_inline_function_p): Likewise.
7953 (edge_badness): Likewise.
7954 (update_edge_key): Likewise.
7955 (inline_small_functions): Likewise. Add dumping of order to two other
7957 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
7958 instead of node uids.
7959 (propagate_pure_const): Likewise.
7960 (propagate_pure_const): Likewise.
7961 * ipa-utils.c (dump_cgraph_node_set): Likewise.
7962 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
7963 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
7965 * tree-pretty-print.c (dump_function_header): Likewise.
7966 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
7967 Print symbol order instead of node uids.
7969 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7971 * config/s390/s390.c (s390_register_move_cost): Don't impose the
7972 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
7974 2013-05-15 Richard Biener <rguenther@suse.de>
7976 PR tree-optimization/57275
7977 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
7978 return value for fail to do runtime alias checks for gather loads.
7980 2013-05-15 Jan Hubicka <jh@suse.cz>
7984 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
7985 weakrefs are not external.
7986 (lto_symtab_merge_decls): Fix thinko when dealing with
7987 non-lto_symtab decls.
7988 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
7989 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
7990 * varpool.c (dump_varpool_node): Dump more flags.
7992 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7994 * config/i386/i386.c (processor_alias_table): Add instruction
7995 FSGSBASE for AMD bdver3 architecture.
7997 2013-05-14 Jakub Jelinek <jakub@redhat.com>
7999 * tree.c (warn_deprecated_use): Print file:line using locus color.
8000 * diagnostic.c (diagnostic_report_current_module): Print file:line
8001 and file:line:column using locus color.
8003 2013-05-14 Mike Stump <mikestump@comcast.net>
8005 * gdbinit.in: Add __null.
8007 2013-05-14 Mike Stump <mikestump@comcast.net>
8009 * recog.h: Rename struct recog_data to Recog_data.
8010 * recog.c: Likewise.
8011 * reload.c (can_reload_into): Likewise.
8012 * config/picochip/picochip.c: Likewise.
8014 2013-05-14 Mike Stump <mikestump@comcast.net>
8016 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8018 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8020 * resource.h (struct resources): Remove unch_memory member.
8021 (CLEAR_RESOURCE): Don't clear unch_memory.
8022 * resource.c (mark_referenced_resources): Don't set it.
8023 (mark_set_resources): Likewise.
8024 (mark_target_live_regs): Don't clear it.
8025 (init_resource_info): Likewise.
8026 * reorg.c (resource_conflicts_p): Don't compare it.
8027 (redundant_insn): Don't set it.
8029 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8031 * emit-rtl.c (next_label): Remove unused function.
8032 (skip_consecutive_labels, link_cc0_insns): Move to ...
8033 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8034 only place where these functions are used, and make them static.
8036 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8038 * fold-const.c (fold_negate_expr): Handle vectors.
8039 (fold_truth_not_expr): Make it static.
8040 (fold_invert_truthvalue): New static function.
8041 (invert_truthvalue_loc): Handle vectors. Do not call
8042 fold_truth_not_expr directly.
8043 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8044 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8045 (fold_comparison): Handle vector constants.
8046 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8047 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8048 * tree.h (fold_truth_not_expr): Remove declaration.
8050 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8052 * config/aarch64/aarch64-simd.md
8053 (aarch64_vcond_internal<mode>): Rename to...
8054 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8055 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8056 float modes. Clarify all iterator modes.
8057 (vcond<mode><mode>): Use new name for vcond expanders.
8058 (vcond<v_cmp_result><mode>): Likewise.
8059 (vcondu<mode><mode>: Likewise.
8060 * config/aarch64/iterators.md (VDQF_COND): New.
8062 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8065 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8066 variable for the shift amount. Check that we shift by non-negative
8069 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8072 * config/arm/arm.h (EPILOGUE_USES): Only return true
8073 for LR_REGNUM after epilogue_completed.
8075 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8077 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8080 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8083 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8084 and Solaris 11+/x86 with gld.
8085 * configure: Regenerate.
8087 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8089 * expmed.c (expand_shift_1): Canonicalize rotates by
8090 constant bitsize / 2 to bitsize - 1.
8091 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8092 case ROTATERT>: Likewise.
8095 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8097 * config/i386/i386.md (rotateinv): New code attr.
8098 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8099 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8100 roll $31, %eax, etc.
8102 2013-05-14 Richard Biener <rguenther@suse.de>
8105 * tree-eh.c (sink_clobbers): Give up for successors with
8106 multiple predecessors and no virtual uses.
8108 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8110 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8111 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8113 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8116 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8117 the case when both op0 and op1 have VOIDmode.
8119 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8121 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8122 in multiply-accumulate mode.
8124 2013-05-13 Guozhi Wei <carrot@google.com>
8126 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8128 2013-05-13 Kai Tietz <ktietz@redhat.com>
8131 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8132 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8133 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8134 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8135 (ix86_expand_prologue): Likewise.
8136 (ix86_expand_split_stack_prologue): Likewise.
8137 (legitimate_pic_address_disp_p): Likewise.
8138 (legitimize_pic_address): Likewise.
8139 (legitimize_tls_address): Likewise.
8140 (legitimize_pe_coff_symbol): Likewise.
8141 (output_pic_addr_const): Likewise.
8142 (construct_plt_address): Likewise.
8143 (ix86_expand_call): Likewise.
8144 (x86_output_mi_thunk): Likewise.
8145 (x86_function_profiler): Likewise.
8147 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
8149 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
8150 similar switch cases.
8151 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
8152 (aarch64_simd_mov_to_<mode>low): Delete.
8153 (aarch64_simd_mov_to_<mode>high): Delete.
8154 (move_lo_quad_<mode>): Add w<-r alternative.
8155 (aarch64_simd_move_hi_quad_<mode>): Likewise.
8156 (aarch64_simd_mov_from_*): Update type attribute.
8157 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
8160 2013-05-13 Jan Hubicka <jh@suse.cz>
8162 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
8163 * config/i386/i386.c (ix86_compute_frame_layout,
8164 ix86_expand_epilogue, emit_i387_cw_initialization,
8165 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
8166 ix86_local_alignment): Fix use of size/speed predicates.
8168 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8170 PR tree-optimization/45216
8171 PR tree-optimization/57157
8172 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
8173 the (-Y) & (B - 1) variant if OP is |.
8174 * expmed.c (expand_shift_1): For rotations by const0_rtx just
8175 return shifted. Use (-op1) & (prec - 1) as other_amount
8176 instead of prec - op1.
8178 2013-05-13 Martin Jambor <mjambor@suse.cz>
8181 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
8182 (ipa_constant_data): New type.
8183 (ipa_jump_func): Use ipa_constant_data to hold information about
8184 constant jump functions.
8185 (ipa_get_jf_constant): Adjust to jump function type changes.
8186 (ipa_get_jf_constant_rdesc): New function.
8187 (ipa_param_descriptor): New field controlled_uses.
8188 (ipa_get_controlled_uses): New function.
8189 (ipa_set_controlled_uses): Likewise.
8190 * ipa-ref.h (ipa_find_reference): Declare.
8191 * ipa-prop.c (ipa_cst_ref_desc): New type.
8192 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
8194 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
8195 New parameter cs. Adjust all callers.
8196 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
8197 (remove_described_reference): New function.
8198 (jfunc_rdesc_usable): Likewise.
8199 (try_make_edge_direct_simple_call): Decrement controlled use count,
8200 attempt to remove reference if it hits zero.
8201 (combine_controlled_uses_counters): New function.
8202 (propagate_controlled_uses): Likewise.
8203 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
8204 (ipa_edge_duplication_hook): Duplicate reference descriptions.
8205 (ipa_print_node_params): Print described use counter.
8206 (ipa_write_jump_function): Adjust to jump function type changes.
8207 (ipa_read_jump_function): New parameter CS, pass it to
8208 ipa_set_jf_constant. Adjust caller.
8209 (ipa_write_node_info): Stream controlled use count
8210 (ipa_read_node_info): Likewise.
8211 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
8213 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
8214 count. Remove cloning-added reference if it reaches zero.
8215 * ipa-ref.c (ipa_find_reference): New function.
8217 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8219 * config/i386/i386.c (processor_target_table): Modified default
8220 alignment values for AMD BD and BT architectures.
8222 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
8224 * tree-vect-generic.c (uniform_vector_p): Move ...
8225 * tree.c (uniform_vector_p): ... here.
8226 * tree.h (uniform_vector_p): Declare it.
8227 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
8230 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8232 PR tree-optimization/57230
8233 * tree-ssa-strlen.c (handle_char_store): Record length for
8234 array store from STRING_CST.
8236 PR tree-optimization/57230
8237 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
8240 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
8242 * config/epiphany/epiphany.c (epiphany_init): Check size of
8243 NUM_MODES_FOR_MODE_SWITCHING.
8244 (epiphany_expand_prologue):
8245 Remove CONFIG_REGNUM initial value handling code.
8246 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
8247 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
8248 (emit_set_fp_mode, epiphany_mode_after): Likewise.
8249 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
8250 Don't return 1 for FP_MODE_NONE.
8251 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
8252 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
8253 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
8254 * config/epiphany/epiphany.md (save_config): New pattern.
8256 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
8258 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
8260 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
8262 * config/i386/i386.md (memory): Handle sseishft1.
8263 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
8264 (*vec_extractv2di_1): Ditto.
8266 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
8268 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
8271 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
8273 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
8274 Add mthumb/march=armv7-a multilib.
8275 Add mthumb/march=armv7-r multilib.
8276 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
8278 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
8280 * config/v850/t-rtems: Add more multilibs.
8282 2013-05-10 Richard Biener <rguenther@suse.de>
8284 PR tree-optimization/57214
8285 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
8286 not propagate from SSA names that occur in abnormal PHI nodes.
8288 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
8290 * stor-layout.c (element_precision): New function.
8291 * machmode.h (element_precision): Declare it.
8292 * tree.c (build_minus_one_cst): New function.
8293 (element_precision): Likewise.
8294 * tree.h (build_minus_one_cst): Declare new function.
8295 (element_precision): Likewise.
8296 * fold-const.c (operand_equal_p): Use element_precision.
8297 (fold_binary_loc): Handle vector types.
8298 * convert.c (convert_to_integer): Use element_precision.
8299 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
8302 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
8304 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
8305 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
8306 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
8307 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8308 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
8309 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
8310 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
8311 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8312 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
8313 (Uuw8): New constraints.
8314 (Usb4): Move into alphabetical order.
8315 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
8316 (sd8_operand, ub8_operand, uw8_operand): New predicates.
8317 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
8318 previously unnamed patterns.
8319 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
8320 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
8321 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
8322 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
8323 of set_attr_alternative/if_then_else. Use extended_mips16 instead
8324 of specific lengths.
8326 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8328 * config/i386/i386.md (rotateinv): New code attr.
8329 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8330 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8331 roll $31, %eax, etc.
8333 PR tree-optimization/45216
8334 PR tree-optimization/57157
8335 * tree-ssa-forwprop.c (simplify_rotate): New function.
8336 (ssa_forward_propagate_and_combine): Call it.
8338 2013-05-10 Richard Biener <rguenther@suse.de>
8340 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
8341 disable peeling when we version for aliasing.
8342 (vector_alignment_reachable_p): Honor explicit user alignment.
8343 (vect_supportable_dr_alignment): Likewise.
8344 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
8345 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
8346 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
8347 then peeling to arrange for the cost-model check to come first.
8349 2013-05-10 Alan Modra <amodra@gmail.com>
8351 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
8352 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
8353 * configure: Regenerate.
8355 2013-05-10 Alan Modra <amodra@gmail.com>
8358 * varasm.c (default_elf_select_section): Move !DECL_P check..
8359 (get_named_section): ..to here before calling get_section_name.
8361 (default_section_type_flags): Add DECL_P check.
8362 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
8363 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
8365 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
8367 * config/epiphany/epiphany.c (epiphany_expand_prologue):
8368 When using gen_stack_adjust_str with a register offset, add a
8369 REG_FRAME_RELATED_EXPR note.
8371 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
8373 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
8374 (*vec_extractv4si_zext_mem): Ditto.
8375 (*vec_extractv2di): Add 0->x and x->x alternatives.
8376 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
8377 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
8379 2013-05-09 Jason Merrill <jason@redhat.com>
8381 N3639 C++1y VLA support
8382 * gimplify.c (gimplify_vla_decl): Don't touch an existing
8385 * tree.c (build_constructor_va): New.
8386 * tree.h: Declare it.
8388 2013-05-09 Martin Jambor <mjambor@suse.cz>
8391 * gimple-fold.c (canonicalize_constructor_val): Call
8392 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
8394 2013-05-09 Jan Hubicka <jh@suse.cz>
8395 Richard Biener <rguenther@suse.de>
8398 * symtab.c (symtab_make_decl_local): Do not add private names.
8400 2013-05-09 Jan Hubicka <jh@suse.cz>
8403 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8404 (unlink_from_assembler_name_hash): Likewise.
8405 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8406 symtab_unregister_node, symtab_initialize_asm_name_hash,
8407 change_decl_assembler_name): Update.
8409 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8411 * config/aarch64/aarch64.md: New movtf split.
8412 (*movtf_aarch64): Update.
8413 (aarch64_movdi_tilow): Handle TF modes and rename to
8414 aarch64_movdi_<mode>low.
8415 (aarch64_movdi_tihigh): Handle TF modes and rename to
8416 aarch64_movdi_<mode>high
8417 (aarch64_movtihigh_di): Handle TF modes and rename to
8418 aarch64_mov<mode>high_di
8419 (aarch64_movtilow_di): Handle TF modes and rename to
8420 aarch64_mov<mode>low_di
8421 (aarch64_movtilow_tilow): Remove spurious whitespace.
8422 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8424 (aarch64_print_operand): Update.
8426 2013-05-09 Alan Modra <amodra@gmail.com>
8428 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8430 * configure: Regenerate.
8432 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8434 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8435 splitter preparation statements.
8436 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8437 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8440 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8442 * gimple-ssa-strength-reduction.c (count_candidates): Change
8443 return value to int.
8444 (analyze_candidates_and_replace): Change type of length to int.
8446 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8448 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8449 (*vec_extract<mode>): Use VI12_128 mode iterator.
8450 (*vec_extract<mode>_mem): Ditto.
8451 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8454 2013-05-08 Diego Novillo <dnovillo@google.com>
8459 2012-08-17 Diego Novillo <dnovillo@google.com>
8462 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8463 * config.in: Regenerate.
8464 * configure: Regenerate.
8465 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8467 2013-05-08 Jan Hubicka <jh@suse.cz>
8470 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8471 * cgraph.h (symtab_node_base): Add unique_name.
8472 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8473 input_overwrite_node, input_varpool_node): Stream unique_name.
8474 * cgraphclones.c (cgraph_create_virtual_clone,
8475 cgraph_function_versioning): Set unique_name.
8476 * ipa.c (function_and_variable_visibility): Set unique_name.
8478 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8480 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8481 (alloc_cand_and_find_basis): Restrict conditional candidate
8482 processing to CAND_MULTs.
8484 2013-05-08 Jan Hubicka <jh@suse.cz>
8487 lto-symtab.c (lto_symtab_symbol_p): New function.
8488 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8489 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8490 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8491 Skip static symbols.
8493 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8495 PR tree-optimization/57200
8496 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8497 Only call inform if the preceding warning_at returns true.
8499 2013-05-07 Han Shen <shenhan@google.com>
8501 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8502 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8503 * common.opt (fstack-protector-strong): New option.
8504 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8505 * doc/invoke.texi (Optimization Options): Document
8506 "-fstack-protector-strong".
8507 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8509 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8511 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8513 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8515 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8516 it is not lhs of the insn.
8517 (setup_reg_equiv): Remove insn having equiv memory which it is not
8519 * lra-constraints.c (process_address): Try to improve generation
8520 code for address base + disp.
8521 (lra_constraints): Make correct the code for checking insn setting
8522 up backward equivalence. Remove insn only if it is in the init
8524 * lra-eliminations.c (update_reg_eliminate): Change return value.
8525 (lra_eliminate): Use the result.
8527 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8529 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8530 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8531 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8532 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8533 register target operands.
8534 (*vec_extractv8hi_sse2): New pattern.
8535 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8536 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8537 (*vec_extract<mode>_mem): New insn and split pattern.
8539 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8541 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8542 (TARGET_ASAN_SHADOW_OFFSET): Define.
8543 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8544 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8546 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8548 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8549 (incr_vec_index): Return -1 if increment not found.
8550 (create_add_on_incoming_edge): Assert if increment not found.
8551 (record_increment): Limit number of increments recorded.
8552 (all_phi_incrs_profitable): Return false if an increment not found.
8553 (replace_profitable_candidates): Don't process increments that were
8555 (analyze_candidates_and_replace): Limit size of incr_vec.
8557 2013-05-07 Richard Biener <rguenther@suse.de>
8559 * calls.c (special_function_p): setjmp-like functions are leaf.
8560 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8561 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8563 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8565 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8566 (aarch64_simd_mov<mode>): New expander.
8567 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8568 (aarch64_simd_mov_to_<mode>high): Likewise.
8569 (aarch64_simd_mov_from_<mode>low): Likewise.
8570 (aarch64_simd_mov_from_<mode>high): Likewise.
8571 (aarch64_dup_lane<mode>): Update.
8572 (aarch64_dup_lanedi): New instruction pattern.
8573 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8574 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8576 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8578 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8579 (replace_mult_candidate): Remove unnecessary argument; remove
8580 unnecessary parameter from call to introduce_cast_before_cand.
8581 (replace_unconditional_candidate): Remove unnecessary parameter
8582 from call to replace_mult_candidate.
8583 (replace_conditional_candidate): Likewise.
8584 (insert_initializers): Use make_temp_ssa_name.
8585 (introduce_cast_before_cand): Remove unnecessary argument; use
8587 (replace_one_candidate): Remove unnecessary argument; remove
8588 unnecessary parameter from calls to introduce_cast_before_cand.
8589 (replace_profitable_candidates): Remove unnecessary parameters
8590 from calls to replace_one_candidate.
8592 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8594 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8595 phi def as possibly hiding a basis for a CAND_ADD whose operands
8596 have been commuted in the analysis.
8597 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8599 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8601 * config/aarch64/aarch64.md
8602 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8603 shift value between 0-4.
8605 2013-05-07 Richard Biener <rguenther@suse.de>
8607 * double-int.h (rshift): New overload.
8608 * double-int.c (rshift): New function.
8609 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8610 (create_reference_ops_from_ref): Remove.
8611 (vn_reference_insert): Use shared ops for constructing the
8612 reference and copy it.
8614 2013-05-07 Richard Biener <rguenther@suse.de>
8617 * tree-eh.c (sink_clobbers): Properly propagate
8618 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8620 2013-05-07 Jakub Jelinek <jakub@redhat.com>
8622 PR tree-optimization/57149
8623 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8624 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8625 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8626 uninit_undefined_value_p instead of ssa_undefined_value_p.
8629 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8630 for modifier == EXPAND_INITIALIZER.
8632 2013-05-07 Anton Blanchard <anton@samba.org>
8634 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8635 for powerpc64 little endian.
8636 * configure: Regenerate.
8638 2013-05-06 Graham Stott <grahams@btinternet.com>
8640 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8641 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8642 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8643 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8645 2013-05-06 Graham Stott <grahams@btinternet.com>
8647 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8648 codes which allow non-lvalues.
8650 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
8652 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8653 components are all 1s.
8654 (integer_minus_onep): New function.
8655 * tree.h (integer_minus_onep): Declare it.
8656 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8657 integer_minus_onep instead of integer_all_onesp.
8659 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8662 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8663 variations of these patterns.
8665 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8667 * config/i386/i386.md (isa): Add x64_sse4 member.
8668 (enabled): Handle x64_sse4.
8669 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8670 instruction for 64bit SSE4_1 targets. Update insn attributes.
8671 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8672 instruction for SSE4_1 targets. Update insn attributes.
8673 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8674 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8675 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8677 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8678 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8679 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8681 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8684 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8686 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8688 * final.c (do_assembler_dialects): Don't handle curly braces and
8689 vertical bar escaped by % as dialect delimiters.
8690 (output_asm_insn): Print curly braces and vertical bar if escaped
8691 by % and ASSEMBLER_DIALECT defined.
8692 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8693 * doc/tm.texi: Regenerated.
8695 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8697 * config/mips/mips.c: Include tree-pass.h.
8698 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8699 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8700 (pass_mips_machine_reorg2): New machine specific pass.
8701 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8702 (mips_option_override): Register the new pass.
8703 * rtl.h (cleanup_barriers): Remove prototype.
8704 (dbr_schedule): Likewise.
8705 * jump.c (cleanup_barriers): Make static.
8706 * reorg.c (dbr_schedule): Likewise.
8708 2013-05-06 Richard Biener <rguenther@suse.de>
8710 PR tree-optimization/57185
8711 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8714 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8717 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8718 between "first_arg" and "insn", not "last" and "insn".
8720 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8722 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8723 (find_candidates_in_block): Re-enable slsr_process_phi.
8724 (create_phi_basis): Fix double counting of candidate adjustment.
8726 2013-05-06 Richard Biener <rguenther@suse.de>
8729 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8730 the edge is also fallthru, preserve it and just clear the
8732 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8733 also complex, preserve that and just clear the fallthru flag.
8734 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8735 update virtual operands.
8737 2013-05-06 Alan Modra <amodra@gmail.com>
8739 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8740 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8741 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8742 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8743 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8744 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8746 2013-05-06 Alan Modra <amodra@gmail.com>
8748 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8749 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8750 (DEFAULT_ASM_ENDIAN): Define.
8751 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8752 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8753 Update -K PIC clause from sysv4.h.
8754 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8755 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8757 2013-05-06 Alan Modra <amodra@gmail.com>
8759 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8760 twice for little-endian.
8761 (ashrdi3_no_power, ashrdi3): Support little-endian.
8763 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8766 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
8767 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
8769 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
8770 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
8771 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
8774 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
8775 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
8777 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
8778 * config/pa/pa.opt: Make mbig-switch a no-op.
8779 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
8780 (CASE_VECTOR_MODE): Always return SImode.
8781 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
8782 for the !TARGET_BIG_SWITCH case.
8783 * config/pa/pa-linux.h: Likewise.
8784 * config/pa/pa-openbsd.h: Likewise.
8785 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
8786 * config/pa/pa.md (short_jump): Remove define_insn.
8787 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
8788 (casesi0): Remove define_insn.
8789 (type): Remove btable_branch.
8790 (pa_combine_type): Likewise.
8791 (in_nullified_branch_delay): Likewise.
8792 (in_call_delay): Likewise.
8793 (define_delay): Likewise.
8794 (define_insn_reservation "Z3"): Likewise.
8795 (define_insn_reservation "Z4"): Likewise.
8796 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
8797 (pa_adjust_insn_length): Remove adjustment for btable branches.
8798 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
8801 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
8803 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8804 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
8805 Add m->r,x alternatives.
8806 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
8807 splitters using SWI48x mode iterator.
8808 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
8809 TARGET_64BIT. Add m->x alternative.
8810 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
8811 Add o->x alternative. Enable for TARGET_SSE.
8812 (sse_storeq): Remove expander.
8813 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
8814 with memory input operand.
8815 (*vec_extractv2di_1 splitter): New.
8816 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
8817 * config/i386/i386.md (ssevecmodelower): New mode attribute.
8819 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8821 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
8822 (INT_LOWPART): Delete.
8823 (extract_MB): Adjust.
8824 (extract_ME): Adjust.
8825 (print_operand): Adjust.
8827 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
8829 * config/rs6000/predicates.md (reg_or_add_cint_operand,
8830 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8831 (reg_or_logical_cint_operand, easy_fp_constant,
8832 logical_const_operand): Delete "CONST_DOUBLE" case.
8833 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
8834 "HOST_BITS_PER_WIDE_INT == 64" test.
8835 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
8836 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
8837 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
8838 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
8840 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
8841 CONST_DOUBLE DImode/VOIDmode case.
8842 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
8843 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8844 CONST_DOUBLE VOIDmode case.
8845 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
8846 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
8847 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
8848 Delete CONST_DOUBLE case.
8849 (splitters for mov FMOVE64 const_double): Delete
8850 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
8851 "HOST_BITS_PER_WIDE_INT >= 64" test.
8852 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
8854 (mov DI const_double): Delete.
8856 2013-05-04 Jakub Jelinek <jakub@redhat.com>
8858 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
8859 on op shows all bits zero in mode of a lowpart subreg, return zero.
8861 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
8864 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
8865 to save TFmode registers and DImode to save TImode registers for
8866 caller save operations.
8867 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
8868 mark being partially clobbered since they only use the first
8871 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
8872 and TDmode only use the upper 64-bits of each VSX register.
8874 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8876 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
8877 (find_candidates_in_block): Disable slsr_process_phi.
8879 2013-05-03 Guozhi Wei <carrot@google.com>
8881 * coverage.c (coverage_obj_init): Move the construction of gcov
8883 (build_init_ctor): ... here.
8885 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8887 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
8888 (slsr_cand_d): Redefine def_phi.
8889 (stride_status, phi_adjust_status, count_phis_status): New enums.
8890 (find_phi_def): New.
8891 (find_basis_for_base_expr): New.
8892 (find_basis_for_candidate): Handle hidden bases.
8893 (alloc_cand_and_find_basis): Handle phi candidates.
8894 (slsr_process_phi): New.
8895 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
8896 (create_mul_imm_cand): Likewise.
8897 (create_add_ssa_cand): Exclude phi base candidates.
8898 (create_add_imm_cand): Likewise.
8899 (slsr_process_cast): Likewise.
8900 (slsr_process_copy): Likewise.
8901 (find_candidates_in_block): Handle phi candidates.
8902 (dump_candidate): Likewise.
8903 (unconditional_cands): Delete.
8904 (unconditional_cands_with_known_stride_p): Delete.
8905 (phi_dependent_cand_p): New.
8906 (cand_increment): Handle phi-dependent candidates.
8907 (replace_dependent): Delete.
8908 (replace_mult_candidate): New.
8909 (replace_unconditional_candidate): New.
8910 (incr_vec_index): Move to avoid forward reference.
8911 (create_add_on_incoming_edge): New.
8912 (create_phi_basis): New.
8913 (replace_dependents): Delete.
8914 (replace_conditional_candidate): New.
8915 (phi_add_costs): New.
8916 (replace_uncond_cands_and_profitable_phis): New.
8917 (record_increment): Handle phi adjustments.
8918 (record_phi_increments): New.
8919 (record_increments): Handle phi adjustments.
8920 (phi_incr_cost): New.
8921 (lowest_cost_path): Handle phis.
8922 (total_savings): Likewise.
8923 (analyze_increments): Likewise.
8924 (ncd_with_phi): New.
8925 (ncd_of_cand_and_phis): New.
8926 (nearest_common_dominator_for_cands): Handle phi increments.
8927 (all_phi_incrs_profitable): New.
8928 (replace_profitable_candidates): Handle phi-dependent candidates.
8929 (analyze_candidates_and_replace): Likewise.
8931 2013-05-03 Teresa Johnson <tejohnson@google.com>
8934 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
8935 do not exceed REG_BR_PROB_BASE.
8937 2013-05-03 Jeff Law <law@redhat.com>
8939 PR tree-optimization/57144
8940 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
8941 operand of the condition will bit into the new type when eliminating
8942 a cast feeding a condition.
8944 2013-05-03 Jakub Jelinek <jakub@redhat.com>
8946 PR rtl-optimization/57130
8947 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
8948 of COMPARE as in_code to the recursive call if needed.
8950 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
8952 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
8953 (enabled): Handle new members.
8954 * config/i386/sse.md (*vec_concatv2si): Merge from
8955 *vec_concatv2si_sse2 and vec_concatv2si_sse.
8956 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
8958 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
8960 PR tree-optimization/57027
8961 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
8962 for fnms opportunity, check we got the prerequisite kind
8963 of tree / gimple before using accessor functions.
8965 2013-05-03 Richard Biener <rguenther@suse.de>
8967 * double-int.h (lshift): New overload without precision
8969 (operator *=, operator +=, operator -=): Move ...
8970 * double-int.c (operator *=, operator +=, operator -=): ... here
8971 and implement more efficiently.
8972 (mul_double_with_sign): Remove.
8973 (lshift_double): Adjust to take unsinged shift argument, push
8974 dispatching code to callers.
8975 (mul_double_wide_with_sign): Add early out for callers that
8976 are not interested in high parts or overflow.
8977 (lshift): New function.
8978 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
8980 (lrotate, rrotate): Use logical shifts.
8981 * expr.c (get_inner_reference): Use lshift.
8982 * fixed-value.c (do_fixed_divide): Likewise.
8983 * tree-dfa.c (get_ref_base_and_extent): Likewise.
8984 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
8985 (indirect_refs_may_alias_p): Likewise.
8986 (stmt_kills_ref_p_1): Likewise.
8988 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8990 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
8992 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
8994 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
8995 scalar form of FABD instruction.
8997 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
8999 * lra-constraints.c (process_alt_operands): Add checking alt
9000 number to choose the best alternative.
9002 2013-05-02 Richard Biener <rguenther@suse.de>
9004 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9005 bitmap and its handling.
9006 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9008 2013-05-02 Richard Biener <rguenther@suse.de>
9011 * tree-inline.c (copy_loops): Properly handle removed loops.
9012 (copy_cfg_body): Mark destination loops for fixup if source
9015 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9018 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9019 generating simple_return for naked functions.
9021 2013-05-02 Martin Jambor <mjambor@suse.cz>
9024 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9025 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9027 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9028 ipa_agg_replacement_value structures.
9029 (known_aggs_to_agg_replacement_list): Likewise.
9030 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9031 (read_agg_replacement_chain): Likewise.
9032 (ipcp_transform_function): Also check that by_ref flags match.
9034 2013-05-02 Richard Biener <rguenther@suse.de>
9036 * graphds.h (struct graph): Add obstack member.
9037 * graphds.c (new_graph): Initialize obstack and allocate
9039 (add_edge): Allocate edge from the obstack.
9040 (free_graph): Free the obstack instead of all edges and vertices.
9042 2013-05-02 Teresa Johnson <tejohnson@google.com>
9044 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9046 * cfg.c (update_bb_profile_for_threading): Ditto.
9047 * tree-inline.c (copy_bb): Ditto.
9048 (copy_edges_for_bb): Ditto.
9049 (initialize_cfun): Ditto.
9050 (copy_cfg_body): Ditto.
9051 (expand_call_inline): Ditto.
9052 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9053 (estimate_node_size_and_time): Ditto.
9054 (inline_merge_summary): Ditto.
9055 * cgraphclones.c (cgraph_clone_edge): Ditto.
9056 (cgraph_clone_node): Ditto.
9057 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9058 (compute_trg_info): Ditto.
9060 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9062 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9063 S reg when fp attribute set.
9064 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9066 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9068 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9070 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9071 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9072 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9074 2013-05-02 Richard Biener <rguenther@suse.de>
9076 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9077 (struct instantiate_cache_entry): New type.
9078 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9079 (struct instantiate_cache_type): New type.
9080 (set_instantiated_value, get_instantiated_value): Remove.
9081 (get_instantiated_value_entry): New function.
9082 (instantiate_scev_name): Use the new cache and adjust.
9083 (instantiate_scev_poly): Adjust.
9084 (instantiate_scev_binary): Likewise.
9085 (instantiate_array_ref): Likewise.
9086 (instantiate_scev_convert): Likewise.
9087 (instantiate_scev_not): Likewise.
9088 (instantiate_scev_3): Likewise.
9089 (instantiate_scev_2): Likewise.
9090 (instantiate_scev_r): Likewise.
9091 (instantiate_scev): Likewise.
9092 (resolve_mixers): Likewise.
9094 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9097 * lra-constraints.c (best_small_class_operands_num): Remove.
9098 (process_alt_operands): Remove small_class_operands_num. Take
9099 small classes operands into losers and only if the operand is not
9100 matched. Modify debugging output.
9101 (curr_insn_transform): Remove best_small_class_operands_num.
9104 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9106 * config/aarch64/aarch64-builtins.c
9107 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9108 * config/aarch64/aarch64-simd-builtins.def
9109 (reduc_splus_): Add new modes.
9110 (reduc_uplus_): New.
9111 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9112 (reduc_uplus_v4sf): Likewise.
9113 (reduc_splus_v4sf): Likewise.
9114 (aarch64_addv<mode>): Likewise.
9115 (reduc_uplus_<mode>): Likewise.
9116 (reduc_splus_<mode>): Likewise.
9117 (aarch64_addvv2di): Likewise.
9118 (reduc_uplus_v2di): Likewise.
9119 (reduc_splus_v2di): Likewise.
9120 (aarch64_addvv2si): Likewise.
9121 (reduc_uplus_v2si): Likewise.
9122 (reduc_splus_v2si): Likewise.
9123 (reduc_<sur>plus_<mode>): New.
9124 (reduc_<sur>plus_v2di): Likewise.
9125 (reduc_<sur>plus_v2si): Likewise.
9126 (reduc_<sur>plus_v4sf): Likewise.
9127 (aarch64_addpv4sf): Likewise.
9128 * config/aarch64/arm_neon.h
9129 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9130 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9131 add UNSPEC_SADDV, UNSPEC_UADDV.
9133 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9135 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9137 * config/aarch64/arm_neon.h
9138 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9140 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9142 * config/aarch64/aarch64-builtins
9143 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
9145 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9147 * config/aarch64/aarch64-simd-builtins.def
9149 (reduc_smin_): Likewise.
9150 (reduc_umax_): Likewise.
9151 (reduc_umin_): Likewise.
9152 (reduc_smax_nan_): Likewise.
9153 (reduc_smin_nan_): Likewise.
9156 (smax): Update for V2SF, V4SF and V2DF modes.
9159 (smin_nan): Likewise.
9160 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
9161 (<su><maxmin><mode>3): ...This, refactor.
9162 (s<maxmin><mode>3): New.
9163 (<maxmin_uns><mode>3): Likewise.
9164 (reduc_<maxmin_uns>_<mode>): Refactor.
9165 (reduc_<maxmin_uns>_v4sf): Likewise.
9166 (reduc_<maxmin_uns>_v2si): Likewise.
9167 (aarch64_<fmaxmin><mode>: Remove.
9168 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
9170 (vmin<q>_f<32,64>): Likewise.
9171 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9173 (su): Add mappings for smax, smin, umax, umin.
9175 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9176 (FMAXMIN): Rename as...
9177 (FMAXMIN_UNS): ...This.
9179 (fmaxminv): Likewise.
9180 (fmaxmin): Likewise.
9182 (maxmin_uns_op): Likewise.
9184 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9186 * config/aarch64/arm_neon.h
9187 (vac<ge, gt><sd>_f<32, 64>): Rename to...
9188 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
9189 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
9191 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9193 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
9194 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
9196 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9198 * config/aarch64/aarch64-simd.md
9199 (vcond<mode>_internal): Handle special cases for constant masks.
9200 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
9201 (vcondu<mode><mode>): Likewise.
9202 (vcond<v_cmp_result><mode>): New.
9204 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9206 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
9207 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
9208 * config/aarch64/aarch64-simd-builtins.def
9209 (cmeq): Update to BUILTIN_VALLDI.
9214 * config/aarch64/arm_neon.h
9215 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
9216 to builtins or C as appropriate.
9218 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9220 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
9222 (cmhi): Rename to...
9224 * config/aarch64/aarch64-simd.md
9225 (simd_mode): Add SF.
9226 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
9227 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
9228 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
9229 (cstore<mode>_neg): ...This.
9230 * config/aarch64/iterators.md
9232 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
9234 (UCOMPARISONS): Likewise.
9235 (optab): Add missing comparisons.
9243 (V_cmp_result): Add DF, SF modes.
9244 (v_cmp_result): Likewise.
9247 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
9249 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
9251 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
9252 define_insn to define_insn_and_split.
9253 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
9254 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
9255 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
9256 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
9257 (thumb2_negscc): Likewise.
9259 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9261 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
9263 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9265 * config/arm/thumb2.md: Remove trailing whitespaces.
9267 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9269 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
9270 Use gen_int_mode rather than GEN_INT.
9272 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
9274 * value-prof.c (stream_in_histogram_value): Remove the strayed
9277 2013-04-30 Richard Biener <rguenther@suse.de>
9280 * cfghooks.c (split_edge): Properly check for the loop latch edge.
9282 2013-04-30 Richard Biener <rguenther@suse.de>
9285 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
9287 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
9289 PR rtl-optimization/56957
9290 PR rtl-optimization/57105
9291 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
9292 variable. Use just INSN_UID for determining whether an insn
9293 should be only disconnected from the insn stream.
9294 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
9296 2013-04-30 Jakub Jelinek <jakub@redhat.com>
9298 PR tree-optimization/57104
9299 * tsan.c (instrument_expr): Don't instrument accesses to
9300 DECL_HARD_REGISTER VAR_DECLs.
9302 2013-04-30 Richard Biener <rguenther@suse.de>
9304 * function.h (loops_for_fn): New inline function.
9305 (set_loops_for_fn): Likewise.
9306 * cfgloop.h (place_new_loop): Add struct function parameter.
9307 (get_loop): Likewise.
9308 (get_loops): Likewise.
9309 (number_of_loops): Likewise.
9311 (fel_init): Likewise.
9312 * cfg.c (get_loop_copy): Adjust.
9313 * cfgloop.c (flow_loops_dump): Likewise.
9314 (record_loop_exits): Likewise.
9315 (verify_loop_structure): Likewise.
9316 * cfgloopanal.c (mark_irreducible_loops): Likewise.
9317 (estimate_reg_pressure_cost): Likewise.
9318 (mark_loop_exit_edges): Likewise.
9319 * cfgloopmanip.c (place_new_loop): Likewise.
9320 (add_loop): Likewise.
9321 (duplicate_loop): Likewise.
9322 * graph.c (draw_cfg_nodes): Likewise.
9323 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
9324 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
9325 (extract_affine_chrec): Likewise.
9326 (build_scop_iteration_domain): Likewise.
9327 * graphite.c (graphite_initialize): Likewise.
9328 * ira-build.c (create_loop_tree_nodes): Likewise.
9329 (more_one_region_p): Likewise.
9330 (rebuild_regno_allocno_maps): Likewise.
9331 (mark_loops_for_removal): Likewise.
9332 (mark_all_loops_for_removal): Likewise.
9333 (remove_unnecessary_regions): Likewise.
9334 (ira_build): Likewise.
9335 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
9336 * loop-init.c (fix_loop_structure): Likewise.
9337 (gate_rtl_move_loop_invariants): Likewise.
9338 (gate_rtl_unswitch): Likewise.
9339 (gate_rtl_unroll_and_peel_loops): Likewise.
9340 (rtl_doloop): Likewise.
9341 * lto-streamer-in.c (input_cfg): Likewise.
9342 * lto-streamer-out.c (output_cfg): Likewise.
9343 * modulo-sched.c (sms_schedule): Likewise.
9344 * predict.c (tree_estimate_probability): Likewise.
9345 (tree_estimate_probability_driver): Likewise.
9346 (estimate_loops): Likewise.
9347 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
9348 (move_sese_region_to_fn): Likewise.
9349 (debug_loop_num): Likewise.
9350 * tree-chrec.c (chrec_evaluate): Likewise.
9351 (hide_evolution_in_other_loops_than_loop): Likewise.
9352 (chrec_component_in_loop_num): Likewise.
9353 (reset_evolution_in_loop): Likewise.
9354 (evolution_function_is_invariant_rec_p): Likewise.
9355 * tree-if-conv.c (main_tree_if_conversion): Likewise.
9356 * tree-inline.c (copy_loops): Likewise.
9357 (copy_cfg_body): Likewise.
9358 (tree_function_versioning): Likewise.
9359 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
9360 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
9362 (add_to_evolution_1): Likewise.
9363 (scev_const_prop): Likewise.
9364 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
9365 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
9366 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
9367 (tree_ssa_lim_initialize): Likewise.
9368 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
9369 (verify_loop_closed_ssa): Likewise.
9370 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
9371 (tree_ssa_loop_im): Likewise.
9372 (tree_ssa_loop_unswitch): Likewise.
9373 (tree_vectorize): Likewise.
9374 (check_data_deps): Likewise.
9375 (tree_ssa_loop_ivcanon): Likewise.
9376 (tree_ssa_loop_bounds): Likewise.
9377 (tree_complete_unroll): Likewise.
9378 (tree_complete_unroll_inner): Likewise.
9379 (tree_parallelize_loops): Likewise.
9380 (tree_ssa_loop_prefetch): Likewise.
9381 (tree_ssa_loop_ivopts): Likewise.
9382 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
9383 * tree-vectorizer.c (vectorize_loops): Likewise.
9385 2013-04-29 Mike Frysinger <vapier@gentoo.org>
9387 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
9388 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
9389 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
9390 with EABI_LINK_SPEC.
9392 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9395 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9398 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9401 * lra-constraints.c (process_alt_operands): Discourage a bit more
9402 using memory for pseudos. Print cost dump for alternatives.
9403 Modify cost values for conflicts with early clobbers.
9404 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9406 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9409 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9411 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9413 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9415 (movdi_aarch64): Support LDR/STR from/to D register.
9417 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9419 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9420 or higher by default.
9422 2013-04-29 Richard Biener <rguenther@suse.de>
9425 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9426 even if not adding abnormal edges for calls that can make
9429 2013-04-29 Richard Biener <rguenther@suse.de>
9432 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9434 (move_stmt_r): Remove redundant checking.
9436 2013-04-29 Teresa Johnson <tejohnson@google.com>
9439 * basic-block.h (apply_scale): New function.
9440 (apply_probability): Use apply_scale.
9441 * gimple-streamer-in.c (input_bb): Ditto.
9442 * lto-streamer-in.c (input_cfg): Ditto.
9443 * lto-cgraph.c (merge_profile_summaries): Ditto.
9444 * tree-optimize.c (execute_fixup_cfg): Ditto.
9445 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9446 (copy_edges_for_bb): Ditto.
9447 (copy_cfg_body): Ditto.
9449 2013-04-29 Tom de Vries <tom@codesourcery.com>
9451 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9452 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9453 (tail_merge_optimize): Handle current_loops == NULL.
9455 2013-04-26 Jeff Law <law@redhat.com>
9457 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9458 (simplify_cond_using_ranges): Generalize code to simplify
9459 COND_EXPRs where one argument is a constant and the other
9460 is an SSA_NAME created by an integral type conversion.
9462 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9464 * config/arm/arm.md (store_minmaxsi): Use only when
9465 optimize_insn_for_size_p.
9467 2013-04-29 Christian Bruel <christian.bruel@st.com>
9470 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9472 2013-04-29 Richard Biener <rguenther@suse.de>
9475 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9476 loop tree make sure to schedule a fixup for the child as well.
9477 (expand_omp_for_generic): Properly add loops.
9478 (expand_omp_for_static_nochunk): Likewise.
9479 (expand_omp_for_static_chunk): Likewise.
9480 (expand_omp_for): For the degenerate case fixup loops.
9481 (expand_omp_sections): Fix default bb placement in loops.
9482 (expand_omp_atomic_pipeline): Properly add loops.
9484 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9486 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9488 2013-04-29 Tom de Vries <tom@codesourcery.com>
9490 * tree-ssa-tail-merge.c: Update header comment.
9492 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9494 * config/aarch64/arm_neon.h
9495 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9496 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9497 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9498 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9499 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9500 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9501 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9502 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9504 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9506 * config/aarch64/aarch64-simd.md
9507 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9508 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9509 fix_trunc, fixuns_trunc.
9510 (ftrunc<VDQF:mode>2): New.
9511 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9512 (fix_trunc_optab): New.
9514 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9516 * config/aarch64/aarch64-builtins.c
9517 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9518 iceilf, lround, iroundf.
9520 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9523 * config/i386/i386.h (enum ix86_tune_indices)
9524 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9525 New, split from X86_TUNE_INTER_UNIT_MOVES.
9526 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9527 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9528 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9529 (TARGET_INTER_UNIT_MOVES): Remove.
9530 * config/i386/i386.c (initial_ix86_tune_features): Update.
9531 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9532 (ix86_expand_convert_uns_didf_sse): Use
9533 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9534 (ix86_expand_vector_init_one_nonzero): Ditto.
9535 (ix86_expand_vector_init_interleave): Ditto.
9536 (inline_secondary_memory_needed): Return true for moves from SSE class
9537 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9538 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9539 * config/i386/constraints.md (Yi, Ym): Depend on
9540 TARGET_INTER_UNIT_MOVES_TO_VEC.
9541 (Yj, Yn): New constraints.
9542 * config/i386/i386.md (*movdi_internal): Change constraints of
9543 operand 1 from Yi to Yj and from Ym to Yn.
9544 (*movsi_internal): Ditto.
9545 (*movdf_internal): Ditto.
9546 (*movsf_internal): Ditto.
9547 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9548 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9549 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9550 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9551 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9552 * config/i386/sse.md (movdi_to_sse): Ditto.
9553 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9554 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9555 TARGET_INTER_UNIT_MOVES.
9556 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9557 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9558 instead of TARGET_INTER_UNIT_MOVES.
9559 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9560 operand 1 from Yi to Yj and from Ym to Yn.
9562 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9564 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9565 (float_truncate_hi_): Likewise.
9566 (float_extend_lo_): Likewise.
9567 (float_truncate_lo_): Likewise.
9568 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9569 (aarch64_float_extend_lo_v2df): Likewise.
9570 (vec_unpacks_hi_v4sf): Likewise.
9571 (aarch64_float_truncate_lo_v2sf): Likewise.
9572 (aarch64_float_truncate_hi_v4sf): Likewise.
9573 (vec_pack_trunc_v2df): Likewise.
9574 (vec_pack_trunc_df): Likewise.
9576 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9578 * config/aarch64/aarch64-builtins.c
9579 (aarch64_fold_builtin): Fold float conversions.
9580 * config/aarch64/aarch64-simd-builtins.def
9581 (floatv2si, floatv4si, floatv2di): New.
9582 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9583 * config/aarch64/aarch64-simd.md
9584 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9585 * config/aarch64/iterators.md (FLOATUORS): New.
9586 (optab): Add float, floatuns.
9587 (su_optab): Likewise.
9589 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9591 * config/aarch64/aarch64-builtins.c
9592 (aarch64_builtin_vectorized_function): Use new names for
9594 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9595 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9596 (fcvtzu): Split as...
9597 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9598 (fcvtas): Split as...
9599 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9600 (fcvtau): Split as...
9601 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9602 (fcvtps): Split as...
9603 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9604 (fcvtpu): Split as...
9605 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9606 (fcvtms): Split as...
9607 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9608 (fcvtmu): Split as...
9609 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9610 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9611 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9612 (lfrintnusf, lfrintnudf): Likewise.
9613 * config/aarch64/aarch64-simd.md
9614 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9616 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9617 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9618 (fcvt_pattern): Likewise.
9620 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9622 * config/aarch64/aarch64-simd.md
9623 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9624 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9626 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9628 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9629 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9630 (vrnd<a,m,n,p>_f32): Implement using builtins.
9631 (vrnd<i,x><q>_f<32, 64>): New.
9633 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9635 * config/aarch64/aarch64-builtins.c
9636 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9637 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9638 (frintz): Rename to...
9640 (frintp): Rename to...
9642 (frintm): Rename to...
9644 (frinti): Rename to...
9645 (nearbyint): ...this.
9646 (frintx): Rename to...
9648 (frinta): Rename to...
9650 * config/aarch64/aarch64-simd.md
9651 (aarch64_frint<frint_suffix><mode>): Delete.
9652 (<frint_pattern><mode>2): Convert to insn.
9653 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9654 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9655 (frint_pattern): Likewise.
9656 (frint_suffix): Likewise.
9658 2013-04-29 Richard Biener <rguenther@suse.de>
9660 PR tree-optimization/57081
9661 * loop-init.c: Include tree-flow.h.
9662 (loop_optimizer_finalize): Free number of iteration estimates.
9663 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9665 2013-04-29 Jakub Jelinek <jakub@redhat.com>
9667 PR tree-optimization/57083
9668 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9669 non-singleton shift count range, zero extend low_bound for uns case.
9671 * config/i386/predicates.md (general_vector_operand): New predicate.
9672 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9673 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9674 if they aren't nonimmediate operands. If their original values
9675 satisfy const_vector_equal_evenodd_p, don't shift them.
9676 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9677 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9678 if not nonimmediate_operand.
9679 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9680 instead of register_operand.
9681 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9683 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
9685 * stor-layout.c (finalize_size_functions): Allocate a structure and
9686 reset cfun before dumping the functions.
9688 2013-04-27 Jakub Jelinek <jakub@redhat.com>
9690 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9693 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9694 use xop_pmacsdqh if uns_p.
9695 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9696 the immediate rotate count.
9698 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9700 * rtl.h (struct rtx_def): Add comment for field jump.
9701 (LRA_SUBREG_P): New macro.
9702 * recog.c (register_operand): Check LRA_SUBREG_P.
9703 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9705 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9706 (lra_final_code_change): Skip subreg change for operators.
9707 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9708 if there are no operand changes.
9709 * lra-constraints.c (curr_insn_set): New.
9710 (match_reload): Set LRA_SUBREG_P.
9711 (emit_spill_move): Ditto.
9712 (check_and_process_move): Use curr_insn_set. Process only single
9713 set insns. Don't initialize sec_mem_p and change_p.
9714 (simplify_operand_subreg): Use LRA_SUBREG_P.
9715 (reg_in_class_p): New function.
9716 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9717 of #ifdef. Add code to remove cycling.
9718 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9719 non-null disp. Reload inner instead of disp when base and index
9720 are null. Try to put lo_sum into register.
9721 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9722 (check_and_process_move): Move code for move cost check to
9723 simple_move_p. Remove equiv_substitution.
9724 (simple_move_p): New function.
9725 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9726 curr_insn_set. Call check_and_process_move only for single set
9727 insns. Use the new function. Move call of check_and_process_move
9728 after operand equiv substitution and address process.
9730 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9733 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9734 with nonlocal goto receivers or returns twice calls, ignore
9735 unininitialized values from abnormal edges to nl goto receiver
9736 or returns twice call.
9738 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9740 PR tree-optimization/57051
9741 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9742 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9745 2013-04-26 Richard Biener <rguenther@suse.de>
9747 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9748 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9749 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9750 (fixup_loop_arrays_after_move): New function.
9751 (move_sese_region_to_fn): Properly outline the loop tree parts
9754 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
9756 * config/i386/i386.md (type, unit): Fix long lines.
9758 2013-04-26 Richard Biener <rguenther@suse.de>
9760 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9761 (lto-streamer-out.o): Likewise.
9762 * cfgloop.c (init_loops_structure): Export, add struct function
9763 argument and adjust.
9764 (flow_loops_find): Adjust.
9765 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9766 (init_loops_structure): Declare.
9767 * lto-streamer-in.c: Include cfgloop.h.
9768 (input_cfg): Input the loop tree.
9769 * lto-streamer-out.c: Include cfgloop.h.
9770 (output_cfg): Output the loop tree.
9771 (output_struct_function_base): Do not drop PROP_loops.
9773 2013-03-26 Richard Biener <rguenther@suse.de>
9775 * tree-cfg.c (execute_build_cfg): Build the loop tree.
9776 (pass_build_cfg): Provide PROP_loops.
9777 (move_sese_region_to_fn): Remove loops that are outlined into fn
9779 * tree-inline.c: Include cfgloop.h.
9780 (initialize_cfun): Do not drop PROP_loops.
9781 (copy_loops): New function.
9782 (copy_cfg_body): Copy loop structure.
9783 (tree_function_versioning): Initialize destination loop tree.
9784 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
9785 (pass_parallelize_loops): Do IL verification.
9786 * loop-init.c (loop_optimizer_init): Fixup loops if required.
9787 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
9788 the CFG make sure we fixup loops as well.
9789 * tree-ssa-tail-merge.c: Include cfgloop.h.
9790 (replace_block_by): When merging loop latches mark loops for fixup.
9791 * lto-streamer-out.c (output_struct_function_base): Drop
9793 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
9794 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
9795 * ipa-split.c: Include cfgloop.h.
9796 (split_function): Add the new return block to the loop tree root.
9797 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
9798 whether we have removed the forwarder block.
9799 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
9800 * cfgloop.h (place_new_loop): Declare.
9801 * cfgloopmanip.c (place_new_loop): Export.
9802 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
9803 (tree-switch-conversion.o): Likewise.
9804 (tree-complex.o): Likewise.
9805 (tree-inline.o): Likewise.
9806 (tree-ssa-tailmerge.o): Likewise.
9807 (ipa-split.o): Likewise.
9808 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
9809 (tree-ssa-copy.o): Likewise.
9810 * tree-switch-conversion.c: Include cfgloop.h
9811 (process_switch): If we emit a bit-test cascade, schedule loops
9813 * tree-complex.c: Include cfgloop.h.
9814 (expand_complex_div_wide): Properly add new basic-blocks to loops.
9815 * asan.c: Include cfgloop.h.
9816 (create_cond_insert_point): Properly add new basic-blocks to
9817 loops, schedule loop fixup.
9818 * cfgloop.c (verify_loop_structure): Check that looks are not
9820 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
9822 (expand_omp_for_generic): Likewise.
9823 (expand_omp_sections): Likewise.
9824 (expand_omp_atomic_pipeline): Schedule loops for fixup.
9825 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
9826 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
9827 is initialized, not when loops are present.
9828 * tree-parloops.c (parallelize_loops): Remove checking here.
9829 * passes.c (init_optimization_passes): Schedule a copy-propagation
9830 pass before complete unrolling of inner loops.
9832 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9834 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
9835 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
9836 (colorize_init): Add argument to _WIN32 version.
9837 * toplev.c: Include diagnostic-color.h.
9838 (process_options): Default to -fdiagnostics-color=auto if
9839 GCC_COLORS env var is in the environment.
9840 * common.opt (fdiagnostics-color=): Add Var and Init.
9841 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
9842 env var is in the environment, the default is auto rather than never.
9844 * diagnostic.h (file_name_as_prefix): Add context argument.
9845 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
9846 the string as locus.
9847 * langhooks.c (lhd_print_error_function): Adjust caller.
9849 2013-04-25 Lawrence Crowl <crowl@google.com>
9851 * var-tracking.c (shared_hash_def::htab):
9852 Change type to hash_table. Update dependent calls and types.
9854 2013-04-25 Lawrence Crowl <crowl@google.com>
9856 * Makefile.in: Update as needed below.
9858 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
9859 Move declaration to after the type's method definitons.
9861 * attribs.c (htab_t scoped_attributes::attribute_hash):
9862 Change type to hash_table. Update dependent calls and types.
9864 * bitmap.c (htab_t bitmap_desc_hash):
9865 Change type to hash_table. Update dependent calls and types.
9867 * cselib.c (htab_t cselib_hash_table):
9868 Change type to hash_table. Update dependent calls and types.
9870 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
9871 (hash_string_slot_node): Move implementation into lto-streamer.h
9872 struct string_slot_hasher.
9873 (eq_string_slot_node): Likewise.
9875 * data-streamer-out.c: Update output_block::string_hash_table
9876 dependent calls and types.
9878 * dwarf2cfi.c (htab_t trace_index):
9879 Change type to hash_table. Update dependent calls and types.
9881 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
9882 Change type to hash_table. Update dependent calls and types.
9883 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
9884 (htab_t optimize_external_refs::map): Likewise.
9885 (htab_t output_comp_unit::extern_map): Likewise.
9886 (htab_t output_comdat_type_unit::extern_map): Likewise.
9887 (htab_t output_macinfo::macinfo_htab): Likewise.
9888 (htab_t optimize_location_lists::htab): Likewise.
9889 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
9891 * except.c (htab_t ehspec_hash_type):
9892 Change type to hash_table. Update dependent calls and types.
9893 (assign_filter_values::ttypes): Likewise.
9894 (assign_filter_values::ehspec): Likewise.
9895 (sjlj_assign_call_site_values::ar_hash): Likewise.
9896 (convert_to_eh_region_ranges::ar_hash): Likewise.
9898 * gcse.c (htab_t pre_ldst_table):
9899 Change type to hash_table. Update dependent calls and types.
9901 * ggc-common.c (htab_t saving_htab):
9902 Change type to hash_table. Update dependent calls and types.
9903 (htab_t loc_hash): Likewise.
9904 (htab_t ptr_hash): Likewise.
9905 (call_count): Rename ggc_call_count.
9906 (call_alloc): Rename ggc_call_alloc.
9907 (loc_descriptor): Rename make_loc_descriptor.
9908 (add_statistics): Rename ggc_add_statistics.
9910 * ggc-common.c (saving_htab):
9911 Change type to hash_table. Update dependent calls and types.
9913 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
9914 (push_gimplify_context): Likewise.
9915 (pop_gimplify_context): Likewise.
9916 (struct gimple_temp_hash_elt): Added.
9917 (struct gimplify_hasher): Likewise.
9918 (struct gimplify_ctx.temp_htab):
9919 Change type to hash_table. Update dependent calls and types.
9921 * gimple-fold.c: Include gimplify-ctx.h.
9923 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
9924 Change type to hash_table. Update dependent calls and types.
9925 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
9926 avoid potential global name collision.
9928 * gimplify.c: Include gimplify-ctx.h.
9929 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
9930 (htab_t gimplify_ctx::temp_htab):
9931 Update dependent calls and types for new type hash_table.
9932 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
9933 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
9935 * gimplify-ctx.h: New.
9936 (struct gimple_temp_hash_elt): Move from gimplify.c.
9937 (class gimplify_hasher): New.
9938 (struct gimplify_ctx): Move from gimple.h.
9939 (htab_t gimplify_ctx::temp_htab):
9940 Change type to hash_table. Update dependent calls and types.
9942 * graphite-clast-to-gimple.c: Include graphite-htab.h.
9943 (htab_t ivs_params::newivs_index):
9944 Change type to hash_table. Update dependent calls and types.
9945 (htab_t ivs_params::params_index): Likewise.
9946 (htab_t print_generated_program::params_index): Likewise.
9947 (htab_t gloog::newivs_index): Likewise.
9948 (htab_t gloog::params_index): Likewise.
9950 * graphite.c: Include graphite-htab.h.
9951 4htab_t graphite_transform_loops::bb_pbb_mapping):
9952 Change type to hash_table. Update dependent calls and types.
9954 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
9955 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
9956 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
9958 * graphite-dependences.c: Include graphite-htab.h.
9959 (loop_is_parallel_p): Change hash table type of parameter.
9961 * graphite-htab.h: New.
9962 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
9963 (extern find_pbb_via_hash): Move from graphite-poly.h.
9964 (extern loop_is_parallel_p): Move from graphite-poly.h.
9965 (extern get_loop_body_pbbs): Move from graphite-poly.h.
9967 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
9968 (extern loop_is_parallel_p): Move to graphite-htab.h.
9969 (extern get_loop_body_pbbs): Move to graphite-htab.h.
9971 * haifa-sched.c (htab_t delay_htab):
9972 Change type to hash_table. Update dependent calls and types.
9973 (htab_t delay_htab_i2): Likewise.
9975 * ira-color.c (htab_t allocno_hard_regs_htab):
9976 Change type to hash_table. Update dependent calls and types.
9978 * ira-costs.c (htab_t cost_classes_htab):
9979 Change type to hash_table. Update dependent calls and types.
9981 * loop-invariant.c (htab_t merge_identical_invariants::eq):
9982 Change type to hash_table. Update dependent calls and types.
9984 * loop-iv.c (htab_t bivs):
9985 Change type to hash_table. Update dependent calls and types.
9987 * loop-unroll.c (htab_t opt_info::insns_to_split):
9988 Change type to hash_table. Update dependent calls and types.
9989 (htab_t opt_info::insns_with_var_to_expand): Likewise.
9991 * lto-streamer.h (struct string_slot): Move from data-streamer.h
9992 (struct string_slot_hasher): New.
9993 (htab_t output_block::string_hash_table):
9994 Change type to hash_table. Update dependent calls and types.
9996 * lto-streamer-in.c (freeing_string_slot_hasher): New.
9997 (htab_t file_name_hash_table):
9998 Change type to hash_table. Update dependent calls and types.
10000 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10003 * lto-streamer.c (htab_t tree_htab):
10004 Change type to hash_table. Update dependent calls and types.
10006 * omp-low.c: Include gimplify-ctx.h.
10008 * passes.c (htab_t name_to_pass_map):
10009 Change type to hash_table. Update dependent calls and types.
10010 (pass_traverse): Rename to passes_pass_traverse.
10012 * plugin.c (htab_t event_tab):
10013 Change type to hash_table. Update dependent calls and types.
10015 * postreload-gcse.c (htab_t expr_table):
10016 Change type to hash_table. Update dependent calls and types.
10017 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10019 * sese.c (debug_rename_map_1): Make extern.
10020 (htab_t copy_bb_and_scalar_dependences::rename_map):
10021 Change type to hash_table. Update dependent calls and types.
10023 * sese.h (extern debug_rename_map): Move to .c file.
10025 * store-motion.c (htab_t store_motion_mems_table):
10026 Change type to hash_table. Update dependent calls and types.
10028 * trans-mem.c (htab_t tm_new_mem_hash):
10029 Change type to hash_table. Update dependent calls and types.
10031 * tree-browser.c (htab_t TB_up_ht):
10032 Change type to hash_table. Update dependent calls and types.
10034 * tree-cfg.c (htab_t discriminator_per_locus):
10035 Change type to hash_table. Update dependent calls and types.
10037 * tree-complex.c: Include tree-hasher.h
10038 (htab_t complex_variable_components):
10039 Change type to hash_table. Update dependent calls and types.
10041 * tree-eh.c (htab_t finally_tree):
10042 Change type to hash_table. Update dependent calls and types.
10044 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10045 struct int_tree_hasher.
10046 (extern int_tree_map_eq): Likewise.
10047 (uid_decl_map_hash): Removed.
10048 (extern decl_tree_map_eq): Likewise.
10050 * tree-hasher.h: New.
10051 (struct int_tree_hasher): New.
10052 (typedef int_tree_htab_type): New.
10054 * tree-inline.c: Include gimplify-ctx.h.
10056 * tree-mudflap.c: Include gimplify-ctx.h.
10058 * tree-parloops.c: Include tree-hasher.h.
10059 (htab_t eliminate_local_variables_stmt::decl_address):
10060 Change type to hash_table. Update dependent calls and types.
10061 (htab_t separate_decls_in_region::decl_copies): Likewise.
10063 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10064 Change type to hash_table. Update dependent calls and types.
10066 * tree-sra.c (candidates):
10067 Change type to hash_table. Update dependent calls and types.
10069 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10071 (int_tree_map_hash): Likewise.
10073 * tree-ssa-dom.c (htab_t avail_exprs):
10074 Change type to hash_table. Update dependent calls and types.
10076 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10077 Change type to hash_table. Update dependent calls and types.
10079 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10080 Change type to hash_table. Update dependent calls and types.
10082 * tree-ssa-phiopt.c (seen_ssa_names):
10083 Change type to hash_table. Update dependent calls and types.
10085 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10086 Change type to hash_table. Update dependent calls and types.
10088 * tree-ssa-uncprop.c (equiv):
10089 Change type to hash_table. Update dependent calls and types.
10091 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10093 PR rtl-optimization/57003
10094 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10095 call note_stores with kill_clobbered_value callback again after
10096 killing regs_invalidated_by_call.
10098 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10100 * config/aarch64/aarch64-simd.md
10101 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10102 (aarch64_simd_bsl<mode>): Likewise.
10103 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10105 2013-04-25 Marek Polacek <polacek@redhat.com>
10107 PR tree-optimization/57066
10108 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10110 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10112 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10114 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10116 * config/aarch64/aarch64-builtins.c
10117 (aarch64_fold_builtin): New.
10118 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10119 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10120 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10121 * config/aarch64/arm_neon.h
10122 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10124 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10125 Tejas Belagod <tejas.belagod@arm.com>
10127 * config/aarch64/aarch64-builtins.c
10128 (aarch64_gimple_fold_builtin): New.
10129 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10130 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10131 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10132 (addvv4sf): Update.
10133 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10135 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10137 * config/aarch64/aarch64.md
10138 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10140 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10142 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10143 (*ngcsi_uxtw): New pattern.
10145 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10146 Julian Brown <julian@codesourcery.com>
10148 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
10149 (TB_DREG): Add T_V4HF.
10150 (v4hf_UP): New macro.
10151 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10152 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10153 Handle initialisation of V4HF. Adjust initialisation of reinterpret
10155 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10156 (arm_vector_mode_supported_p): Handle V4HF.
10157 (arm_mangle_map): Handle V4HFmode.
10158 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
10159 * config/arm/arm_neon_builtins.def: Add entries for
10160 vcvtv4hfv4sf, vcvtv4sfv4hf.
10161 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
10162 (neon_vcvtv4hfv4sf): Likewise.
10163 * config/arm/neon-gen.ml: Handle half-precision floating point
10165 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
10166 * config/arm/arm_neon.h: Regenerate.
10167 * config/arm/neon.ml (type elts): Add F16.
10168 (type vectype): Add T_float16x4, T_floatHF.
10169 (type vecmode): Add V4HF.
10170 (type features): Add Requires_FP_bit feature.
10171 (elt_width): Handle F16.
10172 (elt_class): Likewise.
10173 (elt_of_class_width): Likewise.
10174 (mode_of_elt): Refactor.
10175 (type_for_elt): Handle F16, fix error messages.
10176 (vectype_size): Handle T_float16x4.
10177 (vcvt_sh): New function.
10178 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
10179 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
10180 (string_of_mode): Handle V4HF.
10181 * doc/arm-neon-intrinsics.texi: Regenerate.
10183 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10185 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
10186 format specifier in 'X' case.
10188 2013-04-25 Alan Modra <amodra@gmail.com>
10191 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
10192 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
10193 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
10194 Repeat for many other rotate/shift and mask patterns using subregs.
10195 Name lshiftrt insns.
10196 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
10197 on WORDS_BIG_ENDIAN.
10199 2013-04-25 Alan Modra <amodra@gmail.com>
10201 * config.gcc: Support little-endian powerpc-linux targets.
10202 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
10203 (LINK_OS_LINUX_SPEC): Define.
10204 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
10205 Preserve MASK_LITTLE_ENDIAN.
10206 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
10207 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
10208 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
10209 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
10210 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
10211 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
10212 Correct fp word order for little-endian. Don't shift toc entries
10213 smaller than a word for little-endian.
10214 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
10215 (bswapdi2 splits): Correct low-part subreg for little-endian.
10216 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
10217 low/high where such is correct only for be.
10218 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
10219 little-endian for -mcall-aixdesc.
10221 2013-04-25 Alan Modra <amodra@gmail.com>
10223 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10224 replace_equiv_address_nv.
10226 2013-04-25 Alan Modra <amodra@gmail.com>
10228 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
10230 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10233 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10234 * rtl.h (struct rtx_def): ...
10236 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10238 PR rtl-optimizations/57046
10239 * lra-constraints (split_reg): Set up lra_risky_transformations_p
10240 for multi-reg splits.
10242 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
10244 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
10246 2013-04-24 Sterling Augustine <saugustine@google.com>
10248 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
10249 (comp_dir_string, debug_str_dwo_section): New.
10250 (DEBUG_STR_DWO_SECTION): Rename to ...
10251 (DEBUG_DWO_STR_SECTION): ... this.
10252 (DEBUG_NORM_STR_SECTION): Delete.
10253 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
10254 (DEBUG_STR_DWO_SECTION_FLAGS): New.
10255 (find_AT_string): Move most logic to ...
10256 (find_AT_string_in_table): ... here. New.
10257 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
10258 add_skeleton_AT_string. Delete logic.
10259 (output_skeleton_debug_sections): Remove call to
10260 add_top_level_skeleton_die_attrs.
10261 (add_comp_dir_attribute): Move logic to comp_dir_string.
10262 (dwarf2out_init): Initialize debug_str_dwo_section.
10263 (output_indirect_string): Call find_string_form.
10264 (output_indirect_strings): Rewrite.
10265 (prune_unused_types): Empty skeleton_debug_str_hash.
10266 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
10267 (dwarf2out_finish): Call output_indirect_strings.
10269 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10271 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
10273 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10275 * rtl.h (struct rtx_def): Add comment for field jump.
10276 (LRA_SUBREG_P): New macro.
10277 * recog.c (register_operand): Check LRA_SUBREG_P.
10278 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10280 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10281 (lra_final_code_change): Skip subreg change for operators.
10282 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10283 if there are no operand changes.
10284 * lra-constraints.c (curr_insn_set): New.
10285 (match_reload): Set LRA_SUBREG_P.
10286 (emit_spill_move): Ditto.
10287 (check_and_process_move): Use curr_insn_set. Process only single
10288 set insns. Don't initialize sec_mem_p and change_p.
10289 (simplify_operand_subreg): Use LRA_SUBREG_P.
10290 (reg_in_class_p): New function.
10291 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10292 of #ifdef. Add code to remove cycling.
10293 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10294 non-null disp. Reload inner instead of disp when base and index
10295 are null. Try to put lo_sum into register.
10296 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10297 (check_and_process_move): Move code for move cost check to
10298 simple_move_p. Remove equiv_substitution.
10299 (simple_move_p): New function.
10300 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10301 curr_insn_set. Call check_and_process_move only for single set
10302 insns. Use the new function. Move call of check_and_process_move
10303 after operand equiv substitution and address process.
10305 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
10307 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
10308 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
10309 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
10311 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10313 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
10315 2013-04-24 Marek Polacek <polacek@redhat.com>
10317 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
10318 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
10319 (select_loops_exit_conditions): Likewise.
10320 (number_of_iterations_for_all_loops): Likewise.
10321 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
10322 (scev_analysis): Likewise.
10324 2013-04-02 Catherine Moore <clm@codesourcery.com>
10325 Chao-ying Fu <fu@mips.com>
10327 * config/mips/micromips.md (jraddiusp): New pattern.
10328 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
10329 instruction if possible.
10331 2013-04-24 Alan Modra <amodra@gmail.com>
10333 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
10335 2013-04-24 Julian Brown <julian@codesourcery.com>
10336 Chung-Lin Tang <cltang@codesourcery.com>
10338 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
10339 dependency behavior in enumeration type DIE generation. Add TODO note
10340 to comments about future DW_FORM_sdata/udata re-work of related code.
10342 2013-04-23 Lawrence Crowl <crowl@google.com>
10344 * Makefile.in: Update as needed below.
10346 * hash-table.h (class hash_table):
10347 Correct many methods with parameter types compare_type to the correct
10348 value_type. (Correct code was unlikely to notice the change.)
10349 (hash_table::elements_with_deleted) New.
10350 (class hashtable::iterator): New.
10351 (hashtable::begin()): New.
10352 (hashtable::end()): New.
10353 (FOR_EACH_HASH_TABLE_ELEMENT): New.
10355 * statistics.c (statistics_hashes):
10356 Change type to hash_table. Update dependent calls and types.
10358 * tree-into-ssa.c (var_infos):
10359 Change type to hash_table. Update dependent calls and types.
10361 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
10362 Change type to hash_table. Update dependent calls and types.
10364 * tree-ssa-loop-im.c (struct mem_ref.refs):
10365 Change type to hash_table. Update dependent calls and types.
10367 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
10368 Change type to hash_table. Update dependent calls and types.
10370 * tree-ssa-sccvn.c (vn_tables_s::nary):
10371 Change type to hash_table. Update dependent calls and types.
10372 (vn_tables_s::phis): Likewise.
10373 (vn_tables_s::references): Likewise.
10375 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
10376 (vn_reference_eq): Update parameter and return types.
10378 * tree-ssa-structalias.c (pointer_equiv_class_table):
10379 Change type to hash_table. Update dependent calls and types.
10380 (location_equiv_class_table): Likewise.
10382 * tree-vect-data-refs.c: Consequential changes for making
10383 peeling a hash_table.
10385 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
10386 (destroy_loop_vec_info): Dependent hash_table update.
10388 * tree-vectorizer.h (peeling_htab):
10389 Change type to hash_table. Update dependent calls and types.
10391 2013-04-23 Shiva Chen <shiva0217@gmail.com>
10393 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
10394 to check the register content is equal or not.
10395 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10396 to assign register content record.
10397 * lra-eliminations.c (update_reg_eliminate): Use
10398 lra_update_reg_val_offset to update register content offset.
10399 * lra-int.h (struct lra_reg): Add offset member.
10400 (lra_reg_val_equal_p): New static inline function.
10401 (lra_update_reg_val_offset): New static inline function.
10402 (lra_assign_reg_val): New static inline function.
10403 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10404 to assign register content record.
10405 (initialize_lra_reg_info_element): Initial offset to zero.
10407 2013-04-23 Catherine Moore <clm@codesourcery.com>
10409 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10410 operands. Record compression.
10412 2013-04-23 Xinliang David Li <davidxl@google.com>
10414 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10416 2013-04-23 Richard Biener <rguenther@suse.de>
10418 PR middle-end/57036
10419 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10420 parameter, only add abnormal goto edges from the copied body
10421 if the call could perform abnormal gotos.
10422 (copy_cfg_body): Adjust.
10424 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10426 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10428 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10430 * coretypes.h (gimple_stmt_iterator): Add struct to make
10433 2013-04-23 Richard Biener <rguenther@suse.de>
10435 PR tree-optimization/57026
10436 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10437 from SSA names occuring in abnormal PHI nodes.
10439 2013-04-22 Andi Kleen <ak@linux.intel.com>
10441 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10443 2013-04-22 Andi Kleen <ak@linux.intel.com>
10445 * lto/lto.c (print_lto_report_1): Declare early.
10446 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10448 2013-04-22 Andi Kleen <ak@linux.intel.com>
10450 * common.opt (-flto-report-wpa): Add.
10451 * doc/invoke.texi (-flto-report-wpa): Add.
10452 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10455 2013-04-22 Xinliang David Li <davidxl@google.com>
10457 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10458 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10459 * Makefile.in: New dependency
10461 David Daney <ddaney.cavm@gmail.com>
10463 * configure.ac (gcc_cv_as_micromips_support): Use the
10464 --fatal-warnings option.
10465 * configure: Regenerate.
10467 2013-04-22 Marek Polacek <polacek@redhat.com>
10470 * tsan.c (instrument_expr): Don't instrument expression
10471 in case its size is zero.
10473 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10477 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10479 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10481 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10483 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10484 (gimple_stmt_iterator): New typedef.
10485 * gimple.h (gimple_stmt_iterator): Rename to...
10486 (gimple_stmt_iterator_d): ... This.
10487 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10488 trees be valid for GIMPLE and GENERIC.
10489 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10490 * gimple-fold.c (gimple_fold_call): Call target hook
10491 gimple_fold_builtin.
10492 * hooks.c (hook_bool_gsiptr_false): New.
10493 * hooks.h (hook_bool_gsiptr_false): New.
10494 * target.def (fold_stmt): New.
10495 * doc/tm.texi: Regenerate.
10497 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10500 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10501 a set sp if no stack realignment.
10503 2013-04-22 Nick Clifton <nickc@redhat.com>
10505 * config.gcc (tilegx-linux): Extend extra_objs rather than
10507 (tilepro-linux): Likewise.
10509 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10511 * config/aarch64/aarch64-builtins.c
10513 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10514 (VAR<1-12>): Add MAP parameter.
10515 (BUILTIN_*): Likewise.
10516 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10517 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10518 (aarch64_ushl_n<mode>): Likewise.
10519 (aarch64_sshr_n<mode>): Likewise.
10520 (aarch64_ushr_n<mode>): Likewise.
10521 (aarch64_<maxmin><mode>): Likewise.
10522 (aarch64_sqrt<mode>): Likewise.
10523 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10524 (vshr<q>_n_*): Likewise.
10526 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10528 * config/aarch64/aarch64-builtins.c
10529 (aarch64_simd_builtin_type_mode): Handle SF types.
10531 (BUILTIN_GPF): Define.
10532 (aarch64_init_simd_builtins): Handle SF types.
10533 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10534 (frecps): Likewise.
10535 (frecpx): Likewise.
10536 * config/aarch64/aarch64-simd.md
10537 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10538 (aarch64_frecpe<mode>): New.
10539 (aarch64_frecps<mode>): Likewise.
10540 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10541 (v8type): Add frecp<esx>.
10542 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10543 (aarch64_frecps<mode>): Likewise.
10544 * config/aarch64/iterators.md (FRECP): New.
10545 (frecp_suffix): Likewise.
10546 * config/aarch64/arm_neon.h
10547 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10549 2013-04-22 Christian Bruel <christian.bruel@st.com>
10552 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10553 (REG_CLASS_NAMES): Idem.
10554 (REG_CLASS_CONTENTS): Idem.
10555 (REGCLASS_HAS_FP_REG): Idem.
10556 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10557 (sh_conditional_register_usage): Idem.
10559 2013-04-21 Jeff Law <law@redhat.com>
10561 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10562 (ssa_forward_propagate_and_combine): Use it.
10564 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10566 * lra.c: Update the flow chart diagram.
10568 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10570 PR rtl-optimization/56847
10571 * lra-constraints.c (process_alt_operands): Discourage alternative
10572 with non-matche doffsettable memory constraint fro memory with
10575 2013-04-19 Richard Biener <rguenther@suse.de>
10577 PR tree-optimization/56982
10578 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10580 * gimplify.c (gimplify_call_expr): Notice special calls.
10581 (gimplify_modify_expr): Likewise.
10582 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10583 abnormal control flow receivers.
10584 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10585 in the same way as cfun->has_nonlocal_labels.
10586 (gimple_purge_dead_abnormal_call_edges): Likewise.
10587 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10588 receivers start a basic-block.
10590 2013-04-19 Richard Biener <rguenther@suse.de>
10592 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10594 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10595 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10596 (SLP_TREE_LOAD_PERMUTATION): Add.
10597 (vect_transform_slp_perm_load): Adjust prototype.
10598 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10599 (vect_free_slp_instance): Likewise.
10600 (vect_create_new_slp_node): Likewise.
10601 (vect_supported_slp_permutation_p): Remove.
10602 (vect_slp_rearrange_stmts): Adjust.
10603 (vect_supported_load_permutation_p): Likewise. Inline
10604 vect_supported_slp_permutation_p here.
10605 (vect_analyze_slp_instance): Compute load permutations per
10606 slp node instead of per instance.
10607 (vect_get_slp_defs): Adjust.
10608 (vect_transform_slp_perm_load): Likewise.
10609 (vect_schedule_slp_instance): Remove redundant code.
10610 (vect_schedule_slp): Remove hack for PR56270, add it ...
10611 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10612 CSE loads for SLP. Adjust.
10614 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10616 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10617 spelling in two comments.
10619 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10622 * config/arm/arm.c (load_multiple_sequence): Require SP
10623 as base register for loads if SP is in the register list.
10625 2013-04-19 Martin Jambor <mjambor@suse.cz>
10627 PR tree-optimization/56718
10628 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10629 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10630 and made public. Adjusted all callers.
10631 (ipa_intraprocedural_devirtualization): New function.
10632 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10633 (ipa_intraprocedural_devirtualization): Likewise.
10634 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10636 2013-04-19 Richard Biener <rguenther@suse.de>
10638 PR tree-optimization/57000
10639 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10641 2013-04-19 Terry Guo <terry.guo@arm.com>
10643 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10645 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10646 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10647 (cortex_m4_fmacs): Use new reservations.
10648 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10650 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10652 PR rtl-optimization/56999
10653 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10655 (lra_coalesce): Remove split_origin_bitmap and related code.
10656 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10657 ranges if necessary.
10659 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
10661 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10663 (ix86_expand_call): Remove clobbered_registers array and use
10664 x86_64_ms_sysv_extra_clobbered_registers instead.
10665 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10667 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10669 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10670 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10671 (*call_value_rex64_ms_sysv): Ditto.
10673 2013-04-18 Cary Coutant <ccoutant@google.com>
10675 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10676 parent instead of die_mark.
10678 2013-04-18 Diego Novillo <dnovillo@google.com>
10680 * gimple.c (create_gimple_tmp): New.
10681 (get_expr_type): New.
10682 (build_assign): New.
10683 (build_type_cast): New.
10684 * gimple.h (enum ssa_mode): Define.
10685 (gimple_seq_set_location): New.
10686 * asan.c (build_check_stmt): Change some gimple_build_* calls
10687 to use build_assign and build_type_cast.
10689 2013-04-18 Richard Biener <rguenther@suse.de>
10691 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10692 handle negative step. Remove redundant checks.
10693 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10694 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10695 for negative step and grouped loads fail to vectorize.
10697 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10699 * emit-rtl.c (reset_insn_used_flags): New function.
10700 (reset_all_used_flags): Use it.
10701 (verify_insn_sharing): New function.
10702 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10704 2013-04-18 Jakub Jelinek <jakub@redhat.com>
10706 PR tree-optimization/56984
10707 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10708 and (x >> M) >= N don't register any assertion if N << M is the
10711 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10713 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10714 do not delete_insn insns that have not yet been emitted, only
10715 unlink them with remove_insn.
10716 * df-scan.c (df_insn_delete): Revert r197492.
10718 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10720 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10721 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10723 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10725 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10726 define_insn_and_split.
10727 (and_scc,ior_scc,negscc): Likewise.
10728 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10730 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10732 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10733 can benefit from using a sequence of LDRD instructions in epilogue
10734 instead of a single LDM instruction.
10736 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10739 * doc/extend.texi: Fix typo.
10741 2013-04-17 Richard Biener <rguenther@suse.de>
10743 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10744 (vect_build_slp_tree): ... here.
10745 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10746 match. Remove special-casing of mismatched complex loads.
10747 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10748 re-try the match with swapped commutative operands.
10749 (vect_supported_load_permutation_p): Remove special-casing of
10750 mismatched complex loads.
10751 (vect_analyze_slp_instance): Adjust.
10753 2013-04-17 Richard Biener <rguenther@suse.de>
10755 PR rtl-optimization/56921
10756 * cfgloop.h (struct loop): Add simple_loop_desc member.
10757 (struct niter_desc): Mark with GTY(()).
10758 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10759 * loop-iv.c (get_simple_loop_desc): Likewise.
10760 (free_simple_loop_desc): Likewise.
10763 2013-04-16 Richard Biener <rguenther@suse.de>
10765 PR rtl-optimization/56921
10766 * loop-init.c (pass_rtl_move_loop_invariants): Add
10767 TODO_do_not_ggc_collect to todo_flags_finish.
10768 (pass_rtl_unswitch): Same.
10769 (pass_rtl_unroll_and_peel_loops): Same.
10770 (pass_rtl_doloop): Same.
10772 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
10774 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
10775 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
10776 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
10778 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
10779 * tree-streamer.c (record_common_node): Adjust reference in comment.
10781 2013-04-17 Terry Guo <terry.guo@arm.com>
10783 * config/arm/cortex-m4.md: Add a new bypass.
10785 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10787 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
10789 (*subs_<optab><mode>_multp2): New pattern.
10790 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10791 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
10793 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10795 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
10796 (*subs_mul_imm_<mode>): New pattern.
10798 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
10801 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
10802 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
10803 (vsx_movti_32bit): Change j->wa to O->wa.
10805 2013-04-16 Richard Biener <rguenther@suse.de>
10807 PR rtl-optimization/56921
10808 * loop-init.c (pass_rtl_move_loop_invariants): Add
10809 TODO_do_not_ggc_collect to todo_flags_finish.
10810 (pass_rtl_unswitch): Same.
10811 (pass_rtl_unroll_and_peel_loops): Same.
10812 (pass_rtl_doloop): Same.
10814 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
10816 * config/arm/arm.c (emit_multi_reg_push): New declaration
10817 for an existing function.
10818 (arm_emit_strd_push): New function.
10819 (arm_expand_prologue): Used here.
10820 (arm_emit_ldrd_pop): New function.
10821 (arm_expand_epilogue): Used here.
10822 (arm_get_frame_offsets): Update condition.
10823 (arm_emit_multi_reg_pop): Add a special case for load of a single
10824 register with writeback.
10826 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
10828 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
10831 2013-04-16 Richard Biener <rguenther@suse.de>
10833 PR tree-optimization/56756
10834 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
10835 (first_mem_ref_loc): New.
10836 (execute_sm): Place the load temporarily before a previous
10837 access instead of in the latch edge to ensure its SSA dependencies
10838 are defined at points dominating the load.
10840 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
10842 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
10843 correct fix by moving header and footer insn to the footer of
10844 the merged basic block. Clear BB_END of the merged-away block.
10846 PR middle-end/43631
10847 * emit-rtl.c (make_note_raw): New function.
10848 (link_insn_into_chain): New static inline function.
10849 (add_insn): Use it.
10850 (add_insn_before, add_insn_after): Factor insn chain linking code...
10851 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
10852 using link_insn_into_chain.
10853 (note_outside_basic_block_p): New helper function for emit_note_after
10854 and emit_note_before.
10855 (emit_note_after): Use nobb variant of add_insn_after if the note
10856 should not be contained in a basic block.
10857 (emit_note_before): Use nobb variant of add_insn_before if the note
10858 should not be contained in a basic block.
10859 (emit_note_copy): Use make_note_raw.
10860 (emit_note): Likewise.
10861 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
10862 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
10863 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
10864 the moved barrier the tail of the basic block it follows.
10865 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
10867 2013-04-15 Jakub Jelinek <jakub@redhat.com>
10869 PR tree-optimization/56962
10870 * gimple-ssa-strength-reduction.c (record_increment): Only set
10871 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
10872 either rhs1 or rhs2 is equal to c->base_expr.
10874 2013-04-15 Richard Biener <rguenther@suse.de>
10876 PR tree-optimization/56933
10877 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
10879 (GROUP_READ_WRITE_DEPENDENCE): Remove.
10880 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
10881 * tree-vect-data-refs.c (vect_analyze_group_access): Move
10882 dependence check ...
10883 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
10885 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
10886 GROUP_READ_WRITE_DEPENDENCE.
10888 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10890 * emit-rtl.c (reset_all_used_flags): New function.
10891 (verify_rtl_sharing): Call reset_all_used_flags before and after
10892 performing the checks.
10894 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10896 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
10897 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
10898 * config/arm/constraints.md (De): New constraint.
10899 * config/arm/neon.md (anddi3_neon): Delete.
10900 (neon_vand<mode>): Expand to standard anddi3 pattern.
10901 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
10902 Move earlier in the file.
10903 (neon_inv_logic_op2): Likewise.
10904 (arm_anddi_operand_neon): New predicate.
10906 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10908 * configure.ac (gcc_cv_ld_as_needed): Set
10909 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
10910 Use -z ignore, -z record on *-*-solaris2*.
10911 (HAVE_LD_AS_NEEDED): Update comment.
10912 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
10913 * configure: Regenerate.
10914 * config.in: Regenerate.
10915 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
10916 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
10917 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
10918 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
10919 equivalents. Fix markup.
10920 * doc/tm.texi: Regenerate.
10922 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
10924 * config/i386/i386.opt: New option mstack-protector-guard=.
10925 * config/i386/i386-opts.h: Add enum stack_protector_guard.
10926 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
10927 TARGET_SSP_TLS_GUARD.
10928 * config/i386/i386.c (ix86_option_override_internal): Set
10929 ix86_stack_protector_guard.
10930 * config/i386/i386.md (stack_protect_set): Enable for
10931 TARGET_SSP_TLS_GUARD only.
10932 (stack_protect_set_<mode>): Ditto.
10933 (stack_protect_test): Ditto.
10934 (stack_protect_test_<mode>): Ditto.
10935 * doc/invoke.texi (i386 Option): Document.
10937 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
10940 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
10941 (S_MODES): Set H_MODE bit.
10942 (SF_MODES): Set only S_MODE and SF_MODE bits.
10943 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
10944 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
10945 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
10946 <MODE_FLOAT>: Likewise.
10948 2013-04-15 Joey Ye <joey.ye@arm.com>
10950 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
10952 2013-04-15 Joey Ye <joey.ye@arm.com>
10954 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
10956 (thumb_far_jump_used_p): Count instruction size and set
10959 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
10961 * reorg.c (fill_simple_delay_slots): Reindent block of code.
10962 * resource.c (mark_target_live_regs): Reformat conditional block.
10964 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10966 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
10967 notes, they are emitted only just before final.
10968 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
10970 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
10972 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
10973 * cfgrtl.c (delete_insn): Call it here instead.
10974 * lra-spills.c (lra_final_code_change): Use delete_insn.
10975 * haifa-sched.c (sched_remove_insn): Likewise.
10976 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
10977 returning to the nop pool.
10978 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
10979 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
10981 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10983 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
10984 * doc/tm.texi: Regenerated.
10986 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
10988 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
10991 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
10993 * df-core.c (df_find_def): Compare register numbers.
10994 (df_find_use): Likewise.
10996 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
10999 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11000 lra_in_progress for return.
11002 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11004 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11005 define_insn into define_insn_and_split and emit movsicc patterns.
11007 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11009 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11011 2013-04-12 Richard Biener <rguenther@suse.de>
11013 * tree-pass.h (TODO_do_not_ggc_collect): New.
11014 * passes.c (execute_one_ipa_transform_pass): Honor
11015 TODO_do_not_ggc_collect.
11016 (execute_one_pass): Likewise.
11019 2013-04-10 Richard Biener <rguenther@suse.de>
11021 * passes.c (init_optimization_passes): Remove reload pass.
11022 * ira.c (do_reload): Merge into ...
11024 (rest_of_handle_reload): Remove.
11025 (pass_reload): Likewise.
11026 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11027 of reload for vzeroupper pass placement.
11029 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11031 PR tree-optimization/56918
11032 PR tree-optimization/56920
11033 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11034 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11035 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11036 use wide_mul_with_sign method.
11038 2013-04-12 Richard Biener <rguenther@suse.de>
11040 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11041 not be considered a gimple constant.
11043 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11045 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11046 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11049 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11050 Jakub Jelinek <jakub@redhat.com>
11052 * opts.c: Include diagnostic-color.h.
11053 (common_handle_option): Handle OPT_fdiagnostics_color_.
11054 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11055 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11056 (diagnostic-color.o): New.
11057 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11058 (diagnostic_color_rule): New enum.
11059 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11060 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11061 the location string.
11062 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11063 either NULL, or color kind.
11064 * diagnostic-color.c: New file.
11065 * diagnostic-color.h: New file.
11066 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11068 * doc/invoke.texi (-fdiagnostics-color): Document.
11069 * pretty-print.h (pp_show_color): Define.
11070 (struct pretty_print_info): Add show_color field.
11071 * diagnostic.c: Include diagnostic-color.h.
11072 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11073 macros. Colorize error:, warning: etc. strings and also the location
11075 (diagnostic_show_locus): Colorize the caret line.
11076 * pretty-print.c: Include diagnostic-color.h.
11077 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11078 inside of %< %> quotes or quoted through q format modifier.
11080 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11082 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11084 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11086 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11087 code in CC_NZ mode.
11088 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11091 2013-04-11 Marek Polacek <polacek@redhat.com>
11093 PR tree-optimization/48184
11094 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11096 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11098 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11099 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11100 (skip_simple_arithmetic): Tidy up.
11101 * tree.h (skip_simple_constant_arithmetic): Declare.
11103 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11105 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11107 2013-04-11 Richard Biener <rguenther@suse.de>
11109 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11110 generate vector constants.
11112 2013-04-11 Richard Biener <rguenther@suse.de>
11114 PR tree-optimization/56878
11115 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11116 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11118 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11119 Prefer to align the DR with the most invariant base address.
11121 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11123 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11126 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11128 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11129 floating-point vector comparisons against 0.
11131 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11133 PR tree-optimization/56899
11134 * fold-const.c (extract_muldiv_1): Apply distributive law
11135 only if TYPE_OVERFLOW_WRAPS (ctype).
11137 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11140 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11141 loading has side effect.
11143 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
11145 * config/sparc/sparc.c: Include tree-pass.h.
11146 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
11147 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
11148 head of file. Change return type. Split off gate function.
11149 (sparc_gate_work_around_errata): New function.
11150 (pass_work_around_errata): New pass definition.
11151 (insert_pass_work_around_errata) New pass insert definition to
11152 insert pass_work_around_errata just after delayed-branch scheduling.
11153 (sparc_option_override): Insert the pass.
11154 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
11156 2013-04-10 David S. Miller <davem@davemloft.net>
11158 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
11159 or -mcpu=hypersparc.
11161 * target.def (cstore_mode): New hook.
11162 * target.h: Include insn-codes.h
11163 * targhooks.c: Likewise.
11164 (default_cstore_mode): New function.
11165 * targhooks.h: Declare it.
11166 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
11167 * doc/tm.texi: Rebuild.
11168 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
11169 target hook, rather than inspecting the insn_data.
11170 * config/sparc/sparc.c (sparc_cstore_mode): New function.
11171 (TARGET_CSTORE_MODE): Redefine.
11172 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
11174 * config/sparc/predicates.md (cstore_result_operand): New special
11176 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
11177 Use it for operand 0.
11178 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
11179 (*snesi_special): Likewise.
11180 (*snesi_zero): Likewise.
11181 (*seqsi_zero): Likewise.
11182 (*sltu_insn): Likewise.
11183 (*sgeu_insn): Likewise.
11184 (*seqdi_special): Make operand 0 and comparison operation be of
11186 (*snedi_special): Likewise.
11187 (*snedi_special_vis3): Likewise.
11188 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
11189 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
11190 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
11191 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
11192 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
11193 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
11194 (*sltu_extend_sp64): Likewise.
11195 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
11196 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
11197 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
11198 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
11199 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
11201 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
11203 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
11204 (aarch64_start_file): Use the new function.
11206 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11207 Jason Merrill <jason@redhat.com>
11209 * common.opt: Add -gdwarf.
11210 * opts.c (common_handle_option): Handle it.
11211 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
11213 2013-04-10 Richard Biener <rguenther@suse.de>
11215 * passes.c (execute_todo): Do not call ggc_collect conditional here.
11216 (execute_one_ipa_transform_pass): But unconditionally here.
11217 (execute_one_pass): And here.
11218 (init_optimization_passes): Remove reload pass.
11219 * tree-pass.h (TODO_ggc_collect): Remove.
11220 (pass_reload): Likewise.
11221 * ira.c (do_reload): Merge into ...
11223 (rest_of_handle_reload): Remove.
11224 (pass_reload): Likewise.
11225 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11226 of reload for vzeroupper pass placement.
11227 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
11228 and todo_flags_finish of all passes.
11230 2013-04-10 Richard Biener <rguenther@suse.de>
11232 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
11233 first_const_oprnd field, rename first_def_type to first_op_type.
11234 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
11235 (vect_get_and_check_slp_defs): Always use the type of the
11236 operand. Allow mixed vect_external_def, vect_constant_def types.
11237 (vect_get_constant_vectors): Handle mixed vect_external_def,
11238 vect_constant_def types.
11240 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
11242 PR tree-optimization/55524
11243 * tree-ssa-math-opts.c
11244 (convert_mult_to_fma): Don't use an fms construct
11245 when we don't have an fms operation, but fnma, and it looks
11246 likely that we'll be able to use the latter.
11248 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11250 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
11252 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
11253 inline fail caused by overwritable functions.
11255 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
11257 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
11258 unnecessary bits in the constant power of two case.
11260 2013-04-10 Richard Biener <rguenther@suse.de>
11262 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
11263 broken code swapping operands.
11264 (vect_build_slp_tree): Do not compute load permutations here.
11265 (vect_analyze_slp_instance): Compute load permutations here,
11266 after building the SLP tree.
11268 2013-04-09 Christian Bruel <christian.bruel@st.com>
11270 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
11271 of next/prev_real_insn.
11273 2013-04-09 Jan Hubicka <jh@suse.cz>
11275 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11276 Drop aliased parameter.
11277 (function_and_variable_visibility): Do not handle alias pairs.
11278 * cgraph.c (varpool_externally_visible_p): Update prototype.
11279 * varpool.c (varpool_add_new_variable): Update.
11281 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11283 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
11285 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
11287 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
11289 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
11291 2013-04-09 Marek Polacek <polacek@redhat.com>
11293 PR tree-optimization/48762
11294 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
11296 2013-04-09 Richard Biener <rguenther@suse.de>
11298 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
11300 (vect_build_slp_tree): Likewise.
11301 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
11302 calculating the cost of a SLP instance.
11303 (vect_analyze_slp_instance): Use it from here, after building
11306 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11308 PR middle-end/56883
11309 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
11310 expand_omp_for_static_chunk): Use simple_p = true in
11311 force_gimple_operand_gsi calls when assigning to addressable decls.
11313 2013-04-09 Jeff Law <law@redhat.com>
11315 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
11316 when the boolean was created by converting a wider object which
11317 had a boolean range.
11319 2013-04-09 Richard Biener <rguenther@suse.de>
11321 * tree-vectorizer.h (slp_void_p): Remove.
11322 (slp_tree): Typedef before _slp_tree declaration.
11323 (struct _slp_tree): Use a vector of slp_tree as children.
11324 (vect_get_place_in_interleaving_chain): Remove.
11325 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
11327 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
11329 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
11330 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
11331 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
11332 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
11333 Use slp_node instead of slp_void_p and adjust.
11335 2013-04-09 Richard Biener <rguenther@suse.de>
11337 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
11338 work that is not necessary.
11340 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11342 PR tree-optimization/56854
11343 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
11344 forward into clobber stmts if it would change MEM_REF lhs into
11347 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
11349 * tree.c (type_hash_lookup, type_hash_add): Make static.
11350 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
11352 2013-04-09 Richard Biener <rguenther@suse.de>
11354 * tree.h (unsave_expr_now): Remove.
11355 * tree-inline.c (mark_local_for_remap_r): Remove.
11356 (unsave_expr_1): Likewise.
11357 (unsave_r): Likewise.
11358 (unsave_expr_now): Likewise.
11359 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
11360 (propagate_tree_value): Likewise.
11362 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
11364 * doc/rtl.texi (sequence): Rewrite documentation to match the
11365 current use of SEQUENCE rtl objects.
11366 * rtl.def (SEQUENCE): Likewise.
11368 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
11369 Update documentation.
11370 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
11371 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
11373 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
11375 2013-04-08 Teresa Johnson <tejohnson@google.com>
11377 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
11378 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
11380 (estimate_edge_size_and_time): Add comment to suggest using rounding
11382 (estimate_node_size_and_time): Ditto.
11383 (remap_edge_change_prob): Use helper rounding divide methods.
11384 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
11385 (gimple_mod_pow2_value_transform): Ditto.
11386 (gimple_mod_subtract_transform): Ditto.
11387 (gimple_ic_transform): Ditto.
11388 (gimple_stringops_transform): Ditto.
11389 * stmt.c (conditional_probability): Ditto.
11390 (emit_case_dispatch_table): Ditto.
11391 * lto-cgraph.c (merge_profile_summaries): Ditto.
11392 * tree-optimize.c (execute_fixup_cfg): Ditto.
11393 * cfgcleanup.c (try_forward_edges): Ditto.
11394 * cfgloopmanip.c (scale_loop_profile): Ditto.
11396 (duplicate_loop_to_header_edge): Ditto.
11397 (lv_adjust_loop_entry_edge): Ditto.
11398 * tree-vect-loop.c (vect_transform_loop): Ditto.
11399 * profile.c (compute_branch_probabilities): Ditto.
11400 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11401 * lto-streamer-in.c (input_cfg): Ditto.
11402 * gimple-streamer-in.c (input_bb): Ditto.
11403 * ipa-cp.c (update_profiling_info): Ditto.
11404 (update_specialized_profile): Ditto.
11405 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11406 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11408 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11409 (compute_trg_info): Ditto.
11410 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11411 (purge_dead_edges): Ditto.
11412 * loop-unswitch.c (unswitch_loop): Ditto.
11413 * cgraphclones.c (cgraph_clone_edge): Ditto.
11414 (cgraph_clone_node): Ditto.
11415 * tree-inline.c (copy_bb): Ditto.
11416 (copy_edges_for_bb): Ditto.
11417 (initialize_cfun): Ditto.
11418 (copy_cfg_body): Ditto.
11419 (expand_call_inline): Ditto.
11421 2013-04-08 Kai Tietz <ktietz@redhat.com>
11423 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11424 TARGET_CYGWIN64 by TARGET_64BIT.
11426 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11428 * config/epiphany/epiphany.md (GPR_1): New constant.
11429 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11430 * config/epiphany/epiphany.c (gen_compare_reg):
11431 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11432 is already in place.
11433 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11434 Don't require being called during rtl expansion; If y operlaps r0,
11436 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11437 (epiphany_expand_epilogue): Likewise.
11439 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11440 Don't use CC_FPmode for ORDERED / UNORDERED.
11441 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11443 * config/epiphany/constraints.md (CnL): New constraint.
11444 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11445 * config/epiphany/predicates.md (add_operand): Allow 1024.
11447 * config/epiphany/epiphany.md (logical_op): New code iterator.
11448 (op_mnc): New code attribute.
11449 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11450 (mov_f+1, mov_f+2): New peephole2 patterns.
11452 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11453 (cstoresi4): Also allow re-use of zero result when doing a NE
11454 comparison to a non-zero operand.
11455 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11457 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11458 Use gen_addsi3_i / gen_subsi3_i.
11460 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11464 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11465 contain anything but clobbers, at most one __builtin_stack_restore,
11466 optionally debug stmts and final resx, and if it has at least one
11467 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11468 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11469 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11470 which isn't defaut definition, remove them.
11471 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11472 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11473 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11474 with MEM_REF LHS with SSA_NAME address.
11476 2013-04-08 Jeff Law <law@redhat.com>
11478 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11480 2013-04-08 Richard Biener <rguenther@suse.de>
11482 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11484 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11485 determined vector type.
11486 (vect_analyze_data_refs): Likewise.
11487 (vect_get_new_vect_var): Adjust.
11488 (vect_create_destination_var): Preserve SSA name versions.
11489 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11490 not dump anything here.
11492 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11494 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11495 Add member lr_slot_known.
11496 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11498 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11499 Remove code that sets lr_slot_offset according to what a previous
11500 version of epiphany_emit_save_restore used to do.
11501 (epiphany_emit_save_restore): When doing an lr save or restore,
11502 set/verify lr_slot_known and lr_slot_offset.
11504 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11507 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11510 2013-04-08 Richard Biener <rguenther@suse.de>
11512 * alias.c (find_base_term): Fix thinko in previous change.
11514 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11516 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11517 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11518 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11519 if possible to compute val.
11520 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11521 For QImode integers don't require anything about precision. Use
11522 const_with_all_bytes_same to find out if the constant doesn't have
11523 repeated bytes in it.
11525 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11527 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11530 2013-04-08 Marek Polacek <polacek@redhat.com>
11532 PR rtl-optimization/48182
11533 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11536 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11539 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11540 nuses, make sure we have a label.
11542 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11545 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11546 (rs6000_emit_swdiv_low_precision): Remove.
11547 (rs6000_emit_swdiv): Rewrite to handle between one and four
11548 iterations of Newton-Raphson generally; modify required number of
11549 iterations for some cases.
11550 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11552 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11554 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11555 set-but-unused variable.
11557 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11558 basic blocks of released function bodies garbage-collectable.
11560 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11561 (struct rtl_opt_pass): Add TODO_df_finish.
11563 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11565 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11567 * config/arm/constraints.md (q): New constraint.
11568 * config/arm/ldrdstrd.md: New file.
11569 * config/arm/arm.md (ldrdstrd.md) New include.
11570 (arm_movdi): Use "q" instead of "r" constraint
11571 for double-word memory access.
11572 (movdf_soft_insn): Likewise.
11573 * config/arm/vfp.md (movdi_vfp): Likewise.
11574 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
11575 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
11576 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11577 (mem_ok_for_ldrd_strd): Likewise.
11578 (output_move_double): Update assertion.
11580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11582 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11584 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11586 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11587 define_insn_and_split.
11588 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11590 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11592 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11593 define_insn_and_split.
11594 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
11595 (shiftsi3_compare): New pattern.
11596 (rrx): New pattern.
11597 * config/arm/unspecs.md (UNSPEC_RRX): New.
11599 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11601 * config/arm/arm.md (negdi_extendsidi): New pattern.
11602 (negdi_zero_extendsidi): Likewise.
11604 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11606 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11607 define_insn_and_split.
11608 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11609 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11611 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11613 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11614 define_insn_and_split.
11615 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11616 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11618 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11620 * config/arm/arm.md (subsi3_carryin): New pattern.
11621 (subsi3_carryin_const): Likewise.
11622 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11623 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11625 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11627 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11629 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11631 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
11632 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
11634 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11636 * config/arm/arm.c (arm_expand_builtin): Change fcode
11637 type to unsigned int.
11639 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11641 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11643 2013-04-04 Ian Lance Taylor <iant@google.com>
11645 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11648 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11650 PR middle-end/56729
11651 * df-scan.c (df_insn_delete): Disable failing assert.
11653 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11655 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11656 New function prototype.
11657 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11658 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11659 (arm_builtin_vectorized_function): New function.
11661 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11663 * config/arm/arm_neon_builtins.def: New file.
11664 * config/arm/arm.c (neon_builtin_data): Move contents to
11665 arm_neon_builtins.def.
11666 (enum arm_builtins): Include neon builtin definitions.
11667 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
11668 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
11670 2013-04-04 Marek Polacek <polacek@redhat.com>
11672 PR tree-optimization/48186
11673 * predict.c (maybe_hot_frequency_p): Return false if
11674 HOT_BB_FREQUENCY_FRACTION is 0.
11675 (cgraph_maybe_hot_edge_p): Likewise.
11677 2013-04-04 Richard Biener <rguenther@suse.de>
11679 PR tree-optimization/56826
11680 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11683 2013-04-04 Richard Biener <rguenther@suse.de>
11685 PR tree-optimization/56213
11686 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
11687 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
11689 2013-04-04 Richard Biener <rguenther@suse.de>
11691 PR tree-optimization/56837
11692 * tree-loop-distribution.c (classify_partition): For non-zero
11693 values require that the value has the same precision as its
11694 mode to be useful as memset value.
11696 2013-04-03 Nick Clifton <nickc@redhat.com>
11698 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
11699 (fmssf4): Use fmsf.s on E3V5 architectures.
11700 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11701 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11703 2013-04-03 Jeff Law <law@redhat.com>
11705 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11706 (lra-eliminations.o): Likewise.
11708 2013-04-03 Teresa Johnson <tejohnson@google.com>
11710 * gcov-io.c (compute_working_sets): Moved most of body of old
11711 compute_working_sets here from profile.c.
11712 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
11713 (gcov_working_set_t): Moved typedef here from basic-block.h
11714 (compute_working_set): Declare.
11715 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11716 (get_working_sets): Renamed from compute_working_set,
11717 replace most of body with call to new compute_working_sets.
11718 (get_exec_counts): Replace call to compute_working_sets
11719 to get_working_sets.
11720 * profile.h (get_working_sets): Renamed from compute_working_set.
11721 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
11722 to get_working_sets.
11723 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11724 * gcov-dump.c (dump_working_sets): New function.
11726 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11728 * hwint.c (sext_hwi, zext_hwi): New functions.
11729 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11730 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11731 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11732 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11733 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11734 (sext_hwi, zext_hwi): New functions.
11736 2013-04-03 Jeff Law <law@redhat.com>
11738 PR tree-optimization/56799
11739 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11740 back test for widening conversion erroneously dropped in prior change.
11742 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11745 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11746 instead of next_real_insn.
11748 2013-04-03 Marek Polacek <polacek@redhat.com>
11751 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
11753 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11756 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11758 (thumb1_output_casesi): Likewise.
11759 (thumb2_output_casesi): Likewise.
11761 2013-04-03 Richard Biener <rguenther@suse.de>
11763 PR tree-optimization/56817
11764 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11766 (tree_unroll_loops_completely_1): ... new function to manually
11767 walk the loop tree, properly defering outer loops of unrolled
11768 loops to later iterations.
11770 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11772 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
11773 (vectorizable_load): Likewise.
11774 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
11775 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
11777 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11779 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
11782 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11784 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
11786 2013-04-03 Bin Cheng <bin.cheng@arm.com>
11788 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
11790 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11792 PR tree-optimization/56790
11793 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
11796 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
11798 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
11800 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
11801 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
11804 2013-04-03 Jakub Jelinek <jakub@redhat.com>
11807 * tree.h (force_folding_builtin_constant_p): New decl.
11808 * builtins.c (force_folding_builtin_constant_p): New variable.
11809 (fold_builtin_constant_p): Fold immediately also if
11810 force_folding_builtin_constant_p.
11812 2013-04-03 Richard Biener <rguenther@suse.de>
11814 PR tree-optimization/56812
11815 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
11816 DRs of the same interleaving chain are independent.
11818 2013-04-02 Jason Merrill <jason@redhat.com>
11820 * gdbinit.in (pbb): Use debug fn.
11822 2013-04-02 Lawrence Crowl <crowl@google.com>
11824 * sese.h (struct ivtype_map_elt_s): Remove unused.
11825 (extern debug_ivtype_map): Remove unused.
11826 (extern eq_ivtype_map_elts): Remove unused.
11827 * sese.c (debug_ivtype_map): Removed unused.
11828 (debug_ivtype_map_1): Removed unused.
11829 (debug_ivtype_elt): Remove unused.
11830 (eq_ivtype_map_elts): Remove unused.
11833 2013-04-02 Kai Tietz <ktietz@redhat.com>
11836 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
11837 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
11838 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
11840 (legitimize_pe_coff_symbol): Likewise.
11841 (is_imported_p): New helper-function.
11842 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
11843 for Windows x64 targets.
11844 (ix86_expand_prologue): Optimize for pe-coff targets.
11845 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
11846 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
11847 medium/large code-model.
11848 (legitimize_pic_address): Likewise.
11849 (legitimize_tls_address): Likewise.
11850 (ix86_expand_call): Likewise.
11851 (x86_output_mi_thunk): Likewise.
11852 (get_dllimport_decl): Add new beimport argument.
11853 (construct_plt_address): Don't assert for x64 pe-coff targets.
11854 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
11856 (SYMBOL_FLAG_STUBVAR): New macro.
11857 (SYMBOL_REF_STUBVAR_P): Likewise.
11858 * config/i386/winnt.c (stub_list): New structure.
11859 (stub_head): New local variable.
11860 (i386_pe_record_stub): New function.
11861 (i386_pe_file_end): Emit refptr-stubs.
11863 2013-04-02 Jakub Jelinek <jakub@redhat.com>
11865 PR rtl-optimization/56745
11866 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
11867 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
11870 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
11871 and both of them are MEM_REFs, just compare first argument for
11872 equality and attempt to deal even with differing offsets.
11875 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
11876 of gimple_clobber_p to be MEM_REF.
11877 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
11878 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
11879 after gimplification.
11880 * asan.c (get_mem_ref_of_assignment): Don't instrument
11881 gimple_clobber_p stmts.
11882 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
11883 gimple_clobber_p stmt if they have MEM_REF lhs and
11884 are dead because of another gimple_clobber_p stmt.
11885 * tree-ssa-live.c (clear_unused_block_pointer): Treat
11886 gimple_clobber_p stmts like debug stmts.
11887 (remove_unused_locals): Remove clobbers with MEM_REF lhs
11888 that refer to unused VAR_DECLs or uninitialized values.
11889 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
11890 gimple_clobber_p stmts if they refer to removed parameters.
11891 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
11894 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
11896 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
11897 using SWI48 mode attribute.
11899 2013-04-02 Wei Mi <wmi@google.com>
11901 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
11902 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
11903 *<rotate_insn><mode>3_mask in i386.md.
11905 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
11907 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
11909 2013-04-02 Richard Biener <rguenther@suse.de>
11911 PR tree-optimization/56778
11912 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
11913 Runtime alias tests are not supported for gather loads.
11914 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
11915 stmts referenced from SSA operands before updating SSA form.
11917 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
11918 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11920 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
11921 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
11922 * config/arm/cortex-a53.md: New file.
11923 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
11924 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
11925 * config/arm/arm.c (arm_issue_rate): Likewise.
11926 * config/arm/arm-tune.md: Regenerate
11927 * config/arm/arm-tables.opt: Regenerate.
11928 * config/arm/arm-cores.def: Add cortex-a53.
11930 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
11932 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
11935 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11937 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
11938 scalar load/store operations using B/H registers.
11939 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
11941 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
11943 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
11945 * config/aarch64/aarch64.c
11946 (aarch64_simd_scalar_immediate_valid_for_move): New.
11947 * config/aarch64/aarch64-protos.h
11948 (aarch64_simd_scalar_immediate_valid_for_move): New.
11949 * config/aarch64/constraints.md (Dh, Dq): New.
11950 * config/aarch64/iterators.md (hq): New.
11952 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
11954 * reorg.c (get_branch_condition): Deal with conditional returns.
11955 (fill_simple_delay_slots): Remove dead code dealing with jumps.
11957 2013-04-01 Wei Mi <wmi@google.com>
11959 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
11960 Truncate operand 2 using %b asm operand modifier.
11961 (*<shift_insn><mode>3_mask): Ditto.
11962 (*<rotate_insn><mode>3_mask): Ditto.
11964 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
11966 PR middle-end/56798
11967 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
11969 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
11971 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
11973 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
11975 2013-03-30 Lawrence Crowl <crowl@google.com>
11977 * dse.c (clear_alias_sets): Remove never set.
11978 (disqualified_clear_alias_sets): Remove never set.
11979 (clear_alias_mode_pool): Remove never set.
11980 (dse_step0): Remove condition that is never true.
11981 (canon_address): Remove condition that is never true.
11982 (dse_step7): Remove condition that is never true.
11983 (rest_of_handle_dse): Remove condition that is never true.
11984 (rest_of_handle_dse::did_global): Remove never read from above.
11985 (dse_step2_spill): Remove never called from above.
11986 (dse_step5_spill): Remove never called from above.
11988 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
11990 * doc/md.texi (Standard Names) <casesi>: Update documentation for
11991 JUMP_TABLE_DATA changes.
11992 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
11993 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
11994 (Insns) <jump_table_data>: New entry.
11995 * doc/tm.texi: Regenerate.
11997 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
11999 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12000 for table jump at the end of a basic block using tablejump_p.
12001 * targhooks.c (default_invalid_within_doloop): Likewise.
12002 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12003 target hook implementation that is identical to the default hook.
12004 (rs6000_invalid_within_doloop): Remove.
12006 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12007 unused variable from tablejump_p call.
12009 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12010 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12011 (INSN_DELETED_P): Likewise.
12012 (emit_jump_table_data): New prototype.
12013 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12014 after 4th as unused.
12015 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12016 * sched-vis.c (print_insn): Likewise.
12017 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12018 insn for compatibility with back ends that use next_active_insn to
12019 identify jump table data.
12020 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12021 (remove_insn): Likewise.
12022 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12024 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12025 (emit_jump_table_data): New function.
12027 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12028 basic block, a JUMP_TABLE_DATA never is.
12029 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12030 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12031 off from code handling real insns.
12032 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12033 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12034 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12035 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12036 is not a NONDEBUG_INSN_P.
12037 * ira-costs.c (scan_one_insn): Likewise.
12038 * jump.c (mark_all_labels): Likewise.
12039 (mark_jump_label_1): Likewise.
12040 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12041 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12043 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12044 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12045 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12046 JUMP_TABLE_DATA_P insns.
12047 (calculate_elim_costs_all_insns): Likewise.
12048 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12049 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12050 (delete_output_reload): Code style fixups.
12051 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12052 insn flags on this non-insn.
12053 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12054 as scheduling barriers, for pre-change compatibility.
12055 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12056 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12058 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12059 redundant JUMP_TABLE_DATA_P test.
12060 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12061 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12062 (frv_for_each_packet): Likewise.
12063 * config/i386/i386.c (min_insn_size): Likewise.
12064 (ix86_avoid_jump_mispredicts): Likewise.
12065 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12066 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12067 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12068 (mips16_insn_length): Robustify.
12069 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12070 (mips16_split_long_branches): Likewise.
12071 * config/pa/pa.c (pa_combine_instructions): Likewise.
12072 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12073 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12074 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12075 as contributing to pool range lengths.
12076 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12077 Remove redundant JUMP_TABLE_DATA_P test.
12078 (sh_loop_align): Likewise.
12079 (split_branches): Likewise.
12080 (sh_insn_length_adjustment): Likewise.
12081 * config/spu/spu.c (get_branch_target): Likewise.
12083 2013-03-29 Jan Hubicka <jh@suse.cz>
12085 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12086 gcov streaming; stream hot bb threshold to ltrans.
12087 * predict.c (get_hot_bb_threshold): Break out from ....
12088 (maybe_hot_count_p): ... here.
12089 (set_hot_bb_threshold): New function.
12090 * lto-section-in.c (lto_section_name): Add profile.
12091 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12092 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12093 and data-streamer.h
12094 (histogram_entry): New structure.
12095 (histogram, histogram_pool): New global vars.
12096 (histogram_hash): New structure.
12097 (histogram_hash::hash): New method.
12098 (histogram_hash::equal): Likewise.
12099 (account_time_size): New function.
12100 (cmp_counts): New function.
12101 (dump_histogram): New function.
12102 (ipa_profile_generate_summary): New function.
12103 (ipa_profile_write_summary): New function.
12104 (ipa_profile_read_summary): New function.
12105 (ipa_profile): Decide on threshold.
12106 (pass_ipa_profile): Add ipa_profile_write_summary and
12107 ipa_profile_read_summary.
12108 * Makefile.in (ipa.o): Update dependencies.
12109 * lto-streamer.h (LTO_section_ipa_profile): New section.
12111 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12113 * tree.h (VAR_P): New.
12115 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12118 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12120 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12122 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12123 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12124 (control_flow_insn_p): Likewise.
12125 * cfgrtl.c (duplicate_insn_chain): Likewise.
12126 * final.c (get_attr_length_1): Likewise.
12127 (shorten_branches): Likewise.
12128 (final_scan_insn): Likewise.
12129 * function.c (instantiate_virtual_regs): Likewise.
12130 * gcse.c (insert_insn_end_basic_block): Likewise.
12131 * ira-costs.c (scan_one_insn): Likewise.
12132 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12133 * lra.c (check_rtl): Likewise.
12134 * reload1.c (elimination_costs_in_insn): Likewise.
12135 * reorg.c (follow_jumps): Likewise.
12137 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12138 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12139 (thumb_far_jump_used_p): Likewise.
12140 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12141 (workaround_speculation): Likewise.
12142 (add_sched_insns_for_speculation): Likewise.
12143 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
12144 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12145 (frv_for_each_packet): Likewise.
12146 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
12147 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
12148 (final_emit_insn_group_barriers): Likewise.
12149 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12150 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12151 (mips16_insn_length): Likewise.
12152 * config/pa/pa.c (pa_reorg): Likewise.
12153 (pa_combine_instructions): Likewise.
12154 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
12155 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
12156 (sh_reorg): Likewise.
12157 (split_branches): Likewise.
12158 * config/spu/spu.c (get_branch_target): Likewise.
12160 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
12163 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
12165 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
12166 Fix declaration name.
12168 2013-03-28 Lawrence Crowl <crowl@google.com>
12170 * graphds.h (struct graph.indicies): Remove unused.
12171 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
12172 (SCOP_ORIGINAL_PDDRS): Remove unused.
12173 * sese.h (extern insert_loop_close_phis): Removed unused.
12174 (extern insert_guard_phis): Removed unused.
12175 (extern ivtype_map_elt_info): Removed unused.
12176 (new_ivtype_map_elt): Removed unused.
12177 * sese.c (ivtype_map_elt_info): Removed unused.
12179 2013-03-28 Lawrence Crowl <crowl@google.com>
12181 * Makefile.in: Add several missing include dependences.
12183 (test-dump.o): New. This object is not added to any executable,
12184 but is present for ad-hoc testing.
12186 (debug (const bitmap_head_def &)): New.
12187 (debug (const bitmap_head_def *)): New.
12189 (extern debug (const bitmap_head_def &)): New.
12190 (extern debug (const bitmap_head_def *)): New.
12192 (debug (edge_def &)): New.
12193 (debug (edge_def *)): New.
12195 (debug (basic_block_def &)): New.
12196 (debug (basic_block_def *)): New.
12198 (dump_node (const_tree, int, FILE *)): Correct source file.
12200 (debug (die_struct &)): New.
12201 (debug (die_struct *)): New.
12203 (extern debug (die_struct &)): New.
12204 (extern debug (die_struct *)): New.
12205 * gimple-pretty-print.c
12206 (debug (gimple_statement_d &)): New.
12207 (debug (gimple_statement_d *)): New.
12208 * gimple-pretty-print.h
12209 (extern debug (gimple_statement_d &)): New.
12210 (extern debug (gimple_statement_d *)): New.
12212 (debug (ira_allocno_copy &)): New.
12213 (debug (ira_allocno_copy *)): New.
12214 (debug (ira_allocno &)): New.
12215 (debug (ira_allocno *)): New.
12217 (extern debug (ira_allocno_copy &)): New.
12218 (extern debug (ira_allocno_copy *)): New.
12219 (extern debug (ira_allocno &)): New.
12220 (extern debug (ira_allocno *)): New.
12222 (debug (live_range &)): New.
12223 (debug (live_range *)): New.
12225 (debug (lra_live_range &)): New.
12226 (debug (lra_live_range *)): New.
12228 (debug (lra_live_range &)): New.
12229 (debug (lra_live_range *)): New.
12231 (debug (omega_pb_d &)): New.
12232 (debug (omega_pb_d *)): New.
12234 (extern debug (omega_pb_d &)): New.
12235 (extern debug (omega_pb_d *)): New.
12237 (debug (const rtx_def &)): New.
12238 (debug (const rtx_def *)): New.
12240 (debug_tree (tree): Move within file.
12241 (debug_raw (const tree_node &)): New.
12242 (debug_raw (const tree_node *)): New.
12243 (dump_tree_via_hooks (const tree_node *, int)): New.
12244 (debug (const tree_node &)): New.
12245 (debug (const tree_node *)): New.
12246 (debug_verbose (const tree_node &)): New.
12247 (debug_verbose (const tree_node *)): New.
12248 (debug_head (const tree_node &)): New.
12249 (debug_head (const tree_node *)): New.
12250 (debug_body (const tree_node &)): New.
12251 (debug_body (const tree_node *)): New.
12252 (debug_vec_tree (tree): Move and reimplement in terms of dump.
12253 (debug (vec<tree, va_gc> &)): New.
12254 (debug (vec<tree, va_gc> *)): New.
12256 (extern debug (const rtx_def &)): New.
12257 (extern debug (const rtx_def *)): New.
12259 (debug_raw (simple_bitmap_def &)): New.
12260 (debug_raw (simple_bitmap_def *)): New.
12261 (debug (simple_bitmap_def &)): New.
12262 (debug (simple_bitmap_def *)): New.
12264 (extern debug (simple_bitmap_def &)): New.
12265 (extern debug (simple_bitmap_def *)): New.
12266 (extern debug_raw (simple_bitmap_def &)): New.
12267 (extern debug_raw (simple_bitmap_def *)): New.
12269 (debug (vinsn_def &)): New.
12270 (debug (vinsn_def *)): New.
12271 (debug_verbose (vinsn_def &)): New.
12272 (debug_verbose (vinsn_def *)): New.
12273 (debug (expr_def &)): New.
12274 (debug (expr_def *)): New.
12275 (debug_verbose (expr_def &)): New.
12276 (debug_verbose (expr_def *)): New.
12277 (debug (vec<rtx> &)): New.
12278 (debug (vec<rtx> *)): New.
12280 (extern debug (vinsn_def &)): New.
12281 (extern debug (vinsn_def *)): New.
12282 (extern debug_verbose (vinsn_def &)): New.
12283 (extern debug_verbose (vinsn_def *)): New.
12284 (extern debug (expr_def &)): New.
12285 (extern debug (expr_def *)): New.
12286 (extern debug_verbose (expr_def &)): New.
12287 (extern debug_verbose (expr_def *)): New.
12288 (extern debug (vec<rtx> &)): New.
12289 (extern debug (vec<rtx> *)): New.
12291 (_list_iter_cond_expr): Make inline instead of static.
12293 (debug (sreal &)): New.
12294 (debug (sreal *)): New.
12296 (extern debug (sreal &)): New.
12297 (extern debug (sreal *)): New.
12299 (extern debug_raw (const tree_node &)): New.
12300 (extern debug_raw (const tree_node *)): New.
12301 (extern debug (const tree_node &)): New.
12302 (extern debug (const tree_node *)): New.
12303 (extern debug_verbose (const tree_node &)): New.
12304 (extern debug_verbose (const tree_node *)): New.
12305 (extern debug_head (const tree_node &)): New.
12306 (extern debug_head (const tree_node *)): New.
12307 (extern debug_body (const tree_node &)): New.
12308 (extern debug_body (const tree_node *)): New.
12309 (extern debug (vec<tree, va_gc> &)): New.
12310 (extern debug (vec<tree, va_gc> *)): New.
12312 (debug (struct loop &)): New.
12313 (debug (struct loop *)): New.
12314 (debug_verbose (struct loop &)): New.
12315 (debug_verbose (struct loop *)): New.
12316 * tree-dump.c: Add header dependence.
12318 (extern debug (struct loop &)): New.
12319 (extern debug (struct loop *)): New.
12320 (extern debug_verbose (struct loop &)): New.
12321 (extern debug_verbose (struct loop *)): New.
12323 (debug (data_reference &)): New.
12324 (debug (data_reference *)): New.
12325 (debug (vec<data_reference_p> &)): New.
12326 (debug (vec<data_reference_p> *)): New.
12327 (debug (vec<ddr_p> &)): New.
12328 (debug (vec<ddr_p> *)): New.
12330 (extern debug (data_reference &)): New.
12331 (extern debug (data_reference *)): New.
12332 (extern debug (vec<data_reference_p> &)): New.
12333 (extern debug (vec<data_reference_p> *)): New.
12334 (extern debug (vec<ddr_p> &)): New.
12335 (extern debug (vec<ddr_p> *)): New.
12337 (debug (pt_solution &)): New.
12338 (debug (pt_solution *)): New.
12340 (extern debug (pt_solution &)): New.
12341 (extern debug (pt_solution *)): New.
12343 (debug (_var_map &)): New.
12344 (debug (_var_map *)): New.
12345 (debug (tree_live_info_d &)): New.
12346 (debug (tree_live_info_d *)): New.
12348 (extern debug (_var_map &)): New.
12349 (extern debug (_var_map *)): New.
12350 (extern debug (tree_live_info_d &)): New.
12351 (extern debug (tree_live_info_d *)): New.
12353 2013-03-28 Jan Hubicka <jh@suse.cz>
12355 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
12357 2013-03-28 Ian Bolton <ian.bolton@arm.com>
12359 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
12360 record only when desired or required.
12362 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
12364 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
12365 *vec_extractv2di_1_rex64. Use x64 isa attribute.
12367 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12369 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
12370 (*andsi3_compare0_uxtw): New pattern.
12371 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
12372 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
12374 2013-03-28 Jan Hubicka <jh@suse.cz>
12376 * data-streamer-in.c (streamer_read_gcov_count): New function.
12377 * gimple-streamer-out.c: Include value-prof.h.
12378 (output_gimple_stmt): Output histogram.
12379 (output_bb): Use streamer_write_gcov_count.
12380 * value-prof.c: Include data-streamer.h
12381 (dump_histogram_value): Add HIST_TYPE_MAX.
12382 (stream_out_histogram_value): New function.
12383 (stream_in_histogram_value): New function.
12384 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
12385 (stream_out_histogram_value, stream_in_histogram_value): Declare.
12386 * data-streamer-out.c (streamer_write_gcov_count): New function.
12387 (streamer_write_gcov_count_stream): New function.
12388 * lto-cgraph.c (lto_output_edge): Update counter streaming.
12389 (lto_output_node): Likewise.
12390 (input_node, input_edge): Likewise.
12391 * lto-streamer-out.c (output_cfg): Update streaming.
12392 * lto-streamer-in.c (input_cfg): Likewise.
12393 * data-streamer.h (streamer_write_gcov_count,
12394 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
12395 * gimple-streamer-in.c: Include value-prof.h
12396 (input_gimple_stmt): Input histograms.
12397 (input_bb): Update profile streaming.
12399 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12401 * genmodes.c (emit_max_int): New function.
12402 (emit_insn_modes_h): Added call to emit_max_function.
12403 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12405 * machmode.def: Fixed comment.
12407 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12409 * combine.c (try_combine): Removed useless assert.
12410 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12412 2013-03-28 Marek Polacek <polacek@redhat.com>
12413 Richard Biener <rguenther@suse.de>
12415 PR tree-optimization/56695
12416 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12417 build signed result of a vector comparison.
12418 * tree-cfg.c (verify_gimple_comparison): Check that a result
12419 of a vector comparison has signed type.
12421 2013-03-28 Richard Biener <rguenther@suse.de>
12423 PR tree-optimization/37021
12424 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12425 do not restrict gaps between groups.
12426 * tree-vect-stmts.c (vectorizable_load): Properly account for
12427 a gap between groups.
12429 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12431 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12432 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12435 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12437 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12438 * doc/extend.texi (Named Address Spaces): Ditto.
12439 (Variable Attributes): Ditto.
12441 2013-03-27 Kai Tietz <ktietz@redhat.com>
12443 * config.build: Add support for cygwin x64 target.
12444 * config.gcc: Likewise.
12445 * config.host: Likewise.
12446 * configure.ac: Likewise
12447 * configure: Regenerated.
12449 2013-03-27 Kai Tietz <ktietz@redhat.com>
12451 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12452 * config/i386/t-cygwin-w64: New file.
12453 * config/i386/cygwin-w64.h: New file.
12454 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12455 and add support for x64-cygwin target.
12456 (CPP_SPEC): Likewise.
12457 (CXX_WRAP_SPEC_LIST): Undefine before define.
12458 (LIBGCJ_SONAME): Use 15 as version.
12460 2013-03-27 Richard Biener <rguenther@suse.de>
12462 PR tree-optimization/56716
12463 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12464 dumping for ref nodes.
12466 2013-03-27 Martin Jambor <mjambor@suse.cz>
12468 PR tree-optimization/55334
12469 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12470 restricted pointers to arrays.
12472 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12474 * Makefile.in (.SUFFIXES): Add .cc.
12475 (.c.o): Apply same recipe for implicit rule .cc.o.
12477 2013-03-27 Richard Biener <rguenther@suse.de>
12479 PR tree-optimization/37021
12480 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12481 REALPART/IMAGPART_EXPRs around the supported refs.
12482 * tree-ssa-structalias.c (find_func_aliases): Assume that
12483 floating-point values are not used to transfer pointers.
12485 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12487 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12488 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12489 * doc/tm.texi: Regenerate.
12490 * targhooks.h (default_has_ifunc_p): New.
12491 * targhooks.c (default_has_ifunc_p): Ditto.
12492 * config/linux-protos.h: New file.
12493 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12494 this hook for linux which disables support of indirect functions in
12496 * config/linux-android.c: New file.
12497 * config/t-linux-android.c: Ditto.
12498 * config.gcc: Added new object file linux-android.o.
12499 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
12500 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12501 * varasm.c (do_assemble_alias): Likewise.
12502 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12503 doesn't support indirect functions.
12504 * configure: Regenerate.
12506 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12509 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12510 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12513 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12515 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12517 2013-03-27 Terry Guo <terry.guo@arm.com>
12519 * config/arm/arm-cores.def: Added core cortex-r7.
12520 * config/arm/arm-tune.md: Regenerated.
12521 * config/arm/arm-tables.opt: Regenerated.
12522 * doc/invoke.texi: Added entry for core cortex-r7.
12524 2013-03-27 Walter Lee <walt@tilera.com>
12526 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12527 double-decrement of next_scratch_regno.
12529 2013-03-27 Walter Lee <walt@tilera.com>
12531 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12533 (insn_v1mulus): Ditto.
12534 (insn_v2muls): Ditto.
12536 2013-03-27 Walter Lee <walt@tilera.com>
12538 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12539 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12541 2013-03-27 Walter Lee <walt@tilera.com>
12543 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
12544 (*sibcall_value): Ditto.
12546 2013-03-27 Walter Lee <walt@tilera.com>
12548 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12549 (insn_mnz_v8qi): ... this ...
12550 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12552 (insn_v<n>mnz): Replaced by ...
12553 (insn_v1mnz): ... this ...
12554 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12556 (insn_mz_<mode>): Replaced by ...
12557 (insn_mz_v8qi): ... this ...
12558 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12560 (insn_v<n>mz): Replaced by ...
12561 (insn_v1mz): ... this ...
12562 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12565 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12567 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12569 2013-03-26 Roland McGrath <mcgrathr@google.com>
12571 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12572 than fprintf with a non-constant, non-format string.
12574 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12576 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12577 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12578 operand 0 predicate.
12579 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12580 attribute. Use general_x64nomem_operand as operand 1 predicate.
12581 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12582 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12583 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12584 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12585 (mov<mode>_insv_1): Remove expander. Merge insn with
12586 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12587 Use general_x64nomem_operand as operand 1 predicate.
12588 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12589 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12590 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12591 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12592 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12593 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12594 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12595 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12596 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12597 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12598 (general_x64nomem_operand): Ditto.
12600 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
12602 * config/rtems.opt: Add -pthread option.
12604 2013-03-26 Richard Biener <rguenther@suse.de>
12606 * alias.c (find_base_term): Avoid redundant and not used recursion.
12607 (base_alias_check): Get the initial base term from the caller.
12608 (true_dependence_1): Compute and pass base terms to base_alias_check.
12609 (write_dependence_p): Likewise.
12610 (may_alias_p): Likewise.
12612 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12614 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12615 PC-relative load in SI modes and above only.
12617 2013-03-26 Xinyu Qi <xyqi@marvell.com>
12619 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
12620 * config/arm/iwmmxt.md (WCGR0): Update.
12621 (WCGR1, WCGR2, WCGR3): Likewise.
12623 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12625 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12626 Use x64 and nox64 isa attributes.
12628 2013-03-26 Richard Biener <rguenther@suse.de>
12630 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12631 alignment computations and rely on get_object_alignment_1
12632 for the !TYPE_P case.
12633 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12635 2013-03-26 Walter Lee <walt@tilera.com>
12637 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12638 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12640 2013-03-25 Jeff Law <law@redhat.com>
12642 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12643 check for INTEGRAL_TYPE_P that was missing due to checking in
12644 wrong version of prior patch.
12646 2013-03-25 Walter Lee <walt@tilera.com>
12648 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12649 TILEGX_INSN_SHUFFLEBYTES1.
12650 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12652 (tilegx_builtins): Ditto.
12653 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12655 2013-03-25 Walter Lee <walt@tilera.com>
12657 * config/tilegx/tilegx.md (floatsisf2): New pattern.
12658 (floatunssisf2): New pattern.
12659 (floatsidf2): New pattern.
12660 (floatunssidf2): New pattern.
12662 2013-03-25 Walter Lee <walt@tilera.com>
12664 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12665 tests for constraint J, K, N, P.
12667 2013-03-25 Walter Lee <walt@tilera.com>
12669 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12670 Use indirect/pcrel encoding.
12671 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12674 2013-03-25 Steve Ellcey <sellcey@mips.com>
12676 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12677 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12678 * config/mips/mips.c (mips_option_override): Set IMADD default.
12679 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12680 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12681 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12682 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12683 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12685 2013-03-25 Jeff Law <law@redhat.com>
12687 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12688 slightly to avoid creating and folding useless trees. Simplify
12689 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12691 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
12693 * config/i386/i386.md (*zero_extendsidi2): Merge with
12694 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12695 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12696 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12697 targets. Clarify comment.
12699 2013-03-25 Martin Jambor <mjambor@suse.cz>
12701 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12702 pass-through jump functions differently.
12703 (ipa_read_jump_function): Likewise. Also use setter functions to set
12706 2013-03-25 Martin Jambor <mjambor@suse.cz>
12708 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12709 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12711 (ipa_get_indirect_edge_target): New function.
12712 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12713 ipa_get_indirect_edge_target. Update all callers.
12714 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12715 ipa_get_indirect_edge_target_1 instead of calling
12716 ipa_get_indirect_edge_target.
12717 (create_specialized_node): Pass aggvlas to
12718 ipcp_discover_new_direct_edges.
12720 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12722 * config/arm/arm.md (f_sels, f_seld): New types.
12723 (*cmov<mode>): New pattern.
12724 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12727 2013-03-25 Kai Tietz <ktietz@redhat.com>
12729 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12730 POSIX-printf for mingw-hosted builds.
12732 2013-03-25 Richard Biener <rguenther@suse.de>
12734 PR middle-end/56694
12735 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12736 must-not-throw stmt location.
12738 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12740 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12741 Emit load-acquire versions when acq is true.
12742 (arm_emit_store_exclusive): Add rel parameter.
12743 Emit store-release versions when rel is true.
12744 (arm_split_compare_and_swap): Use acquire-release instructions
12746 of barriers when appropriate.
12747 (arm_split_atomic_op): Likewise.
12748 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12749 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12750 (VUNSPEC_SLX): Likewise.
12751 (VUNSPEC_LDA): Likewise.
12752 (VUNSPEC_STL): Likewise.
12753 * config/arm/sync.md (atomic_load<mode>): New pattern.
12754 (atomic_store<mode>): Likewise.
12755 (arm_load_acquire_exclusive<mode>): Likewise.
12756 (arm_load_acquire_exclusivesi): Likewise.
12757 (arm_load_acquire_exclusivedi): Likewise.
12758 (arm_store_release_exclusive<mode>): Likewise.
12760 2013-03-25 Catherine Moore <clm@codesourcery.com>
12762 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12763 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12764 * config/mip/predicates.md (lwsp_swsp_operand,
12765 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12766 sb16_operand, db4_operand, db7_operand, ib3_operand,
12767 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
12768 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
12769 andi16_operand): New predicates.
12770 * config/mips/mips.md (compression): New attribute.
12771 (enabled): New attribute.
12772 (length): Consider compression in computing length.
12773 (shift_compression): New code attribute.
12774 (*add<mode>3): New operands. Record compression.
12775 (sub<mode>3): Likewise.
12776 (one_cmpl<mode>2): Likewise.
12777 (*and<mode>3): Likewise.
12778 (*ior<mode>3): Likewise.
12779 (unnamed pattern for xor): Likewise.
12780 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
12781 (*<optab><mode>3): Likewise.
12782 (*mov<mode>_internal: Likewise.
12783 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
12784 (mips_unsigned_immediate_p): New.
12785 (umips_lwsp_swsp_address_p): New.
12786 (m16_based_address_p): New.
12787 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
12788 (mips_unsigned_immediate_p): New prototype.
12789 (lwsp_swsp_address_p): New prototype.
12790 (m16_based_address_p): New prototype.
12791 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
12792 (mips_signed_immediate_p): New function.
12793 (m16_based_address_p): New function.
12794 (lwsp_swsp_address_p): New function.
12795 (mips_print_operand_punctuation): Recognize short delay slot insns
12796 for microMIPS.add<mode>3"
12798 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12801 * config/arm/iterators.md (v_cmp_result): New mode attribute.
12802 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
12804 2013-03-25 Richard Biener <rguenther@suse.de>
12806 PR tree-optimization/56689
12807 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
12810 2013-03-25 Richard Biener <rguenther@suse.de>
12812 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
12814 (memory_references): Likewise.
12815 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
12816 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
12817 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
12818 (gather_mem_refs_in_loops): Fold into ...
12819 (analyze_memory_references): ... this. Move initialization
12820 to tree_ssa_lim_initialize.
12821 (fill_always_executed_in): Rename to ...
12822 (fill_always_executed_in_1): ... this.
12823 (fill_always_executed_in): Move contains_call computation to
12824 this new function from ...
12825 (tree_ssa_lim_initialize): ... here.
12826 (tree_ssa_lim): Call fill_always_executed_in.
12828 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12830 * postreload.c (reload_combine): Fix code detecting returns.
12832 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
12834 * function.c (emit_use_return_register_into_block): On cc0 targets,
12835 do not emit the sequence between cc0 setter and user.
12837 2013-03-25 Kai Tietz <ktietz@redhat.com>
12839 * config/i386/predicates.md (local_symbolic_operand): Interpret
12840 dll-imported symbols as none-local.
12842 2013-03-25 Richard Biener <rguenther@suse.de>
12844 * tree-ssa-loop-im.c (struct depend): Remove.
12845 (struct lim_aux_data): Make depends a vec of gimples.
12846 (free_lim_aux_data): Adjust.
12847 (add_dependency): Likewise.
12848 (set_level): Likewise.
12850 2013-03-25 Richard Biener <rguenther@suse.de>
12852 PR middle-end/56434
12853 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
12854 the pointer returned by calls with ECF_MALLOC set.
12856 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12858 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
12860 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
12862 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
12863 using MMXMODE mode iterator.
12864 (*move<mode>_internal): Merge with *movv2sf_internal and
12865 *movv2sf_internal_rex64 using MMXMODE mode iterator.
12867 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12869 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
12870 (record_last_mem_set_info): Likewise.
12872 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
12873 of XNEWVEC followed by memset.
12874 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
12876 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
12878 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
12879 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
12880 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
12881 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
12882 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
12883 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
12884 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
12885 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
12886 BARRIER_P instead of GET_CODE.
12888 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12890 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
12891 inaccuracy in the probing code.
12893 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
12894 (ctrapdi4): Likewise.
12896 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12898 * calls.c (expand_call): Add missing guard to code handling return
12899 of non-BLKmode structures in MSB.
12900 * function.c (expand_function_end): Likewise.
12902 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
12904 * combine.c (try_combine): Adjust comment. Do not add the set of
12905 insn #0 if the destination indirectly is set or dies in insn #2.
12906 Tidy up code to distribute a new note.
12908 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12910 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
12911 also for alternatives 16 and 17.
12913 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
12915 * config/i386/sse.md (*mov<mode>_internal): Merge with
12916 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
12917 Emit insn template depending on type attribute. Use
12918 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
12919 movd instead of movq mnemonic for interunit moves. Rewrite mode
12920 attribute calculation. Remove unit attribute calculation.
12921 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
12922 Set prefix_data16 attribute for DImode ssemov types.
12923 Use Ym instead of y for SSE-MMX conversion alternatives.
12924 Reorder operand constraints.
12926 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
12928 * df.h (df_insn_delete): Adjust prototype.
12929 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
12930 and let it decide whether mark the basic block dirty.
12931 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
12932 * df-scan.c (df_insn_info_delete): New helper function, split
12933 off from df_insn_delete.
12934 (df_scan_free_bb_info): Use it.
12935 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
12937 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
12938 that the insn is actually an insn and it has a non-NULL basic block.
12939 Do not mark basic block dirty if only deleting a DEBUG_INSN.
12941 2013-03-22 Richard Biener <rguenther@suse.de>
12943 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
12945 (mem_ref_alloc): Do not allocate them.
12946 (refs_independent_p): Do not query or maintain a cache.
12948 2013-03-22 Richard Biener <rguenther@suse.de>
12950 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
12951 (gather_mem_refs_in_loops): Do not compute it.
12952 (analyze_memory_references): Do not allocate it.
12953 (tree_ssa_lim_finalize): Do not free it.
12954 (for_all_locs_in_loop): Do not query all_refs_in_loop.
12956 2013-03-22 Richard Biener <rguenther@suse.de>
12958 * is-a.h (as_a): Use gcc_checking_assert.
12960 2013-03-22 Ian Bolton <ian.bolton@arm.com>
12962 * config/aarch64/aarch64.c (aarch64_print_operand): New
12963 format specifier for printing a constant in hex.
12964 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
12965 format specifier for printing second operand.
12967 2013-03-22 Richard Biener <rguenther@suse.de>
12969 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
12971 (gather_mem_refs_in_loops): Perform store accumulation here.
12972 (create_vop_ref_mapping_loop): Remove.
12973 (create_vop_ref_mapping): Likewise.
12974 (analyze_memory_references): Initialize refs_stored_in_loop.
12975 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
12976 (record_indep_loop): Remove.
12977 (record_dep_loop): New function.
12978 (ref_indep_loop_p_1): Adjust to only walk over references
12979 in the loop, not its subloops.
12980 (ref_indep_loop_p): Rename to ...
12981 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
12982 maintaining a more fine-grained cache.
12983 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
12984 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
12986 2013-03-22 Richard Biener <rguenther@suse.de>
12988 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
12989 (struct mem_ref): Make accesses_in_loop a vec of a vec of
12990 aggregate mem_ref_loc.
12991 (free_mem_ref_locs): Inline into ...
12992 (memref_free): ... this and adjust.
12993 (mem_ref_alloc): Adjust.
12994 (mem_ref_locs_alloc): Remove.
12995 (record_mem_ref_loc): Adjust.
12996 (get_all_locs_in_loop): Rewrite into ...
12997 (for_all_locs_in_loop): ... this iterator.
12998 (rewrite_mem_ref_loc): New functor.
12999 (rewrite_mem_refs): Use for_all_locs_in_loop.
13000 (sm_set_flag_if_changed): New functor.
13001 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13002 (ref_always_accessed): New functor.
13003 (ref_always_accessed_p): Use for_all_locs_in_loop.
13005 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13007 * tree-pass.h (PROP_gimple_lvec): New.
13008 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13009 (init_optimization_passes): Move pass_lower_vector.
13010 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13012 (pass_lower_vector): Provide PROP_gimple_lvec.
13013 (pass_lower_vector_ssa): Likewise.
13014 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13016 2013-03-21 Mark Wielaard <mjw@redhat.com>
13018 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13020 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13022 * config/i386/i386.md (*movdi_internal): Disparage slightly
13023 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13024 conversion alternatives.
13026 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13028 PR middle-end/48087
13029 * diagnostic.def (DK_WERROR): New kind.
13030 * diagnostic.h (werrorcount): Define.
13031 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13032 promoted to DK_ERROR, increment DK_WERROR counter instead of
13034 * toplev.c (toplev_main): Call print_ignored_options even if
13035 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13036 even if just werrorcount is non-zero.
13039 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13041 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13042 (string_cst_pool_decl): New function.
13043 (optimize_one_addr_into_implicit_ptr): New function.
13044 (resolve_addr_in_expr): Optimize DWARF location expression
13045 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13046 which doesn't live in memory, but has DW_AT_location or
13047 DW_AT_const_value, or refers to a string literal, into
13048 DW_OP_GNU_implicit_pointer.
13049 (optimize_location_into_implicit_ptr): New function.
13050 (resolve_addr): If removing DW_AT_location of a variable because
13051 it was DW_OP_addr of address of the variable, but the variable doesn't
13052 live in memory, try to emit const value attribute for the initializer.
13054 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13056 * tree.h (VECTOR_TYPE_P): New macro.
13057 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13058 TYPE_MODE): Use it.
13059 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13060 VEC_COND_EXPR cannot be lvalues.
13061 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13063 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13065 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13066 Restrict the transformation to equal modes.
13068 2013-03-21 Richard Biener <rguenther@suse.de>
13070 PR tree-optimization/39326
13071 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13072 (MEM_ANALYZABLE): Adjust.
13073 (record_mem_ref_loc): Move bitmap ops ...
13074 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13075 unanalyzable refs, do not record locations for it.
13076 (analyze_memory_references): Allocate ref zero as shared
13078 (refs_independent_p): Do not test for unanalyzed mems here.
13079 (ref_indep_loop_p_1): Special-case disambiguation against
13080 the unanalyzed ref.
13081 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13083 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13085 * config/arm/arm-protos.h (tune_params): Add
13086 prefer_neon_for_64bits field.
13087 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13088 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13089 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13090 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13091 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13092 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13093 (arm_option_override): Handle -mneon-for-64bits new option.
13094 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13095 (prefer_neon_for_64bits): Declare new variable.
13096 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13097 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13098 (arch_enabled): Handle new arch types. Remove support for onlya8
13100 (one_cmpldi2): Use new arch names.
13101 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13102 * config/arm/arm.opt (mneon-for-64bits): Add option.
13103 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13104 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13105 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13107 * doc/invoke.texi (-mneon-for-64bits): Document.
13109 2013-03-21 Richard Biener <rguenther@suse.de>
13111 PR tree-optimization/39326
13112 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13113 (sort_bbs_in_loop_postorder_cmp): New function.
13114 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13116 2013-03-21 Richard Biener <rguenther@suse.de>
13118 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13119 (vect_insert_into_interleaving_chain): Likewise.
13120 (vect_drs_dependent_in_basic_block): Inline ...
13121 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13123 (vect_analyze_data_ref_dependence): ... here. Simplify.
13124 (vect_check_interleaving): Simplify.
13125 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13126 (vect_slp_analyze_data_ref_dependences): ... this new function.
13127 (dr_group_sort_cmp): New function.
13128 (vect_analyze_data_ref_accesses): Compute data-reference groups
13129 here instead of in vect_analyze_data_ref_dependence. Use
13130 a more efficient algorithm.
13131 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13132 vect_slp_analyze_data_ref_dependences. Call
13133 vect_analyze_data_ref_accesses earlier.
13134 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13135 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13136 (vect_slp_analyze_data_ref_dependences): New prototype.
13138 2013-03-21 Richard Biener <rguenther@suse.de>
13140 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13141 ref is stored in the loop.
13142 (find_refs_for_sm): Walk only over all stores.
13143 (store_motion_loop): Allocate from lim_bitmap_obstack.
13144 (store_motion): Likewise.
13146 2013-03-21 Richard Biener <rguenther@suse.de>
13148 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
13149 Update virtual SSA form.
13151 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13153 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
13154 * configure: Regenerate.
13155 * config.in: Regenerate.
13156 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
13157 if !HAVE_LD_EH_FRAME_CIEV3.
13159 2013-03-21 Richard Biener <rguenther@suse.de>
13161 * tree-cfg.c (verify_expr_no_block): New function.
13162 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
13163 nor DECL_VALUE_EXPR have locations with associated blocks.
13164 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
13165 (clear_unused_block_pointer): Remove code dealing with
13166 blocks in DECL_DEBUG_EXPR locations.
13168 2013-03-21 Richard Biener <rguenther@suse.de>
13170 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
13171 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
13172 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
13173 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
13174 instead of DECL_DEBUG_EXPR_IS_FROM.
13175 * gimplify.c (gimplify_modify_expr): Likewise.
13176 * tree-cfg.c (verify_expr_location_1): Likewise.
13177 * tree-complex.c (create_one_component_var): Likewise.
13178 * tree-sra.c (create_access_replacement): Likewise.
13179 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
13180 (clear_unused_block_pointer): Likewise.
13181 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
13182 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
13183 * var-tracking.c (var_debug_decl): Likewise.
13184 (track_expr_p): Likewise.
13185 * tree-inline.c (add_local_variables): Likewise. Set
13186 DECL_HAS_DEBUG_EXPR_P after copying it.
13187 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
13188 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
13190 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13193 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
13194 * configure: Regenerate.
13195 * config.in: Regenerate.
13196 * config/i386/i386.md (*movdf_internal): Use
13197 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13198 movd instead of movq mnemonic for interunit moves.
13199 (*movdi_internal): Ditto.
13201 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13203 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
13204 (abd<mode>_3): New pattern.
13205 (aba<mode>_3): New pattern.
13206 (fabd<mode>_3): New pattern.
13208 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13210 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
13211 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
13212 occurrence of REGISTER_PREFIX as its empty string.
13214 2013-03-20 Jeff Law <law@redhat.com>
13216 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
13217 addititional equivalences for equality comparisons between an SSA_NAME
13218 and a constant where the SSA_NAME was set from a widening conversion.
13220 2013-03-20 Walter Lee <walt@tilera.com>
13222 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
13224 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13226 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
13227 depending on type attribute.
13228 (*movti_internal): Ditto.
13229 (*movtf_internal): Ditto.
13230 (*movxf_internal): Ditto.
13231 (*movdf_internal): Ditto.
13232 (*movsf_internal): Ditto.
13234 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13236 * config/i386/i386.md (*movti_internal): Set prefix attribute to
13237 maybe_vex for sselog1 and ssemov types.
13238 (*movdi_internal): Reorder operand constraints.
13239 (*movsi_internal): Ditto. Set prefix attribute to
13240 maybe_vex for sselog1 and ssemov types.
13241 (*movtf_internal): Set prefix attribute to maybe_vex
13242 for sselog1 and ssemov types.
13243 (*movdf_internal): Ditto. Set prefix_data16 attribute for
13244 DImode ssemov types. Reorder operand constraints.
13245 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
13246 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
13247 attribute for SImode ssemov types. Reorder operand constraints.
13249 2013-03-20 Martin Jambor <mjambor@suse.cz>
13251 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
13252 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
13254 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
13256 * config/rs6000/predicates.md (indexed_address, update_address_mem
13257 update_indexed_address_mem): New predicates.
13258 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
13259 attribute for load/store instructions.
13260 * config/rs6000/dfp.md (movsd_store): Likewise.
13261 (movsd_load): Likewise.
13262 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
13263 (unnamed HI->DI extend define_insn): Likewise.
13264 (unnamed SI->DI extend define_insn): Likewise.
13265 (unnamed QI->SI extend define_insn): Likewise.
13266 (unnamed QI->HI extend define_insn): Likewise.
13267 (unnamed HI->SI extend define_insn): Likewise.
13268 (unnamed HI->SI extend define_insn): Likewise.
13269 (extendsfdf2_fpr): Likewise.
13270 (movsi_internal1): Likewise.
13271 (movsi_internal1_single): Likewise.
13272 (movhi_internal): Likewise.
13273 (movqi_internal): Likewise.
13274 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
13275 attribute for load/store instructions.
13276 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
13278 (mov<mode>_softfloat): Likewise.
13279 (mov<mode>_hardfloat32): Likewise.
13280 (mov<mode>_hardfloat64): Likewise.
13281 (mov<mode>_softfloat64): Likewise.
13282 (movdi_internal32): Likewise.
13283 (movdi_internal64): Likewise.
13284 (probe_stack_<mode>): Likewise.
13286 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
13288 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
13289 floating point, and decimal floating point to reload iterator.
13291 * config/rs6000/constraints.md (wl constraint): New constraints to
13292 return FLOAT_REGS if certain options are used to reduce the number
13293 of separate patterns that exist in the file.
13294 (wx constraint): Likewise.
13295 (wz constraint): Likewise.
13297 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13298 -mdebug=reg, print wg, wl, wx, and wz constraints.
13299 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13300 Initialize the reload functions for 64-bit binary/decimal floating
13302 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13303 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13304 create the buffer on the stack to overcome not having a 32-bit
13306 (rs6000_emit_move): Likewise.
13307 (rs6000_secondary_memory_needed_rtx): Likewise.
13308 (rs6000_alloc_sdmode_stack_slot): Likewise.
13309 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13310 via xxlxor, just like DFmode 0.0.
13312 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
13313 define as 1 if we are running on a power7 or newer.
13314 (enum r6000_reg_class_enum): Add new constraints.
13316 * config/rs6000/dfp.md (movsd): Delete, combine with binary
13317 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13318 with other moves by using conditional constraits (wg). Use LFIWZX
13319 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
13320 (movsd splitter): Likewise.
13321 (movsd_hardfloat): Likewise.
13322 (movsd_softfloat): Likewise.
13324 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
13325 binary and decimal floating point moves.
13326 (fmove_ok): New attributes to combine binary and decimal floating
13327 point moves, and to combine power6x (mfpgpr) moves along normal
13329 (real_value_to_target): Likewise.
13330 (f32_lr): Likewise.
13331 (f32_lm): Likewise.
13332 (f32_li): Likewise.
13333 (f32_sr): Likewise.
13334 (f32_sm): Likewise.
13335 (f32_si): Likewise.
13336 (movsf): Combine binary and decimal floating point moves. Combine
13337 power6x (mfpgpr) moves with other moves by using conditional
13338 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
13339 (mov<mode> for SFmode/SDmode); Likewise.
13340 (SFmode/SDmode splitters): Likewise.
13341 (movsf_hardfloat): Likewise.
13342 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
13343 (movsf_softfloat): Likewise.
13344 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
13346 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
13347 wx and wz constraints.
13349 * config/rs6000/constraints.md (wg constraint): New constraint to
13350 return FLOAT_REGS if -mmfpgpr (power6x) was used.
13352 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
13355 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13356 -mdebug=reg, print wg, wl, wx, and wz constraints.
13357 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13358 Initialize the reload functions for 64-bit binary/decimal floating
13360 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13361 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13362 create the buffer on the stack to overcome not having a 32-bit
13364 (rs6000_emit_move): Likewise.
13365 (rs6000_secondary_memory_needed_rtx): Likewise.
13366 (rs6000_alloc_sdmode_stack_slot): Likewise.
13367 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13368 via xxlxor, just like DFmode 0.0.
13370 * config/rs6000/dfp.md (movdd): Delete, combine with binary
13371 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13372 with other moves by using conditional constraits (wg). Use LFIWZX
13373 and STFIWX for loading SDmode on power7.
13374 (movdd splitters): Likewise.
13375 (movdd_hardfloat32): Likewise.
13376 (movdd_softfloat32): Likewise.
13377 (movdd_hardfloat64_mfpgpr): Likewise.
13378 (movdd_hardfloat64): Likewise.
13379 (movdd_softfloat64): Likewise.
13381 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
13382 64-bit binary and decimal floating point moves.
13383 (FMOVE64X): Likewise.
13384 (movdf): Combine 64-bit binary and decimal floating point moves.
13385 Combine power6x (mfpgpr) moves with other moves by using
13386 conditional constraits (wg).
13387 (mov<mode> for DFmode/DDmode): Likewise.
13388 (DFmode/DDmode splitters): Likewise.
13389 (movdf_hardfloat32): Likewise.
13390 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
13391 (movdf_softfloat32): Likewise.
13392 (movdf_hardfloat64_mfpgpr): Likewise.
13393 (movdf_hardfloat64): Likewise.
13394 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
13395 (movdf_softfloat64): Likewise.
13396 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13397 (reload_<mode>_load): Move to later in the file so they aren't in
13398 the middle of the floating point move insns.
13399 (reload_<mode>_store): Likewise.
13401 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13404 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13405 constraint if -mdebug=reg.
13406 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13407 Enable using dd reload support if needed.
13409 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13410 binary and decimal floating point moves in rs6000.md.
13411 (movtd_internal): Likewise.
13413 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13414 decimal floating point moves.
13416 (movtf_internal): Likewise.
13417 (mov<mode>_internal, TDmode/TFmode): Likewise.
13418 (movtf_softfloat): Likewise.
13419 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13421 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13422 movdi_internal64, using wg constraint for move direct operations.
13423 (movdi_internal64): Likewise.
13425 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13426 MODES_TIEABLE_P for selected modes. Print the numerical value of
13427 the various virtual registers. Use GPR/FPR first/last values,
13428 instead of hard coding the register numbers. Print which modes
13429 have reload functions registered.
13430 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13431 settings before/after setting cpu, target and subtarget settings.
13432 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13433 and for secondary reload failures in rs6000_secondary_reload_inner.
13434 (rs6000_secondary_reload_fail): Likewise.
13435 (rs6000_secondary_reload_inner): Likewise.
13437 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13438 macros for first/last GPR and FPR registers.
13439 (LAST_GPR_REGNO): Likewise.
13440 (FIRST_FPR_REGNO): Likewise.
13441 (LAST_FPR_REGNO): Likewise.
13443 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13444 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13445 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13446 (vcond<mode><mode>): Likewise.
13447 (vcondu<mode><mode>): Likewise.
13448 (vector_gtu<mode>): Likewise.
13449 (vector_gte<mode>): Likewise.
13450 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13451 to prevent the compiler from converting DImode operations to TImode.
13452 (ior<mode>3): Likewise.
13453 (and<mode>3): Likewise.
13454 (one_cmpl<mode>2): Likewise.
13455 (nor<mode>3): Likewise.
13456 (andc<mode>3): Likewise.
13458 * config/rs6000/constraints.md (wt constraint): New constraint
13459 that returns VSX_REGS if TImode is allowed in VSX registers.
13461 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13462 constant under VSX.
13464 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13465 similar to TImode, but it is restricted to being in the GPRs.
13467 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13468 TImode to occupy a single VSX register.
13470 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13471 -mvsx-timode for power7/power8.
13472 (power7 cpu): Likewise.
13473 (power8 cpu): Likewise.
13475 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13476 sure that TFmode/TDmode take up two registers if they are ever
13477 allowed in the upper VSX registers.
13478 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13480 (rs6000_init_hard_regno_mode_ok): Likewise.
13481 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13482 constraint. Print if LRA is turned on.
13483 (rs6000_option_override_internal): Give an error if -mvsx-timode
13484 and VSX is not enabled.
13485 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13486 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13487 to reg+offset addressing. Use PTImode when checking offset
13488 addresses for validity.
13489 (reg_offset_addressing_ok_p): Likewise.
13490 (rs6000_legitimate_offset_address_p): Likewise.
13491 (rs6000_legitimize_address): Likewise.
13492 (rs6000_legitimize_reload_address): Likewise.
13493 (rs6000_legitimate_address_p): Likewise.
13494 (rs6000_eliminate_indexed_memrefs): Likewise.
13495 (rs6000_emit_move): Likewise.
13496 (rs6000_secondary_reload): Likewise.
13497 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13498 reloads to fpr registers to continue to use reg+offset addressing,
13499 but 64-bit reloads to altivec registers need reg+reg addressing.
13500 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13501 it. Treat LO_SUM like a PLUS operation.
13502 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13503 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13504 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13505 registers to share a register with a smaller sized type, since VSX
13506 puts scalars in the upper 64-bits.
13507 (print_operand): Add support for PTImode.
13508 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13509 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13510 registers, but don't have arithmetic support.
13511 (rs6000_memory_move_cost): Add test for VSX.
13512 (rs6000_opt_masks): Add -mvsx-timode.
13514 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13517 (VSr): Use wt constraint for TImode.
13518 (VSv): Drop TImode support.
13519 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13520 (vsx_movti_64bit): Likewise.
13521 (vsx_movti_32bit): Likewise.
13522 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13523 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13524 one '?' on the appropriate output constraint. Do not allow TImode
13525 logical operations on 32-bit systems.
13526 (vsx_ior<mode>3): Likewise.
13527 (vsx_xor<mode>3): Likewise.
13528 (vsx_one_cmpl<mode>2): Likewise.
13529 (vsx_nor<mode>3): Likewise.
13530 (vsx_andc<mode>3): Likewise.
13531 (vsx_concat_<mode>): Likewise.
13532 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13534 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13535 OPTION_MASK_VSX_TIMODE.
13536 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13537 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13539 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13540 (TI2 iterator): New iterator for TImode, PTImode.
13541 (wd mode attribute): Add values for vector types.
13542 (movti_string): Replace TI move operations with operations for TImode
13543 and PTImode. Add support for TImode being allowed in VSX registers.
13544 (mov<mode>_string, TImode/PTImode): Likewise.
13545 (movti_ppc64): Likewise.
13546 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13547 (TI mode splitters): Likewise.
13549 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13552 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
13554 PR tree-optimization/56355
13555 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13556 Also handle integers with undefined overflow.
13558 2013-03-20 Catherine Moore <clm@codesourcery.com>
13559 Maciej W. Rozycki <macro@codesourcery.com>
13560 Tom de Vries <tom@codesourcery.com>
13561 Nathan Sidwell <nathan@codesourcery.com>
13562 Iain Sandoe <iain@codesourcery.com>
13563 Nathan Froyd <froydnj@codesourcery.com>
13564 Chao-ying Fu <fu@mips.com>
13566 * doc/extend.texi: (micromips, nomicromips, nocompression):
13567 Document new function attributes.
13568 * doc/invoke.texi (minterlink-compressed, mmicromips,
13569 m14k, m14ke, m14kec): Document new options.
13570 (minterlink-mips16): Update documentation.
13571 * doc/md.texi (ZC, ZD): Document new constraints.
13572 * configure.ac (gcc_cv_as_micromips): Check if linker
13573 supports the .set micromips directive.
13574 * configure: Regenerate.
13575 * config.in: Regenerate.
13576 * config/mips/mips-tables.opt: Regenerate.
13577 * config/mips/micromips.md: New file.
13578 * constraints.md (ZC, ZD): New constraints.
13579 * config/mips/predicates.md (movep_src_register): New predicate.
13580 (movep_src_operand): New predicate.
13581 (non_volatile_mem_operand): New predicate.
13582 * config/mips/mips.md (multimem): New type.
13583 (length): Differentiate between 17-bit and 18-bit branch offsets.
13584 (MOVEP1, MOVEP2): New mode iterator.
13585 (mov_<load>l): Use ZC constraint.
13586 (mov_<load>r): Likewise.
13587 (mov_<store>l): Likewise.
13588 (mov_<store>r): Likewise.
13589 (*branch_equality<mode>_inverted): Add microMIPS support.
13590 (*branch_equality<mode>): Likewise.
13591 (*jump_absolute): Likewise.
13592 (indirect_jump_<mode>): Likewise.
13593 (tablejump_<mode>): Likewise.
13594 (<optab>_internal): Likewise.
13595 (sibcall_internal): Likewise.
13596 (sibcall_value_internal): Likewise.
13597 (prefetch): Use constraint ZD.
13598 * config/mips/mips.opt (minterlink-compressed): New option.
13599 (minterlink-mips16): Now an alias for minterlink-compressed.
13600 (mmicromips): New option.
13601 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13602 (compare_and_swap_12): Likewise.
13603 (sync_add<mode>): Likewise.
13604 (sync_<optab>_12): Likewise.
13605 (sync_old_<optab>_12): Likewise.
13606 (sync_new_<optab>_12): Likewise.
13607 (sync_nand_12): Likewise.
13608 (sync_old_nand_12): Likewise.
13609 (sync_new_nand_12): Likewise.
13610 (sync_sub<mode>): Likewise.
13611 (sync_old_add<mode>): Likewise.
13612 (sync_old_sub<mode>): Likewise.
13613 (sync_new_add<mode>): Likewise.
13614 (sync_new_sub<mode>): Likewise.
13615 (sync_<optab><mode>): Likewise.
13616 (sync_old_<optab><mode>): Likewise.
13617 (sync_new_<optab><mode>): Likewise.
13618 (sync_nand<mode>): Likewise.
13619 (sync_old_nand<mode>): Likewise.
13620 (sync_new_nand<mode>): Likewise.
13621 (sync_lock_test_and_set<mode>): Likewise.
13622 (test_and_set_12): Likewise.
13623 (atomic_compare_and_swap<mode>): Likewise.
13624 (atomic_exchange<mode>_llsc): Likewise.
13625 (atomic_fetch_add<mode>_llsc): Likewise.
13626 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13627 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13628 (umips_save_restore_pattern_p): Likewise.
13629 (umips_load_store_pair_p): Likewise.
13630 (umips_output_load_store_pair): Likewise.
13631 (umips_movep_target_p): Likewise.
13632 (umips_12bit_offset_address_p): Likewise.
13633 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13634 (mips_base_mips16): Rename this...
13635 (mips_base_compression_flags): ...to this. Update all uses.
13636 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13637 (mips_mips16_decl_p): Delete.
13638 (mips_nomips16_decl_p): Delete.
13639 (mips_get_compress_on_flags): New function.
13640 (mips_get_compress_off_flags): New function.
13641 (mips_get_compress_mode): New function.
13642 (mips_get_compress_on_name): New function.
13643 (mips_get_compress_off_name): New function.
13644 (mips_insert_attributes): Support multiple compression types.
13645 (mips_merge_decl_attributes): Likewise.
13646 (umips_12bit_offset_address_p): New function.
13647 (mips_start_function_definition): Emit .set micromips directive.
13648 (mips_call_may_need_jalx_p): New function.
13649 (mips_function_ok_for_sibcall): Add microMIPS support.
13650 (mips_print_operand_punctuation): Support short delay slots and
13652 (umips_swm_mask, umips_swm_encoding): New.
13653 (umips_build_save_restore): New function.
13654 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13655 (was_mips16_p): Remove.
13656 (old_compression_mode): New.
13657 (mips_set_compression_mode): New function.
13658 (mips_set_current_function): Add microMIPS support.
13659 (mips_option_override): Likewise.
13660 (umips_save_restore_pattern_p): New function.
13661 (umips_output_save_restore): New function.
13662 (umips_load_store_pair_p_1): New function.
13663 (umips_load_store_pair_p): New function.
13664 (umips_output_load_store_pair_1): New function.
13665 (umips_output_load_store_pair): New function.
13666 (umips_movep_target_p) New function.
13667 (mips_prepare_pch_save): Add microMIPS support.
13668 * config/mips/mips.h (TARGET_COMPRESSION): New.
13669 (TARGET_CPU_CPP_BUILTINS): Update macro
13670 to use new compression flags and to support microMIPS.
13671 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13672 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13673 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13674 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13675 (ASM_SPEC): Support mmicromips and mno-micromips.
13676 (M16STORE_REG_P): New macro.
13677 (MIPS_CALL): Support TARGET_MICROMIPS.
13678 (MICROMIPS_J): New macro.
13679 (mips_base_mips16): Rename this...
13680 (mips_base_compression_flags): ...to this.
13681 (UMIPS_12BIT_OFFSET_P): New macro.
13682 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13683 (MULTILIB_DIRNAMES): Likewise.
13684 2013-03-20 Richard Biener <rguenther@suse.de>
13686 PR tree-optimization/56661
13687 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13688 the result does not have to be distinct.
13690 2013-03-20 Richard Biener <rguenther@suse.de>
13692 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13695 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13696 Steven Bosscher <steven@gcc.gnu.org>
13698 PR rtl-optimization/56605
13699 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13701 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13704 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13705 that require movd instead of movq.
13707 2013-03-20 Richard Biener <rguenther@suse.de>
13709 * tree-ssa-structalias.c (struct variable_info): Add pointer
13710 to the first field of an aggregate with sub-vars. Make
13711 this and the pointer to the next subfield its ID.
13712 (vi_next): New function.
13713 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13714 storedanything_id, integer_id): Increment by one.
13715 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13716 get_call_clobber_vi): Adjust.
13717 (solution_set_expand): Simplify and speedup.
13718 (solution_set_add): Inline into ...
13719 (set_union_with_increment): ... this. Adjust accordingly.
13720 (do_sd_constraint): Likewise.
13721 (do_ds_constraint): Likewise.
13722 (do_complex_constraint): Simplify.
13723 (build_pred_graph): Adjust.
13724 (solve_graph): Likewise. Simplify and speedup.
13725 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13726 get_constraint_for_component_ref, get_constraint_for_1,
13727 first_vi_for_offset, first_or_preceding_vi_for_offset,
13728 create_function_info_for, create_variable_info_for_1,
13729 create_variable_info_for, intra_create_variable_infos): Adjust.
13730 (init_base_vars): Push NULL for ID zero.
13731 (compute_points_to_sets): Adjust.
13733 2013-03-20 Richard Biener <rguenther@suse.de>
13735 * cfgloop.c (verify_loop_structure): Streamline and avoid
13736 ICEing on corrupt loop tree.
13737 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13740 2013-03-20 Richard Biener <rguenther@suse.de>
13742 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13743 check whether an SSA update is needed.
13745 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13747 * config/mips/constraints.md (T): Rename to...
13751 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13752 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13754 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13756 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13757 (*subsi3_carryin_uxtw): Likewise.
13759 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13761 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13762 (*rorsi3_insn_uxtw): Likewise.
13764 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13766 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
13767 (*extrsi5_insn_uxtw): Likewise.
13769 2013-03-19 Richard Biener <rguenther@suse.de>
13771 PR tree-optimization/56273
13772 * passes.c (init_optimization_passes): Move second VRP after DOM.
13774 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
13776 * config/i386/i386.md (*movti_internal): Merge from
13777 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
13778 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
13779 nox64 isa attributes.
13781 2013-03-18 Richard Biener <rguenther@suse.de>
13783 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
13785 (merge_node_constraints): Likewise.
13786 (build_succ_graph): Likewise.
13787 (valid_graph_edge): Inline into single caller.
13788 (unify_nodes): Likewise. Use bitmap_set_bit return value
13790 (scc_visit): Fix formatting and variable use.
13791 (do_sd_constraint): Use gcc_checking_assert.
13792 (do_ds_constraint): Likewise.
13793 (do_complex_constraint): Likewise.
13794 (condense_visit): Likewise. Cleanup.
13795 (dump_pred_graph): New function.
13796 (perform_var_substitution): Dump the pred-graph before
13797 variable substitution.
13798 (find_equivalent_node): Use gcc_checking_assert.
13799 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
13801 2013-03-18 Richard Biener <rguenther@suse.de>
13803 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13804 Remove cond_expr_stmt_list argument and do not gimplify the
13806 (vect_loop_versioning): Adjust.
13807 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
13808 Cleanup to use less temporaries.
13809 (vect_create_data_ref_ptr): Cleanup.
13811 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13813 PR tree-optimization/56635
13814 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
13815 require types_compatible_p types.
13817 2013-03-18 Nick Clifton <nickc@redhat.com>
13819 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
13820 spurious backslash.
13822 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
13823 Add missing line to comment describing function.
13825 2013-03-18 Richard Biener <rguenther@suse.de>
13827 PR tree-optimization/56210
13828 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
13829 Handle string / character search functions.
13830 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
13832 2013-03-18 Richard Biener <rguenther@suse.de>
13834 PR middle-end/56483
13835 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
13836 and implement properly.
13837 * gimple.h (gimple_cond_single_var_p): Remove.
13839 2013-03-18 Richard Biener <rguenther@suse.de>
13841 * tree-data-ref.h (find_data_references_in_loop): Declare.
13842 * tree-data-ref.c (get_references_in_stmt): Use a stack
13843 vector pre-allocated in the callers.
13844 (find_data_references_in_stmt): Adjust.
13845 (graphite_find_data_references_in_stmt): Likewise.
13846 (create_rdg_vertices): Likewise.
13847 (find_data_references_in_loop): Export.
13848 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
13849 Compute dependences here...
13850 (vect_analyze_data_refs): ...not here. When we encounter
13851 a non-vectorizable data reference in basic-block vectorization
13852 truncate the data reference vector. Do not bother to
13853 fixup data-dependence information for gather loads.
13854 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
13855 of data references, as reported.
13857 2013-03-18 Richard Biener <rguenther@suse.de>
13859 PR tree-optimization/3713
13860 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
13861 has_constants and expr.
13862 (stmt_has_constants): Properly valueize SSA names when deciding
13863 whether the stmt has constants.
13865 2013-03-18 Richard Biener <rguenther@suse.de>
13867 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
13868 whole function when there is nothing to do.
13869 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
13870 * tree-vectorizer.c (vectorize_loops): Update virtual and
13871 loop-closed SSA once.
13872 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
13874 2013-03-18 Richard Biener <rguenther@suse.de>
13876 PR middle-end/56113
13877 * domwalk.c (bb_postorder): New global static.
13878 (cmp_bb_postorder): New function.
13879 (walk_dominator_tree): Replace scheme imposing an order for
13880 visiting dominator sons by one sorting them at the time they
13881 are pushed on the stack.
13883 2013-03-18 Richard Biener <rguenther@suse.de>
13885 PR tree-optimization/39326
13886 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
13887 (struct mem_ref): Replace mem member with ao_ref typed member.
13888 (MEM_ANALYZABLE): Adjust.
13889 (memref_eq): Likewise.
13890 (mem_ref_alloc): Likewise.
13891 (gather_mem_refs_stmt): Likewise.
13892 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
13893 (execute_sm_if_changed_flag_set): Adjust.
13894 (execute_sm): Likewise.
13895 (ref_always_accessed_p): Likewise.
13896 (refs_independent_p): Likewise.
13897 (can_sm_ref_p): Likewise.
13899 2013-03-18 Jakub Jelinek <jakub@redhat.com>
13902 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
13903 return 1 even for !unsignedp.
13905 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13907 * config/i386/i386.md (isa): Add x64 and nox64.
13908 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
13909 (*pushtf): Enable *roF alternative for x64 isa only.
13910 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
13911 mode attribute of integer alternatives to DImode for TARGET_64BIT.
13912 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
13913 (*movtf_internal): Merge from *movtf_internal_rex64 and
13914 *movtf_internal_sse. Use x64 and nox64 isa attributes.
13915 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
13916 nox64 isa attributes.
13917 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
13918 nox64 isa attributes.
13919 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
13921 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13923 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13925 2013-03-17 Jakub Jelinek <jakub@redhat.com>
13928 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
13929 class names. Remove trailing comma after "ALL_REGS".
13931 2013-03-16 Jan Hubicka <jh@suse.cz>
13933 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
13934 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
13935 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
13936 of cgraph_get_create_node.
13937 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
13939 2013-03-16 Jason Merrill <jason@redhat.com>
13942 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
13943 with DW_AT_default_value.
13945 2013-03-16 Jakub Jelinek <jakub@redhat.com>
13947 * BASE-VER: Set to 4.9.0.
13949 2013-03-14 Andi Kleen <ak@linux.intel.com>
13952 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
13953 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
13954 Document _x* TSX intrinsics.
13956 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
13957 David Holsgrove <david.holsgrove@xilinx.com>
13959 * configure.ac: Add MicroBlaze TLS support detection.
13960 * configure: Regenerate.
13961 * config/microblaze/microblaze-protos.h
13962 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
13963 symbol_mentioned_p, label_mentioned_p): Add prototypes.
13964 * config/microblaze/microblaze.c (microblaze_address_type): Add
13965 ADDRESS_TLS and tls_reloc address types.
13966 (microblaze_address_info): Add tls_reloc.
13967 (TARGET_HAVE_TLS): Define.
13968 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
13969 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
13970 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
13971 load_tls_operand, microblaze_call_tls_get_addr,
13972 microblaze_legitimize_tls_address): New functions.
13973 (microblaze_classify_unspec): Handle UNSPEC_TLS.
13974 (get_base_reg): Use microblaze_tls_symbol_p.
13975 (microblaze_classify_address): Handle TLS.
13976 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
13977 label_mentioned_p and microblaze_tls_referenced_p.
13978 (microblaze_legitimize_address): Handle TLS.
13979 (microblaze_address_insns): Handle ADDRESS_TLS.
13980 (pic_address_needs_scratch): Handle TLS.
13981 (print_operand_address): Handle TLS.
13982 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
13983 (microblaze_expand_move): Handle TLS.
13984 (microblaze_legitimate_constant_p): Check
13985 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
13986 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
13987 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
13988 (PIC_OFFSET_TABLE_REGNUM): Set.
13989 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
13990 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
13991 (addsi3, movsi_internal2, movdf_internal): Update constraints
13992 * config/microblaze/predicates.md (arith_plus_operand): Define
13993 (move_operand): Redefine as move_src_operand,
13994 check microblaze_tls_referenced_p.
13996 2013-03-14 Ian Bolton <ian.bolton@arm.com>
13998 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
13999 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14001 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14003 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14006 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14008 PR tree-optimization/53265
14009 * common.opt (Waggressive-loop-optimizations): New option.
14010 * tree-ssa-loop-niter.c: Include tree-pass.h.
14011 (do_warn_aggressive_loop_optimizations): New function.
14012 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14013 if number_of_latch_executions returned constant.
14014 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14015 early. If number_of_latch_executions returned constant, set
14016 nb_iterations_upper_bound back to it.
14017 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14019 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14020 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14022 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14023 (MULTILIB_OSDIRNAMES): Set.
14024 * genmultilib: If defaultosdirname doesn't start with :: , set
14025 defaultosdirname2 instead, clear it and emit two . multilib_raw
14026 entries instead of just one.
14028 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14030 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14031 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14032 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14033 (SUBTARGET_OVERRIDE_OPTIONS): New.
14035 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14038 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14039 (musermode): Convert to Var(TARGET_USERMODE).
14040 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14041 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14042 * config/sh/sh.c (sh_option_override): Use
14043 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14044 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14046 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14048 (udivsi3_i4_single, divsi3_i4_single): Use
14049 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14051 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14053 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14056 2013-03-13 Richard Biener <rguenther@suse.de>
14058 PR tree-optimization/56608
14059 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14060 calls when vectorizing basic-blocks.
14062 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14065 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14068 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14070 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14072 2013-03-11 Jan Hubicka <jh@suse.cz>
14075 * lto-streamer-out.c (output_symbol_p): Skip references from
14076 constructors of external variables.
14078 2013-03-11 Jan Hubicka <jh@suse.cz>
14080 PR middle-end/56571
14081 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14083 * emit-rtl.c (verify_rtx_sharing): Likewise.
14084 (copy_insn_1): Likewise.
14085 * rtl.c (copy_rtx): Likewise.
14087 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14090 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14091 output_operand_lossage message.
14093 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14096 * arm.c (shift_op): Validate RTL pattern on the fly.
14097 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14100 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14103 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14104 new call patterns that clobber %fr12.
14105 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14106 split and postreload patterns.
14107 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14108 registers %fr12 and %fr12R as call used.
14110 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14112 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14113 (canon_address, record_store, replace_read, check_mem_read_rtx,
14114 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14115 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14116 rest_of_handle_dse): Likewise.
14118 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14120 PR middle-end/56524
14121 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14123 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14124 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14125 (save_optabs_if_changed): Replace with...
14126 (init_tree_optimization_optabs): ...this.
14127 * optabs.c (save_optabs_if_changed): Rename to...
14128 (init_tree_optimization_optabs): ...this. Take the optimization node
14129 as argument. Do nothing if the base optabs are already correct.
14130 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14131 to recompute optabs.
14132 * function.h (function): Remove optabs field.
14133 * function.c (invoke_set_current_function_hook): Call
14134 init_tree_optimization_optabs. Use the result to initialize
14137 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14139 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14140 if GTMA_HAS_NO_INSTRUMENTATION.
14141 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14142 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14143 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
14144 * gimple-pretty-print.c (dump_gimple_transaction): Handle
14145 GTMA_HAS_NO_INSTRUMENTATION.
14147 2013-03-08 Jakub Jelinek <jakub@redhat.com>
14149 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
14152 2013-03-08 Marek Polacek <polacek@redhat.com>
14153 Jakub Jelinek <jakub@redhat.com>
14155 PR tree-optimization/56478
14156 * predict.c (is_comparison_with_loop_invariant_p): Change the
14157 type of loop_step to tree.
14158 (predict_loops): Adjust.
14159 (predict_iv_comparison): Perform the computations on double_ints.
14161 2013-03-08 Richard Biener <rguenther@suse.de>
14163 PR tree-optimization/56570
14164 * tree-cfg.c (verify_expr_location_1): Verify locations for
14166 * tree-sra.c (create_access_replacement): Strip locations
14167 from DECL_DEBUG_EXPRs.
14169 2013-03-08 Richard Biener <rguenther@suse.de>
14171 * tree-inline.c (expand_call_inline): Do not associate
14172 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
14173 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
14175 2013-03-08 Richard Biener <rguenther@suse.de>
14177 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
14178 or block changes with -Og. Fix for location / block encoding
14179 changes and PHI arguments with locations.
14181 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
14183 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
14185 (struct output_info): Likewise.
14186 (register_overhead): Remove bad gcc_assert.
14187 (bitmap_find_bit): If there is only a single bitmap element, do not
14188 count a miss as a search.
14189 (print_statistics): Update for counter type changes.
14190 (dump_bitmap_statistics): Likewise. Print headers such that they
14191 are properly lined up with the printed counters.
14193 2013-03-07 Jakub Jelinek <jakub@redhat.com>
14195 PR tree-optimization/56559
14196 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
14197 check that it has only a single use.
14199 2013-03-07 Richard Biener <rguenther@suse.de>
14201 * doc/invoke.texi (fwhole-program): Discourage use in combination
14204 2013-03-06 Jakub Jelinek <jakub@redhat.com>
14206 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
14208 PR tree-optimization/56539
14209 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
14210 instead of GSI_CONTINUE_LINKING as last argument to
14211 force_gimple_operand_gsi. Adjust function comment.
14213 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
14216 PR middle-end/56548
14217 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
14218 promoted mode, convert the result back to the original mode.
14220 2013-03-06 Richard Biener <rguenther@suse.de>
14222 PR middle-end/56294
14223 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
14224 (insert_updated_phi_nodes_compare_uids): New function.
14225 (update_ssa): Sort symbols_to_rename after UID before
14226 traversing it to insert PHI nodes.
14228 2013-03-06 Richard Biener <rguenther@suse.de>
14230 PR middle-end/50494
14231 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
14232 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
14235 2013-02-13 Richard Biener <rguenther@suse.de>
14238 * varasm.c (output_constant_def_1): Get the decl representing
14239 the constant as argument.
14240 (output_constant_def): Wrap output_constant_def_1.
14241 (make_decl_rtl): Use output_constant_def_1 with the decl
14242 representing the constant.
14243 (build_constant_desc): Optionally re-use a decl already
14244 representing the constant.
14245 (tree_output_constant_def): Adjust.
14247 2013-03-06 Joey Ye <joey.ye@arm.com>
14250 * gcc.c (convert_white_space): New function.
14251 (main): Handles white space in function name.
14253 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
14256 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
14257 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
14258 to SH_DIV_CALL_TABLE for TARGET_SH2.
14259 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
14261 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
14262 call-table options.
14264 2013-03-05 Sterling Augustine <saugustine@google.com>
14265 Cary Coutant <ccoutant@google.com>
14268 * dwarf2out.c (resolve_addr): Don't call
14269 remove_loc_list_addr_table_entries a second time for the same
14272 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14275 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
14276 (avoid_complex_debug_insns): New function.
14277 (expand_debug_locations): Call it.
14279 PR rtl-optimization/56484
14280 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
14281 lifetimes of hard registers on small register class machines.
14283 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14285 * config/microblaze/microblaze-protos.h: Rename
14286 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
14287 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
14289 (microblaze_fast_interrupt_function_p): New function.
14290 (microblaze_is_interrupt_handler): Rename to
14291 microblaze_is_interrupt_variant and add fast_interrupt check.
14292 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
14293 (save_restore_insns): Likewise.
14294 (compute_frame_size): Likewise.
14295 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
14296 (microblaze_globalize_label): Likewise.
14297 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
14298 * config/microblaze/microblaze.md: Use wrapper
14299 microblaze_is_interrupt_variant.
14301 2013-03-05 Kai Tietz <ktietz@redhat.com>
14303 * sdbout.c (sdbout_one_type): Switch to current function's section
14304 supporting cold/hot.
14306 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14308 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
14311 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14313 PR middle-end/56461
14314 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
14315 if VALGRIND_GET_VBITS is defined, temporarily make object
14316 memory all defined, and restore previous valgrind addressability
14317 and definability afterwards. Free this_object at the end.
14319 PR middle-end/56461
14320 * lra.c (lra): Call lra_clear_live_ranges if live_p,
14321 right before calling lra_create_live_ranges, also call it
14322 when clearing live_p. Only call lra_clear_live_ranges
14323 at the end if live_p.
14325 PR middle-end/56461
14326 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
14328 2013-03-05 Richard Biener <rguenther@suse.de>
14330 PR tree-optimization/56521
14331 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
14334 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
14337 * except.h (remove_unreachable_eh_regions): New prototype.
14338 * except.c (remove_eh_handler_splicer): New function, split out
14339 of remove_eh_handler.
14340 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
14341 warning about running it on many EH regions one at a time.
14342 (remove_unreachable_eh_regions_worker): New function, walk the
14343 EH tree in depth-first order and remove non-marked regions.
14344 (remove_unreachable_eh_regions): New function.
14345 * tree-eh.c (mark_reachable_handlers): New function, split out
14346 from remove_unreachable_handlers.
14347 (remove_unreachable_handlers): Use mark_reachable_handlers and
14348 remove_unreachable_eh_regions.
14349 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
14350 and remove_unreachable_eh_regions.
14352 2013-03-05 Richard Biener <rguenther@suse.de>
14354 PR middle-end/56525
14355 * loop-init.c (fix_loop_structure): Remove loops in two stages,
14356 not freeing them until the end.
14358 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14360 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
14362 2013-03-05 Richard Biener <rguenther@suse.de>
14364 PR tree-optimization/56270
14365 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
14366 of loads after scheduling an SLP instance.
14368 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14370 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
14372 (check_gcc_parallelize): Run guality.exp as a separate job from
14373 vect.exp with unsorted.exp and $(dg_target_exps) separately from
14374 struct-layout-1.exp with stackalign.exp.
14376 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
14378 PR middle-end/56461
14379 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
14380 load_index sbitmap even if some bit in it isn't set.
14382 PR middle-end/56461
14383 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
14384 (discover_iteration_bound_by_body_walk): Change queues to
14385 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
14386 spelling in comment. Call safe_push on queues[bound_index] directly.
14387 Release queues[queue_index] in every iteration unconditionally.
14388 Release bounds vector.
14390 PR middle-end/56461
14391 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
14392 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
14393 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14396 2013-03-05 Richard Biener <rguenther@suse.de>
14399 * tree-inline.c (remap_blocks_to_null): New function.
14400 (expand_call_inline): When expanding a call stmt without
14401 an associated block inline remap all callee blocks to NULL.
14403 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14405 PR rtl-optimization/56494
14406 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14407 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14408 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14410 PR middle-end/56461
14411 * sel-sched-ir.c (free_sched_pools): Release
14412 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14413 if succs_info_pool.max_top isn't -1.
14416 * opts.c (opts_obstack, opts_concat): Moved to...
14417 * opts-common.c (opts_obstack, opts_concat): ... here.
14419 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14421 PR middle-end/56461
14422 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14424 2013-03-04 Martin Jambor <mjambor@suse.cz>
14426 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14427 all appropriate places.
14429 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14431 PR tree-optimization/56424
14432 * ipa-split.c (split_function): Do not set the RSO flag if result is
14433 not by reference and its type is a register type.
14435 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14437 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14438 (microblaze_legitimate_pic_operand): Likewise
14439 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14440 new function microblaze_legitimate_pic_operand
14441 * config/microblaze/microblaze-protos.h
14442 (microblaze_legitimate_pic_operand): Declare.
14444 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14446 * config/microblaze/predicates.md (call_insn_simple_operand):
14447 New predicate for supported rtx code types.
14448 * config/microblaze/microblaze.md (call_internal1): Use
14449 call_insn_simple_operand predicate.
14451 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14453 PR middle-end/56461
14454 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14455 partitions.ordered_remove.
14457 PR middle-end/56461
14458 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14459 vec_oprnds0.create (1) for modifier == NONE.
14461 PR middle-end/56461
14462 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14463 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14464 vec_oprnds1 right before pushing anything to it for
14467 PR middle-end/56461
14468 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14469 set nbbs to 0 instead of having separate code path.
14470 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14471 instead of false as last argument if returning NULL.
14473 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14475 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14476 the attribute is now called "target" instead of "option".
14477 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14478 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14479 attribute/pragma name for TARGET_OPTION_VALID_P and
14480 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14481 * doc/tm.texi: Regenerated.
14483 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14485 * config/microblaze/microblaze.c:
14486 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14487 * config/microblaze/microblaze.h: Add -mxl-reorder to
14489 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14490 instructions emitted if TARGET_REORDER.
14491 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14492 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14495 2013-03-01 Xinliang David Li <davidxl@google.com>
14497 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14500 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14502 PR middle-end/56461
14503 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14504 vector even when returning true. Fix up function comment formatting.
14506 PR middle-end/56461
14507 * ira-build.c (ira_loop_nodes_count): New variable.
14508 (create_loop_tree_nodes): Initialize it.
14509 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14511 PR middle-end/56461
14512 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14513 method on dr_chain and result_chain.
14514 * tree-vect-stmts.c (vectorizable_store): Only call
14515 result_chain.create if j == 0.
14517 PR middle-end/56461
14518 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14519 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14520 before overwriting it.
14522 2013-03-01 Tobias Burnus <burnus@net-b.de>
14524 * doc/extended.texi (C Extensions): Change order in @menu
14526 (Other MIPS Built-in Functions): Move last MIPS entry before
14527 "picoChip Built-in Functions".
14528 (SH Built-in Functions): Move after RX Built-in Functions.
14529 * doc/gcc.texi (Introduction): Change order in @menu
14531 * doc/md.texi (Constraints): Ditto.
14532 * gty.texi (Type Information): Ditto.
14533 (User-provided marking routines for template types): Make
14535 * doc/invoke.texi (AArch64 Options): Move before
14536 "Adapteva Epiphany Options".
14538 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14539 Jakub Jelinek <jakub@redhat.com>
14542 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14543 no_address_safety_analysis attribute.
14544 * doc/extend.texi (no_address_safety_attribute): Rename to
14545 no_sanitize_address attribute, mention no_address_safety_analysis
14546 attribute as deprecated alias.
14548 2013-02-28 Jakub Jelinek <jakub@redhat.com>
14550 PR middle-end/56461
14551 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14552 type to vec<vec<tree> > *.
14553 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14554 to be vec<tree> instead of vec<tree> *, set vec_defs
14555 to vNULL and call vec_defs.create (number_of_vects), adjust other
14557 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14558 vectorizable_condition): Adjust vect_get_slp_defs callers.
14560 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14562 * config/aarch64/aarch64.c
14563 (aarch64_float_const_representable): Remove unused variable.
14565 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14567 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14569 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14571 * config/aarch64/aarch64-builtins.c
14572 (aarch64_init_simd_builtins): Make static.
14574 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14576 * config/aarch64/aarch64.c
14577 (aarch64_simd_make_constant): Make static.
14579 2013-02-28 Martin Jambor <mjambor@suse.cz>
14581 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14582 with no initialization to the RHS of debug statements.
14584 2013-02-28 Martin Jambor <mjambor@suse.cz>
14586 PR tree-optimization/56294
14587 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14589 (get_access_replacement): Do not call create_access_replacement.
14590 Assert a replacement exists.
14591 (get_repl_default_def_ssa_name): Create the replacement declaration
14594 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14596 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14597 final_end_function.
14599 2013-02-28 Marek Polacek <polacek@redhat.com>
14601 PR rtl-optimization/56466
14602 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14603 if we're changing a loop.
14604 (peel_loops_completely): Likewise.
14606 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14609 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14611 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14614 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14615 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14616 INTX_FTYPE_FX, FX_FTYPE_INTX.
14617 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14619 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14621 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14622 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14623 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14624 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14625 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14626 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14627 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14628 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14629 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14630 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14631 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14632 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14633 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14634 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14635 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14636 (avrxmega6): Increase max flash segments from 5 to 6.
14637 * config/avr/t-multilib: Regenerate.
14638 * config/avr/avr-tables.opt: Regenerate.
14639 * doc/avr-mmcu.texi: Regenerate.
14641 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14643 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14644 (avr_device_to_arch): Rename to avr_device_to_ld.
14645 (avr_device_to_as): New prototype.
14646 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14647 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14648 * config/avr/driver-avr.c (avr_device_to_as): New.
14649 (avr_device_to_arch): Rename to avr_device_to_ld.
14651 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14653 PR middle-end/56461
14654 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14655 method on dr_chain and result_chain.
14657 PR middle-end/56461
14658 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14659 pointer_set_destroy on not_executed_last_iteration.
14661 PR middle-end/56461
14662 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
14664 PR middle-end/56461
14665 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14666 FOR_EACH_DEFINED_FUNCTION when freeing state.
14668 PR middle-end/56461
14669 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14671 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14674 PR middle-end/56461
14675 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14676 known_aggs[i].items and release known_aggs vector.
14678 PR middle-end/56461
14679 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14681 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14683 * config/microblaze/microblaze.c (microblaze_emit_compare):
14684 Use xor for EQ/NE comparisions.
14685 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14686 (cbranchsf4): Adjust operator to comparison_operator.
14688 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14690 PR middle-end/56461
14691 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14693 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14694 vec_safe_push, always update *slot.
14695 (redirect_edge_var_map_clear): Use vec_free.
14696 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14697 (free_var_map_entry): Use vec_free.
14698 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14699 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14701 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
14703 PR middle-end/45472
14704 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14705 when the may_trap_p bit of the exprs being merged differs.
14706 Reorder tests for speculativeness in the logical and operator.
14708 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14710 * incpath.c (add_standard_paths): Use reconcat instead of concat
14711 where appropriate and avoid leaking memory.
14713 * opts.h: Include obstack.h.
14714 (opts_concat): New prototype.
14715 (opts_obstack): New declaration.
14716 * opts.c (opts_concat): New function.
14717 (opts_obstack): New variable.
14718 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14719 (finish_options): Use opts_concat instead of concat
14720 and XOBNEWVEC instead of XNEWVEC.
14721 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14722 generate_option): Likewise.
14723 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14724 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14727 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14728 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14730 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14732 PR middle-end/56461
14733 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14735 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14737 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14738 (arm_block_move_unaligned_straight): Likewise.
14739 (arm_adjust_block_mem): Likewise.
14741 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14744 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14745 temp, cond and label.
14746 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14749 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14750 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14753 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14754 of prologue/epilogue functions.
14757 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14758 Remove unused variable cfa_offset.
14759 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14762 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14763 type promotion to unsigned.
14766 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
14767 for HOST_WIDE_INT of 32 bit / same size as int.
14768 (arm_block_move_unaligned_straight): Likewise.
14769 (arm_adjust_block_mem): Likewise.
14772 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
14775 2013-02-26 Marek Polacek <polacek@redhat.com>
14777 PR tree-optimization/56426
14778 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
14780 2013-02-26 Richard Biener <rguenther@suse.de>
14783 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
14784 unused variable loops.
14786 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14788 PR tree-optimization/56448
14789 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
14790 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
14791 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
14792 later operands of the references, or even first operand for
14793 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
14795 PR tree-optimization/56443
14796 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
14797 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
14798 to type_for_mode langhook.
14800 2013-02-25 Matt Turner <mattst88@gmail.com>
14802 * doc/invoke.texi: Document r4700.
14804 2013-02-25 Richard Biener <rguenther@suse.de>
14806 PR tree-optimization/56175
14807 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
14809 (simplify_bitwise_binary): ... here. Also guard the conversion
14810 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
14812 2013-02-25 Catherine Moore <clm@codesourcery.com>
14815 2013-02-24 Catherine Moore <clm@codesourcery.com>
14816 Maciej W. Rozycki <macro@codesourcery.com>
14817 Tom de Vries <tom@codesourcery.com>
14818 Nathan Sidwell <nathan@codesourcery.com>
14819 Iain Sandoe <iain@codesourcery.com>
14820 Nathan Froyd <froydnj@codesourcery.com>
14821 Chao-ying Fu <fu@mips.com>
14823 * doc/extend.texi: (micromips, nomicromips, nocompression):
14824 Document new function attributes.
14825 * doc/invoke.texi (minterlink-compressed, mmicromips,
14826 m14k, m14ke, m14kec): Document new options.
14827 (minterlink-mips16): Update documentation.
14828 * doc/md.texi (ZC, ZD): Document new constraints.
14829 * configure.ac (gcc_cv_as_micromips): Check if linker
14830 supports the .set micromips directive.
14831 * configure: Regenerate.
14832 * config.in: Regenerate.
14833 * config/mips/mips-tables.opt: Regenerate.
14834 * config/mips/micromips.md: New file.
14835 * constraints.md (ZC, AD): New constraints.
14836 * config/mips/predicates.md (movep_src_register): New predicate.
14837 (movep_src_operand): New predicate.
14838 (non_volatile_mem_operand): New predicate.
14839 * config/mips/mips.md (multimem): New type.
14840 (length): Differentiate between 17-bit and 18-bit branch offsets.
14841 (MOVEP1, MOVEP2): New mode iterator.
14842 (mov_<load>l): Use ZC constraint.
14843 (mov_<load>r): Likewise.
14844 (mov_<store>l): Likewise.
14845 (mov_<store>r): Likewise.
14846 (*branch_equality<mode>_inverted): Add microMIPS support.
14847 (*branch_equality<mode>): Likewise.
14848 (*jump_absolute): Likewise.
14849 (indirect_jump_<mode>): Likewise.
14850 (tablejump_<mode>): Likewise.
14851 (<optab>_internal): Likewise.
14852 (sibcall_internal): Likewise.
14853 (sibcall_value_internal): Likewise.
14854 (prefetch): Use constraint ZD.
14855 * config/mips/mips.opt (minterlink-compressed): New option.
14856 (minterlink-mips16): Now an alias for minterlink-compressed.
14857 (mmicromips): New option.
14858 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14859 (compare_and_swap_12): Likewise.
14860 (sync_add<mode>): Likewise.
14861 (sync_<optab>_12): Likewise.
14862 (sync_old_<optab>_12): Likewise.
14863 (sync_new_<optab>_12): Likewise.
14864 (sync_nand_12): Likewise.
14865 (sync_old_nand_12): Likewise.
14866 (sync_new_nand_12): Likewise.
14867 (sync_sub<mode>): Likewise.
14868 (sync_old_add<mode>): Likewise.
14869 (sync_old_sub<mode>): Likewise.
14870 (sync_new_add<mode>): Likewise.
14871 (sync_new_sub<mode>): Likewise.
14872 (sync_<optab><mode>): Likewise.
14873 (sync_old_<optab><mode>): Likewise.
14874 (sync_new_<optab><mode>): Likewise.
14875 (sync_nand<mode>): Likewise.
14876 (sync_old_nand<mode>): Likewise.
14877 (sync_new_nand<mode>): Likewise.
14878 (sync_lock_test_and_set<mode>): Likewise.
14879 (test_and_set_12): Likewise.
14880 (atomic_compare_and_swap<mode>): Likewise.
14881 (atomic_exchange<mode>_llsc): Likewise.
14882 (atomic_fetch_add<mode>_llsc): Likewise.
14883 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14884 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14885 (umips_save_restore_pattern_p): Likewise.
14886 (umips_load_store_pair_p): Likewise.
14887 (umips_output_load_store_pair): Likewise.
14888 (umips_movep_target_p): Likewise.
14889 (umips_12bit_offset_address_p): Likewise.
14890 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14891 (mips_base_mips16): Rename this...
14892 (mips_base_compression_flags): ...to this. Update all uses.
14893 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14894 (mips_mips16_decl_p): Delete.
14895 (mips_nomips16_decl_p): Delete.
14896 (mips_get_compress_on_flags): New function.
14897 (mips_get_compress_off_flags): New function.
14898 (mips_get_compress_mode): New function.
14899 (mips_get_compress_on_name): New function.
14900 (mips_get_compress_off_name): New function.
14901 (mips_insert_attributes): Support multiple compression types.
14902 (mips_merge_decl_attributes): Likewise.
14903 (umips_12bit_offset_address_p): New function.
14904 (mips_start_function_definition): Emit .set micromips directive.
14905 (mips_call_may_need_jalx_p): New function.
14906 (mips_function_ok_for_sibcall): Add microMIPS support.
14907 (mips_print_operand_punctuation): Support short delay slots and
14909 (umips_swm_mask, umips_swm_encoding): New.
14910 (umips_build_save_restore): New function.
14911 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14912 (was_mips16_p): Remove.
14913 (old_compression_mode): New.
14914 (mips_set_compression_mode): New function.
14915 (mips_set_current_function): Add microMIPS support.
14916 (mips_option_override): Likewise.
14917 (umips_save_restore_pattern_p): New function.
14918 (umips_output_save_restore): New function.
14919 (umips_load_store_pair_p_1): New function.
14920 (umips_load_store_pair_p): New function.
14921 (umips_output_load_store_pair_1): New function.
14922 (umips_output_load_store_pair): New function.
14923 (umips_movep_target_p) New function.
14924 (mips_prepare_pch_save): Add microMIPS support.
14925 * config/mips/mips.h (TARGET_COMPRESSION): New.
14926 (TARGET_CPU_CPP_BUILTINS): Update macro
14927 to use new compression flags and to support microMIPS.
14928 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14929 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14930 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14931 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14932 (ASM_SPEC): Support mmicromips and mno-micromips.
14933 (M16STORE_REG_P): New macro.
14934 (MIPS_CALL): Support TARGET_MICROMIPS.
14935 (MICROMIPS_J): New macro.
14936 (mips_base_mips16): Rename this...
14937 (mips_base_compression_flags): ...to this.
14938 (UMIPS_12BIT_OFFSET_P): New macro.
14939 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14940 (MULTILIB_DIRNAMES): Likewise.
14942 2013-02-25 Tom de Vries <tom@codesourcery.com>
14944 PR rtl-optimization/56131
14945 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
14946 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
14947 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
14949 2013-02-25 Tobias Burnus <burnus@net-b.de>
14951 * doc/invoke.texi (-fsanitize=): Move from optimization
14952 to debugging options.
14954 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14956 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
14958 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
14959 Alexander Monakov <amonakov@ispras.ru>
14961 PR middle-end/56077
14962 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
14963 flush pending lists also on non-jumps. Adjust comment.
14965 2013-02-24 Catherine Moore <clm@codesourcery.com>
14966 Maciej W. Rozycki <macro@codesourcery.com>
14967 Tom de Vries <tom@codesourcery.com>
14968 Nathan Sidwell <nathan@codesourcery.com>
14969 Iain Sandoe <iain@codesourcery.com>
14970 Nathan Froyd <froydnj@codesourcery.com>
14971 Chao-ying Fu <fu@mips.com>
14973 * doc/extend.texi: (micromips, nomicromips, nocompression):
14974 Document new function attributes.
14975 * doc/invoke.texi (minterlink-compressed, mmicromips,
14976 m14k, m14ke, m14kec): Document new options.
14977 (minterlink-mips16): Update documentation.
14978 * doc/md.texi (ZC, ZD): Document new constraints.
14979 * configure.ac (gcc_cv_as_micromips): Check if linker
14980 supports the .set micromips directive.
14981 * configure: Regenerate.
14982 * config.in: Regenerate.
14983 * config/mips/mips-tables.opt: Regenerate.
14984 * config/mips/micromips.md: New file.
14985 * constraints.md (ZC, AD): New constraints.
14986 * config/mips/predicates.md (movep_src_register): New predicate.
14987 (movep_src_operand): New predicate.
14988 (non_volatile_mem_operand): New predicate.
14989 * config/mips/mips.md (multimem): New type.
14990 (length): Differentiate between 17-bit and 18-bit branch offsets.
14991 (MOVEP1, MOVEP2): New mode iterator.
14992 (mov_<load>l): Use ZC constraint.
14993 (mov_<load>r): Likewise.
14994 (mov_<store>l): Likewise.
14995 (mov_<store>r): Likewise.
14996 (*branch_equality<mode>_inverted): Add microMIPS support.
14997 (*branch_equality<mode>): Likewise.
14998 (*jump_absolute): Likewise.
14999 (indirect_jump_<mode>): Likewise.
15000 (tablejump_<mode>): Likewise.
15001 (<optab>_internal): Likewise.
15002 (sibcall_internal): Likewise.
15003 (sibcall_value_internal): Likewise.
15004 (prefetch): Use constraint ZD.
15005 * config/mips/mips.opt (minterlink-compressed): New option.
15006 (minterlink-mips16): Now an alias for minterlink-compressed.
15007 (mmicromips): New option.
15008 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15009 (compare_and_swap_12): Likewise.
15010 (sync_add<mode>): Likewise.
15011 (sync_<optab>_12): Likewise.
15012 (sync_old_<optab>_12): Likewise.
15013 (sync_new_<optab>_12): Likewise.
15014 (sync_nand_12): Likewise.
15015 (sync_old_nand_12): Likewise.
15016 (sync_new_nand_12): Likewise.
15017 (sync_sub<mode>): Likewise.
15018 (sync_old_add<mode>): Likewise.
15019 (sync_old_sub<mode>): Likewise.
15020 (sync_new_add<mode>): Likewise.
15021 (sync_new_sub<mode>): Likewise.
15022 (sync_<optab><mode>): Likewise.
15023 (sync_old_<optab><mode>): Likewise.
15024 (sync_new_<optab><mode>): Likewise.
15025 (sync_nand<mode>): Likewise.
15026 (sync_old_nand<mode>): Likewise.
15027 (sync_new_nand<mode>): Likewise.
15028 (sync_lock_test_and_set<mode>): Likewise.
15029 (test_and_set_12): Likewise.
15030 (atomic_compare_and_swap<mode>): Likewise.
15031 (atomic_exchange<mode>_llsc): Likewise.
15032 (atomic_fetch_add<mode>_llsc): Likewise.
15033 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15034 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15035 (umips_save_restore_pattern_p): Likewise.
15036 (umips_load_store_pair_p): Likewise.
15037 (umips_output_load_store_pair): Likewise.
15038 (umips_movep_target_p): Likewise.
15039 (umips_12bit_offset_address_p): Likewise.
15040 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15041 (mips_base_mips16): Rename this...
15042 (mips_base_compression_flags): ...to this. Update all uses.
15043 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15044 (mips_mips16_decl_p): Delete.
15045 (mips_nomips16_decl_p): Delete.
15046 (mips_get_compress_on_flags): New function.
15047 (mips_get_compress_off_flags): New function.
15048 (mips_get_compress_mode): New function.
15049 (mips_get_compress_on_name): New function.
15050 (mips_get_compress_off_name): New function.
15051 (mips_insert_attributes): Support multiple compression types.
15052 (mips_merge_decl_attributes): Likewise.
15053 (umips_12bit_offset_address_p): New function.
15054 (mips_start_function_definition): Emit .set micromips directive.
15055 (mips_call_may_need_jalx_p): New function.
15056 (mips_function_ok_for_sibcall): Add microMIPS support.
15057 (mips_print_operand_punctuation): Support short delay slots and
15059 (umips_swm_mask, umips_swm_encoding): New.
15060 (umips_build_save_restore): New function.
15061 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15062 (was_mips16_p): Remove.
15063 (old_compression_mode): New.
15064 (mips_set_compression_mode): New function.
15065 (mips_set_current_function): Add microMIPS support.
15066 (mips_option_override): Likewise.
15067 (umips_save_restore_pattern_p): New function.
15068 (umips_output_save_restore): New function.
15069 (umips_load_store_pair_p_1): New function.
15070 (umips_load_store_pair_p): New function.
15071 (umips_output_load_store_pair_1): New function.
15072 (umips_output_load_store_pair): New function.
15073 (umips_movep_target_p) New function.
15074 (mips_prepare_pch_save): Add microMIPS support.
15075 * config/mips/mips.h (TARGET_COMPRESSION): New.
15076 (TARGET_CPU_CPP_BUILTINS): Update macro
15077 to use new compression flags and to support microMIPS.
15078 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15079 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15080 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15081 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15082 (ASM_SPEC): Support mmicromips and mno-micromips.
15083 (M16STORE_REG_P): New macro.
15084 (MIPS_CALL): Support TARGET_MICROMIPS.
15085 (MICROMIPS_J): New macro.
15086 (mips_base_mips16): Rename this...
15087 (mips_base_compression_flags): ...to this.
15088 (UMIPS_12BIT_OFFSET_P): New macro.
15089 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15090 (MULTILIB_DIRNAMES): Likewise.
15092 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15095 * target-globals.c (save_target_globals): For init_reg_sets and
15096 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15098 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15100 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15101 * config/aarch64/t-aarch64
15102 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15104 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15106 PR inline-asm/56148
15107 * lra-constraints.c (process_alt_operands): Reload operand
15108 conflicting with earlier clobber only if no more other conflicting
15111 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15114 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15115 if not linking a shared library.
15117 2013-02-22 Seth LaForge <sethml@google.com>
15119 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15121 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15123 * config/arm/arm.md (split for extendsidi): Update condition.
15124 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15125 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15126 (qhs_zextenddi_cstr): Likewise.
15128 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15130 PR middle-end/56420
15131 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15132 avoid signed wrapping.
15133 (expand_mult): Handle properly multiplication by
15134 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15135 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15136 in the compiler if coeff is HOST_WIDE_INT_MIN.
15137 (expand_divmod): Don't make ext_op1 static, change it's type to
15138 uhwi. Avoid undefined behavior in -INTVAL (op1).
15140 PR rtl-optimization/50339
15141 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15143 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
15144 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
15145 into splitting_ashiftrt field.
15146 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
15148 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
15151 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
15153 PR middle-end/56108
15154 * trans-mem.c (execute_tm_mark): Do not expand transactions that
15155 are sure to go irrevocable.
15157 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
15159 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
15160 scalars are valid operands.
15162 2013-02-21 Martin Jambor <mjambor@suse.cz>
15164 PR tree-optimization/56310
15165 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
15166 only matching indices and non-negative final offsets.
15167 (intersect_aggregates_with_edge): Pass src_idx to
15168 agg_replacements_to_vector. Pass src_idx insstead of index to
15169 intersect_with_agg_replacements.
15171 2013-02-21 Martin Jambor <mjambor@suse.cz>
15173 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
15174 instead of hard-wired defaults.
15176 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
15178 * doc/invoke.texi (MIPS Options): Update documentation of the
15179 floating-point multiply-accumulate instruction restrictions.
15181 2013-02-21 Kostya Serebryany <kcc@google.com>
15183 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
15184 asan_shadow_offset on x86_64 linux.
15186 2013-02-21 Richard Biener <rguenther@suse.de>
15188 PR tree-optimization/56415
15190 2013-02-11 Richard Biener <rguenther@suse.de>
15192 PR tree-optimization/56273
15193 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15196 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15199 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
15202 PR inline-asm/56405
15203 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
15204 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
15206 2013-02-20 Jan Hubicka <jh@suse.cz>
15208 PR tree-optimization/56265
15209 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
15210 when target is referenced for first time.
15212 2013-02-20 Richard Biener <rguenther@suse.de>
15214 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
15215 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
15216 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
15217 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
15218 not return anything.
15219 (rename_ssa_copies): Do not remove unused locals.
15220 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
15221 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
15222 * passes.c (execute_function_todo): Do not schedule unused locals
15223 removal if cleanup_tree_cfg did something.
15224 * tree-ssa-live.c (remove_unused_locals): Dump statistics
15225 about the number of removed locals.
15227 2013-02-20 Richard Biener <rguenther@suse.de>
15229 PR tree-optimization/56398
15230 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
15232 2013-02-20 Martin Jambor <mjambor@suse.cz>
15234 PR tree-optimization/55334
15235 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
15236 restricted pointers to arrays.
15238 2013-02-20 Richard Biener <rguenther@suse.de>
15239 Jakub Jelinek <jakub@redhat.com>
15241 PR tree-optimization/56396
15242 * tree-ssa-ccp.c (n_const_val): New static variable.
15243 (get_value): Return NULL for SSA names we don't have a lattice
15245 (ccp_initialize): Initialize n_const_val.
15246 * tree-ssa-copy.c (n_copy_of): New static variable.
15247 (init_copy_prop): Initialize n_copy_of.
15248 (get_value): Return NULL_TREE for SSA names we don't have a
15251 2013-02-20 Martin Jambor <mjambor@suse.cz>
15253 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
15255 2013-02-20 Richard Biener <rguenther@suse.de>
15257 * genpreds.c (write_lookup_constraint): Do not compare first
15258 letter of the constraint again.
15260 2013-02-20 Richard Biener <rguenther@suse.de>
15262 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
15264 (get_use_iv_cost): Terminate hashtable walk when coming across
15267 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
15269 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
15270 reassociation for avx2 targets.
15272 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15274 * config/microblaze/microblaze.c: microblaze_has_clz = 0
15275 Add version check for v8.10.a to enable microblaze_has_clz
15276 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
15277 version and TARGET_PATTERN_COMPARE check
15278 * config/microblaze/microblaze.md: New clzsi2 instruction
15280 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15282 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
15283 function before branching.
15285 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15287 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
15289 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
15291 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15293 PR middle-end/55889
15294 * sel-sched.c: Include ira.h.
15295 (implicit_clobber_conflict_p): New function.
15296 (moveup_expr): Use it.
15297 * Makefile.in (sel-sched.o): Depend on ira.h.
15299 2013-02-19 Richard Biener <rguenther@suse.de>
15301 PR tree-optimization/56384
15302 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
15303 (vn_hash_type): Split out from ...
15304 (vn_hash_constant_with_type): ... here.
15305 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
15306 (vn_phi_eq): Compare types from vn_phi_s structure.
15307 (vn_phi_lookup): Populate vn_phi_s type.
15308 (vn_phi_insert): Likewise.
15310 2013-02-19 Jakub Jelinek <jakub@redhat.com>
15312 PR tree-optimization/56350
15313 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
15314 if haven't found reduction or nested cycle operand, rather than
15315 asserting we must find it.
15317 PR tree-optimization/56381
15318 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
15321 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
15322 Jakub Jelinek <jakub@redhat.com>
15325 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
15326 (swap_optab_enable): Same.
15327 (init_all_optabs): Use argument instead of global.
15328 * tree.h (struct tree_optimization_option): New field target_optabs.
15329 * expr.h (init_all_optabs): Add argument to prototype.
15330 (TREE_OPTIMIZATION_OPTABS): New.
15331 (save_optabs_if_changed): Protoize.
15332 * optabs.h: Declare this_fn_optabs.
15333 * optabs.c (save_optabs_if_changed): New.
15334 Declare this_fn_optabs.
15335 (init_optabs): Add argument to init_all_optabs() call.
15336 * function.c (invoke_set_current_function_hook): Handle per
15338 * function.h (struct function): New field optabs.
15339 * config/mips/mips.c (mips_set_mips16_mode): Handle when
15340 optimization_current_node has changed.
15341 * target-globals.h (save_target_globals_default_opts): Protoize.
15342 * target-globals.c (save_target_globals_default_opts): New.
15344 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15347 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
15348 registers %fr12 and %fr12R as call used.
15351 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
15352 and HImode, require all displacements to be an integer multiple of
15354 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
15355 only allow QImode and HImode when reload is in progress and strict is
15356 true. Likewise for symbolic addresses. Use base14_operand to check
15357 displacements in REG+BASE addresses.
15359 2013-02-18 Richard Biener <rguenther@suse.de>
15361 PR tree-optimization/56366
15362 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15363 handle sign-conversion of outer-loop initial induction value.
15365 2013-02-18 Richard Biener <rguenther@suse.de>
15367 PR middle-end/56349
15368 * cfghooks.c (merge_blocks): If we merge a latch into another
15369 block adjust references to it.
15370 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
15371 (verify_loop_structure): Verify that a recorded latch is in fact
15374 2013-02-18 Richard Biener <rguenther@suse.de>
15376 PR tree-optimization/56321
15377 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
15378 order SSA name release and virtual operand unlinking.
15380 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15382 * config/microblaze/microblaze.md (save_stack_block): Define.
15383 (restore_stack_block): Likewise.
15385 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15387 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
15388 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
15389 * config/microblaze/microblaze.c (microblaze_option_override):
15390 Bail out early for PIC modes when target does not support PIC.
15392 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15394 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
15395 Replace with a microblaze version.
15396 (microblaze_trampoline_init): Adapt for microblaze.
15397 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15400 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15401 Dodji Seketeli <dodji@redhat.com>
15404 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15405 (instrument_mem_region_access): Do not forget to always put
15406 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15407 0) statement, even for cases where either 'base' or 'base + len'
15408 are not instrumented -- because they have been previously
15409 instrumented. Simplify the logic by putting all the statements
15410 instrument 'base + len' inside a sequence, and then insert that
15411 sequence right before the current insertion point. Then, to
15412 instrument 'base + len', just get an iterator on that statement.
15413 And do not forget to update the pointer to iterator the function
15414 received as argument.
15416 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15418 PR rtl-optimization/56348
15419 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15421 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15423 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15424 (clean_graph_dump_file): Pass base to start_graph_dump.
15426 2013-02-14 Richard Henderson <rth@redhat.com>
15429 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15431 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15433 * collect2-aix.h: Define F_LOADONLY.
15435 2013-02-14 Richard Biener <rguenther@suse.de>
15438 * varasm.c (output_constant_def_1): Get the decl representing
15439 the constant as argument.
15440 (output_constant_def): Wrap output_constant_def_1.
15441 (make_decl_rtl): Use output_constant_def_1 with the decl
15442 representing the constant.
15443 (build_constant_desc): Optionally re-use a decl already
15444 representing the constant.
15445 (tree_output_constant_def): Adjust.
15447 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15450 * asan.c (instrument_builtin_call): Really put the length of the
15451 second source argument into src1_len.
15453 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15455 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15456 argument. If it is false, don't create edge from then_bb to
15458 (insert_if_then_before_iter): Pass true to it.
15459 (build_check_stmt): Pass false to it.
15460 (transform_statements): Flush hash table only on extended basic
15461 block boundaries, rather than at the beginning of every bb.
15462 Don't flush hash table on nonfreeing_call_p calls.
15463 * tree-flow.h (nonfreeing_call_p): New prototype.
15464 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15466 2013-02-13 David S. Miller <davem@davemloft.net>
15468 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15470 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15473 * ira.c (max_regno_before_ira): Move from ...
15475 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15478 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15480 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15482 2013-02-13 Richard Biener <rguenther@suse.de>
15485 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15486 globals in MEM_REFs.
15488 2013-02-13 Richard Biener <rguenther@suse.de>
15490 * loop-init.c (loop_optimizer_init): Clear loop state when
15491 re-initializing preserved loops.
15492 * loop-unswitch.c (unswitch_single_loop): Return whether
15493 we unswitched the loop. Do not verify loop state here.
15494 (unswitch_loops): When we unswitched a loop discover new loops.
15496 2013-02-13 Kostya Serebryany <kcc@google.com>
15498 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15500 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15502 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15504 Avoid instrumenting duplicated memory access in the same basic block
15505 * Makefile.in (asan.o): Add new dependency on hash-table.h
15506 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15507 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15508 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15509 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15510 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15511 (get_mem_ref_of_assignment): New functions.
15512 (get_mem_refs_of_builtin_call): Extract from
15513 instrument_builtin_call and tweak a little bit to make it fit with
15515 (instrument_builtin_call): Use the new
15516 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15517 of is_gimple_builtin_call.
15518 (instrument_derefs, instrument_mem_region_access): Insert the
15519 instrumented memory reference into the hash table.
15520 (maybe_instrument_assignment): Renamed instrument_assignment into
15521 this, and change it to advance the iterator when instrumentation
15522 actually happened and return true in that case. This makes it
15523 homogeneous with maybe_instrument_assignment, and thus give a
15524 chance to callers to be more 'regular'.
15525 (transform_statements): Clear the memory reference hash table
15526 whenever we enter a new BB, when we cross a function call, or when
15527 we are done transforming statements. Use
15528 maybe_instrument_assignment instead of instrumentation. No more
15529 need to special case maybe_instrument_assignment and advance the
15530 iterator after calling it; it's now handled just like
15531 maybe_instrument_call. Update comment.
15533 2013-02-13 Richard Biener <rguenther@suse.de>
15535 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15536 Fix loop discovery code.
15538 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15540 PR inline-asm/56148
15541 * lra-constraints.c (process_alt_operands): Match early clobber
15542 operand with itself. Check conflicts with earlyclobber only if
15543 the operand is not reloaded. Prefer to reload conflicting operand
15544 if earlyclobber and matching operands are the same.
15546 2013-02-12 Richard Biener <rguenther@suse.de>
15549 * lto-streamer-out.c (write_symbol): Do not output symbols
15550 for hard register variables.
15552 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
15555 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15556 (umulsidi3_insn, mulsidi3_insn): New insns.
15558 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
15560 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15561 (struct tune_params): Add vec_costs field.
15562 * config/arm/arm.c (arm_builtin_vectorization_cost)
15563 (arm_add_stmt_cost): New functions.
15564 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15565 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15566 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15567 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15568 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15569 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15570 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15572 2013-02-12 Richard Biener <rguenther@suse.de>
15575 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15576 decls again if possible.
15578 2013-02-12 Richard Biener <rguenther@suse.de>
15580 PR middle-end/56288
15581 * tree-ssa.c (verify_ssa_name): Fix check, move
15582 SSA_NAME_IN_FREE_LIST check up.
15584 2013-02-12 Jakub Jelinek <jakub@redhat.com>
15585 Steven Bosscher <steven@gcc.gnu.org>
15587 PR rtl-optimization/56151
15588 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15589 equal to op0 or op1, and last_insn pattern is CODE operation
15590 with MEM dest and one of the operands matches that MEM.
15592 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15594 * doc/extend.texi: Document Function Multiversioning and "default"
15595 parameter string to target attribute.
15596 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15597 target attribute parameter is "default".
15598 (ix86_compare_version_priority): Remove checks for target attribute.
15599 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15600 Remove check for target attribute equal to NULL. Add assert.
15601 (ix86_generate_version_dispatcher_body): Change error to sorry.
15603 2013-02-11 Iain Sandoe <iain@codesourcery.com>
15604 Jack Howarth <howarth@bromo.med.uc.edu>
15605 Patrick Marlier <patrick.marlier@gmail.com>
15608 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15609 define ENDFILE_SPEC as TM_DESTRUCTOR.
15610 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15612 2013-02-11 Alexander Potapenko <glider@google.com>
15613 Jack Howarth <howarth@bromo.med.uc.edu>
15614 Jakub Jelinek <jakub@redhat.com>
15617 * config/darwin.c (cdtor_record): Rename ctor_record.
15618 (sort_cdtor_records): Rename sort_ctor_records.
15619 (finalize_dtors): New routine to sort destructors by
15620 priority before use in assemble_integer.
15621 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15623 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
15625 PR rtl-optimization/56275
15626 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15627 offset is non-negative and less than cmode size before
15628 calling simplify_subreg.
15630 2013-02-11 Richard Biener <rguenther@suse.de>
15632 PR tree-optimization/56264
15633 * cfgloop.h (fix_loop_structure): Adjust prototype.
15634 * loop-init.c (fix_loop_structure): Return the number of
15635 newly discovered loops.
15636 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15637 are discovered, do a full loop-closed SSA rewrite.
15639 2013-02-11 Richard Biener <rguenther@suse.de>
15641 PR tree-optimization/56273
15642 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15644 (check_array_ref): Fix missing newline in dumps.
15645 (search_for_addr_array): Likewise.
15647 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
15649 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15651 2013-02-09 Jakub Jelinek <jakub@redhat.com>
15654 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15656 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15658 PR rtl-optimization/56246
15659 * lra-constraints.c (simplify_operand_subreg): Try to reuse
15661 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15662 constraints are satisfied.
15664 2013-02-08 Jeff Law <law@redhat.com>
15667 * emit-rtl.c (reg_is_parm_p): New function.
15668 * regs.h (reg_is_parm_p): New prototype.
15669 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15670 callee-clobbered registers.
15672 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15675 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15676 If there is no implicit builtin declaration, just return NULL.
15678 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15680 * config/i386/sse.md (FMAMODEM): New mode iterator.
15681 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15682 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15684 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15686 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15687 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15688 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15690 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15692 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15693 (microblaze*-*-elf): Likewise.
15694 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15696 * config/microblaze/microblaze-c.c: Add builtin defines for
15697 _LITTLE_ENDIAN and _BIG_ENDIAN.
15698 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15699 add to TARGET_DEFAULT flags.
15700 Expand ASM_SPEC and LINK_SPEC.
15701 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15702 * config/microblaze/microblaze.md: Update extendsidi2 and
15703 movdi_internal instructions to use low-order / high-order reg
15705 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15706 options and inversemask / mask of LITTLE_ENDIAN.
15707 * config/microblaze/t-microblaze: Expand multilib options to
15708 include mlittle-endian (le) and update exceptions patterns.
15710 2013-02-08 Jakub Jelinek <jakub@redhat.com>
15712 PR rtl-optimization/56195
15713 * lra-constraints.c (get_reload_reg): Don't reuse regs
15714 if they have smaller mode than requested, if they have
15715 wider mode than requested, try to return a SUBREG.
15717 PR tree-optimization/56250
15718 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15719 if type is unsigned and code isn't MULT_EXPR.
15721 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15723 PR tree-optimization/56064
15724 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15725 bits according to mode.
15726 * fixed-value.h (fixed_from_double_int)
15727 (const_fixed_from_double_int): Adjust comments.
15729 2013-02-08 Richard Biener <rguenther@suse.de>
15732 * lto-streamer.h (struct data_in): Remove current_file, current_line
15733 and current_col members.
15734 * lto-streamer-out.c (lto_output_location): Stream changed bits
15735 en-block for efficiency.
15736 * lto-streamer-in.c (clear_line_info): Remove.
15737 (lto_input_location): Cache current file, line and column
15738 globally via local statics. Read changed bits en-block.
15739 (input_function): Do not call clear_line_info.
15740 (lto_read_body): Likewise.
15741 (lto_input_toplevel_asms): Likewise.
15743 2013-02-08 Michael Matz <matz@suse.de>
15745 PR tree-optimization/52448
15746 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15747 (nt_call_phase): New static.
15748 (add_or_mark_expr): Only mark accesses with newer phase than any
15750 (nonfreeing_call_p): New.
15751 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15752 (nt_fini_block): Keep blocks marked as visited.
15753 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15755 2013-02-08 Richard Biener <rguenther@suse.de>
15757 * ira.c (ira): Free broken dominator information.
15759 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15761 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15763 2013-02-08 Marek Polacek <polacek@redhat.com>
15765 * cfgloop.c (verify_loop_structure): Add more checking of headers.
15767 2013-02-08 Richard Biener <rguenther@suse.de>
15769 PR middle-end/56181
15770 * cfgloop.h (flow_loops_find): Adjust.
15771 (bb_loop_header_p): Declare.
15772 * cfgloop.c (bb_loop_header_p): New function split out from ...
15773 (flow_loops_find): ... here. Adjust function signature,
15774 support incremental loop structure update.
15775 (verify_loop_structure): Cleanup. Verify a loop is a loop.
15776 * cfgloopmanip.c (fix_loop_structure): Move ...
15777 * loop-init.c (fix_loop_structure): ... here.
15778 (apply_loop_flags): Split out from ...
15779 (loop_optimizer_init): ... here.
15780 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
15781 in incremental mode, only remove dead loops here.
15783 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15786 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
15787 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
15788 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
15789 (*round<mode>3.libgcc): New insns for fixed-modes.
15790 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
15791 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
15792 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
15793 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
15794 implementations. Define to __builtin_avr_absFX,
15795 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
15796 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
15797 __builtin_avr_countlsFX, respectively.
15798 * config/avr/avr-c.c (target.h): Include it.
15799 (enum avr_builtin_id): New enum.
15800 (avr_resolve_overloaded_builtin): New static function.
15801 (avr_register_target_pragmas): Use it to set
15802 targetm.resolve_overloaded_builtin.
15803 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
15804 tree nodes used by DEF_BUILTIN.
15805 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
15806 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
15807 <AVR_BUILTIN_xxBITS>: Same.
15809 2013-02-08 Richard Biener <rguenther@suse.de>
15811 * cfgloop.c (verify_loop_structure): Properly handle
15812 a loop exiting to another loop header.
15813 * ira-int.h (ira_loops): Remove.
15814 * ira.c (ira_loops): Remove.
15815 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
15816 (do_reload): Use loop_optimizer_finalize.
15817 * ira-build.c (create_loop_tree_nodes): Use get_loops and
15818 number_of_loops to access the loop tree.
15819 (more_one_region_p): Likewise.
15820 (finish_loop_tree_nodes): Likewise.
15821 (rebuild_regno_allocno_maps): Likewise.
15822 (mark_loops_for_removal): Likewise.
15823 (mark_all_loops_for_removal): Likewise.
15824 (remove_unnecessary_regions): Likewise.
15825 (ira_build): Likewise.
15826 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15828 2013-02-08 Richard Biener <rguenther@suse.de>
15830 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
15831 * ipa-pure-const.c (analyze_function): Avoid calling
15832 mark_irreducible_loops twice.
15833 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
15835 2013-02-07 David S. Miller <davem@davemloft.net>
15837 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
15839 * var-tracking.c (vt_add_function_parameter): Test the presence of
15840 HAVE_window_save properly and do not remap argument registers when
15841 we have a leaf function.
15843 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
15846 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
15848 * config/i386/i386.c (ix86_print_operand): Ditto.
15850 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15852 * lra-constraints.c (process_alt_operands): Fix recently added comment.
15854 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
15856 PR rtl-optimization/56225
15857 * lra-constraints.c (process_alt_operands): Check that reload hard
15858 reg can hold value for strict_low_part.
15860 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15863 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
15864 dwarf2out_end_function.
15865 (in_first_function_p, maybe_at_text_label_p,
15866 first_loclabel_num_not_at_text_label): New variables.
15867 (dwarf2out_var_location): In the first function find out
15868 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
15869 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
15872 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
15874 PR rtl-optimization/56178
15875 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
15876 SUBREG of a register. Tidy up related block of code.
15877 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
15878 note if the source is a register or a SUBREG of a register.
15880 2013-02-07 Jakub Jelinek <jakub@redhat.com>
15883 * config/rs6000/rs6000.md (ptrm): New mode attr.
15884 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
15885 call_value_indirect_aix<pttrsize>,
15886 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
15889 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
15891 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
15892 if -bnortl. Convert to strcmp and strncmp.
15894 2013-02-07 Alan Modra <amodra@gmail.com>
15897 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
15898 addresses won't wrap when offsetting.
15899 (rs6000_secondary_reload): Provide secondary reloads needed for
15900 wrapping LO_SUM addresses.
15902 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
15904 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
15905 MACH, just __MACH__.
15907 2013-02-06 Richard Biener <rguenther@suse.de>
15909 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
15910 instead of calling fix_loop_structure.
15912 2013-02-06 Jakub Jelinek <jakub@redhat.com>
15914 PR middle-end/56217
15915 * omp-low.c (use_pointer_for_field): Return false if
15916 lower_send_shared_vars doesn't generate any copy-out code.
15918 2013-02-06 Tom de Vries <tom@codesourcery.com>
15920 PR rtl-optimization/56131
15921 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
15922 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
15923 of the label is NULL. Add comment.
15925 2013-02-05 Jakub Jelinek <jakub@redhat.com>
15927 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
15930 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
15931 (STATIC_LIBTSAN_LIBS): Likewise.
15932 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
15933 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
15934 is defined, don't add anything else beyond that.
15935 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
15936 (LINK_COMMAND_SPEC): Use them.
15938 PR tree-optimization/56205
15939 * tree-stdarg.c (check_all_va_list_escapes): Return true if
15940 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
15941 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
15943 2013-02-05 Richard Biener <rguenther@suse.de>
15945 PR tree-optimization/53342
15946 PR tree-optimization/53185
15947 * tree-vectorizer.h (vect_check_strided_load): Remove.
15948 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
15949 not disallow peeling for vectorized strided loads.
15950 (vect_check_strided_load): Make static and simplify.
15951 (vect_analyze_data_refs): Adjust.
15952 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
15953 correctly when vectorizing strided loads.
15955 2013-02-05 Richard Biener <rguenther@suse.de>
15957 * doc/install.texi: Refer to ISL, not PPL.
15959 2013-02-05 Jan Hubicka <jh@suse.cz>
15961 PR tree-optimization/55789
15962 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
15964 2013-02-05 Jan Hubicka <jh@suse.cz>
15966 PR tree-optimization/55789
15967 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
15968 the dead call anyway.
15970 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
15973 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
15975 2013-02-04 Alexander Potapenko <glider@google.com>
15976 Jack Howarth <howarth@bromo.med.uc.edu>
15977 Jakub Jelinek <jakub@redhat.com>
15980 * config/darwin.c (sort_ctor_records): Stabilized qsort
15981 on constructor priority by using original position.
15982 (finalize_ctors): New routine to sort constructors by
15983 priority before use in assemble_integer.
15984 (machopic_asm_out_constructor): Use finalize_ctors if needed.
15986 2013-02-04 Jakub Jelinek <jakub@redhat.com>
15989 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
15990 about visibility on artificial decls.
15991 * config/sol2.c (solaris_assemble_visibility): Likewise.
15993 2013-02-04 Kai Tietz <ktietz@redhat.com>
15996 * config/i386/i386.c (function_value_ms_64): Add additional valtype
15997 argument and improve checking of return-argument types for 16-byte
15999 (ix86_function_value_1): Add additional valtype argument on call
16000 of function_value_64.
16001 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16002 handling infunction_value_64 function.
16004 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16006 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16008 2013-02-04 Richard Biener <rguenther@suse.de>
16010 PR tree-optimization/56188
16011 * tree-ssa-structalias.c (label_visit): Consider case with
16012 initially non-empty points-to set.
16013 (perform_var_substitution): Dump node mapping and clean up.
16015 2013-02-04 Richard Guenther <rguenther@suse.de>
16018 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16019 node prevail as last resort.
16020 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16021 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16023 2013-02-04 Richard Biener <rguenther@suse.de>
16025 PR tree-optimization/56113
16026 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16028 (equiv_class_lookup_or_add): ... this.
16029 (label_visit): Adjust and fix error in previous patch.
16030 (perform_var_substitution): Adjust.
16032 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16034 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16035 * config/sh/sh4-300.md: Likewise.
16036 * config/sh/sh4a.md: Likewise.
16037 * config/sh/constraints.md: Likewise.
16038 * config/sh/sh.md: Likewise.
16039 * config/sh/netbsd-elf.h: Likewise.
16040 * config/sh/predicates.md: Likewise.
16041 * config/sh/sh-protos.h: Likewise.
16042 * config/sh/ushmedia.h: Likewise.
16043 * config/sh/linux.h: Likewise.
16044 * config/sh/sh.c: Likewise.
16045 * config/sh/superh.h: Likewise.
16046 * config/sh/elf.h: Likewise.
16047 * config/sh/sh4.md: Likewise.
16048 * config/sh/sh.h: Likewise.
16050 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16052 * config/pa/constraints.md: Adjust unused letters. Change "T"
16053 constraint to match_test floating_point_store_memory_operand().
16054 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16055 (base14_operand): New.
16056 (floating_point_store_memory_operand): New.
16057 (integer_store_memory_operand): Revise to use base14_operand and
16058 reg_plus_base_memory_operand.
16059 (move_dest_operand): Allow symbolic_memory_operands.
16060 (symbolic_memory_operand): Check for LO_SOM.
16061 (symbolic_operand): Change default case to break.
16062 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16063 CONST_DOUBLE values to be reloaded by putting them into memory when
16064 the destination is a floating point register.
16065 (movdf): Remove code to handle CONST_DOUBLE.
16067 (reload_indf_r1): New.
16068 (reload_insf_r1): New.
16069 Consistently use "Q" and "T" constraints with integer and floating
16070 point move instructions, respectively.
16071 (movdi): Remove FAIL.
16072 Change predicate for source operand unamed DImode move from
16073 general_operand to move_src_operand.
16074 (umulsidi3): Change predicate for destination operand to
16076 Likewise for similar unamed patterns.
16077 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16078 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16079 (hppa_legitimize_address): Simplify mask calculation.
16080 (pa_emit_move_sequence): Revised handling of secondary reloads from
16081 REG+D addresses for floating point loads and stores. Directly handle
16082 loading CONST0_RTX (mode) to a floating point register.
16083 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16084 to floating point registers. Don't restrict secondary reloads to
16085 floating point registers to integer modes. Revise some comments and
16087 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16088 (pa_legitimate_address_p): New.
16089 (pa_legitimize_reload_address): New.
16090 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16091 (STRICT_REG_OK_FOR_BASE_P): New.
16092 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16093 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16095 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16096 Andrew Dixie <andrewd@gentrack.com>
16098 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16101 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16103 * expmed.c (extract_bit_field_1): Pass the full width of the
16104 structure to get_best_reg_extraction_insn.
16106 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16109 * configure.ac (use_cxa_atexit): Add AIX.
16110 * configure: Regenerate.
16112 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16114 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16117 * final.c (need_profile_function): New variable.
16118 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16119 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16120 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16121 notes, targetm.asm_out.function_prologue doesn't emit anything,
16122 HAVE_prologue and profiler should be emitted before prologue,
16123 set need_profile_function instead of emitting it.
16124 (final_scan_insn): If need_profile_function, emit
16125 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16126 NOTE_INSN_FUNCTION_BEG note.
16128 2013-02-01 Richard Henderson <rth@redhat.com>
16130 * config/rs6000/rs6000.md (smulditi3): New.
16133 * config/alpha/alpha.md (umulditi3): New.
16135 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16137 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16138 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16140 2013-02-01 Richard Biener <rguenther@suse.de>
16142 PR tree-optimization/56113
16143 * tree-ssa-structalias.c (label_visit): Reduce work for
16144 single-predecessor nodes.
16146 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
16148 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
16149 range isn't testing for zero.
16151 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
16153 PR middle-end/56113
16154 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
16156 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
16157 Nick Clifton <nickc@redhat.com>
16159 * config/v850/constraints.md (Q): Define as a memory constraint.
16160 * config/v850/predicates.md (label_ref_operand): New predicate.
16161 (e3v5_shift_operand): New predicate.
16162 (ior_operator): New predicate.
16163 * config/v850/t-v850: Add e3v5 multilib.
16164 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
16165 (v850_gen_movdi): Prototype.
16166 * config/v850/v850.c: Add support for e3v5 architecture.
16167 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
16169 (construct_save_jarl): Add e3v5 long JARL support.
16170 (v850_adjust_insn_length): New function. Adjust length of call
16171 insns when using e3v5 instructions.
16172 (v850_gen_movdi): New function: Generate instructions to move a
16174 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
16175 (CPP_SPEC): Define __v850e3v5__ as appropriate.
16176 (TARGET_USE_FPU): Enable for e3v5.
16177 (CONST_OK_FOR_W): New macro.
16178 (ADJUST_INSN_LENGTH): Define.
16179 * config/v850/v850.md (UNSPEC_LOOP): Define.
16180 (attr cpu): Add v850e3v5.
16181 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
16182 (movdi): New pattern.
16183 (movdi_internal): New pattern.
16184 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
16185 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
16186 (cstoresf4): Likewise.
16187 (cstoredf4): Likewise.
16188 (insv): New pattern.
16189 (rotlso3_a): New pattern.
16190 (rotlsi3_b): New pattern
16191 (rotlsi3_v850e3v5): New pattern.
16192 (doloop_begin): New pattern.
16193 (fix_loop_counter): New pattern.
16194 (doloop_end): New pattern.
16195 (branch_normal): Add e3v5 long branch support.
16196 (branch_invert): Likewise.
16197 (branch_z_normal): Likewise.
16198 (branch_z_invert): Likewise.
16199 (branch_nz_normal): Likewise.
16200 (branch_nz_invert): Likewise.
16201 (call_internal_short): Add e3v5 register-indirect JARL support.
16202 (call_internal_long): Likewise.
16203 (call_value_internal_short): Likewise.
16204 (call_value_internal_long): Likewise.
16205 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
16206 (mloop): New option.
16207 * config.gcc: Add support for configuring v840e3v5 target.
16208 * doc/invoke.texi: Document new v850 specific command line options.
16210 2013-01-31 Paul Koning <ni1d@arrl.net>
16214 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
16215 children if parent is a class.
16216 (prune_unused_types_prune): Don't add DW_AT_declaration.
16218 2013-01-31 Richard Biener <rguenther@suse.de>
16220 PR tree-optimization/56157
16221 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
16222 match up operand with SLP child.
16224 2013-01-31 Jason Merrill <jason@redhat.com>
16227 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
16228 parameters the first time.
16229 (gen_scheduled_generic_parms_dies): Check completeness here.
16231 2013-01-31 Richard Biener <rguenther@suse.de>
16233 PR middle-end/53073
16234 * common.opt (faggressive-loop-optimizations): New flag,
16235 enabled by default.
16236 * doc/invoke.texi (faggressive-loop-optimizations): Document.
16237 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
16238 infer_loop_bounds_from_undefined by it.
16240 2013-01-31 Richard Biener <rguenther@suse.de>
16242 PR tree-optimization/56150
16243 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
16244 visit virtual operands.
16245 (find_uses_to_rename_bb): Likewise.
16247 2013-01-31 Richard Biener <rguenther@suse.de>
16249 PR tree-optimization/56150
16250 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
16251 mixed store non-store stmts.
16253 2013-01-30 Jakub Jelinek <jakub@redhat.com>
16256 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
16257 LIBASAN_EARLY_SPEC is defined.
16258 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
16259 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
16261 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
16264 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
16265 invalid args instead of ICEing on it.
16266 (ix86_valid_target_attribute_tree): Return error_mark_node if
16267 ix86_valid_target_attribute_inner_p failed.
16268 (ix86_valid_target_attribute_p): Return false only if
16269 ix86_valid_target_attribute_tree returned error_mark_node. Allow
16270 target("default") attribute.
16271 (sorted_attr_string): Change argument from const char * to tree,
16272 merge in all target attribute arguments rather than just one.
16273 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
16274 instead of free. Avoid using strcat.
16275 (ix86_mangle_function_version_assembler_name): Mangle
16276 target("default") as if no target attribute is present. Adjust
16277 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
16278 instead of xmalloc and XDELETEVEC instead of free.
16279 (ix86_function_versions): Don't return true if one of the decls
16280 doesn't have target attribute. If they don't and one of the decls
16281 is DECL_FUNCTION_VERSIONED, report an error. Adjust
16282 sorted_attr_string caller. Use XDELETEVEC instead of free.
16283 (ix86_supports_function_versions): Remove.
16284 (make_name): Fix up formatting.
16285 (make_dispatcher_decl): Remove resolver_name and its initialization.
16286 Avoid leaking memory.
16287 (is_function_default_version): Return true if there is
16288 target("default") attribute rather than no target attribute at all.
16289 (make_resolver_func): Avoid leaking memory.
16290 (ix86_generate_version_dispatcher_body): Likewise.
16291 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
16292 * target.def (supports_function_versions): Remove.
16293 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
16294 * doc/tm.texi: Regenerated.
16296 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
16298 PR rtl-optimization/56144
16299 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
16300 for values with side effects.
16302 2013-01-30 Richard Biener <rguenther@suse.de>
16304 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
16305 (sparseset_pop): Likewise.
16306 * cfganal.c (compute_idf): Likewise. Increase work-stack size
16307 to be able to use quick_push in the worker loop.
16309 2013-01-30 Marek Polacek <polacek@redhat.com>
16311 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
16313 2013-01-30 Richard Biener <rguenther@suse.de>
16316 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
16318 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
16320 PR tree-optimization/56064
16321 * fixed-value.c (fixed_from_double_int): New function.
16322 * fixed-value.h (fixed_from_double_int): New prototype.
16323 (const_fixed_from_double_int): New static inline function.
16324 * fold-const.c (native_interpret_fixed): New static function.
16325 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
16326 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
16327 (native_encode_fixed): New static function.
16328 (native_encode_expr) <FIXED_CST>: Use it.
16329 (native_interpret_int): Move double_int worker code to...
16330 * double-int.c (double_int::from_buffer): ...this new static method.
16331 * double-int.h (double_int::from_buffer): Prototype it.
16333 2013-01-30 Richard Biener <rguenther@suse.de>
16335 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
16336 New pointer-map and obstack.
16337 (init_alias_vars): Allocate pointer-map and obstack.
16338 (delete_points_to_sets): Free them.
16339 (find_what_var_points_to): Cache result.
16340 (find_what_p_points_to): Adjust for changed interface of
16341 find_what_var_points_to.
16342 (compute_points_to_sets): Likewise.
16343 (ipa_pta_execute): Likewise.
16345 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16347 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
16348 * configure: Regenerate.
16349 * config.in: Regenerate.
16350 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
16351 #nobits/#progbits if supported.
16353 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
16356 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
16357 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
16358 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
16360 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16362 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
16363 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
16365 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16367 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
16368 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
16370 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16372 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
16374 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
16375 * config/arm/cortex-a7.md: New bypasses using
16376 arm_mac_accumulator_is_result.
16378 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16380 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
16381 (cortex_a7_neon_mla): Likewise.
16382 (cortex_a7_fpfmad): New reservation.
16383 (cortex_a7_fpmacs): Use ffmas and update required units.
16384 (cortex_a7_fpmuld): Update required units and latency.
16385 (cortex_a7_fpmacd): Likewise.
16386 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
16387 (cortex_a7_neon). Likewise.
16388 (bypass) Update participating units.
16390 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16392 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
16393 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
16395 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16396 (vfp_fmul): Use ffmad.
16397 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16398 (cortex_r4_fmacd): Use ffmad.
16399 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16400 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16401 (cortex_a9_fmacd): Use ffmad.
16402 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16403 (cortex_a8_vfp_macd): Use ffmad.
16404 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16405 (cortex_a5_fpmacd): Use ffmad.
16406 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16407 (cortex_a15_vfp_macd): Use ffmad.
16408 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16410 2013-01-29 Jason Merrill <jason@redhat.com>
16413 * varasm.c (default_assemble_visibility): Don't warn about
16414 visibility on artificial decls.
16416 2013-01-29 Richard Biener <rguenther@suse.de>
16418 PR tree-optimization/56113
16419 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16421 (label_visit): Free duplicate bitmaps and record the leader instead.
16422 (perform_var_substitution): Adjust.
16424 2013-01-29 Richard Biener <rguenther@suse.de>
16426 PR tree-optimization/55270
16427 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16428 the CFG, schedule loops for fixup.
16430 2013-01-29 Nick Clifton <nickc@redhat.com>
16432 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16435 2013-01-28 Leif Ekblad <leif@rdos.net>
16437 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16438 * config/i386/i386.h (TARGET_RDOS): New macro.
16439 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16440 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16441 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16442 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16443 DEFAULT_LARGE_SECTION_THRESHOLD.
16444 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16445 * config/i386/rdos.h: New file.
16446 * config/i386/rdos64.h: New file.
16448 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16451 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16454 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16456 PR rtl-optimization/56117
16457 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16458 call cselib_lookup_from_insn on the MEM before calling
16459 add_insn_mem_dependence.
16461 2013-01-28 Richard Biener <rguenther@suse.de>
16463 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16464 to a stmt that didn't have one.
16465 (copy_phis_for_bb): Likewise for PHI arguments.
16466 (copy_debug_stmt): Likewise for debug stmts.
16468 2013-01-28 Richard Biener <rguenther@suse.de>
16470 PR tree-optimization/56034
16471 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16472 (partition_builtin_p): Adjust.
16473 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16474 it is the last partition.
16475 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16476 up the vertex for the definition.
16477 (classify_partition): Classify whether a partition is a
16478 PKIND_REDUCTION, thus has uses outside of the loop.
16479 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16480 Merge all PKIND_REDUCTION partitions into the last partition.
16481 (tree_loop_distribution): Seed partitions from reductions as well.
16483 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16485 PR tree-optimization/56125
16486 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16487 pow(x,c) into sqrt(x) * powi(x, n/2) or
16488 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16489 optimizing for size.
16490 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16491 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16494 PR tree-optimization/56094
16495 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16496 to UNKNOWN_LOCATION while gimplifying expr.
16498 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16501 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16502 operand 0 in movabs insn template for -masm=intel asm alternative.
16503 (*movabs<mode>_2): Ditto for operand 1.
16505 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16508 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16511 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16513 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16516 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16518 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16519 Undef to avoid warning.
16521 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16523 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16524 * configure: Regenerate.
16526 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16528 PR tree-optimization/56098
16529 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16530 for stmts with volatile ops.
16531 (cond_store_replacement): Don't optimize if assign has volatile ops.
16532 (cond_if_else_store_replacement_1): Don't optimize if either
16533 then_assign or else_assign have volatile ops.
16534 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16537 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16539 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16541 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16543 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16544 missing ':' in asm example.
16546 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16548 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16549 entries into lane and laneq entries.
16550 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16551 Remove AdvSIMD scalar modes.
16552 (aarch64_sq<r>dmulh_laneq<mode>): New.
16553 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16555 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16556 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16557 * config/aarch64/iterators.md (VCOND): New.
16560 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16563 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16564 Add NULL LIBNAME argument to existing definitions.
16565 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16566 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16567 * config/avr/avr.c (DEF_BUILTIN): Same.
16568 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16569 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16570 implementation is available (DECL_ASSEMBLER_NAME is set).
16571 (avr_fold_absfx): New static function.
16572 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16573 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16574 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16575 AVR_BUILTIN_ABSLLK.
16576 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16577 (abshk, absk, abslk, absllk): Provide as static inline functions.
16579 2013-01-25 Marek Polacek <polacek@redhat.com>
16581 PR tree-optimization/56035
16582 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16584 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
16586 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16587 (*movtf_internal_rex64): Add (!o,C) alternative
16588 (*movxf_internal_rex64): Ditto.
16589 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16591 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
16593 * doc/invoke.texi: fix typo.
16594 * doc/objc.texi: fix typo.
16596 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16598 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16599 for the first two alternatives.
16601 2013-01-24 Diego Novillo <dnovillo@google.com>
16603 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
16604 (ggc-zone.o): Remove.
16605 * configure.ac: Remove option --with-gc.
16606 * configure: Re-generate.
16607 * doc/install.texi: Remove documentation for --with-gc.
16608 * gengtype.c (write_enum_defn): Remove. Update all users.
16609 (write_Types_process_field): Remove generation of gt_e_* argument.
16610 (output_type_enum): Remove. Update all users.
16611 (write_enum_defn): Remove. Update all users.
16612 (enum alloc_zone): Remove. Update all users.
16613 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
16614 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16615 Update all callers.
16616 (struct ptr_data): Remove field TYPE. Update all users.
16617 (gt_pch_note_object): Remove argument TYPE. Update all users.
16618 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16620 * ggc-none.c (ggc_alloc_typed_stat): Remove.
16621 (struct alloc_zone): Remove.
16622 (ggc_internal_alloc_zone_stat): Remove.
16623 (ggc_internal_cleared_alloc_zone_stat): Remove.
16624 * ggc-page.c (ggc_alloc_typed_stat): Remove.
16625 (ggc_pch_count_object): Remove last argument. Update all users.
16626 (ggc_pch_alloc_object): Remove last argument. Update all users.
16627 (struct alloc_zone): Remove.
16628 * ggc-zone.c: Remove.
16629 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
16630 (struct alloc_zone): Remove.
16631 (ggc_alloc_typed_stat): Remove.
16632 (ggc_alloc_typed): Remove.
16633 (ggc_splay_alloc): Remove first argument.
16634 (rtl_zone): Remove. Update all users.
16635 (tree_zone): Remove. Update all users.
16636 (tree_id_zone): Remove. Update all users.
16637 (ggc_internal_zone_alloc_stat): Remove. Update all users.
16638 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
16639 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
16640 * tree-ssanames.c: Remove references to zone allocator in comments.
16642 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
16644 * config/avr/avr.c (avr_out_fract): Make register numbers that
16645 might be outside of source operand signed.
16647 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
16649 * config/i386/constraints.md (Yf): New constraint.
16650 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16651 of f constraint to conditionaly disable x87 register preferences.
16652 (*movdf_internal): Ditto.
16653 (*movsf_internal): Ditto.
16655 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16657 PR inline-asm/55934
16658 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16659 that have operands with impossible constraints.
16660 Add a FIXME for a speed-up opportunity.
16661 * lra-constraints.c (process_alt_operands): Verify that a class
16662 selected from constraints on asms is valid for the operand mode.
16663 (curr_insn_transform): Remove incorrect comment.
16665 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
16667 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16668 TOC operand is a valid symbol ref in the constant pool.
16670 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16672 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
16674 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
16677 * config/avr/stdfix.h: New file.
16678 * t-avr (stdfix-gcc.h): New rule to build it.
16679 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16681 2013-01-23 Kostya Serebryany <kcc@google.com>
16683 * config/darwin.h: remove dependency on
16684 CoreFoundation (asan on Mac OS).
16686 2013-01-23 Jakub Jelinek <jakub@redhat.com>
16689 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16690 instead of cmpdi_operand for first comparison operand.
16691 Don't assert that comparison operands aren't both constants.
16693 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16695 * doc/install.texi (Downloading the Source): Update references to
16696 downloading separate components.
16698 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16700 * doc/extend.texi (__int128): Improve grammar.
16702 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
16705 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16706 alternative to (o,r).
16707 (*movdi_internal_rex64): Remove (!o,n) alternative.
16708 (DImode immediate->memory splitter): Remove.
16709 (DImode immediate->memory peephole2): Remove.
16710 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16711 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16712 alternative to (!o,*r).
16713 (*movtf_internal_sse): New pattern.
16714 (*movxf_internal_rex64): New pattern.
16715 (*movxf_internal): Disable for TARGET_64BIT.
16716 (*movdf_internal_rex64): Remove (!o,F) alternative.
16718 2013-01-22 Jakub Jelinek <jakub@redhat.com>
16720 PR middle-end/56074
16721 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16722 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16723 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16724 stmt locations where LOCATION_LOCUS of the stmt location is
16725 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16728 * config/i386/i386.md (UNSPEC_STOS): New.
16729 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16730 *strsetqi_1): Add UNSPEC_STOS.
16732 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16735 * doc/invoke.texi: Remove left over -Wsynth example.
16737 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16739 PR tree-optimization/56051
16740 * fold-const.c (fold_binary_loc): Don't fold
16741 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16742 a narrowing conversion, or widening conversion from signed
16745 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16747 PR rtl-optimization/56023
16748 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16749 dependent on debug instruction.
16751 2013-01-21 Martin Jambor <mjambor@suse.cz>
16753 PR middle-end/56022
16754 * function.c (allocate_struct_function): Call
16755 invoke_set_current_function_hook earlier.
16757 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16759 * reload1.c (init_reload): Only initialize reload_obstack
16760 during the first call.
16762 2013-01-21 Marek Polacek <polacek@redhat.com>
16764 * cfgloop.c (verify_loop_structure): Fix up grammar.
16766 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
16768 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
16769 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
16771 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16774 * config/arm/marvell-pj4.md: Update copyright year.
16775 Fix up use of alu to alu_reg and simple_alu_imm.
16777 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16779 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
16781 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
16784 * lra-constraints.c (curr_insn_transform): Don't reuse original
16785 insn for secondary memory move when memory mode should be different.
16787 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16789 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
16790 atomic_storedi_1): New patterns.
16792 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
16794 btver2 pipeline descriptions.
16795 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
16797 * config/i386/i386.md (btver2_decode): New type attributes.
16798 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
16800 * config/i386/btver2.md: New file describing btver2 pipelines.
16802 2013-01-19 Andrew Pinski <apinski@cavium.com>
16804 PR tree-optimization/52631
16805 * tree-ssa-sccvn (visit_use): Before looking up the original
16806 statement, try looking up the simplified expression.
16808 2013-01-19 Anthony Green <green@moxielogic.com>
16810 * config/moxie/moxie.c (moxie_expand_prologue): Set
16811 current_function_static_stack_size.
16813 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16815 PR tree-optimization/56029
16816 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
16817 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
16819 2013-01-18 Sharad Singhai <singhai@google.com>
16821 PR tree-optimization/55995
16822 * dumpfile.c (dump_loc): Print location only if available.
16823 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
16825 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
16828 * lra-constraints.c (curr_insn_transform): Reuse original insn for
16829 secondary memory move.
16830 (inherit_reload_reg): Use rclass instead of cl for
16831 check_secondary_memory_needed_p.
16833 2013-01-18 Jakub Jelinek <jakub@redhat.com>
16835 PR middle-end/56015
16836 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
16837 the case where writing real complex part of target modifies op1.
16839 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
16841 * config/aarch64/aarch64-simd.md
16842 (aarch64_vcond_internal<mode>): Handle unordered cases.
16843 * config/aarch64/iterators.md (v_cmp_result): New.
16845 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
16846 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16848 * config/arm/marvell-pj4.md: New file.
16849 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
16850 * config/arm/arm.md (generic_sched): Add marvell_pj4.
16851 (generic_vfp): Likewise.
16852 * config/arm/arm-cores.def: Add marvell-pj4.
16853 * config/arm/arm-tune.md: Regenerate.
16854 * config/arm/arm-tables.opt: Regenerate.
16855 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
16856 * doc/invoke.texi: Document marvell-pj4.
16858 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
16860 * config/aarch64/arm_neon.h: Map scalar types to standard types.
16862 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16867 * var-tracking.c (negative_power_of_two_p): New.
16868 (global_get_addr_cache, local_get_addr_cache): New.
16869 (get_addr_from_global_cache, get_addr_from_local_cache): New.
16870 (vt_canonicalize_addr): Rewrite using the above. Adjust the
16872 (vt_stack_offset_p): Remove.
16873 (vt_canon_true_dep): Always canonicalize loc's address.
16874 (clobber_overlapping_mems): Make sure we have a MEM.
16875 (local_get_addr_clear_given_value): New.
16876 (val_reset): Clear local cached entries.
16877 (compute_bb_dataflow): Create and release the local cache.
16878 Disable duplicate MEMs clobbering.
16879 (emit_notes_in_bb): Clobber MEMs likewise.
16880 (vt_emit_notes): Create and release the local cache.
16881 (vt_initialize, vt_finalize): Create and release the global
16882 cache, respectively.
16883 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
16885 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16887 PR libmudflap/53359
16888 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
16889 not found in the symtab.
16891 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
16894 PR rtl-optimization/55547
16895 PR rtl-optimization/53827
16898 * alias.c (offset_overlap_p): New, factored out of...
16899 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
16900 the conservative special case for symbolic constants. Don't
16901 adjust zero sizes on alignment.
16903 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
16905 PR rtl-optimization/52573
16906 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
16907 REG_UNUSED for the same register.
16909 2013-01-17 Richard Biener <rguenther@suse.de>
16910 Marek Polacek <polacek@redhat.com>
16912 PR rtl-optimization/55833
16913 * loop-unswitch.c (unswitch_loops): Move loop verification...
16914 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
16915 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
16916 Set it to true when we're removing a loop from hierarchy tree in
16917 an irreducible region.
16918 (fix_bb_placements): Adjust caller.
16919 (fix_loop_placements): Likewise.
16921 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
16923 * config/avr/builtins.def (DEF_BUILTIN): Factor out
16924 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
16925 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
16926 Remove ID. Adjust comments.
16927 * config/avr/avr-c.c (avr_builtin_name): Remove.
16928 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
16929 * config/avr/avr.c (avr_tolower): New static function.
16930 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
16931 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
16932 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
16935 2013-01-17 Jan Hubicka <jh@suse.cz>
16937 PR tree-optimization/55273
16938 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
16940 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
16943 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
16944 store through atomic_store<mode>_1.
16945 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
16947 2013-01-17 Martin Jambor <mjambor@suse.cz>
16949 PR tree-optimizations/55264
16950 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
16951 for virtual methods.
16952 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
16953 virtual methods before inlining is over.
16954 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
16956 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
16959 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
16961 PR rtl-optimization/56005
16962 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
16963 pending reads for prefetch.
16965 2013-01-16 Ian Bolton <ian.bolton@arm.com>
16967 * config/aarch64/aarch64.md
16968 (*cstoresi_neg_uxtw): New pattern.
16969 (*cmovsi_insn_uxtw): New pattern.
16970 (*<optab>si3_uxtw): New pattern.
16971 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
16972 (*<optab>si3_insn_uxtw): New pattern.
16973 (*bswapsi2_uxtw): New pattern.
16975 2013-01-16 Richard Biener <rguenther@suse.de>
16977 * tree-inline.c (tree_function_versioning): Remove set but
16978 never used variable.
16980 2013-01-16 Richard Biener <rguenther@suse.de>
16982 PR tree-optimization/55964
16983 * tree-flow.h (rename_variables_in_loop): Remove.
16984 (rename_variables_in_bb): Likewise.
16985 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
16986 (copy_loop_before): Adjust and delete update-ssa status.
16987 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
16988 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
16989 (rename_variables_in_loop): Remove.
16990 (slpeel_update_phis_for_duplicate_loop): Likewise.
16991 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
16992 use available cfg machinery instead of duplicating it.
16993 Update PHI nodes and perform poor-mans SSA update here.
16994 (slpeel_tree_peel_loop_to_edge): Adjust.
16996 2013-01-16 Richard Biener <rguenther@suse.de>
16998 PR tree-optimization/54767
16999 PR tree-optimization/53465
17000 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17001 (vrp_visit_phi_node): For PHI arguments coming via backedges
17002 drop all symbolical range information.
17003 (execute_vrp): Compute backedges.
17005 2013-01-16 Richard Biener <rguenther@suse.de>
17007 * doc/install.texi: Update CLooG and ISL requirements to
17010 2013-01-16 Christian Bruel <christian.bruel@st.com>
17013 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17014 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17015 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17017 2013-01-16 DJ Delorie <dj@redhat.com>
17019 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17020 (UNSPECV_SP_SWITCH_E): New.
17021 (sp_switch_1): Change to an unspec.
17022 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17025 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17027 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17028 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17029 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17030 (expand_mem_thread_fence): Ditto.
17031 (expand_mem_signal_fence): Ditto.
17032 (expand_atomic_load): Ditto.
17033 (expand_atomic_store): Ditto.
17035 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17037 PR rtl-optimization/55547
17038 PR rtl-optimization/53827
17041 * alias.c (memrefs_conflict_p): Set sizes to negative after
17044 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17047 * function.c (thread_prologue_and_epilogue_insns): Always
17048 add crtl->drap_reg to set_up_by_prologue.set, even if
17049 stack_realign_drap is false.
17051 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17053 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17054 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17055 *call): Fix indention.
17057 2013-01-15 Tom de Vries <tom@codesourcery.com>
17060 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17063 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17065 PR rtl-optimization/55153
17066 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17068 2013-01-15 Martin Jambor <mjambor@suse.cz>
17070 PR tree-optimization/55920
17071 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17072 accesses as grp_to_be_debug_replaced.
17074 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17076 PR tree-optimization/55920
17077 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17078 there is non-useless type conversion needed from debug rhs to lhs,
17079 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17081 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17082 Mikael Pettersson <mikpe@it.uu.se>
17085 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17087 (ASM_OUTPUT_CASE_LABEL): Remove.
17088 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17089 * final.c (shorten_branches): Update alignment of labels before
17090 jump tables if CASE_VECTOR_SHORTEN_MODE.
17092 2013-01-15 Richard Biener <rguenther@suse.de>
17095 * system.h: Do not include gmp.h for building host tools.
17097 2013-01-15 Richard Biener <rguenther@suse.de>
17099 PR middle-end/55882
17100 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17101 account for bitpos when computing alignment.
17103 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17105 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17106 (ix86_target_macros_internal): Likewise.
17108 * config/i386/i386.c (m_CORE2I7): Removed.
17109 (m_CORE_HASWELL): New macro.
17110 (m_CORE_ALL): Likewise.
17111 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17112 (initial_ix86_arch_features): Likewise.
17113 (processor_target_table): Initializations for Core avx2.
17114 (cpu_names): New names "core-avx2".
17115 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17116 PROCESSOR_CORE_HASWELL.
17117 (ix86_issue_rate): New case.
17118 (ia32_multipass_dfa_lookahead): Likewise.
17119 (ix86_sched_init_global): Likewise.
17121 * config/i386/i386.h (TARGET_HASWELL): New macro.
17122 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17123 (processor_type): New PROCESSOR_HASWELL.
17125 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17127 PR tree-optimization/55955
17128 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17129 *SHIFT_EXPR and *ROTATE_EXPR codes.
17131 PR tree-optimization/48766
17132 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17133 -ftrapv disable -fwrapv.
17135 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17138 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17139 etc. to 1 and not to __flash.
17140 Use LL suffix for __INT24_MAX__ with -mint8.
17141 Use ULL suffix for __UINT24_MAX__ with -mint8.
17143 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17145 * config/avr/avr-arch.h
17146 (struct base_arch_s): Use typedef avr_arch_t instead.
17147 (struct arch_info_s): Use typedef avr_arch_info_t instead.
17148 (struct mcu_type_s): Use typedef avr_mcu_t instead.
17149 * config/avr/avr.c: Same.
17150 * config/avr/avr-devices.c: Same.
17151 * config/avr/driver-avr.c: Same.
17152 * config/avr/gen-avr-mmcu-texi.c: Same.
17153 * config/avr/avr-mcus.def: Adjust comment.
17155 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
17157 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
17158 * config/aarch64/iterators.md (VALLDI): New.
17160 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
17161 Andi Kleen <ak@linux.intel.com>
17164 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
17165 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
17168 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17170 * config/avr/avr-stdint.h: Remove trailing blanks.
17171 * config/avr/avr-log.h: Same.
17172 * config/avr/avr-arch.h: Same.
17173 * config/avr/avr-devices.c: Same.
17174 * config/avr/avr-dimode.md: Same.
17175 * config/avr/predicates.md: Same.
17176 * config/avr/avr-c.c: Same. And fix typo.
17178 * config/avr/avr-protos.h: Same. And:
17179 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
17180 (init_cumulative_args): Rename to avr_init_cumulative_args.
17181 (expand_prologue): Rename to avr_expand_prologue.
17182 (expand_epilogue): Rename to avr_expand_epilogue.
17183 (adjust_insn_length): Rename to avr_adjust_insn_length.
17184 (notice_update_cc): Rename to avr_notice_update_cc.
17185 (final_prescan_insn): Rename to avr_final_prescan_insn.
17186 * config/avr/avr.c: Same.
17187 * config/avr/avr.h: Same.
17188 * config/avr/avr.md: Remove trailing blanks.
17189 (prologue): Use avr_expand_prologue.
17190 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
17192 2013-01-14 Richard Biener <rguenther@suse.de>
17194 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
17195 verify_location, collect_subblocks): New functions.
17196 (verify_gimple_in_cfg): Verify that locations only reference
17197 BLOCKs in the functions BLOCK tree.
17199 2013-01-14 Richard Biener <rguenther@suse.de>
17201 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
17203 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
17205 (insert_out_of_ssa_copy_on_edge): Likewise.
17206 (rewrite_close_phi_out_of_ssa): Likewise.
17207 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
17209 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
17210 propagated constants.
17211 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
17214 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17216 * config/avr/avr-modes.def: Add GPL copyright notice.
17218 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
17220 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
17221 MEMMODEL_MASK to determine memory model.
17222 (atomic_store<mode>): Ditto from operands[2].
17223 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
17225 2013-01-13 Jakub Jelinek <jakub@redhat.com>
17228 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
17229 (fold_gimple_assign): Don't call unshare_expr here.
17230 (fold_ctor_reference): Call unshare_expr.
17232 2013-01-13 Terry Guo <terry.guo@arm.com>
17234 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
17235 * doc/fragments.texi: Document MULTILIB_REUSE.
17236 * gcc.c (multilib_reuse): New internal spec.
17237 (set_multilib_dir): Also search multilib from multilib_reuse.
17238 * genmultilib (tmpmultilib3): Refactor code.
17239 (tmpmultilib4): Ditto.
17240 (multilib_reuse): New multilib argument.
17242 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
17244 * Makefile.in: Update copyright.
17246 2013-01-12 Tom de Vries <tom@codesourcery.com>
17248 PR middle-end/55890
17249 * calls.c (expand_call): Check if arg_nr is valid.
17251 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17253 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
17254 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
17255 documentation. Add missing '__' in front of
17256 __builtin_ia32_packssdw256.
17258 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17261 * config/s390/s390.c (s390_preferred_reload_class): Do not return
17262 NO_REGS for larl operands.
17263 (s390_reload_larl_operand): Use s390_load_address instead of
17266 2013-01-11 Richard Biener <rguenther@suse.de>
17268 * tree-cfg.c (verify_node_sharing_1): Split out from ...
17269 (verify_node_sharing): ... here.
17270 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
17272 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
17274 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
17275 Substitute TREECHECKING.
17276 * configure: Regenerate.
17277 * Makefile.in (TREECHECKING): New.
17279 2013-01-11 Richard Guenther <rguenther@suse.de>
17281 PR tree-optimization/44061
17282 * tree-vrp.c (extract_range_basic): Compute zero as
17283 value-range for __builtin_constant_p of function parameters.
17285 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
17287 Update copyright years.
17289 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
17291 PR rtl-optimization/55672
17292 * lra-eliminations.c (mark_not_eliminable): Permit addition with
17293 const to be eliminable.
17295 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
17297 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
17298 * configure: Regenerate.
17300 2013-01-10 Richard Biener <rguenther@suse.de>
17302 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
17304 2013-01-10 Richard Biener <rguenther@suse.de>
17307 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
17308 locations for virtual PHI arguments.
17309 (rewrite_update_phi_arguments): Likewise.
17311 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
17313 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
17316 2013-01-10 Jakub Jelinek <jakub@redhat.com>
17318 PR tree-optimization/55921
17319 * tree-complex.c (expand_complex_asm): New function.
17320 (expand_complex_operations_1): Call it for GIMPLE_ASM.
17322 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17325 * config/s390/s390.c (s390_symref_operand_p)
17326 (s390_loadrelative_operand_p): Merge the two functions.
17327 (s390_check_qrst_address, print_operand_address): Add parameters
17328 to s390_loadrelative_operand_p invokation.
17329 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
17330 (s390_reload_larl_operand, s390_secondary_reload): Use
17331 s390_loadrelative_operand_p instead of s390_symref_operand_p.
17332 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
17334 2013-01-09 Mike Stump <mikestump@comcast.net>
17336 * dse.c (record_store): Remove unnecessary assert.
17338 2013-01-09 Jan Hubicka <jh@suse.cz>
17340 PR tree-optimization/55569
17341 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
17342 * cfgloop.h (scale_loop_profile): Likewise.
17344 2013-01-09 Jan Hubicka <jh@suse.cz>
17347 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
17349 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
17351 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
17353 PR middle-end/55114
17354 * expr.h (maybe_emit_group_store): Declare.
17355 * expr.c (maybe_emit_group_store): New function.
17356 * builtins.c (expand_builtin_int_roundingfn): Call it.
17357 (expand_builtin_int_roundingfn_2): Likewise.
17359 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
17361 PR rtl-optimization/55829
17362 * lra-constraints.c (match_reload): Add code for absent output.
17363 (curr_insn_transform): Add code for reloads of matched inputs
17366 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
17368 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
17369 attribute of movddup insn to DF.
17370 (*vec_interleave_lowv2df): Ditto.
17371 (vec_dupv2df): Ditto.
17373 2013-01-09 Jan Hubicka <jh@suse.cz>
17375 PR tree-optimiation/55875
17376 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
17377 EVERY_ITERATION parameter.
17378 (number_of_iterations_exit): Check if exit is executed every iteration.
17379 (idx_infer_loop_bounds): Similarly here.
17380 (n_of_executions_at_most): Simplify
17381 to only test for cases where statement is dominated by the
17382 particular bound; handle correctly the "postdominance" test.
17383 (scev_probably_wraps_p): Use max loop iterations info
17384 as a global bound first.
17386 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
17387 Nick Clifton <nickc@redhat.com>
17389 * config/v850/v850.md (cbranchsf4): New pattern.
17390 (cstoresf4): New pattern.
17391 (cbranchdf4): New pattern.
17392 (cstoredf4): New pattern.
17393 (movsicc): Disallow floating point comparisons.
17394 (cmpsf_le_insn): Fix order of operators.
17395 (cmpsf_lt_insn): Likewise.
17396 (cmpsf_eq_insn): Likewise.
17397 (cmpdf_le_insn): Likewise.
17398 (cmpdf_lt_insn): Likewise.
17399 (cmpdf_eq_insn): Likewise.
17400 (cmpsf_ge_insn): Use LE comparison.
17401 (cmpdf_ge_insn): Likewise.
17402 (cmpsf_gt_insn): Use LT comparison.
17403 (cmpdf_gt_insn): Likewise.
17404 (cmpsf_ne_insn): Delete pattern.
17405 (cmpdf_ne_insn): Delete pattern.
17406 * config/v850/v850.c (v850_gen_float_compare): Use
17407 gen_cmpdf_eq_insn for NE comparison.
17408 (v850_float_z_comparison_operator)
17409 (v850_float_nz_comparison_operator): Move from here ...
17410 * config/v850/predicates.md: ... to here. Move GT and GE
17411 comparisons into v850_float_z_comparison_operator.
17412 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17414 (v850_float_nz_comparison_operator): Likewise.
17416 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17418 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17419 with calls to gen_insvsi/gen_insvdi.
17421 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17423 * config/i386/i386.c (initial_ix86_tune_features): Set up
17424 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17426 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17427 Jakub Jelinek <jakub@redhat.com>
17429 PR tree-optimization/48189
17430 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17431 If nitercst is 0, don't predict the exit edge.
17433 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17435 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17436 in asm_fprintf with reg_names.
17437 (aarch64_print_operand_address): Likewise.
17438 (aarch64_return_addr): Likewise.
17439 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17441 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17443 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17444 (INT_U6_BITS): Likewise.
17445 * config/pa/predicates.md (uint6_operand): New predicate.
17446 (shift5_operand, shift6_operand): Likewise.
17447 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17449 (lshrdi3): Use shift6_operand.
17450 (shrpsi4, shrpdi4): New insn patterns.
17451 (extzv): Delete expander.
17452 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17453 predicates in unamed zero extract patterns. Tighten common constraint.
17454 (extv): Delete expander.
17455 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17456 predicates in unamed sign extract patterns. Tighten common constraint.
17457 (insv): Delete expander.
17458 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17459 predicates in unamed insert patterns. Tighten common constraint.
17460 Change uint32_operand predicate to uint6_operand predicate in unamed
17461 DImode pattern to insert constant values of type 1...1xxxx.
17463 2013-01-04 Jan Hubicka <jh@suse.cz>
17465 PR tree-optimization/55823
17466 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17469 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17470 Uros Bizjak <ubizjak@gmail.com>
17472 PR rtl-optimization/55845
17473 * df-problems.c (can_move_insns_across): Stop scanning at
17474 volatile_insn_p source instruction or give up if
17475 across_from .. across_to range contains any volatile_insn_p
17478 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17480 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17481 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17483 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17484 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17486 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17489 * asan.c (asan_clear_shadow): New function.
17490 (asan_emit_stack_protection): Use it.
17492 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17494 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17495 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17496 with tab instead of space.
17498 2013-01-08 Nick Clifton <nickc@redhat.com>
17500 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17501 register bank 0 at the start of an interrupt handler.
17502 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17505 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17507 * config/aarch64/aarch64-simd.md
17508 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17509 (aarch64_simd_bsl): Likewise.
17510 (aarch64_vcond_internal<mode>): Likewise.
17511 (vcond<mode><mode>): Likewise.
17512 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17513 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17515 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17517 * config/aarch64/aarch64-builtins.c
17518 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17520 2013-01-08 Martin Jambor <mjambor@suse.cz>
17523 * tree-sra.c (analyze_access_subtree): Return true also after
17524 potentially creating a debug-only replacement.
17526 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17528 PR middle-end/55890
17529 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17531 PR tree-optimization/54120
17532 * tree-vrp.c (range_fits_type_p): Don't allow
17533 src_precision < precision from signed vr to unsigned_p
17534 if vr->min or vr->max is negative.
17535 (simplify_float_conversion_using_ranges): Test can_float_p
17536 against CODE_FOR_nothing.
17538 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17539 Richard Biener <rguenther@suse.de>
17541 PR middle-end/55851
17542 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17543 types instead of just INTEGER_TYPE types.
17545 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
17547 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17550 2013-01-07 Steve Ellcey <sellcey@mips.com>
17553 * config/mips/mips.opt: Change mad to mmad to match documentation.
17555 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17558 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17559 .progmemx.data now.
17561 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17564 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17565 (avr_addrspace_t): Add .section_name field.
17566 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17568 (avr_addrspace): Same. Initialize .section_name. Remove last
17569 NULL entry. Put __memx into .progmemx.data.
17570 (progmem_section_prefix): Remove.
17571 (avr_asm_init_sections): No need to initialize progmem_section.
17572 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17573 section name prefix.
17574 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17575 retrieve the progmem section.
17576 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17577 boundary to run over avr_addrspace[].
17578 (avr_register_target_pragmas): Ditto.
17580 2013-01-06 Jakub Jelinek <jakub@redhat.com>
17582 * varasm.c (output_constant_def_contents): For asan_protect_global
17583 protected strings, adjust DECL_ALIGN if needed, before testing for
17585 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17586 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17588 (output_object_block): For asan protected decls, emit asan padding
17589 after their contents.
17590 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17591 (asan_finish_file): Test it here instead.
17593 2013-01-07 Nick Clifton <nickc@redhat.com>
17594 Matthias Klose <doko@debian.org>
17595 Doug Kwan <dougkwan@google.com>
17596 H.J. Lu <hongjiu.lu@intel.com>
17599 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17601 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17603 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17605 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
17607 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17609 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17612 * doc/install.texi (Cross-Compiler-Specific Options): Document
17615 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17617 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17618 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17619 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17620 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17621 vqmovun_high_s64): Fix source operand number and update copyright.
17623 2013-01-07 Richard Biener <rguenther@suse.de>
17625 PR middle-end/55890
17626 * gimple.h (gimple_call_builtin_p): New overload.
17627 * gimple.c (validate_call): New function.
17628 (gimple_call_builtin_p): Likewise.
17629 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17630 Use gimple_call_builtin_p.
17631 (find_func_clobbers): Likewise.
17632 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17633 (strlen_optimize_stmt): Likewise.
17635 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17637 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17638 (vld1q_dup_*): Likewise.
17639 (vld1_*): Likewise.
17640 (vld1q_*): Likewise.
17641 (vld1_lane_*): Likewise.
17642 (vld1q_lane_*): Likewise.
17644 2013-01-07 Richard Biener <rguenther@suse.de>
17646 * lto-streamer.h (LTO_minor_version): Bump to 2.
17648 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17650 * config/aarch64/aarch64-protos.h
17651 (aarch64_const_double_zero_rtx_p): Rename to...
17652 (aarch64_float_const_zero_rtx_p): ...this.
17653 (aarch64_float_const_representable_p): New.
17654 (aarch64_output_simd_mov_immediate): Likewise.
17655 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17656 move immediate case.
17657 * config/aarch64/aarch64.c
17658 (aarch64_const_double_zero_rtx_p): Rename to...
17659 (aarch64_float_const_zero_rtx_p): ...this.
17660 (aarch64_print_operand): Allow printing of new constants.
17661 (aarch64_valid_floating_const): New.
17662 (aarch64_legitimate_constant_p): Check for valid floating-point
17664 (aarch64_simd_valid_immediate): Likewise.
17665 (aarch64_vect_float_const_representable_p): New.
17666 (aarch64_float_const_representable_p): Likewise.
17667 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17668 (aarch64_output_simd_mov_immediate): New.
17669 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17670 (*movdf_aarch64): Likewise.
17671 * config/aarch64/constraints.md (Ufc): New.
17672 (Y): call aarch64_float_const_zero_rtx.
17673 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17675 2013-01-07 Richard Biener <rguenther@suse.de>
17677 PR tree-optimization/55888
17678 PR tree-optimization/55862
17679 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17680 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17681 not if it is contained therein.
17683 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17685 * config/avr/t-avr: Typo.
17687 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17690 * config/avr/t-avr: Don't automatically rebuild
17691 $(srcdir)/config/avr/t-multilib
17692 $(srcdir)/config/avr/avr-tables.opt
17693 $(srcdir)/doc/avr-mmcu.texi
17694 (avr-mcus): New phony target to build them on request.
17695 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17696 * avr/avr-mcus.def: Adjust comments.
17698 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
17700 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17702 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17704 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17706 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17708 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17710 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
17712 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17713 to generate profiling.
17714 * config/rs6000/aix64.h (LIB_SPEC): Same.
17716 2013-01-04 Andrew Pinski <apinski@cavium.com>
17718 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17720 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17722 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
17724 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17726 (ix86_expand_move): Ditto.
17727 (ix86_zero_extend_to_Pmode): Ditto.
17728 (ix86_expand_call): Ditto.
17729 (ix86_expand_special_args_builtin): Ditto.
17730 (ix86_expand_builtin): Ditto.
17732 2013-01-04 Richard Biener <rguenther@suse.de>
17734 PR tree-optimization/55862
17735 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17736 translating them through PHI nodes.
17738 2013-01-04 Martin Jambor <mjambor@suse.cz>
17740 PR tree-optimization/55755
17741 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17742 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17744 2013-01-04 Marek Polacek <polacek@redhat.com>
17746 PR middle-end/55859
17747 * opts.c (default_options_optimization): Clarify error message.
17749 2013-01-04 Richard Biener <rguenther@suse.de>
17751 PR middle-end/55863
17752 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17755 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17758 * config/pa/pa.md (movsi): Revert previous change.
17759 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17762 2013-01-03 Richard Henderson <rth@redhat.com>
17764 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17765 after eliminating TLS symbols.
17767 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
17770 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
17771 * graphite-poly.c (debug_gmp_value): Likewise.
17773 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
17776 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
17777 selected code model, define __code_mode_small__, __code_model_medium__,
17778 __code_model_large__, __code_model_32__ or __code_model_kernel__.
17779 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
17780 xchg temporary register with %k. Declare temporary register as
17782 [__x86_64__]: For medium and large code models, preserve %rbx register.
17784 2013-01-03 Richard Biener <rguenther@suse.de>
17786 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
17787 (dump_subscript): Adjust.
17788 (finalize_ddr_dependent): Do not dump redundant info.
17789 (analyze_siv_subscript): Adjust.
17790 (subscript_dependence_tester): Likewise.
17791 (compute_affine_dependence): Likewise.
17793 2013-01-03 Richard Biener <rguenther@suse.de>
17796 2013-01-03 Richard Biener <rguenther@suse.de>
17798 PR tree-optimization/55857
17799 * tree-vect-stmts.c (vectorizable_load): Do not setup
17800 re-alignment for invariant loads.
17802 2013-01-02 Richard Biener <rguenther@suse.de>
17804 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17805 invariant load do not generate a vector load from the scalar location.
17807 2013-01-03 Richard Biener <rguenther@suse.de>
17809 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
17810 for not vectorizing.
17811 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
17812 not build INDIRECT_REFs, call get_name once only.
17813 (vect_create_data_ref_ptr): Likewise. Dump base object kind
17814 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
17816 2013-01-03 Richard Biener <rguenther@suse.de>
17818 PR tree-optimization/55857
17819 * tree-vect-stmts.c (vectorizable_load): Do not setup
17820 re-alignment for invariant loads.
17822 2013-01-03 Richard Biener <rguenther@suse.de>
17825 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
17826 prefer a built-in decl.
17828 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17830 * gcc.c (process_command): Update copyright notice dates.
17831 * gcov.c (print_version): Likewise.
17832 * gcov-dump.c (print_version): Likewise.
17834 PR rtl-optimization/55838
17835 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
17836 iv0.step, iv1.step and step.
17838 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17839 Marc Glisse <marc.glisse@inria.fr>
17841 PR tree-optimization/55832
17842 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
17843 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
17844 integer_{one,zero}_node.
17846 2013-01-03 Jakub Jelinek <jakub@redhat.com>
17849 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
17850 * var-tracking.c (reverse_op): Don't add reverse ops to
17851 VALUEs that have already
17852 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
17854 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
17856 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
17858 2013-01-02 Teresa Johnson <tejohnson@google.com>
17860 * dumpfile.c (dump_loc): Print filename with location.
17861 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
17862 new location_t parameter to emit complete unroll message with
17863 new dump framework.
17864 (canonicalize_loop_induction_variables): Compute loops location
17865 and pass to try_unroll_loop_completely.
17866 * loop-unroll.c (report_unroll_peel): New function.
17867 (peel_loops_completely): Use new dump format with location
17868 for main dumpfile message, and invoke report_unroll_peel on success.
17869 (decide_unrolling_and_peeling): Ditto.
17870 (decide_peel_once_rolling): Remove old dumpfile message subsumed
17871 by report_unroll_peel.
17872 (decide_peel_completely): Ditto.
17873 (decide_unroll_constant_iterations): Ditto.
17874 (decide_unroll_runtime_iterations): Ditto.
17875 (decide_peel_simple): Ditto.
17876 (decide_unroll_stupid): Ditto.
17877 * cfgloop.c (get_loop_location): New function.
17878 * cfgloop.h (get_loop_location): Declare.
17880 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17882 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
17885 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17887 PR middle-end/55198
17888 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
17889 BLKmode objects when EXPAND_MEMORY is specified.
17891 2013-01-02 Sriraman Tallam <tmsriram@google.com>
17893 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
17895 (fold_builtin_cpu): Do not share cpu model decls across statements.
17897 2013-01-02 Jason Merrill <jason@redhat.com>
17900 * tree.c (build_array_type_1): Revert earlier change.
17902 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
17904 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
17906 * config/aarch64/aarch64-tune.md: Re-generate.
17908 2013-01-02 Richard Biener <rguenther@suse.de>
17910 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
17911 invariant load do not generate a vector load from the scalar location.
17913 2013-01-02 Richard Biener <rguenther@suse.de>
17916 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
17917 * configure: Regenerate.
17919 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17921 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
17922 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
17923 (expand_builtin_int_roundingfn_2): Keep the original target around
17924 for the fallback case.
17926 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
17928 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
17929 to be clear for sign changes.
17931 2013-01-01 Jan Hubicka <jh@suse.cz>
17933 * ipa-inline-analysis.c: Fix formatting.
17935 2013-01-01 Jakub Jelinek <jakub@redhat.com>
17937 PR tree-optimization/55831
17938 * tree-vect-loop.c (get_initial_def_for_induction): Use
17939 gsi_after_labels instead of gsi_start_bb.
17941 Copyright (C) 2013 Free Software Foundation, Inc.
17943 Copying and distribution of this file, with or without modification,
17944 are permitted in any medium without royalty provided the copyright
17945 notice and this notice are preserved.