1 2013-08-16 Jan Hubicka <jh@suse.cz>
4 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
6 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11 2013-08-16 David Malcolm <dmalcolm@redhat.com>
13 * gengtype.c (type_for_name): Add special-case support for
14 locating types within the "gcc::" namespace.
15 (open_base_files): Emit a "using namespace gcc" directive.
17 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
20 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
21 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
23 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
24 array instead of each individual operand as a separate argument.
25 (emit_fusion_gpr_load): Likewise.
26 (expand_fusion_gpr_load): Add new function declaration.
28 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
29 signature to have the operands passed as an array, instead of as
30 separate arguments. Allow ZERO_EXTEND to be in the memory
31 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
32 depend on the register live/dead flags when peepholes are run.
33 (expand_fusion_gpr_load): New function to be called from the
34 peephole2 pass, to change the register that addis sets to be the
36 (emit_fusion_gpr_load): Change the calling signature to have the
37 operands passed as an array, instead of as separate arguments.
38 Allow ZERO_EXTEND to be in the memory address, and also
39 SIGN_EXTEND if -mpower8-fusion-sign.
41 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
43 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
44 adjust the register addis loads up in the peephole2 pass. Do not
45 depend on the register live/dead state when the peephole pass is
48 2013-08-16 David Malcolm <dmalcolm@redhat.com>
50 * gengtype.c (create_user_defined_type): Ensure that the kind
51 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
52 declaration is seen before the GTY((user)) marking.
54 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
57 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
59 2013-08-16 Jan Hubicka <jh@suse.cz>
61 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
62 arugment expected_type.
63 (gimple_fold_call): Use it.
64 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
65 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
66 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
67 (try_make_edge_direct_virtual_call): Likewise.
68 * tree.c (obj_type_ref_class): New.
69 * tree.h (obj_type_ref_class): Use it.
71 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
73 * sched-vis.c (rtl_slim_pp_initialized): Remove.
74 (rtl_slim_pp): Likewise.
75 (init_rtl_slim_pretty_print): Likewise.
76 (dump_value_slim): Don't call it. Use local pretty printer.
77 (dump_insn_slim): Likewise.
78 (dump_rtl_slim): Likewise.
79 (str_pattern_slim): Likewise.
80 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
83 2013-08-16 Jakub Jelinek <jakub@redhat.com>
85 PR tree-optimization/58164
86 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
87 walk gimple_goto_dest of GIMPLE_GOTO.
89 PR tree-optimization/58165
90 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
91 bi_call must be the last stmt in a bb, don't split_block, instead
92 use fallthru edge from it and give up if there is none.
93 Release conds vector when returning early.
95 2013-08-14 Xinliang David Li <davidxl@google.com>
97 * config/i386/i386.c (ix86_option_override_internal):
98 Remove unused variable and field.
100 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
103 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
104 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
105 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
106 and Linux, correct BLKmode alignment when 128-bit alignment is
107 required and compatibility flag is not set.
108 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
109 for zero-size arguments when compatibility flag is not set.
111 2013-08-14 Jakub Jelinek <jakub@redhat.com>
113 PR tree-optimization/58145
114 * tree-sra.c (build_ref_for_offset): If prev_base has
115 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
117 2013-08-14 Xinliang David Li <davidxl@google.com>
119 * config/i386/i386.c (ix86_option_override_internal):
120 Fix uninitialized variable error.
122 2013-08-14 Xinliang David Li <davidxl@google.com>
124 * config/i386/i386.opt: Define two new options.
125 * config/i386/x86-tune.def: Add arch selector field in macros.
126 * config/i386/i386.h: Adjust macro definition.
127 * config/i386/i386.c (ix86_option_override_internal):
129 (parse_mtune_ctrl_str): New function.
130 (set_ix86_tune_features): New function.
131 (ix86_function_specific_restore): Call the new helper function.
133 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
135 PR rtl-optimization/57662
136 * sel-sched.c (code_motion_process_successors): When the current insn
137 is removed after the recursive traversal, break from the loop.
138 Add comments and debug printouts.
140 2013-08-14 Jakub Jelinek <jakub@redhat.com>
141 Alexandre Oliva <aoliva@redhat.com>
144 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
145 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
146 there also UNSPEC_PLTOFF.
148 2013-08-14 Marek Polacek <polacek@redhat.com>
150 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
151 NUM_CONDITIONS bit positions.
153 2013-08-13 Cary Coutant <ccoutant@google.com>
155 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
156 (attr_checksum): Hash vector contents instead of pointer.
157 (attr_checksum_ordered): Likewise.
159 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
161 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
162 when Pmode != word_mode. Add length_address attribute.
163 (sse3_monitor_<mode>): Merge from sse3_monitor and
164 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
165 Pmode != word_mode. Update insn length attribute.
166 * config/i386/i386.c (ix86_option_override_internal): Update
167 ix86_gen_monitor selection for merged sse3_monitor insn.
169 2013-08-13 Julian Brown <julian@codesourcery.com>
171 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
172 perform invalid legitimization on greater-than-word-size modes for
175 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
177 * ira.c (setup_class_translate_array): Use aclass instead of cl
178 for classes not fully covered by allocno classes.
180 2013-08-13 Jakub Jelinek <jakub@redhat.com>
182 PR tree-optimization/57661
183 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
184 * tree-inline.c (tree_function_versioning): Initialize it.
185 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
186 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
187 that is not being copied.
190 * asan.c (instrument_strlen_call): Fix typo in comment.
191 Use char * type even for the lhs of POINTER_PLUS_EXPR.
193 2013-08-13 Steve Ellcey <sellcey@mips.com>
195 * config/mips/mips.md (prefetch): Use lw instead of ld on
196 loongson in 32bit mode.
198 2013-08-13 Nick Clifton <nickc@redhat.com>
200 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
202 2013-08-13 Jan Hubicka <jh@suse.cz>
204 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
205 introduced edge; fix typo in sanity check.
206 (cgraph_resolve_speculation): Export; improve diagnostic.
207 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
208 speculation at type mismatch.
209 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
210 (cgraph_resolve_speculation): Declare.
211 (symtab_can_be_discarded): New function.
212 * value-prof.c (gimple_ic_transform): Remove actual transform code.
213 * ipa-inline-transform.c (speculation_removed): New global var.
214 (clone_inlined_nodes): See if speculation can be removed.
215 (inline_call): If speculations was removed, we growths may not match.
216 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
217 (speculation_useful_p): New function.
218 (resolve_noninline_speculation): New function.
219 (inline_small_functions): Resolve useless speculations.
220 * ipa-inline.h (speculation_useful_p): Declare
221 * ipa.c (can_replace_by_local_alias): Simplify.
222 (ipa_profile): Produce speculative calls in non-lto, too;
223 add simple cost model; produce local aliases.
225 2013-08-13 David Malcolm <dmalcolm@redhat.com>
227 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
230 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
232 * config/i386/i386.c (ix86_function_versions): Use error + inform.
234 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
236 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
237 iterator instead of X87MODEF.
239 2013-08-12 Perez Read <netfirewall@gmail.com>
242 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
243 operand 0 for intel asm alternative.
244 (*movabs<mode>_2): Ditto for operand 1.
246 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
248 * config/aarch64/arm_none.h
249 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
251 2013-08-12 Nick Clifton <nickc@redhat.com>
253 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
255 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
257 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
258 expand for QI/HImode operand to produce more effictive code for
259 unsigned char(short) --> float(double) conversion.
261 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
263 * doc/invoke.texi: Mention that -ftls-model does not force the final
266 2013-08-12 Marek Polacek <polacek@redhat.com>
267 Marc Glisse <marc.glisse@inria.fr>
269 PR tree-optimization/57980
270 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
271 when creating -1 constant.
273 2013-08-10 Jan Hubicka <jh@suse.cz>
275 Workaround binutils PR14342.
276 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
277 (gimple_init_edge_profiler): Likewise.
278 (gimple_gen_ic_func_profiler): Likewise.
280 2013-08-09 Jan Hubicka <jh@suse.cz>
282 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
284 2013-08-09 Xinliang David Li <davidxl@google.com>
286 * config/i386/stringop.def: New file.
287 * config/i386/stringop.opt: New file.
288 * config/i386/i386-opts.h: Include stringopt.def.
289 * config/i386/i386.opt: Include stringopt.opt.
290 * config/i386/i386.c (ix86_option_override_internal):
291 Override default size based stringop inline strategies with options.
292 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
295 2013-08-09 Jan Hubicka <jh@suse.cz>
297 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
299 2013-08-09 Jan Hubicka <jh@suse.cz>
301 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
303 (dump_cgraph_node): Dump profile-id.
304 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
305 and common_target_probability.
306 * lto-cgraph.c (lto_output_edge): Stream common targets.
307 (lto_output_node): Stream profile ids.
308 (input_node): Stream profile ids.
309 (input_edge): Stream common targets.
310 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
311 * ipa.c: Include value-prof.h
312 (ipa_profile_generate_summary): Turn indirect call statement histograms
314 (ipa_profile): Turn common targets into speculative edges.
316 2013-08-09 Jan Hubicka <jh@suse.cz>
318 * cgraph.h (cgraph_node): Add profile_id.
319 * value-prof.c (cgraph_node_map): Turn into pointer_map.
320 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
321 (del_node_map): Update.
322 (find_func_by_funcdef_no): Replace by ...
323 (find_func_by_profile_id): ... this one.
324 (gimple_ic_transform): Do not remove useful histograms when
325 speculation is not done; dump info when indirect call removal
327 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
328 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
329 (__gcov_indirect_call_profiler_v2): .. this one.
330 * profile.h (init_node_map): Update.
331 * coverage.c (coverage_compute_profile_id): New function.
332 * coverage.h (coverage_compute_profile_id): Declare.
333 * tree-profile.c (init_ic_make_global_vars): Make
334 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
335 (gimple_init_edge_profiler): Update prototype of
336 __gcov_indirect_call_profiler.
337 (gimple_gen_ic_func_profiler): Simplify.
338 (tree_profiling): Use init_node_map
340 2013-08-09 Jan Hubicka <jh@suse.cz>
342 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
343 non-speculative refs.
344 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
345 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
346 (cgraph_set_call_stmt): Likewise.
347 (cgraph_create_edge_1): Fix release checking compilatoin;
349 (cgraph_free_edge): Free indirect info.
350 (cgraph_turn_edge_to_speculative): New function.
351 (cgraph_speculative_call_info): New function.
352 (cgraph_make_edge_direct): Return direct edge; handle speculation.
353 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
354 (dump_cgraph_node): Dump speculation.
355 (verify_edge_count_and_frequency): Accept speculative edges.
356 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
357 (verify_cgraph_node): Handle speculation.
358 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
359 (cgraph_set_call_stmt): Update prototype.
360 (cgraph_make_edge_direct): Update prototype.
361 (cgraph_speculative_call_info): Declare.
362 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
363 to change; update call of ipa_find_references.
364 * ipa-ref.c (ipa_record_reference): Fix return value; clear
365 lto_stmt_uid and speculative flags.
366 (ipa_dump_references): Dump speculation.
367 (ipa_clone_references): Clone speculative flag.
368 (ipa_clone_referring): Likewise.
369 (ipa_clone_ref): New function.
370 (ipa_find_reference): Look into lto_stmt_uids
371 (ipa_clear_stmts_in_references): Do not clear speculative calls.
372 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
373 (ipa_find_reference): Update declaration.
374 (ipa_clone_ref): Declare.
375 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
376 stream speculative flag.
377 (lto_output_ref): Stream statements uids and speculation.
378 (input_ref): Likewise.
379 (input_edge): Stream speuclation.
380 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
381 (cgraph_set_call_stmt_including_clones): Handle speculation.
382 * ipa-inline.c (heap_edge_removal_hook): New function.
383 (inline_small_functions): Register it.
384 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
385 also initialize refs.
386 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
388 (try_make_edge_direct_simple_call): Likewise.
389 (try_make_edge_direct_simple_call): Likewise.
390 (update_indirect_edges_after_inlining): Likewise.
391 (remove_described_reference): Look proper lto_stmt_uid.
392 (propagate_controlled_uses): Likewise.
393 (propagate_controlled_uses): Liekwise.
394 * tree-inline.c (copy_bb): Copy speculative edges.
395 (redirect_all_calls): New function.
396 (copy_cfg_body): Do redirection after loop info is updated.
397 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
399 2013-08-09 Jan Hubicka <jh@suse.cz>
401 * lto-streamer-out.c (output_function): Renumber PHIs.
402 * lto-streamer-in.c (input_function): Likewise.
404 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
406 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
407 (get_lane_unsigned): Likewise.
408 (dup_lane_scalar): Likewise.
409 (get_lane): enable for VALL.
410 * config/aarch64/aarch64-simd.md
411 (aarch64_dup_lane_scalar<mode>): Remove.
412 (aarch64_get_lane_signed<mode>): Likewise.
413 (aarch64_get_lane_unsigned<mode>): Likewise.
414 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
415 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
416 (aarch64_get_lane<mode>): Enable for all vector modes.
417 (aarch64_get_lanedi): Remove misleading constraints.
418 * config/aarch64/arm_neon.h
419 (__aarch64_vget_lane_any): Define.
420 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
421 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
422 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
423 * config/aarch64/iterators.md (VDQQH): New.
427 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
429 * configure.ac: Add GAS check for LEON instructions on SPARC.
430 * configure: Regenerate.
431 * config.in: Likewise.
432 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
434 * config/sparc/sparc.opt (LEON, LEON3): New masks.
435 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
437 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
438 (AS_LEON_FLAG): New macro.
439 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
440 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
441 Deal with LEON and LEON3 for the memory model.
442 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
443 (atomic_compare_and_swap<mode>_1): Likewise.
444 (*atomic_compare_and_swap<mode>_1): Likewise.
446 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
448 * config/arm/neon.md (vcond): Fix floating-point vector
449 comparisons against 0.
451 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
453 * lra-constraints.c (emit_spill_move): Remove assert.
454 (process_alt_operands): Add more debugging
455 output. Increase reject for spilling into memory. Decrease
456 reject for reloading scratch.
457 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
459 2013-08-08 Steve Ellcey <sellcey@mips.com>
461 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
462 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
463 micromips incompatible. Add nan2008.
464 (MULTILIB_DIRNAMES): Add nan2008.
465 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
466 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
467 and micromips incompatible. Add nan2008.
468 (MULTILIB_DIRNAMES): Add nan2008.
469 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
471 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
473 PR rtl-optimization/58079
474 * combine.c (combine_simplify_rtx): Avoid using SUBST if
475 simplify_comparison has widened a comparison with an integer.
477 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
479 * config/arm/neon.md (movmisalign<mode>): Disable when we
480 don't allow unaligned accesses.
481 (*movmisalign<mode>_neon_store): Likewise.
482 (*movmisalign<mode>_neon_load): Likewise.
483 (*movmisalign<mode>_neon_store): Likewise.
484 (*movmisalign<mode>_neon_load): Likewise.
486 2013-08-08 Jan Hubicka <jh@suse.cz>
488 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
489 (make_pass_rebuild_cgraph_edges): Also clear references.
490 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
491 * ipa-inline-transform.c (inline_transform): Remove all references
493 * cgraphunit.c (expand_function): Remove all references after
495 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
496 (ipa_find_reference): Rewrite to iterator.
497 (remove_stmt_references): Likewise.
498 (ipa_clear_stmts_in_references): New function.
499 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
500 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
502 * ipa-split.c (split_function): Remove references in split function.
504 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
507 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
508 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
510 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
513 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
514 suggested mode for the assignment isn't compatible with the
517 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
520 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
522 2013-08-07 Xinliang David Li <davidxl@google.com>
524 * config/i386/i386.opt: New option -mtune-ctrl=.
525 * config/i386/x86-tune.def: New file.
526 * config/i386/i386.h: include x86-tune.def.
527 * config/i386/i386.c (ix86_option_override_internal):
528 Parsing -mtune-ctrl= option and set tune features.
530 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
533 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
534 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
537 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
539 * rtl.h (update_alignments): Declare.
540 * final.c (grow_label_align): New function extracted from...
541 (shorten_branches): ...here. Call it.
542 (update_alignments): New function.
543 * reorg.c (sibling_labels): New variable.
544 (get_label_before): Add SIBLING parameter. If it is non-zero, push
545 the new label along with it onto the sibling_labels vector.
546 (fill_simple_delay_slots): Adjust call to get_label_before.
547 (fill_slots_from_thread): Likewise.
548 (relax_delay_slots): Likewise.
549 (make_return_insns): Likewise.
550 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
552 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
554 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
555 document its semantics.
556 (diagnostic_report_diagnostic): Adjust accordingly.
558 2013-08-07 David Malcolm <dmalcolm@redhat.com>
560 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
561 (sparc_option_override): ...and port to new C++ pass API.
562 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
564 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
566 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
568 2013-08-06 Caroline Tice <cmtice@google.com>
570 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
571 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
572 * tree-pass.h: Add pass_vtable_verify.
573 * varasm.c (assemble_variable): Add code to properly set the comdat
574 section and name for the .vtable_map_vars section.
575 (assemble_vtyv_preinit_initializer): New function.
576 (default_sectin_type_flags): Make sure .vtable_map_vars section has
578 * output.h: Add function decl for assemble_vtv_preinit_initializer.
579 * vtable-verify.c: New file.
580 * vtable-verify.h: New file.
581 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
582 initialiation levels.
583 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
584 * passes.def: Insert pass_vtable_verify.
585 * aclocal.m4: Reorder includes.
586 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
587 -fvtv-counts options.
588 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
589 as appropriate, if -fvtable-verify=... is used.
590 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
591 -fvtable-verify=... is used.
592 * Makefile.in (OBJS): Add vtable-verify.o to list.
593 (vtable-verify.o): Add new build rule.
594 (GTFILES): Add vtable-verify.c to list.
595 * common.opt (fvtable-verify=): New flag.
596 (vtv_priority): Values for fvtable-verify= flag.
597 (fvtv-counts): New flag.
598 (fvtv-debug): New flag.
599 * tree.h (save_vtable_map_decl): New extern function decl.
601 2013-08-07 David Malcolm <dmalcolm@redhat.com>
603 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
604 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
605 (pass_data_rl78_devirt): ...new pass_data instance and...
606 (make_pass_rl78_devirt): ...new function.
607 (rl78_asm_file_start): Port pass registration to new C++ API.
609 2013-08-07 David Malcolm <dmalcolm@redhat.com>
611 * coretypes.h (rtl_opt_pass): Add.
613 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
614 (epiphany_init): Port to new C++ pass API.
615 (epiphany_optimize_mode_switching): Likewise.
616 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
617 (pass_manager::get_pass_mode_switching): New.
618 (pass_manager::get_pass_peephole2): New.
619 * mode-switching.c (pass_mode_switching): Add clone method.
620 * recog.c (pass_peephole2): Add clone method.
621 (pass_split_all_insns): Add clone method.
623 2013-08-06 David Malcolm <dmalcolm@redhat.com>
625 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
626 (mips_option_override): ...here, porting to new C++ API for passes.
628 2013-08-06 Jan Hubicka <jh@suse.cz>
630 * cgraph.c (cgraph_get_body): New function based on lto.c
632 * cgraph.h (cgraph_get_body): Declare.
633 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
635 * cgraphunit.c (expand_function): Get body prior expanding.
636 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
637 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
639 * passes.c (do_per_function_toporder): Get body.
640 * tree-inline.c (expand_call_inline): Get body prior inlining it.
641 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
643 2013-08-06 Martin Jambor <mjambor@suse.cz>
646 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
647 is not re-finalized. Rename second parameter to no_collect.
649 2013-08-06 Martin Jambor <mjambor@suse.cz>
652 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
653 MEM_REF has proper alignment information.
655 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
658 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
660 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
661 rtx (*) (rtx, ...) with insn_gen_fn.
662 * genoutput.c (output_insn_data): Cast gen_? function pointers to
663 insn_gen_fn::stored_funcptr. Add initializer braces.
665 2013-08-05 David Malcolm <dmalcolm@redhat.com>
667 Rewrite how instances of passes are cloned to remove assumptions
668 about their sizes (thus allowing pass subclasses to have
669 additional data fields, albeit non-GC-managed ones at this point).
671 * passes.c (make_pass_instance): Now that passes have clone
672 methods, rewrite this function to eliminate XNEW and memcpy
673 calls that used hardcoded sizes. Since this function no longer
674 creates pass instances, rename it to...
675 (add_pass_instance): ...this. Document the old way that passes were
676 numbered and flagged, and rework this function to continue using it.
677 (next_pass_1): Add an initial_pass argument for use by
679 (position_pass): When adding multiple instances of a pass, use
680 the pass's clone method, rather than relying on the XNEW/memcpy
681 within the former make_pass_instance (now add_pass_instance).
682 (pass_manager::pass_manager): When invoking next_pass_1, also supply
683 the initial instance of the current pass within the pass manager.
685 2013-08-05 David Malcolm <dmalcolm@redhat.com>
687 This is the automated part of the conversion of passes from C
688 structs to C++ classes.
690 Patch autogenerated by refactor_passes.py from
691 https://github.com/davidmalcolm/gcc-refactoring-scripts
692 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
694 * asan.c (pass_asan): Convert from a global struct to a subclass of
695 gimple_opt_pass along with...
696 (pass_data_asan): ...new pass_data instance and...
697 (make_pass_asan): ...new function.
698 (pass_asan_O0): Convert from a global struct to a subclass of
699 gimple_opt_pass along with...
700 (pass_data_asan_O0): ...new pass_data instance and...
701 (make_pass_asan_O0): ...new function.
702 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
703 subclass of rtl_opt_pass along with...
704 (pass_data_inc_dec): ...new pass_data instance and...
705 (make_pass_inc_dec): ...new function.
706 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
707 a subclass of rtl_opt_pass along with...
708 (pass_data_reorder_blocks): ...new pass_data instance and...
709 (make_pass_reorder_blocks): ...new function.
710 (pass_duplicate_computed_gotos): Convert from a global struct to a
711 subclass of rtl_opt_pass along with...
712 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
713 (make_pass_duplicate_computed_gotos): ...new function.
714 (pass_partition_blocks): Convert from a global struct to a subclass of
715 rtl_opt_pass along with...
716 (pass_data_partition_blocks): ...new pass_data instance and...
717 (make_pass_partition_blocks): ...new function.
718 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
719 struct to a subclass of rtl_opt_pass along with...
720 (pass_data_branch_target_load_optimize1): ...new pass_data instance
722 (make_pass_branch_target_load_optimize1): ...new function.
723 (pass_branch_target_load_optimize2): Convert from a global struct to a
724 subclass of rtl_opt_pass along with...
725 (pass_data_branch_target_load_optimize2): ...new pass_data instance
727 (make_pass_branch_target_load_optimize2): ...new function.
728 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
729 of rtl_opt_pass along with...
730 (pass_data_jump): ...new pass_data instance and...
731 (make_pass_jump): ...new function.
732 (pass_jump2): Convert from a global struct to a subclass of
733 rtl_opt_pass along with...
734 (pass_data_jump2): ...new pass_data instance and...
735 (make_pass_jump2): ...new function.
736 * cfgexpand.c (pass_expand): Convert from a global struct to a
737 subclass of rtl_opt_pass along with...
738 (pass_data_expand): ...new pass_data instance and...
739 (make_pass_expand): ...new function.
740 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
741 of rtl_opt_pass along with...
742 (pass_data_free_cfg): ...new pass_data instance and...
743 (make_pass_free_cfg): ...new function.
744 (pass_into_cfg_layout_mode): Convert from a global struct to a
745 subclass of rtl_opt_pass along with...
746 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
747 (make_pass_into_cfg_layout_mode): ...new function.
748 (pass_outof_cfg_layout_mode): Convert from a global struct to a
749 subclass of rtl_opt_pass along with...
750 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
751 (make_pass_outof_cfg_layout_mode): ...new function.
752 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
753 struct to a subclass of gimple_opt_pass along with...
754 (pass_data_build_cgraph_edges): ...new pass_data instance and...
755 (make_pass_build_cgraph_edges): ...new function.
756 (pass_rebuild_cgraph_edges): Convert from a global struct to a
757 subclass of gimple_opt_pass along with...
758 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
759 (make_pass_rebuild_cgraph_edges): ...new function.
760 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
761 subclass of gimple_opt_pass along with...
762 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
764 (make_pass_remove_cgraph_callee_edges): ...new function.
765 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
766 struct to a subclass of rtl_opt_pass along with...
767 (pass_data_stack_adjustments): ...new pass_data instance and...
768 (make_pass_stack_adjustments): ...new function.
769 * combine.c (pass_combine): Convert from a global struct to a subclass
770 of rtl_opt_pass along with...
771 (pass_data_combine): ...new pass_data instance and...
772 (make_pass_combine): ...new function.
773 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
774 global struct to a subclass of rtl_opt_pass along with...
775 (pass_data_compare_elim_after_reload): ...new pass_data instance
777 (make_pass_compare_elim_after_reload): ...new function.
778 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
779 of rtl_opt_pass along with...
780 (pass_data_rtl_cprop): ...new pass_data instance and...
781 (make_pass_rtl_cprop): ...new function.
782 * cse.c (pass_cse): Convert from a global struct to a subclass of
783 rtl_opt_pass along with...
784 (pass_data_cse): ...new pass_data instance and...
785 (make_pass_cse): ...new function.
786 (pass_cse2): Convert from a global struct to a subclass of
787 rtl_opt_pass along with...
788 (pass_data_cse2): ...new pass_data instance and...
789 (make_pass_cse2): ...new function.
790 (pass_cse_after_global_opts): Convert from a global struct to a
791 subclass of rtl_opt_pass along with...
792 (pass_data_cse_after_global_opts): ...new pass_data instance and...
793 (make_pass_cse_after_global_opts): ...new function.
794 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
795 of rtl_opt_pass along with...
796 (pass_data_ud_rtl_dce): ...new pass_data instance and...
797 (make_pass_ud_rtl_dce): ...new function.
798 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
799 rtl_opt_pass along with...
800 (pass_data_fast_rtl_dce): ...new pass_data instance and...
801 (make_pass_fast_rtl_dce): ...new function.
802 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
803 a subclass of rtl_opt_pass along with...
804 (pass_data_df_initialize_opt): ...new pass_data instance and...
805 (make_pass_df_initialize_opt): ...new function.
806 (pass_df_initialize_no_opt): Convert from a global struct to a
807 subclass of rtl_opt_pass along with...
808 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
809 (make_pass_df_initialize_no_opt): ...new function.
810 (pass_df_finish): Convert from a global struct to a subclass of
811 rtl_opt_pass along with...
812 (pass_data_df_finish): ...new pass_data instance and...
813 (make_pass_df_finish): ...new function.
814 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
815 rtl_opt_pass along with...
816 (pass_data_rtl_dse1): ...new pass_data instance and...
817 (make_pass_rtl_dse1): ...new function.
818 (pass_rtl_dse2): Convert from a global struct to a subclass of
819 rtl_opt_pass along with...
820 (pass_data_rtl_dse2): ...new pass_data instance and...
821 (make_pass_rtl_dse2): ...new function.
822 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
823 subclass of rtl_opt_pass along with...
824 (pass_data_dwarf2_frame): ...new pass_data instance and...
825 (make_pass_dwarf2_frame): ...new function.
826 * except.c (pass_set_nothrow_function_flags): Convert from a global
827 struct to a subclass of rtl_opt_pass along with...
828 (pass_data_set_nothrow_function_flags): ...new pass_data instance
830 (make_pass_set_nothrow_function_flags): ...new function.
831 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
832 subclass of rtl_opt_pass along with...
833 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
835 (make_pass_convert_to_eh_region_ranges): ...new function.
836 * final.c (pass_compute_alignments): Convert from a global struct to a
837 subclass of rtl_opt_pass along with...
838 (pass_data_compute_alignments): ...new pass_data instance and...
839 (make_pass_compute_alignments): ...new function.
840 (pass_final): Convert from a global struct to a subclass of
841 rtl_opt_pass along with...
842 (pass_data_final): ...new pass_data instance and...
843 (make_pass_final): ...new function.
844 (pass_shorten_branches): Convert from a global struct to a subclass of
845 rtl_opt_pass along with...
846 (pass_data_shorten_branches): ...new pass_data instance and...
847 (make_pass_shorten_branches): ...new function.
848 (pass_clean_state): Convert from a global struct to a subclass of
849 rtl_opt_pass along with...
850 (pass_data_clean_state): ...new pass_data instance and...
851 (make_pass_clean_state): ...new function.
852 * function.c (pass_instantiate_virtual_regs): Convert from a global
853 struct to a subclass of rtl_opt_pass along with...
854 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
855 (make_pass_instantiate_virtual_regs): ...new function.
856 (pass_leaf_regs): Convert from a global struct to a subclass of
857 rtl_opt_pass along with...
858 (pass_data_leaf_regs): ...new pass_data instance and...
859 (make_pass_leaf_regs): ...new function.
860 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
861 subclass of rtl_opt_pass along with...
862 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
864 (make_pass_thread_prologue_and_epilogue): ...new function.
865 (pass_match_asm_constraints): Convert from a global struct to a
866 subclass of rtl_opt_pass along with...
867 (pass_data_match_asm_constraints): ...new pass_data instance and...
868 (make_pass_match_asm_constraints): ...new function.
869 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
870 subclass of rtl_opt_pass along with...
871 (pass_data_rtl_fwprop): ...new pass_data instance and...
872 (make_pass_rtl_fwprop): ...new function.
873 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
874 rtl_opt_pass along with...
875 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
876 (make_pass_rtl_fwprop_addr): ...new function.
877 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
878 rtl_opt_pass along with...
879 (pass_data_rtl_pre): ...new pass_data instance and...
880 (make_pass_rtl_pre): ...new function.
881 (pass_rtl_hoist): Convert from a global struct to a subclass of
882 rtl_opt_pass along with...
883 (pass_data_rtl_hoist): ...new pass_data instance and...
884 (make_pass_rtl_hoist): ...new function.
885 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
886 subclass of gimple_opt_pass along with...
887 (pass_data_lower_cf): ...new pass_data instance and...
888 (make_pass_lower_cf): ...new function.
889 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
890 from a global struct to a subclass of gimple_opt_pass along with...
891 (pass_data_strength_reduction): ...new pass_data instance and...
892 (make_pass_strength_reduction): ...new function.
893 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
894 of rtl_opt_pass along with...
895 (pass_data_rtl_ifcvt): ...new pass_data instance and...
896 (make_pass_rtl_ifcvt): ...new function.
897 (pass_if_after_combine): Convert from a global struct to a subclass of
898 rtl_opt_pass along with...
899 (pass_data_if_after_combine): ...new pass_data instance and...
900 (make_pass_if_after_combine): ...new function.
901 (pass_if_after_reload): Convert from a global struct to a subclass of
902 rtl_opt_pass along with...
903 (pass_data_if_after_reload): ...new pass_data instance and...
904 (make_pass_if_after_reload): ...new function.
905 * init-regs.c (pass_initialize_regs): Convert from a global struct to
906 a subclass of rtl_opt_pass along with...
907 (pass_data_initialize_regs): ...new pass_data instance and...
908 (make_pass_initialize_regs): ...new function.
909 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
910 of ipa_opt_pass_d along with...
911 (pass_data_ipa_cp): ...new pass_data instance and...
912 (make_pass_ipa_cp): ...new function.
913 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
914 global struct to a subclass of gimple_opt_pass along with...
915 (pass_data_inline_parameters): ...new pass_data instance and...
916 (make_pass_inline_parameters): ...new function.
917 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
918 subclass of gimple_opt_pass along with...
919 (pass_data_early_inline): ...new pass_data instance and...
920 (make_pass_early_inline): ...new function.
921 (pass_ipa_inline): Convert from a global struct to a subclass of
922 ipa_opt_pass_d along with...
923 (pass_data_ipa_inline): ...new pass_data instance and...
924 (make_pass_ipa_inline): ...new function.
925 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
926 struct to a subclass of gimple_opt_pass along with...
927 (pass_data_local_pure_const): ...new pass_data instance and...
928 (make_pass_local_pure_const): ...new function.
929 (pass_ipa_pure_const): Convert from a global struct to a subclass of
930 ipa_opt_pass_d along with...
931 (pass_data_ipa_pure_const): ...new pass_data instance and...
932 (make_pass_ipa_pure_const): ...new function.
933 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
934 to a subclass of ipa_opt_pass_d along with...
935 (pass_data_ipa_reference): ...new pass_data instance and...
936 (make_pass_ipa_reference): ...new function.
937 * ipa-split.c (pass_split_functions): Convert from a global struct to
938 a subclass of gimple_opt_pass along with...
939 (pass_data_split_functions): ...new pass_data instance and...
940 (make_pass_split_functions): ...new function.
941 (pass_feedback_split_functions): Convert from a global struct to a
942 subclass of gimple_opt_pass along with...
943 (pass_data_feedback_split_functions): ...new pass_data instance and...
944 (make_pass_feedback_split_functions): ...new function.
945 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
946 global struct to a subclass of simple_ipa_opt_pass along with...
947 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
949 (make_pass_ipa_function_and_variable_visibility): ...new function.
950 (pass_ipa_free_inline_summary): Convert from a global struct to a
951 subclass of simple_ipa_opt_pass along with...
952 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
953 (make_pass_ipa_free_inline_summary): ...new function.
954 (pass_ipa_whole_program_visibility): Convert from a global struct to a
955 subclass of ipa_opt_pass_d along with...
956 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
958 (make_pass_ipa_whole_program_visibility): ...new function.
959 (pass_ipa_profile): Convert from a global struct to a subclass of
960 ipa_opt_pass_d along with...
961 (pass_data_ipa_profile): ...new pass_data instance and...
962 (make_pass_ipa_profile): ...new function.
963 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
964 ipa_opt_pass_d along with...
965 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
966 (make_pass_ipa_cdtor_merge): ...new function.
967 * ira.c (pass_ira): Convert from a global struct to a subclass of
968 rtl_opt_pass along with...
969 (pass_data_ira): ...new pass_data instance and...
970 (make_pass_ira): ...new function.
971 (pass_reload): Convert from a global struct to a subclass of
972 rtl_opt_pass along with...
973 (pass_data_reload): ...new pass_data instance and...
974 (make_pass_reload): ...new function.
975 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
976 subclass of rtl_opt_pass along with...
977 (pass_data_cleanup_barriers): ...new pass_data instance and...
978 (make_pass_cleanup_barriers): ...new function.
979 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
980 of rtl_opt_pass along with...
981 (pass_data_loop2): ...new pass_data instance and...
982 (make_pass_loop2): ...new function.
983 (pass_rtl_loop_init): Convert from a global struct to a subclass of
984 rtl_opt_pass along with...
985 (pass_data_rtl_loop_init): ...new pass_data instance and...
986 (make_pass_rtl_loop_init): ...new function.
987 (pass_rtl_loop_done): Convert from a global struct to a subclass of
988 rtl_opt_pass along with...
989 (pass_data_rtl_loop_done): ...new pass_data instance and...
990 (make_pass_rtl_loop_done): ...new function.
991 (pass_rtl_move_loop_invariants): Convert from a global struct to a
992 subclass of rtl_opt_pass along with...
993 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
994 (make_pass_rtl_move_loop_invariants): ...new function.
995 (pass_rtl_unswitch): Convert from a global struct to a subclass of
996 rtl_opt_pass along with...
997 (pass_data_rtl_unswitch): ...new pass_data instance and...
998 (make_pass_rtl_unswitch): ...new function.
999 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1000 subclass of rtl_opt_pass along with...
1001 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1003 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1004 (pass_rtl_doloop): Convert from a global struct to a subclass of
1005 rtl_opt_pass along with...
1006 (pass_data_rtl_doloop): ...new pass_data instance and...
1007 (make_pass_rtl_doloop): ...new function.
1008 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1009 a subclass of rtl_opt_pass along with...
1010 (pass_data_lower_subreg): ...new pass_data instance and...
1011 (make_pass_lower_subreg): ...new function.
1012 (pass_lower_subreg2): Convert from a global struct to a subclass of
1013 rtl_opt_pass along with...
1014 (pass_data_lower_subreg2): ...new pass_data instance and...
1015 (make_pass_lower_subreg2): ...new function.
1016 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1017 struct to a subclass of ipa_opt_pass_d along with...
1018 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1019 (make_pass_ipa_lto_gimple_out): ...new function.
1020 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1021 of ipa_opt_pass_d along with...
1022 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1023 (make_pass_ipa_lto_finish_out): ...new function.
1024 * mode-switching.c (pass_mode_switching): Convert from a global struct
1025 to a subclass of rtl_opt_pass along with...
1026 (pass_data_mode_switching): ...new pass_data instance and...
1027 (make_pass_mode_switching): ...new function.
1028 * modulo-sched.c (pass_sms): Convert from a global struct to a
1029 subclass of rtl_opt_pass along with...
1030 (pass_data_sms): ...new pass_data instance and...
1031 (make_pass_sms): ...new function.
1032 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1033 subclass of gimple_opt_pass along with...
1034 (pass_data_expand_omp): ...new pass_data instance and...
1035 (make_pass_expand_omp): ...new function.
1036 (pass_lower_omp): Convert from a global struct to a subclass of
1037 gimple_opt_pass along with...
1038 (pass_data_lower_omp): ...new pass_data instance and...
1039 (make_pass_lower_omp): ...new function.
1040 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1041 of gimple_opt_pass along with...
1042 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1043 (make_pass_diagnose_omp_blocks): ...new function.
1044 * passes.c (pass_early_local_passes): Convert from a global struct to
1045 a subclass of simple_ipa_opt_pass along with...
1046 (pass_data_early_local_passes): ...new pass_data instance and...
1047 (make_pass_early_local_passes): ...new function.
1048 (pass_all_early_optimizations): Convert from a global struct to a
1049 subclass of gimple_opt_pass along with...
1050 (pass_data_all_early_optimizations): ...new pass_data instance and...
1051 (make_pass_all_early_optimizations): ...new function.
1052 (pass_all_optimizations): Convert from a global struct to a subclass
1053 of gimple_opt_pass along with...
1054 (pass_data_all_optimizations): ...new pass_data instance and...
1055 (make_pass_all_optimizations): ...new function.
1056 (pass_all_optimizations_g): Convert from a global struct to a subclass
1057 of gimple_opt_pass along with...
1058 (pass_data_all_optimizations_g): ...new pass_data instance and...
1059 (make_pass_all_optimizations_g): ...new function.
1060 (pass_rest_of_compilation): Convert from a global struct to a subclass
1061 of rtl_opt_pass along with...
1062 (pass_data_rest_of_compilation): ...new pass_data instance and...
1063 (make_pass_rest_of_compilation): ...new function.
1064 (pass_postreload): Convert from a global struct to a subclass of
1065 rtl_opt_pass along with...
1066 (pass_data_postreload): ...new pass_data instance and...
1067 (make_pass_postreload): ...new function.
1068 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1069 subclass of rtl_opt_pass along with...
1070 (pass_data_gcse2): ...new pass_data instance and...
1071 (make_pass_gcse2): ...new function.
1072 * postreload.c (pass_postreload_cse): Convert from a global struct to
1073 a subclass of rtl_opt_pass along with...
1074 (pass_data_postreload_cse): ...new pass_data instance and...
1075 (make_pass_postreload_cse): ...new function.
1076 * predict.c (pass_profile): Convert from a global struct to a subclass
1077 of gimple_opt_pass along with...
1078 (pass_data_profile): ...new pass_data instance and...
1079 (make_pass_profile): ...new function.
1080 (pass_strip_predict_hints): Convert from a global struct to a subclass
1081 of gimple_opt_pass along with...
1082 (pass_data_strip_predict_hints): ...new pass_data instance and...
1083 (make_pass_strip_predict_hints): ...new function.
1084 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1085 of rtl_opt_pass along with...
1086 (pass_data_peephole2): ...new pass_data instance and...
1087 (make_pass_peephole2): ...new function.
1088 (pass_split_all_insns): Convert from a global struct to a subclass of
1089 rtl_opt_pass along with...
1090 (pass_data_split_all_insns): ...new pass_data instance and...
1091 (make_pass_split_all_insns): ...new function.
1092 (pass_split_after_reload): Convert from a global struct to a subclass
1093 of rtl_opt_pass along with...
1094 (pass_data_split_after_reload): ...new pass_data instance and...
1095 (make_pass_split_after_reload): ...new function.
1096 (pass_split_before_regstack): Convert from a global struct to a
1097 subclass of rtl_opt_pass along with...
1098 (pass_data_split_before_regstack): ...new pass_data instance and...
1099 (make_pass_split_before_regstack): ...new function.
1100 (pass_split_before_sched2): Convert from a global struct to a subclass
1101 of rtl_opt_pass along with...
1102 (pass_data_split_before_sched2): ...new pass_data instance and...
1103 (make_pass_split_before_sched2): ...new function.
1104 (pass_split_for_shorten_branches): Convert from a global struct to a
1105 subclass of rtl_opt_pass along with...
1106 (pass_data_split_for_shorten_branches): ...new pass_data instance
1108 (make_pass_split_for_shorten_branches): ...new function.
1109 * ree.c (pass_ree): Convert from a global struct to a subclass of
1110 rtl_opt_pass along with...
1111 (pass_data_ree): ...new pass_data instance and...
1112 (make_pass_ree): ...new function.
1113 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1114 subclass of rtl_opt_pass along with...
1115 (pass_data_stack_regs): ...new pass_data instance and...
1116 (make_pass_stack_regs): ...new function.
1117 (pass_stack_regs_run): Convert from a global struct to a subclass of
1118 rtl_opt_pass along with...
1119 (pass_data_stack_regs_run): ...new pass_data instance and...
1120 (make_pass_stack_regs_run): ...new function.
1121 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1122 subclass of rtl_opt_pass along with...
1123 (pass_data_cprop_hardreg): ...new pass_data instance and...
1124 (make_pass_cprop_hardreg): ...new function.
1125 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1126 subclass of rtl_opt_pass along with...
1127 (pass_data_reginfo_init): ...new pass_data instance and...
1128 (make_pass_reginfo_init): ...new function.
1129 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1130 of rtl_opt_pass along with...
1131 (pass_data_regmove): ...new pass_data instance and...
1132 (make_pass_regmove): ...new function.
1133 * regrename.c (pass_regrename): Convert from a global struct to a
1134 subclass of rtl_opt_pass along with...
1135 (pass_data_regrename): ...new pass_data instance and...
1136 (make_pass_regrename): ...new function.
1137 * reorg.c (pass_delay_slots): Convert from a global struct to a
1138 subclass of rtl_opt_pass along with...
1139 (pass_data_delay_slots): ...new pass_data instance and...
1140 (make_pass_delay_slots): ...new function.
1141 (pass_machine_reorg): Convert from a global struct to a subclass of
1142 rtl_opt_pass along with...
1143 (pass_data_machine_reorg): ...new pass_data instance and...
1144 (make_pass_machine_reorg): ...new function.
1145 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1146 of rtl_opt_pass along with...
1147 (pass_data_sched): ...new pass_data instance and...
1148 (make_pass_sched): ...new function.
1149 (pass_sched2): Convert from a global struct to a subclass of
1150 rtl_opt_pass along with...
1151 (pass_data_sched2): ...new pass_data instance and...
1152 (make_pass_sched2): ...new function.
1153 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1154 to a subclass of rtl_opt_pass along with...
1155 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1156 (make_pass_stack_ptr_mod): ...new function.
1157 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1158 to a subclass of rtl_opt_pass along with...
1159 (pass_data_rtl_store_motion): ...new pass_data instance and...
1160 (make_pass_rtl_store_motion): ...new function.
1161 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1162 of gimple_opt_pass along with...
1163 (pass_data_tracer): ...new pass_data instance and...
1164 (make_pass_tracer): ...new function.
1165 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1166 to a subclass of gimple_opt_pass along with...
1167 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1168 (make_pass_diagnose_tm_blocks): ...new function.
1169 (pass_lower_tm): Convert from a global struct to a subclass of
1170 gimple_opt_pass along with...
1171 (pass_data_lower_tm): ...new pass_data instance and...
1172 (make_pass_lower_tm): ...new function.
1173 (pass_tm_init): Convert from a global struct to a subclass of
1174 gimple_opt_pass along with...
1175 (pass_data_tm_init): ...new pass_data instance and...
1176 (make_pass_tm_init): ...new function.
1177 (pass_tm_mark): Convert from a global struct to a subclass of
1178 gimple_opt_pass along with...
1179 (pass_data_tm_mark): ...new pass_data instance and...
1180 (make_pass_tm_mark): ...new function.
1181 (pass_tm_edges): Convert from a global struct to a subclass of
1182 gimple_opt_pass along with...
1183 (pass_data_tm_edges): ...new pass_data instance and...
1184 (make_pass_tm_edges): ...new function.
1185 (pass_tm_memopt): Convert from a global struct to a subclass of
1186 gimple_opt_pass along with...
1187 (pass_data_tm_memopt): ...new pass_data instance and...
1188 (make_pass_tm_memopt): ...new function.
1189 (pass_ipa_tm): Convert from a global struct to a subclass of
1190 simple_ipa_opt_pass along with...
1191 (pass_data_ipa_tm): ...new pass_data instance and...
1192 (make_pass_ipa_tm): ...new function.
1193 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1194 subclass of gimple_opt_pass along with...
1195 (pass_data_call_cdce): ...new pass_data instance and...
1196 (make_pass_call_cdce): ...new function.
1197 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1198 subclass of gimple_opt_pass along with...
1199 (pass_data_build_cfg): ...new pass_data instance and...
1200 (make_pass_build_cfg): ...new function.
1201 (pass_split_crit_edges): Convert from a global struct to a subclass of
1202 gimple_opt_pass along with...
1203 (pass_data_split_crit_edges): ...new pass_data instance and...
1204 (make_pass_split_crit_edges): ...new function.
1205 (pass_warn_function_return): Convert from a global struct to a
1206 subclass of gimple_opt_pass along with...
1207 (pass_data_warn_function_return): ...new pass_data instance and...
1208 (make_pass_warn_function_return): ...new function.
1209 (pass_warn_function_noreturn): Convert from a global struct to a
1210 subclass of gimple_opt_pass along with...
1211 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1212 (make_pass_warn_function_noreturn): ...new function.
1213 (pass_warn_unused_result): Convert from a global struct to a subclass
1214 of gimple_opt_pass along with...
1215 (pass_data_warn_unused_result): ...new pass_data instance and...
1216 (make_pass_warn_unused_result): ...new function.
1217 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1218 a subclass of gimple_opt_pass along with...
1219 (pass_data_merge_phi): ...new pass_data instance and...
1220 (make_pass_merge_phi): ...new function.
1221 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1222 a subclass of gimple_opt_pass along with...
1223 (pass_data_lower_complex): ...new pass_data instance and...
1224 (make_pass_lower_complex): ...new function.
1225 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1226 gimple_opt_pass along with...
1227 (pass_data_lower_complex_O0): ...new pass_data instance and...
1228 (make_pass_lower_complex_O0): ...new function.
1229 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1230 subclass of gimple_opt_pass along with...
1231 (pass_data_lower_eh): ...new pass_data instance and...
1232 (make_pass_lower_eh): ...new function.
1233 (pass_refactor_eh): Convert from a global struct to a subclass of
1234 gimple_opt_pass along with...
1235 (pass_data_refactor_eh): ...new pass_data instance and...
1236 (make_pass_refactor_eh): ...new function.
1237 (pass_lower_resx): Convert from a global struct to a subclass of
1238 gimple_opt_pass along with...
1239 (pass_data_lower_resx): ...new pass_data instance and...
1240 (make_pass_lower_resx): ...new function.
1241 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1242 of gimple_opt_pass along with...
1243 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1244 (make_pass_lower_eh_dispatch): ...new function.
1245 (pass_cleanup_eh): Convert from a global struct to a subclass of
1246 gimple_opt_pass along with...
1247 (pass_data_cleanup_eh): ...new pass_data instance and...
1248 (make_pass_cleanup_eh): ...new function.
1249 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1250 to a subclass of simple_ipa_opt_pass along with...
1251 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1252 (make_pass_ipa_lower_emutls): ...new function.
1253 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1254 a subclass of gimple_opt_pass along with...
1255 (pass_data_if_conversion): ...new pass_data instance and...
1256 (make_pass_if_conversion): ...new function.
1257 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1258 subclass of gimple_opt_pass along with...
1259 (pass_data_build_ssa): ...new pass_data instance and...
1260 (make_pass_build_ssa): ...new function.
1261 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1262 global struct to a subclass of gimple_opt_pass along with...
1263 (pass_data_loop_distribution): ...new pass_data instance and...
1264 (make_pass_loop_distribution): ...new function.
1265 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1266 subclass of gimple_opt_pass along with...
1267 (pass_data_mudflap_1): ...new pass_data instance and...
1268 (make_pass_mudflap_1): ...new function.
1269 (pass_mudflap_2): Convert from a global struct to a subclass of
1270 gimple_opt_pass along with...
1271 (pass_data_mudflap_2): ...new pass_data instance and...
1272 (make_pass_mudflap_2): ...new function.
1273 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1274 subclass of gimple_opt_pass along with...
1275 (pass_data_mudflap_1): ...new pass_data instance and...
1276 (make_pass_mudflap_1): ...new function.
1277 (pass_mudflap_2): Convert from a global struct to a subclass of
1278 gimple_opt_pass along with...
1279 (pass_data_mudflap_2): ...new pass_data instance and...
1280 (make_pass_mudflap_2): ...new function.
1281 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1282 gimple_opt_pass along with...
1283 (pass_data_nrv): ...new pass_data instance and...
1284 (make_pass_nrv): ...new function.
1285 (pass_return_slot): Convert from a global struct to a subclass of
1286 gimple_opt_pass along with...
1287 (pass_data_return_slot): ...new pass_data instance and...
1288 (make_pass_return_slot): ...new function.
1289 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1290 to a subclass of gimple_opt_pass along with...
1291 (pass_data_object_sizes): ...new pass_data instance and...
1292 (make_pass_object_sizes): ...new function.
1293 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1294 global struct to a subclass of gimple_opt_pass along with...
1295 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1297 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1298 (pass_fixup_cfg): Convert from a global struct to a subclass of
1299 gimple_opt_pass along with...
1300 (pass_data_fixup_cfg): ...new pass_data instance and...
1301 (make_pass_fixup_cfg): ...new function.
1302 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1303 (make_pass_mudflap_1): ...new function.
1304 (pass_mudflap_2): Replace declaration with that of...
1305 (make_pass_mudflap_2): ...new function.
1306 (pass_asan): Replace declaration with that of...
1307 (make_pass_asan): ...new function.
1308 (pass_asan_O0): Replace declaration with that of...
1309 (make_pass_asan_O0): ...new function.
1310 (pass_tsan): Replace declaration with that of...
1311 (make_pass_tsan): ...new function.
1312 (pass_tsan_O0): Replace declaration with that of...
1313 (make_pass_tsan_O0): ...new function.
1314 (pass_lower_cf): Replace declaration with that of...
1315 (make_pass_lower_cf): ...new function.
1316 (pass_refactor_eh): Replace declaration with that of...
1317 (make_pass_refactor_eh): ...new function.
1318 (pass_lower_eh): Replace declaration with that of...
1319 (make_pass_lower_eh): ...new function.
1320 (pass_lower_eh_dispatch): Replace declaration with that of...
1321 (make_pass_lower_eh_dispatch): ...new function.
1322 (pass_lower_resx): Replace declaration with that of...
1323 (make_pass_lower_resx): ...new function.
1324 (pass_build_cfg): Replace declaration with that of...
1325 (make_pass_build_cfg): ...new function.
1326 (pass_early_tree_profile): Replace declaration with that of...
1327 (make_pass_early_tree_profile): ...new function.
1328 (pass_cleanup_eh): Replace declaration with that of...
1329 (make_pass_cleanup_eh): ...new function.
1330 (pass_sra): Replace declaration with that of...
1331 (make_pass_sra): ...new function.
1332 (pass_sra_early): Replace declaration with that of...
1333 (make_pass_sra_early): ...new function.
1334 (pass_early_ipa_sra): Replace declaration with that of...
1335 (make_pass_early_ipa_sra): ...new function.
1336 (pass_tail_recursion): Replace declaration with that of...
1337 (make_pass_tail_recursion): ...new function.
1338 (pass_tail_calls): Replace declaration with that of...
1339 (make_pass_tail_calls): ...new function.
1340 (pass_tree_loop): Replace declaration with that of...
1341 (make_pass_tree_loop): ...new function.
1342 (pass_tree_loop_init): Replace declaration with that of...
1343 (make_pass_tree_loop_init): ...new function.
1344 (pass_lim): Replace declaration with that of...
1345 (make_pass_lim): ...new function.
1346 (pass_tree_unswitch): Replace declaration with that of...
1347 (make_pass_tree_unswitch): ...new function.
1348 (pass_predcom): Replace declaration with that of...
1349 (make_pass_predcom): ...new function.
1350 (pass_iv_canon): Replace declaration with that of...
1351 (make_pass_iv_canon): ...new function.
1352 (pass_scev_cprop): Replace declaration with that of...
1353 (make_pass_scev_cprop): ...new function.
1354 (pass_empty_loop): Replace declaration with that of...
1355 (make_pass_empty_loop): ...new function.
1356 (pass_record_bounds): Replace declaration with that of...
1357 (make_pass_record_bounds): ...new function.
1358 (pass_graphite): Replace declaration with that of...
1359 (make_pass_graphite): ...new function.
1360 (pass_graphite_transforms): Replace declaration with that of...
1361 (make_pass_graphite_transforms): ...new function.
1362 (pass_if_conversion): Replace declaration with that of...
1363 (make_pass_if_conversion): ...new function.
1364 (pass_loop_distribution): Replace declaration with that of...
1365 (make_pass_loop_distribution): ...new function.
1366 (pass_vectorize): Replace declaration with that of...
1367 (make_pass_vectorize): ...new function.
1368 (pass_slp_vectorize): Replace declaration with that of...
1369 (make_pass_slp_vectorize): ...new function.
1370 (pass_complete_unroll): Replace declaration with that of...
1371 (make_pass_complete_unroll): ...new function.
1372 (pass_complete_unrolli): Replace declaration with that of...
1373 (make_pass_complete_unrolli): ...new function.
1374 (pass_parallelize_loops): Replace declaration with that of...
1375 (make_pass_parallelize_loops): ...new function.
1376 (pass_loop_prefetch): Replace declaration with that of...
1377 (make_pass_loop_prefetch): ...new function.
1378 (pass_iv_optimize): Replace declaration with that of...
1379 (make_pass_iv_optimize): ...new function.
1380 (pass_tree_loop_done): Replace declaration with that of...
1381 (make_pass_tree_loop_done): ...new function.
1382 (pass_ch): Replace declaration with that of...
1383 (make_pass_ch): ...new function.
1384 (pass_ccp): Replace declaration with that of...
1385 (make_pass_ccp): ...new function.
1386 (pass_phi_only_cprop): Replace declaration with that of...
1387 (make_pass_phi_only_cprop): ...new function.
1388 (pass_build_ssa): Replace declaration with that of...
1389 (make_pass_build_ssa): ...new function.
1390 (pass_build_alias): Replace declaration with that of...
1391 (make_pass_build_alias): ...new function.
1392 (pass_build_ealias): Replace declaration with that of...
1393 (make_pass_build_ealias): ...new function.
1394 (pass_dominator): Replace declaration with that of...
1395 (make_pass_dominator): ...new function.
1396 (pass_dce): Replace declaration with that of...
1397 (make_pass_dce): ...new function.
1398 (pass_dce_loop): Replace declaration with that of...
1399 (make_pass_dce_loop): ...new function.
1400 (pass_cd_dce): Replace declaration with that of...
1401 (make_pass_cd_dce): ...new function.
1402 (pass_call_cdce): Replace declaration with that of...
1403 (make_pass_call_cdce): ...new function.
1404 (pass_merge_phi): Replace declaration with that of...
1405 (make_pass_merge_phi): ...new function.
1406 (pass_split_crit_edges): Replace declaration with that of...
1407 (make_pass_split_crit_edges): ...new function.
1408 (pass_pre): Replace declaration with that of...
1409 (make_pass_pre): ...new function.
1410 (pass_profile): Replace declaration with that of...
1411 (make_pass_profile): ...new function.
1412 (pass_strip_predict_hints): Replace declaration with that of...
1413 (make_pass_strip_predict_hints): ...new function.
1414 (pass_lower_complex_O0): Replace declaration with that of...
1415 (make_pass_lower_complex_O0): ...new function.
1416 (pass_lower_complex): Replace declaration with that of...
1417 (make_pass_lower_complex): ...new function.
1418 (pass_lower_vector): Replace declaration with that of...
1419 (make_pass_lower_vector): ...new function.
1420 (pass_lower_vector_ssa): Replace declaration with that of...
1421 (make_pass_lower_vector_ssa): ...new function.
1422 (pass_lower_omp): Replace declaration with that of...
1423 (make_pass_lower_omp): ...new function.
1424 (pass_diagnose_omp_blocks): Replace declaration with that of...
1425 (make_pass_diagnose_omp_blocks): ...new function.
1426 (pass_expand_omp): Replace declaration with that of...
1427 (make_pass_expand_omp): ...new function.
1428 (pass_expand_omp_ssa): Replace declaration with that of...
1429 (make_pass_expand_omp_ssa): ...new function.
1430 (pass_object_sizes): Replace declaration with that of...
1431 (make_pass_object_sizes): ...new function.
1432 (pass_strlen): Replace declaration with that of...
1433 (make_pass_strlen): ...new function.
1434 (pass_fold_builtins): Replace declaration with that of...
1435 (make_pass_fold_builtins): ...new function.
1436 (pass_stdarg): Replace declaration with that of...
1437 (make_pass_stdarg): ...new function.
1438 (pass_early_warn_uninitialized): Replace declaration with that of...
1439 (make_pass_early_warn_uninitialized): ...new function.
1440 (pass_late_warn_uninitialized): Replace declaration with that of...
1441 (make_pass_late_warn_uninitialized): ...new function.
1442 (pass_cse_reciprocals): Replace declaration with that of...
1443 (make_pass_cse_reciprocals): ...new function.
1444 (pass_cse_sincos): Replace declaration with that of...
1445 (make_pass_cse_sincos): ...new function.
1446 (pass_optimize_bswap): Replace declaration with that of...
1447 (make_pass_optimize_bswap): ...new function.
1448 (pass_optimize_widening_mul): Replace declaration with that of...
1449 (make_pass_optimize_widening_mul): ...new function.
1450 (pass_warn_function_return): Replace declaration with that of...
1451 (make_pass_warn_function_return): ...new function.
1452 (pass_warn_function_noreturn): Replace declaration with that of...
1453 (make_pass_warn_function_noreturn): ...new function.
1454 (pass_cselim): Replace declaration with that of...
1455 (make_pass_cselim): ...new function.
1456 (pass_phiopt): Replace declaration with that of...
1457 (make_pass_phiopt): ...new function.
1458 (pass_forwprop): Replace declaration with that of...
1459 (make_pass_forwprop): ...new function.
1460 (pass_phiprop): Replace declaration with that of...
1461 (make_pass_phiprop): ...new function.
1462 (pass_tree_ifcombine): Replace declaration with that of...
1463 (make_pass_tree_ifcombine): ...new function.
1464 (pass_dse): Replace declaration with that of...
1465 (make_pass_dse): ...new function.
1466 (pass_nrv): Replace declaration with that of...
1467 (make_pass_nrv): ...new function.
1468 (pass_rename_ssa_copies): Replace declaration with that of...
1469 (make_pass_rename_ssa_copies): ...new function.
1470 (pass_sink_code): Replace declaration with that of...
1471 (make_pass_sink_code): ...new function.
1472 (pass_fre): Replace declaration with that of...
1473 (make_pass_fre): ...new function.
1474 (pass_check_data_deps): Replace declaration with that of...
1475 (make_pass_check_data_deps): ...new function.
1476 (pass_copy_prop): Replace declaration with that of...
1477 (make_pass_copy_prop): ...new function.
1478 (pass_vrp): Replace declaration with that of...
1479 (make_pass_vrp): ...new function.
1480 (pass_uncprop): Replace declaration with that of...
1481 (make_pass_uncprop): ...new function.
1482 (pass_return_slot): Replace declaration with that of...
1483 (make_pass_return_slot): ...new function.
1484 (pass_reassoc): Replace declaration with that of...
1485 (make_pass_reassoc): ...new function.
1486 (pass_rebuild_cgraph_edges): Replace declaration with that of...
1487 (make_pass_rebuild_cgraph_edges): ...new function.
1488 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
1489 (make_pass_remove_cgraph_callee_edges): ...new function.
1490 (pass_build_cgraph_edges): Replace declaration with that of...
1491 (make_pass_build_cgraph_edges): ...new function.
1492 (pass_local_pure_const): Replace declaration with that of...
1493 (make_pass_local_pure_const): ...new function.
1494 (pass_tracer): Replace declaration with that of...
1495 (make_pass_tracer): ...new function.
1496 (pass_warn_unused_result): Replace declaration with that of...
1497 (make_pass_warn_unused_result): ...new function.
1498 (pass_diagnose_tm_blocks): Replace declaration with that of...
1499 (make_pass_diagnose_tm_blocks): ...new function.
1500 (pass_lower_tm): Replace declaration with that of...
1501 (make_pass_lower_tm): ...new function.
1502 (pass_tm_init): Replace declaration with that of...
1503 (make_pass_tm_init): ...new function.
1504 (pass_tm_mark): Replace declaration with that of...
1505 (make_pass_tm_mark): ...new function.
1506 (pass_tm_memopt): Replace declaration with that of...
1507 (make_pass_tm_memopt): ...new function.
1508 (pass_tm_edges): Replace declaration with that of...
1509 (make_pass_tm_edges): ...new function.
1510 (pass_split_functions): Replace declaration with that of...
1511 (make_pass_split_functions): ...new function.
1512 (pass_feedback_split_functions): Replace declaration with that of...
1513 (make_pass_feedback_split_functions): ...new function.
1514 (pass_strength_reduction): Replace declaration with that of...
1515 (make_pass_strength_reduction): ...new function.
1516 (pass_ipa_lower_emutls): Replace declaration with that of...
1517 (make_pass_ipa_lower_emutls): ...new function.
1518 (pass_ipa_function_and_variable_visibility): Replace declaration with
1520 (make_pass_ipa_function_and_variable_visibility): ...new function.
1521 (pass_ipa_tree_profile): Replace declaration with that of...
1522 (make_pass_ipa_tree_profile): ...new function.
1523 (pass_early_local_passes): Replace declaration with that of...
1524 (make_pass_early_local_passes): ...new function.
1525 (pass_ipa_whole_program_visibility): Replace declaration with that
1527 (make_pass_ipa_whole_program_visibility): ...new function.
1528 (pass_ipa_lto_gimple_out): Replace declaration with that of...
1529 (make_pass_ipa_lto_gimple_out): ...new function.
1530 (pass_ipa_increase_alignment): Replace declaration with that of...
1531 (make_pass_ipa_increase_alignment): ...new function.
1532 (pass_ipa_inline): Replace declaration with that of...
1533 (make_pass_ipa_inline): ...new function.
1534 (pass_ipa_free_lang_data): Replace declaration with that of...
1535 (make_pass_ipa_free_lang_data): ...new function.
1536 (pass_ipa_free_inline_summary): Replace declaration with that of...
1537 (make_pass_ipa_free_inline_summary): ...new function.
1538 (pass_ipa_cp): Replace declaration with that of...
1539 (make_pass_ipa_cp): ...new function.
1540 (pass_ipa_reference): Replace declaration with that of...
1541 (make_pass_ipa_reference): ...new function.
1542 (pass_ipa_pure_const): Replace declaration with that of...
1543 (make_pass_ipa_pure_const): ...new function.
1544 (pass_ipa_pta): Replace declaration with that of...
1545 (make_pass_ipa_pta): ...new function.
1546 (pass_ipa_lto_finish_out): Replace declaration with that of...
1547 (make_pass_ipa_lto_finish_out): ...new function.
1548 (pass_ipa_tm): Replace declaration with that of...
1549 (make_pass_ipa_tm): ...new function.
1550 (pass_ipa_profile): Replace declaration with that of...
1551 (make_pass_ipa_profile): ...new function.
1552 (pass_ipa_cdtor_merge): Replace declaration with that of...
1553 (make_pass_ipa_cdtor_merge): ...new function.
1554 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
1556 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1557 (pass_init_datastructures): Replace declaration with that of...
1558 (make_pass_init_datastructures): ...new function.
1559 (pass_fixup_cfg): Replace declaration with that of...
1560 (make_pass_fixup_cfg): ...new function.
1561 (pass_expand): Replace declaration with that of...
1562 (make_pass_expand): ...new function.
1563 (pass_instantiate_virtual_regs): Replace declaration with that of...
1564 (make_pass_instantiate_virtual_regs): ...new function.
1565 (pass_rtl_fwprop): Replace declaration with that of...
1566 (make_pass_rtl_fwprop): ...new function.
1567 (pass_rtl_fwprop_addr): Replace declaration with that of...
1568 (make_pass_rtl_fwprop_addr): ...new function.
1569 (pass_jump): Replace declaration with that of...
1570 (make_pass_jump): ...new function.
1571 (pass_jump2): Replace declaration with that of...
1572 (make_pass_jump2): ...new function.
1573 (pass_lower_subreg): Replace declaration with that of...
1574 (make_pass_lower_subreg): ...new function.
1575 (pass_cse): Replace declaration with that of...
1576 (make_pass_cse): ...new function.
1577 (pass_fast_rtl_dce): Replace declaration with that of...
1578 (make_pass_fast_rtl_dce): ...new function.
1579 (pass_ud_rtl_dce): Replace declaration with that of...
1580 (make_pass_ud_rtl_dce): ...new function.
1581 (pass_rtl_dce): Replace declaration with that of...
1582 (make_pass_rtl_dce): ...new function.
1583 (pass_rtl_dse1): Replace declaration with that of...
1584 (make_pass_rtl_dse1): ...new function.
1585 (pass_rtl_dse2): Replace declaration with that of...
1586 (make_pass_rtl_dse2): ...new function.
1587 (pass_rtl_dse3): Replace declaration with that of...
1588 (make_pass_rtl_dse3): ...new function.
1589 (pass_rtl_cprop): Replace declaration with that of...
1590 (make_pass_rtl_cprop): ...new function.
1591 (pass_rtl_pre): Replace declaration with that of...
1592 (make_pass_rtl_pre): ...new function.
1593 (pass_rtl_hoist): Replace declaration with that of...
1594 (make_pass_rtl_hoist): ...new function.
1595 (pass_rtl_store_motion): Replace declaration with that of...
1596 (make_pass_rtl_store_motion): ...new function.
1597 (pass_cse_after_global_opts): Replace declaration with that of...
1598 (make_pass_cse_after_global_opts): ...new function.
1599 (pass_rtl_ifcvt): Replace declaration with that of...
1600 (make_pass_rtl_ifcvt): ...new function.
1601 (pass_into_cfg_layout_mode): Replace declaration with that of...
1602 (make_pass_into_cfg_layout_mode): ...new function.
1603 (pass_outof_cfg_layout_mode): Replace declaration with that of...
1604 (make_pass_outof_cfg_layout_mode): ...new function.
1605 (pass_loop2): Replace declaration with that of...
1606 (make_pass_loop2): ...new function.
1607 (pass_rtl_loop_init): Replace declaration with that of...
1608 (make_pass_rtl_loop_init): ...new function.
1609 (pass_rtl_move_loop_invariants): Replace declaration with that of...
1610 (make_pass_rtl_move_loop_invariants): ...new function.
1611 (pass_rtl_unswitch): Replace declaration with that of...
1612 (make_pass_rtl_unswitch): ...new function.
1613 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
1614 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1615 (pass_rtl_doloop): Replace declaration with that of...
1616 (make_pass_rtl_doloop): ...new function.
1617 (pass_rtl_loop_done): Replace declaration with that of...
1618 (make_pass_rtl_loop_done): ...new function.
1619 (pass_web): Replace declaration with that of...
1620 (make_pass_web): ...new function.
1621 (pass_cse2): Replace declaration with that of...
1622 (make_pass_cse2): ...new function.
1623 (pass_df_initialize_opt): Replace declaration with that of...
1624 (make_pass_df_initialize_opt): ...new function.
1625 (pass_df_initialize_no_opt): Replace declaration with that of...
1626 (make_pass_df_initialize_no_opt): ...new function.
1627 (pass_reginfo_init): Replace declaration with that of...
1628 (make_pass_reginfo_init): ...new function.
1629 (pass_inc_dec): Replace declaration with that of...
1630 (make_pass_inc_dec): ...new function.
1631 (pass_stack_ptr_mod): Replace declaration with that of...
1632 (make_pass_stack_ptr_mod): ...new function.
1633 (pass_initialize_regs): Replace declaration with that of...
1634 (make_pass_initialize_regs): ...new function.
1635 (pass_combine): Replace declaration with that of...
1636 (make_pass_combine): ...new function.
1637 (pass_if_after_combine): Replace declaration with that of...
1638 (make_pass_if_after_combine): ...new function.
1639 (pass_ree): Replace declaration with that of...
1640 (make_pass_ree): ...new function.
1641 (pass_partition_blocks): Replace declaration with that of...
1642 (make_pass_partition_blocks): ...new function.
1643 (pass_match_asm_constraints): Replace declaration with that of...
1644 (make_pass_match_asm_constraints): ...new function.
1645 (pass_regmove): Replace declaration with that of...
1646 (make_pass_regmove): ...new function.
1647 (pass_split_all_insns): Replace declaration with that of...
1648 (make_pass_split_all_insns): ...new function.
1649 (pass_fast_rtl_byte_dce): Replace declaration with that of...
1650 (make_pass_fast_rtl_byte_dce): ...new function.
1651 (pass_lower_subreg2): Replace declaration with that of...
1652 (make_pass_lower_subreg2): ...new function.
1653 (pass_mode_switching): Replace declaration with that of...
1654 (make_pass_mode_switching): ...new function.
1655 (pass_sms): Replace declaration with that of...
1656 (make_pass_sms): ...new function.
1657 (pass_sched): Replace declaration with that of...
1658 (make_pass_sched): ...new function.
1659 (pass_ira): Replace declaration with that of...
1660 (make_pass_ira): ...new function.
1661 (pass_reload): Replace declaration with that of...
1662 (make_pass_reload): ...new function.
1663 (pass_clean_state): Replace declaration with that of...
1664 (make_pass_clean_state): ...new function.
1665 (pass_branch_prob): Replace declaration with that of...
1666 (make_pass_branch_prob): ...new function.
1667 (pass_value_profile_transformations): Replace declaration with that
1669 (make_pass_value_profile_transformations): ...new function.
1670 (pass_postreload_cse): Replace declaration with that of...
1671 (make_pass_postreload_cse): ...new function.
1672 (pass_gcse2): Replace declaration with that of...
1673 (make_pass_gcse2): ...new function.
1674 (pass_split_after_reload): Replace declaration with that of...
1675 (make_pass_split_after_reload): ...new function.
1676 (pass_branch_target_load_optimize1): Replace declaration with that
1678 (make_pass_branch_target_load_optimize1): ...new function.
1679 (pass_thread_prologue_and_epilogue): Replace declaration with that
1681 (make_pass_thread_prologue_and_epilogue): ...new function.
1682 (pass_stack_adjustments): Replace declaration with that of...
1683 (make_pass_stack_adjustments): ...new function.
1684 (pass_peephole2): Replace declaration with that of...
1685 (make_pass_peephole2): ...new function.
1686 (pass_if_after_reload): Replace declaration with that of...
1687 (make_pass_if_after_reload): ...new function.
1688 (pass_regrename): Replace declaration with that of...
1689 (make_pass_regrename): ...new function.
1690 (pass_cprop_hardreg): Replace declaration with that of...
1691 (make_pass_cprop_hardreg): ...new function.
1692 (pass_reorder_blocks): Replace declaration with that of...
1693 (make_pass_reorder_blocks): ...new function.
1694 (pass_branch_target_load_optimize2): Replace declaration with that
1696 (make_pass_branch_target_load_optimize2): ...new function.
1697 (pass_leaf_regs): Replace declaration with that of...
1698 (make_pass_leaf_regs): ...new function.
1699 (pass_split_before_sched2): Replace declaration with that of...
1700 (make_pass_split_before_sched2): ...new function.
1701 (pass_compare_elim_after_reload): Replace declaration with that of...
1702 (make_pass_compare_elim_after_reload): ...new function.
1703 (pass_sched2): Replace declaration with that of...
1704 (make_pass_sched2): ...new function.
1705 (pass_stack_regs): Replace declaration with that of...
1706 (make_pass_stack_regs): ...new function.
1707 (pass_stack_regs_run): Replace declaration with that of...
1708 (make_pass_stack_regs_run): ...new function.
1709 (pass_df_finish): Replace declaration with that of...
1710 (make_pass_df_finish): ...new function.
1711 (pass_compute_alignments): Replace declaration with that of...
1712 (make_pass_compute_alignments): ...new function.
1713 (pass_duplicate_computed_gotos): Replace declaration with that of...
1714 (make_pass_duplicate_computed_gotos): ...new function.
1715 (pass_variable_tracking): Replace declaration with that of...
1716 (make_pass_variable_tracking): ...new function.
1717 (pass_free_cfg): Replace declaration with that of...
1718 (make_pass_free_cfg): ...new function.
1719 (pass_machine_reorg): Replace declaration with that of...
1720 (make_pass_machine_reorg): ...new function.
1721 (pass_cleanup_barriers): Replace declaration with that of...
1722 (make_pass_cleanup_barriers): ...new function.
1723 (pass_delay_slots): Replace declaration with that of...
1724 (make_pass_delay_slots): ...new function.
1725 (pass_split_for_shorten_branches): Replace declaration with that of...
1726 (make_pass_split_for_shorten_branches): ...new function.
1727 (pass_split_before_regstack): Replace declaration with that of...
1728 (make_pass_split_before_regstack): ...new function.
1729 (pass_convert_to_eh_region_ranges): Replace declaration with that
1731 (make_pass_convert_to_eh_region_ranges): ...new function.
1732 (pass_shorten_branches): Replace declaration with that of...
1733 (make_pass_shorten_branches): ...new function.
1734 (pass_set_nothrow_function_flags): Replace declaration with that of...
1735 (make_pass_set_nothrow_function_flags): ...new function.
1736 (pass_dwarf2_frame): Replace declaration with that of...
1737 (make_pass_dwarf2_frame): ...new function.
1738 (pass_final): Replace declaration with that of...
1739 (make_pass_final): ...new function.
1740 (pass_rtl_seqabstr): Replace declaration with that of...
1741 (make_pass_rtl_seqabstr): ...new function.
1742 (pass_release_ssa_names): Replace declaration with that of...
1743 (make_pass_release_ssa_names): ...new function.
1744 (pass_early_inline): Replace declaration with that of...
1745 (make_pass_early_inline): ...new function.
1746 (pass_inline_parameters): Replace declaration with that of...
1747 (make_pass_inline_parameters): ...new function.
1748 (pass_update_address_taken): Replace declaration with that of...
1749 (make_pass_update_address_taken): ...new function.
1750 (pass_convert_switch): Replace declaration with that of...
1751 (make_pass_convert_switch): ...new function.
1752 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
1753 to a subclass of simple_ipa_opt_pass along with...
1754 (pass_data_ipa_tree_profile): ...new pass_data instance and...
1755 (make_pass_ipa_tree_profile): ...new function.
1756 * tree-sra.c (pass_sra_early): Convert from a global struct to a
1757 subclass of gimple_opt_pass along with...
1758 (pass_data_sra_early): ...new pass_data instance and...
1759 (make_pass_sra_early): ...new function.
1760 (pass_sra): Convert from a global struct to a subclass of
1761 gimple_opt_pass along with...
1762 (pass_data_sra): ...new pass_data instance and...
1763 (make_pass_sra): ...new function.
1764 (pass_early_ipa_sra): Convert from a global struct to a subclass of
1765 gimple_opt_pass along with...
1766 (pass_data_early_ipa_sra): ...new pass_data instance and...
1767 (make_pass_early_ipa_sra): ...new function.
1768 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
1769 subclass of gimple_opt_pass along with...
1770 (pass_data_ccp): ...new pass_data instance and...
1771 (make_pass_ccp): ...new function.
1772 (pass_fold_builtins): Convert from a global struct to a subclass of
1773 gimple_opt_pass along with...
1774 (pass_data_fold_builtins): ...new pass_data instance and...
1775 (make_pass_fold_builtins): ...new function.
1776 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
1777 subclass of gimple_opt_pass along with...
1778 (pass_data_copy_prop): ...new pass_data instance and...
1779 (make_pass_copy_prop): ...new function.
1780 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
1781 global struct to a subclass of gimple_opt_pass along with...
1782 (pass_data_rename_ssa_copies): ...new pass_data instance and...
1783 (make_pass_rename_ssa_copies): ...new function.
1784 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
1785 subclass of gimple_opt_pass along with...
1786 (pass_data_dce): ...new pass_data instance and...
1787 (make_pass_dce): ...new function.
1788 (pass_dce_loop): Convert from a global struct to a subclass of
1789 gimple_opt_pass along with...
1790 (pass_data_dce_loop): ...new pass_data instance and...
1791 (make_pass_dce_loop): ...new function.
1792 (pass_cd_dce): Convert from a global struct to a subclass of
1793 gimple_opt_pass along with...
1794 (pass_data_cd_dce): ...new pass_data instance and...
1795 (make_pass_cd_dce): ...new function.
1796 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
1797 subclass of gimple_opt_pass along with...
1798 (pass_data_dominator): ...new pass_data instance and...
1799 (make_pass_dominator): ...new function.
1800 (pass_phi_only_cprop): Convert from a global struct to a subclass of
1801 gimple_opt_pass along with...
1802 (pass_data_phi_only_cprop): ...new pass_data instance and...
1803 (make_pass_phi_only_cprop): ...new function.
1804 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
1805 subclass of gimple_opt_pass along with...
1806 (pass_data_dse): ...new pass_data instance and...
1807 (make_pass_dse): ...new function.
1808 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
1809 a subclass of gimple_opt_pass along with...
1810 (pass_data_forwprop): ...new pass_data instance and...
1811 (make_pass_forwprop): ...new function.
1812 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
1813 struct to a subclass of gimple_opt_pass along with...
1814 (pass_data_tree_ifcombine): ...new pass_data instance and...
1815 (make_pass_tree_ifcombine): ...new function.
1816 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
1817 subclass of gimple_opt_pass along with...
1818 (pass_data_ch): ...new pass_data instance and...
1819 (make_pass_ch): ...new function.
1820 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
1821 subclass of gimple_opt_pass along with...
1822 (pass_data_tree_loop): ...new pass_data instance and...
1823 (make_pass_tree_loop): ...new function.
1824 (pass_tree_loop_init): Convert from a global struct to a subclass of
1825 gimple_opt_pass along with...
1826 (pass_data_tree_loop_init): ...new pass_data instance and...
1827 (make_pass_tree_loop_init): ...new function.
1828 (pass_lim): Convert from a global struct to a subclass of
1829 gimple_opt_pass along with...
1830 (pass_data_lim): ...new pass_data instance and...
1831 (make_pass_lim): ...new function.
1832 (pass_tree_unswitch): Convert from a global struct to a subclass of
1833 gimple_opt_pass along with...
1834 (pass_data_tree_unswitch): ...new pass_data instance and...
1835 (make_pass_tree_unswitch): ...new function.
1836 (pass_predcom): Convert from a global struct to a subclass of
1837 gimple_opt_pass along with...
1838 (pass_data_predcom): ...new pass_data instance and...
1839 (make_pass_predcom): ...new function.
1840 (pass_vectorize): Convert from a global struct to a subclass of
1841 gimple_opt_pass along with...
1842 (pass_data_vectorize): ...new pass_data instance and...
1843 (make_pass_vectorize): ...new function.
1844 (pass_graphite): Convert from a global struct to a subclass of
1845 gimple_opt_pass along with...
1846 (pass_data_graphite): ...new pass_data instance and...
1847 (make_pass_graphite): ...new function.
1848 (pass_graphite_transforms): Convert from a global struct to a subclass
1849 of gimple_opt_pass along with...
1850 (pass_data_graphite_transforms): ...new pass_data instance and...
1851 (make_pass_graphite_transforms): ...new function.
1852 (pass_check_data_deps): Convert from a global struct to a subclass of
1853 gimple_opt_pass along with...
1854 (pass_data_check_data_deps): ...new pass_data instance and...
1855 (make_pass_check_data_deps): ...new function.
1856 (pass_iv_canon): Convert from a global struct to a subclass of
1857 gimple_opt_pass along with...
1858 (pass_data_iv_canon): ...new pass_data instance and...
1859 (make_pass_iv_canon): ...new function.
1860 (pass_scev_cprop): Convert from a global struct to a subclass of
1861 gimple_opt_pass along with...
1862 (pass_data_scev_cprop): ...new pass_data instance and...
1863 (make_pass_scev_cprop): ...new function.
1864 (pass_record_bounds): Convert from a global struct to a subclass of
1865 gimple_opt_pass along with...
1866 (pass_data_record_bounds): ...new pass_data instance and...
1867 (make_pass_record_bounds): ...new function.
1868 (pass_complete_unroll): Convert from a global struct to a subclass of
1869 gimple_opt_pass along with...
1870 (pass_data_complete_unroll): ...new pass_data instance and...
1871 (make_pass_complete_unroll): ...new function.
1872 (pass_complete_unrolli): Convert from a global struct to a subclass of
1873 gimple_opt_pass along with...
1874 (pass_data_complete_unrolli): ...new pass_data instance and...
1875 (make_pass_complete_unrolli): ...new function.
1876 (pass_parallelize_loops): Convert from a global struct to a subclass
1877 of gimple_opt_pass along with...
1878 (pass_data_parallelize_loops): ...new pass_data instance and...
1879 (make_pass_parallelize_loops): ...new function.
1880 (pass_loop_prefetch): Convert from a global struct to a subclass of
1881 gimple_opt_pass along with...
1882 (pass_data_loop_prefetch): ...new pass_data instance and...
1883 (make_pass_loop_prefetch): ...new function.
1884 (pass_iv_optimize): Convert from a global struct to a subclass of
1885 gimple_opt_pass along with...
1886 (pass_data_iv_optimize): ...new pass_data instance and...
1887 (make_pass_iv_optimize): ...new function.
1888 (pass_tree_loop_done): Convert from a global struct to a subclass of
1889 gimple_opt_pass along with...
1890 (pass_data_tree_loop_done): ...new pass_data instance and...
1891 (make_pass_tree_loop_done): ...new function.
1892 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
1893 struct to a subclass of gimple_opt_pass along with...
1894 (pass_data_cse_reciprocals): ...new pass_data instance and...
1895 (make_pass_cse_reciprocals): ...new function.
1896 (pass_cse_sincos): Convert from a global struct to a subclass of
1897 gimple_opt_pass along with...
1898 (pass_data_cse_sincos): ...new pass_data instance and...
1899 (make_pass_cse_sincos): ...new function.
1900 (pass_optimize_bswap): Convert from a global struct to a subclass of
1901 gimple_opt_pass along with...
1902 (pass_data_optimize_bswap): ...new pass_data instance and...
1903 (make_pass_optimize_bswap): ...new function.
1904 (pass_optimize_widening_mul): Convert from a global struct to a
1905 subclass of gimple_opt_pass along with...
1906 (pass_data_optimize_widening_mul): ...new pass_data instance and...
1907 (make_pass_optimize_widening_mul): ...new function.
1908 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
1909 subclass of gimple_opt_pass along with...
1910 (pass_data_phiopt): ...new pass_data instance and...
1911 (make_pass_phiopt): ...new function.
1912 (pass_cselim): Convert from a global struct to a subclass of
1913 gimple_opt_pass along with...
1914 (pass_data_cselim): ...new pass_data instance and...
1915 (make_pass_cselim): ...new function.
1916 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
1917 subclass of gimple_opt_pass along with...
1918 (pass_data_phiprop): ...new pass_data instance and...
1919 (make_pass_phiprop): ...new function.
1920 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
1921 subclass of gimple_opt_pass along with...
1922 (pass_data_pre): ...new pass_data instance and...
1923 (make_pass_pre): ...new function.
1924 (pass_fre): Convert from a global struct to a subclass of
1925 gimple_opt_pass along with...
1926 (pass_data_fre): ...new pass_data instance and...
1927 (make_pass_fre): ...new function.
1928 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
1929 subclass of gimple_opt_pass along with...
1930 (pass_data_reassoc): ...new pass_data instance and...
1931 (make_pass_reassoc): ...new function.
1932 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
1933 subclass of gimple_opt_pass along with...
1934 (pass_data_sink_code): ...new pass_data instance and...
1935 (make_pass_sink_code): ...new function.
1936 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
1937 subclass of gimple_opt_pass along with...
1938 (pass_data_strlen): ...new pass_data instance and...
1939 (make_pass_strlen): ...new function.
1940 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
1941 struct to a subclass of gimple_opt_pass along with...
1942 (pass_data_build_alias): ...new pass_data instance and...
1943 (make_pass_build_alias): ...new function.
1944 (pass_build_ealias): Convert from a global struct to a subclass of
1945 gimple_opt_pass along with...
1946 (pass_data_build_ealias): ...new pass_data instance and...
1947 (make_pass_build_ealias): ...new function.
1948 (pass_ipa_pta): Convert from a global struct to a subclass of
1949 simple_ipa_opt_pass along with...
1950 (pass_data_ipa_pta): ...new pass_data instance and...
1951 (make_pass_ipa_pta): ...new function.
1952 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
1953 subclass of gimple_opt_pass along with...
1954 (pass_data_uncprop): ...new pass_data instance and...
1955 (make_pass_uncprop): ...new function.
1956 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
1957 global struct to a subclass of gimple_opt_pass along with...
1958 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
1959 (make_pass_late_warn_uninitialized): ...new function.
1960 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
1961 to a subclass of gimple_opt_pass along with...
1962 (pass_data_init_datastructures): ...new pass_data instance and...
1963 (make_pass_init_datastructures): ...new function.
1964 (pass_early_warn_uninitialized): Convert from a global struct to a
1965 subclass of gimple_opt_pass along with...
1966 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
1967 (make_pass_early_warn_uninitialized): ...new function.
1968 (pass_update_address_taken): Convert from a global struct to a
1969 subclass of gimple_opt_pass along with...
1970 (pass_data_update_address_taken): ...new pass_data instance and...
1971 (make_pass_update_address_taken): ...new function.
1972 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
1973 struct to a subclass of gimple_opt_pass along with...
1974 (pass_data_release_ssa_names): ...new pass_data instance and...
1975 (make_pass_release_ssa_names): ...new function.
1976 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
1977 subclass of gimple_opt_pass along with...
1978 (pass_data_stdarg): ...new pass_data instance and...
1979 (make_pass_stdarg): ...new function.
1980 * tree-switch-conversion.c (pass_convert_switch): Convert from a
1981 global struct to a subclass of gimple_opt_pass along with...
1982 (pass_data_convert_switch): ...new pass_data instance and...
1983 (make_pass_convert_switch): ...new function.
1984 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
1985 to a subclass of gimple_opt_pass along with...
1986 (pass_data_tail_recursion): ...new pass_data instance and...
1987 (make_pass_tail_recursion): ...new function.
1988 (pass_tail_calls): Convert from a global struct to a subclass of
1989 gimple_opt_pass along with...
1990 (pass_data_tail_calls): ...new pass_data instance and...
1991 (make_pass_tail_calls): ...new function.
1992 * tree-vect-generic.c (pass_lower_vector): Convert from a global
1993 struct to a subclass of gimple_opt_pass along with...
1994 (pass_data_lower_vector): ...new pass_data instance and...
1995 (make_pass_lower_vector): ...new function.
1996 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
1997 gimple_opt_pass along with...
1998 (pass_data_lower_vector_ssa): ...new pass_data instance and...
1999 (make_pass_lower_vector_ssa): ...new function.
2000 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2001 to a subclass of gimple_opt_pass along with...
2002 (pass_data_slp_vectorize): ...new pass_data instance and...
2003 (make_pass_slp_vectorize): ...new function.
2004 (pass_ipa_increase_alignment): Convert from a global struct to a
2005 subclass of simple_ipa_opt_pass along with...
2006 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2007 (make_pass_ipa_increase_alignment): ...new function.
2008 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2009 gimple_opt_pass along with...
2010 (pass_data_vrp): ...new pass_data instance and...
2011 (make_pass_vrp): ...new function.
2012 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2013 subclass of simple_ipa_opt_pass along with...
2014 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2015 (make_pass_ipa_free_lang_data): ...new function.
2016 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2017 gimple_opt_pass along with...
2018 (pass_data_tsan): ...new pass_data instance and...
2019 (make_pass_tsan): ...new function.
2020 (pass_tsan_O0): Convert from a global struct to a subclass of
2021 gimple_opt_pass along with...
2022 (pass_data_tsan_O0): ...new pass_data instance and...
2023 (make_pass_tsan_O0): ...new function.
2024 * var-tracking.c (pass_variable_tracking): Convert from a global
2025 struct to a subclass of rtl_opt_pass along with...
2026 (pass_data_variable_tracking): ...new pass_data instance and...
2027 (make_pass_variable_tracking): ...new function.
2028 * web.c (pass_web): Convert from a global struct to a subclass of
2029 rtl_opt_pass along with...
2030 (pass_data_web): ...new pass_data instance and...
2031 (make_pass_web): ...new function.
2032 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2033 declaration with that of...
2034 (make_pass_mode_switch_use): ...new function.
2035 (pass_resolve_sw_modes): Replace declaration with that of...
2036 (make_pass_resolve_sw_modes): ...new function.
2037 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2038 from a global struct to a subclass of rtl_opt_pass along with...
2039 (pass_data_mode_switch_use): ...new pass_data instance and...
2040 (make_pass_mode_switch_use): ...new function.
2041 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2042 from a global struct to a subclass of rtl_opt_pass along with...
2043 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2044 (make_pass_resolve_sw_modes): ...new function.
2045 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2046 struct to a subclass of rtl_opt_pass along with...
2047 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2048 (make_pass_insert_vzeroupper): ...new function.
2049 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2050 global struct to a subclass of rtl_opt_pass along with...
2051 (pass_data_work_around_errata): ...new pass_data instance and...
2052 (make_pass_work_around_errata): ...new function.
2053 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2054 struct to a subclass of rtl_opt_pass along with...
2055 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2056 (make_pass_mips_machine_reorg2): ...new function.
2058 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2060 * passes.c (pass_manager::operator new): New.
2062 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2064 Handwritten part of conversion of passes to C++ classes.
2066 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2067 (toplev.o): Add dep on PASS_MANAGER_H.
2068 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2069 of early local pases to reflect this moving from a global to a
2070 member of gcc::pass_manager.
2071 (cgraph_add_new_function): Likewise.
2072 * lto-cgraph.c (lto_output_node): Update for conversion of
2073 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2074 * passes.c (opt_pass::clone): New.
2075 (opt_pass::gate): New.
2076 (opt_pass::execute): New.
2077 (opt_pass::opt_pass): New.
2078 (pass_manager::execute_early_local_passes): New.
2079 (pass_manager::execute_pass_mode_switching): new.
2080 (finish_optimization_passes): Convert to...
2081 (pass_manager::finish_optimization_passes): ...this.
2082 (finish_optimization_passes): Update for conversion of passes to
2084 (register_dump_files_1): Use has_gate since we cannot portably
2085 check a vtable entry against NULL.
2086 (dump_one_pass): Likewise.
2087 (ipa_write_summaries_2): Likewise.
2088 (ipa_write_optimization_summaries_1): Likewise.
2089 (ipa_read_summaries_1): Likewise.
2090 (ipa_read_optimization_summaries_1): Likewise.
2091 (execute_ipa_stmt_fixups): Likewise.
2092 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2093 pass-creation functions rather than wiring up globals, and
2094 storing the results in fields of pass_manager generated using
2096 (pass_manager::dump_profile_report): Update for conversion of
2097 passes to C++ classes.
2098 (pass_manager::execute_ipa_summary_passes): Likewise.
2099 (execute_one_ipa_transform_pass): Likewise.
2100 (execute_one_pass): Use has_gate and has_execute since we cannot
2101 portably check a vtable entry against NULL.
2102 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2103 (pass_manager): Use pass-instances.def to add fields for the
2104 various pass instances.
2105 * toplev.c (finalize): Update for move of
2106 finish_optimization_passes to a method of gcc::pass_manager.
2107 * toplev.h (finish_optimization_passes): Move to method of class
2109 * tree-pass.h (struct pass_data): New.
2110 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2111 (opt_pass::gate): Convert to virtual function.
2112 (opt_pass::~opt_pass): New.
2113 (opt_pass::clone): New.
2114 (opt_pass::execute): Convert to virtual function.
2115 (opt_pass::opt_pass): New.
2116 (opt_pass::ctxt_): new.
2117 (gimple_opt_pass): Convert to subclass of opt_pass.
2118 (gimple_opt_pass::gimple_opt_pass): New.
2119 (rtl_opt_pass): Convert to subclass of opt_pass.
2120 (rtl_opt_pass::rtl_opt_pass): New.
2121 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2122 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2123 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2124 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2125 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2126 invocation of pass_mode_switching to reflect this moving from a
2127 global to a member of gcc::pass_manager.
2128 (ix86_option_override): Rework how pass_insert_vzeroupper is
2129 added to the pass_manager to reflect autogenerated changes.
2130 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2132 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2134 PR rtl-optimization/57708
2135 * recog.c (peep2_find_free_register): Validate all regs in a
2138 2013-08-05 Jan Hubicka <jh@suse.cz>
2141 * cgraph.c (verify_cgraph_node): Accept local flags from other
2143 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2144 (function_and_variable_visibility): Likewise.
2145 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2147 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2149 * graph.c (init_graph_slim_pretty_print): Remove.
2150 (print_graph_cfg): Do not call it. Use local pretty printer.
2151 (start_graph_dump): Likewise.
2153 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2155 * gimple-pretty-print.c (buffer): Remove.
2156 (initialized): Likewise.
2157 (maybe_init_pretty_print): Likewise.
2158 (print_gimple_stmt): Do not call it. Use non-static local
2159 pretty_printer variable.
2160 (print_gimple_expr): Likewise.
2161 (print_gimple_seq): Likewise.
2162 (gimple_dump_bb): Likewise.
2164 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2166 * asan.c (asan_pp): Remove.
2167 (asan_pp_initialized): Likewise.
2168 (asan_pp_initialize): Likewise.
2169 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2170 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2171 (asan_add_global): Likewise.
2173 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2175 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2176 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2177 * pretty-print.c (pp_formatted_text_data): Likewise.
2178 (pp_indent): Rename from pp_base_indent.
2179 (pp_format): Rename from pp_base_format.
2180 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2181 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2182 (pp_flush): Rename from pp_base_flush.
2183 (pp_set_line_maximum_length): Rename from
2184 pp_base_set_line_maximum_length.
2185 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2186 (pp_set_prefix): Rename from pp_base_set_prefix.
2187 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2188 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2189 (pp_append_text): Rename from pp_base_append_text.
2190 (pp_formatted_text): Rename from pp_base_formatted_text.
2191 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2192 (pp_remaining_character_count_for_line): Rename from
2193 pp_base_remaining_character_count_for_line.
2194 (pp_newline): Rename from pp_base_newline.
2195 (pp_character): Rename from pp_base_character.
2196 (pp_string): Rename from pp_base_string.
2197 (pp_maybe_space): Rename from pp_base_maybe_space.
2198 * asan.c (asan_pp_string): Adjust.
2199 (asan_emit_stack_protection): Likewise.
2200 (asan_add_global): Likewise.
2201 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2202 * tree-mudflap.c (mf_varname_tree): Likewise.
2203 * tree-pretty-print.c (pp_tree_identifier): Rename from
2204 pp_base_tree_identifier.
2205 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2206 Declare as function.
2208 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2210 * pretty-print.h (pp_bar_bar): New.
2211 (pp_ampersand_ampersand): Likewise.
2212 (pp_less_equal): Likewise.
2213 (pp_greater_equal): Likewise.
2214 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2215 printer functions instead of pp_string or operators and punctuators.
2216 (dump_gimple_call): Likewise.
2217 (dump_gimple_omp_for): Likewise.
2218 (dump_gimple_transaction): Likewise.
2219 (dump_gimple_phi): Likewise.
2220 (pp_gimple_stmt_1): Likewise.
2221 * sched-vis.c (print_insn): Likewise.
2222 * tree-mudflap.c (mf_varname_tree): Likewise.
2223 * tree-pretty-print.c (dump_block_node): Likewise.
2224 (dump_generic_node): Likewise.
2226 2013-08-02 Jan Hubicka <jh@suse.cz>
2228 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2230 * lto-streamer-out.c (tree_is_indexable): Results decls and
2231 parm decls are not indexable.
2232 (DFS_write_tree_body): Do not follow args and results.
2233 (hash_tree): Likewise.
2234 (output_functions): Rearrange so struct function is needed
2235 only when real body is output; be able to also ouptut abstract
2236 functions; output DECL_ARGUMENTS and DECL_RESULT.
2237 (lto_output): When not in WPA, ale store abstract functions.
2238 (write_symbol): Do not care about RESULT_DECL.
2239 (output_symbol_p): Handle correctly sbtract decls.
2240 * lto-streamer-in.c (input_function): Rearrange so struct
2241 function can be NULL at entry; allow streaming of
2242 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2243 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2244 sanity check during LTO.
2245 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2246 RESULT_DECl and DECL_ARGUMENTS.
2247 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2250 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2252 * pretty-print.h (pp_underscore): New.
2254 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2255 printer functions instead of pp_character.
2256 (dump_binary_rhs): Likewise.
2257 (dump_ternary_rhs): Likewise.
2258 (dump_gimple_call_args): Likewise.
2259 (pp_points_to_solution): Likewise.
2260 (dump_gimple_call): Likewise.
2261 (dump_gimple_switch): Likewise.
2262 (dump_gimple_cond): Likewise.
2263 (dump_gimple_bind): Likewise.
2264 (dump_gimple_try): Likewise.
2265 (dump_gimple_omp_for): Likewise.
2266 (dump_gimple_omp_continue): Likewise.
2267 (dump_gimple_omp_single): Likewise.
2268 (dump_gimple_omp_sections): Likewise.
2269 (dump_gimple_omp_block): Likewise.
2270 (dump_gimple_omp_critical): Likewise.
2271 (dump_gimple_transaction): Likewise.
2272 (dump_gimple_asm): Likewise.
2273 (dump_gimple_phi): Likewise.
2274 (dump_gimple_omp_parallel): Likewise.
2275 (dump_gimple_omp_task): Likewise.
2276 (dump_gimple_omp_atomic_load): Likewise.
2277 (dump_gimple_omp_atomic_store): Likewise.
2278 (dump_gimple_mem_ops): Likewise.
2279 (pp_gimple_stmt_1): Likewise.
2280 (pp_cfg_jump): Likewise.
2281 (dump_implicit_edges): Likewise.
2282 (gimple_dump_bb_for_graph): Likewise.
2283 * graph.c (draw_cfg_node): Likewise.
2284 * langhooks.c (lhd_print_error_function): Likewise.
2285 * sched-vis.c (print_exp): Likewise.
2286 (print_value): Likewise.
2287 (print_pattern): Likewise.
2288 (print_insn): Likewise.
2289 (rtl_dump_bb_for_graph): Likewise.
2290 * tree-pretty-print.c (dump_function_declaration): Likewise.
2291 (dump_array_domain): Likewise.
2292 (dump_omp_clause): Likewise.
2293 (dump_location): Likewise.
2294 (dump_generic_node): Likewise.
2295 (print_struct_decl): Likewise.
2296 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2298 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2300 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2301 candidate table when replacing a candidate statement.
2302 (replace_rhs_if_not_dup): Likewise.
2303 (replace_one_candidate): Likewise.
2305 2013-08-02 Jan Hubicka <jh@suse.cz>
2306 Martin Liska <marxin.liska@gmail.com>
2308 * cgraphunit.c (add_new_function): Fix logic when adding from
2310 (assemble_thunk): Rename to ...
2311 (expand_thunk); .. this one; export; get it working with
2312 general functions; make produced gimple valid.
2313 * cgraph.h (expand_thunk): Declare.
2315 2013-08-02 Jan Hubicka <jh@suse.cz>
2317 * ipa-cp.c (gather_context_independent_values): Use
2318 ipa_get_param_move_cost.
2319 (get_replacement_map): Remove PARAM; move parameter folding
2321 (create_specialized_node): Update.
2322 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2323 assert that we have gimple body; update move_cost.
2324 (count_formal_params): Assert that we have gimple body.
2325 (ipa_dump_param): New function.
2326 (ipa_alloc_node_params): Break out from ...
2327 (ipa_initialize_node_params): ... here.
2328 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2329 (ipa_write_node_info): Stream move costs.
2330 (ipa_read_node_info): Read move costs.
2331 (ipa_update_after_lto_read): Do not recompute node params.
2332 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2333 (ipa_get_param): Check we are not in WPA.
2334 (ipa_get_param_move_cost): New.
2335 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2336 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2337 parm numbers to be present.
2339 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2341 PR rtl-optimization/58048
2342 * lra-constraints.c (process_alt_operands): Don't check asm
2343 operand on register.
2345 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2347 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2348 the implied StoreLoad barrier for atomic operations if before.
2350 2013-08-02 Jan Hubicka <jh@suse.cz>
2351 Martin Liska <marxin.liska@gmail.com>
2353 * cgraph.c (cgraph_function_body_availability): Do not check
2355 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2356 symtab_node_availability): Declare.
2357 * ipa.c (can_replace_by_local_alias): New.
2358 (function_and_variable_visibility): Use it.
2359 * symtab.c (symtab_for_node_and_aliases,
2360 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2362 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2364 PR rtl-optimization/57963
2365 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2366 (lra_constraints): Use them.
2368 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2370 * config/arm/types.md (define_attr "type"): Add "load_acq"
2372 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2374 (cortex_a53_store1): Likewise.
2376 2013-08-01 Jan Hubicka <jh@suse.cz>
2378 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
2379 partitions are not needed.
2381 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
2383 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
2384 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
2385 MAYBE_NON_Q_CLASS_P where appropriate.
2387 2013-08-01 Jan Hubicka <jh@suse.cz>
2389 * cgraph.h (release_function_body): Declare.
2390 * tree.c (free_lang_data_in_decl): Free, parameters and return values
2391 of unused delcarations.
2393 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2395 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
2396 RTL form when subtracting a constant.
2398 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2400 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
2401 Generate canonical plus rtx with negated immediate instead of minus
2403 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
2405 2013-08-01 Jan Hubicka <jh@suse.cz>
2407 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
2408 (cgraph_release_function_body): Likewise.
2409 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2410 * cgraph.h (cgrpah_node): Rename abstract_and_needed
2411 to used_as_abstract_origin.
2412 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
2413 symbols used as abstract origins.
2414 * cgraphunit.c (analyze_functions): Update.
2415 * ipa.c (symtab_remove_unreachable_nodes): Recompute
2416 used_as_abstract_origin.
2417 * tree-inline.c (tree_function_versioning): Update
2418 used_as_abstract_origin; be ready for DECL_RESULT and
2419 DECL_ARGUMENTS to be NULL.
2421 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
2422 for abstract functions.
2423 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
2426 2013-08-01 Jan Hubicka <jh@suse.cz>
2428 * profile.c (compute_value_histograms): Fix thinko.
2430 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2432 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
2433 aarch-common-protos.h to extra_headers.
2434 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2435 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
2436 * config/aarch64/t-aarch64 (aarch-common.o): Define.
2438 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
2440 * config/aarch64/aarch64.md (define_attr "type"): Delete.
2441 Include "../arm/types.md". Define "type" attribute for all patterns.
2442 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
2445 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
2447 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
2448 to support power8 load fusion.
2449 (fusion_gpr_mem_load): Likewise.
2451 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
2453 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
2454 declarations for power8 load fusion.
2455 (emit_fusion_gpr_load): Likewise.
2457 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
2458 tuning for power8, turn on fusion mode by default. Turn on sign
2459 extending fusion mode if normal fusion mode is on, and we are at
2461 (fusion_gpr_load_p): New function, return true if we can fuse an
2462 addis instruction with a dependent load to a GPR.
2463 (emit_fusion_gpr_load): Emit the instructions for power8 load
2466 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
2467 (VSX load fusion peepholes): New peepholes to fuse together an
2468 addi instruction with a VSX load instruction.
2470 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
2471 peepholes to fuse an addis instruction with a load to a GPR base
2472 register. If we are supporting sign extending fusions, convert
2473 sign extending loads to zero extending loads and add an explicit
2476 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2478 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
2479 aarch-common-protos.h to extra_headers.
2480 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
2481 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
2482 (arm_early_store_addr_dep): Likewise.
2483 (arm_no_early_alu_shift_dep): Likewise.
2484 (arm_no_early_alu_shift_value_dep): Likewise.
2485 (arm_no_early_mul_dep): Likewise.
2486 (arm_no_early_store_addr_dep): Likewise.
2487 (arm_mac_accumulator_is_mul_result): Likewise.
2488 (arm_mac_accumulator_is_result): Likewise.
2489 * config/arm/aarch-common.c: ... here. New file.
2490 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
2492 (arm_early_store_addr_dep): Likewise.
2493 (arm_no_early_alu_shift_dep): Likewise.
2494 (arm_no_early_alu_shift_value_dep): Likewise.
2495 (arm_no_early_mul_dep): Likewise.
2496 (arm_no_early_store_addr_dep): Likewise.
2497 (arm_mac_accumulator_is_mul_result): Likewise.
2498 (arm_mac_accumulator_is_result): Likewise.
2499 * config/arm/aarch-common-protos.h: ... here. New file.
2500 * config/arm/t-arm (aarch-common.o): Define.
2502 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
2504 * config/arm/arm.md: Include new file "types.md".
2505 (define_attr "type"): Move from here to ...
2506 (define_attr "mul32"): Likewise.
2507 (define_attr "mul64"): Likewise.
2508 * config/arm/types.md: ... here. New file.
2510 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
2512 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
2513 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
2515 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2517 * gen-pass-instances.awk: Fix offset of substr().
2519 2013-07-31 David Malcolm <dmalcolm@redhat.com>
2521 * Makefile.in (pass-instances.def): New.
2522 (passes.o): Replace dependency on passes.def with one on
2525 * gen-pass-instances.awk: New.
2527 * passes.c (pass_manager::pass_manager): Use pass-instances.def
2528 rather than passes.def, updating local definition of NEXT_PASS
2529 macro to add an extra NUM parameter (currently unused).
2531 2013-07-30 David Malcolm <dmalcolm@redhat.com>
2533 * Makefile.in (PASS_MANAGER_H): New.
2534 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
2535 (passes.o): Likewise.
2536 (statistics.o): Likewise.
2537 (cgraphunit.o): Likewise.
2538 (context.o): Depend on PASS_MANAGER_H.
2540 * pass_manager.h: New.
2542 * cgraphunit.c (cgraph_add_new_function): Update for moves
2543 of globals to fields of pass_manager.
2544 (analyze_function): Likewise.
2545 (expand_function): Likewise.
2546 (ipa_passes): Likewise.
2547 (compile): Likewise.
2549 * context.c (context::context): New.
2550 * context.h (context::context): New.
2551 (context::get_passes): New.
2552 (context::passes_): New.
2554 * lto-cgraph.c (input_node): Update for moves of globals to
2555 fields of pass_manager.
2557 * passes.c (all_passes): Remove, in favor of a field of the
2558 same name within the new class pass_manager.
2559 (all_small_ipa_passes): Likewise.
2560 (all_lowering_passes): Likewise.
2561 (all_regular_ipa_passes): Likewise.
2562 (all_late_ipa_passes): Likewise.
2563 (all_lto_gen_passes): Likewise.
2564 (passes_by_id): Likewise.
2565 (passes_by_id_size): Likewise.
2566 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2567 the new class pass_manager.
2568 (set_pass_for_id): Convert to...
2569 (pass_manager::set_pass_for_id): ...method.
2570 (get_pass_for_id): Convert to...
2571 (pass_manager::get_pass_for_id): ...method.
2572 (register_one_dump_file): Move body of implementation into...
2573 (pass_manager::register_one_dump_file): ...here.
2574 (register_dump_files_1): Convert to...
2575 (pass_manager::register_dump_files_1): ...method.
2576 (register_dump_files): Convert to...
2577 (pass_manager::register_dump_files): ...method.
2578 (create_pass_tab): Update for moves of globals to fields of
2580 (dump_passes): Move body of implementation into...
2581 (pass_manager::dump_passes): ...here.
2582 (register_pass): Move body of implementation into...
2583 (pass_manager::register_pass): ...here.
2584 (init_optimization_passes): Convert into...
2585 (pass_manager::pass_manager): ...constructor for new
2586 pass_manager class, and initialize the pass_lists array.
2587 (check_profile_consistency): Update for moves of globals to
2588 fields of pass_manager.
2589 (dump_profile_report): Move body of implementation into...
2590 (pass_manager::dump_profile_report): ...here.
2591 (ipa_write_summaries_1): Update for moves of pass lists from
2592 being globals to fields of pass_manager.
2593 (ipa_write_optimization_summaries): Likewise.
2594 (ipa_read_summaries): Likewise.
2595 (ipa_read_optimization_summaries): Likewise.
2596 (execute_all_ipa_stmt_fixups): Likewise.
2598 * statistics.c (statistics_fini): Update for moves of globals to
2599 fields of pass_manager.
2601 * toplev.c (general_init): Replace call to
2602 init_optimization_passes with construction of the pass_manager
2605 * tree-pass.h (all_passes): Remove, in favor of a field of the
2606 same name within the new class pass_manager.
2607 (all_small_ipa_passes): Likewise.
2608 (all_lowering_passes): Likewise.
2609 (all_regular_ipa_passes): Likewise.
2610 (all_lto_gen_passes): Likewise.
2611 (all_late_ipa_passes): Likewise.
2612 (passes_by_id): Likewise.
2613 (passes_by_id_size): Likewise.
2614 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
2615 the new class pass_manager.
2616 (get_pass_for_id): Remove.
2618 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2620 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
2623 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
2625 * arm.md (mulhi3): New expand pattern.
2627 2013-07-30 Jan Hubicka <jh@suse.cz>
2628 Martin Liska <marxin.liska@gmail.com>
2630 * profile.c (compute_value_histograms): Do not ICE when
2631 there is mismatch only on some counters.
2633 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2635 PR rtl-optimization/57637
2636 * function.c (move_insn_for_shrink_wrap): Also check the
2637 GEN set of the LIVE problem for the liveness analysis
2638 if it exists, otherwise give up.
2640 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2642 PR tree-optimization/57993
2643 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
2644 replaced statement in the candidate table.
2645 (phi_add_costs): Return infinite cost when the hidden basis does
2646 not dominate all phis on which the candidate is dependent.
2647 (replace_one_candidate): Record replaced statement in the
2650 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
2652 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
2653 (ashlv2si3): New expander.
2654 (*ashlv2si3_i): New define_insn_and_split.
2655 * predicates.md (float_operation): Allow patterns with three
2658 PR rtl-optimization/58021
2659 * mode-switching.c (create_pre_exit): Always split off preceding
2660 insns if we are not at the basic block head.
2662 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
2664 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
2665 (UCLIBC_DYNAMIC_LINKER): New macro.
2666 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
2668 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
2669 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
2671 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
2672 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
2673 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
2674 for SF and DF modes. Use ieee_quad_format for TF mode.
2675 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
2676 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
2677 (mips_option_override): Handle `-mnan=legacy'.
2678 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
2679 `-mabs=2008' and `-mnan=2008'.
2680 (OPTION_DEFAULT_SPECS): Add "nan" default.
2681 (ASM_SPEC): Handle `-mnan='.
2682 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
2683 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
2684 comment accordingly.
2685 (neg<mode>2): Likewise.
2686 * config/mips/mips.opt (mabs, mnan): New options.
2687 * doc/install.texi (Configuration): Document `--with-nan=' option.
2688 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
2690 (MIPS Options): Document them.
2691 * config.gcc <mips*-*-*>: Handle `--with-nan='.
2692 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
2693 * configure: Regenerate.
2694 * config.in: Regenerate.
2696 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2698 * config/i386/i386.md (float post-reload splitters): Do not check
2699 for subregs of SSE registers.
2701 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
2702 H.J. Lu <hongjiu.lu@intel.com>
2706 * config/i386/i386.md (post-reload splitter
2707 to avoid partial SSE reg dependency stalls): New pattern.
2709 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
2711 * config/s390/s390.md ("movcc"): Swap load and store instructions.
2713 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
2715 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
2716 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
2718 2013-07-26 Cary Coutant <ccoutant@google.com>
2720 * dwarf2out.c (die_checksum_ordered): Don't include template
2721 instantiations in signature.
2722 (is_template_parameter): New function.
2723 (is_template_instantiation): New function.
2724 (generate_skeleton_bottom_up): Don't include template instantiations
2726 (generate_skeleton): Likewise.
2727 (break_out_comdat_types): Move recursive call to break out nested
2729 (prune_unused_types_mark_generic_parms_dies): Call
2730 is_template_parameter.
2732 2013-07-26 Ian Bolton <ian.bolton@arm.com>
2734 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
2735 uses vector registers.
2736 * config/aarch64/iterators.md: Add attributes rtn and vas.
2738 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2739 Richard Earnshaw <richard.earnshaw@arm.com>
2741 * combine.c (simplify_comparison): Re-canonicalize operands
2743 * config/arm/arm.md (movcond_addsi): New splitter.
2745 2013-07-25 Sterling Augustine <saugustine@google.com>
2747 * dwarf2out.c (size_of_pubnames): Move code to...
2748 (include_pubname_in_output): ...here. New.
2749 (want_pubnames): Rearrange.
2750 (output_pubnames): Call include_pubname_in_output. Move assertion.
2752 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2754 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
2756 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
2759 * doc/extend.texi: Remove obsolete builtins. Fix
2760 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
2762 2013-07-25 Jan Hubicka <jh@suse.cz>
2764 * cgraph.c (release_function_body): Break out from ...
2765 (cgraph_release_function_body): ... this one; also release DECL_RESULT
2767 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
2768 old_tree in the map.
2769 (create_specialized_node): Update.
2770 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
2772 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
2773 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
2774 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
2775 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
2778 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2780 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
2781 addsi3_carryin_alt2_<optab>): Correct output template.
2783 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2785 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
2786 Adjust for arm_restrict_it.
2787 Remove trailing whitespace.
2789 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
2791 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
2792 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
2794 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
2796 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
2798 PR rtl-optimization/57960
2799 * lra-constraints.c (process_alt_operands): Use the right mode
2800 when checking strict_low.
2802 2013-07-25 Jan Hubicka <jh@suse.cz>
2804 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
2805 * cgraph.c (cgraph_remove_node): Do not release function body
2806 when in cgraph streaming.
2807 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
2808 in other partitions are not considered reachable; fix handling of
2811 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2813 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
2815 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2820 * config/arm/arm.md ("*sibcall_insn): Replace use of
2821 Ss with US. Adjust output for v5 and v4t.
2822 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
2824 * config/arm/constraints.md ("Ss"): Rename to US.
2826 2013-07-25 Terry Guo <terry.guo@arm.com>
2828 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
2829 shift_add/shift_sub0/shift_sub1 RTXs.
2831 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
2832 Anton Blanchard <anton@au1.ibm.com>
2834 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
2835 (altivec_vpks<VI_char>ss): Likewise.
2836 (altivec_vpks<VI_char>us): Likewise.
2837 (altivec_vpku<VI_char>us): Likewise.
2838 (altivec_vpku<VI_char>um): Likewise.
2840 2013-07-24 David Malcolm <dmalcolm@redhat.com>
2842 Introduce context class.
2844 * Makefile.in (CONTEXT_H): New.
2845 (OBJS): Add context.o.
2846 (toplev.o): Add CONTEXT_H to dependencies.
2849 * toplev.c (general_init): Create the singleton gcc::context instance.
2855 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
2857 PR rtl-optimization/57968
2858 * mode-switching.c (create_pre_exit): Allow instructions that
2859 don't set a return register to need a non-exit mode.
2861 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2862 Anton Blanchard <anton@au1.ibm.com>
2864 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
2865 operands to vperm for little endian.
2866 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
2867 of lvsl to create the control mask for a vperm for little endian.
2869 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2870 Anton Blanchard <anton@au1.ibm.com>
2872 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2873 two operands for little-endian.
2875 2013-07-23 Steve Ellcey <sellcey@mips.com>
2877 * config/mips/mips.c (mips_case_values_threshold): New.
2878 (TARGET_CASE_VALUES_THRESHOLD): Define.
2880 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2881 Anton Blanchard <anton@au1.ibm.com>
2883 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
2884 selection of field for vector splat in little endian mode.
2886 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2888 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
2889 expanders to rs6000.md.
2890 (ior<mode>3): Likewise.
2891 (and<mode>3): Likewise.
2892 (one_cmpl<mode>2): Likewise.
2893 (nor<mode>3): Likewise.
2894 (andc<mode>3): Likewise.
2895 (eqv<mode>3): Likewise.
2896 (nand<mode>3): Likewise.
2897 (orc<mode>3): Likewise.
2899 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
2902 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
2903 to split multi-word logical operations.
2904 (rs6000_split_logical_di): Likewise.
2905 (rs6000_split_logical): Likewise.
2907 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
2908 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
2909 and allow TImode operations in 32-bit.
2910 (vsx_and<mode>3_64bit): Likewise.
2911 (vsx_ior<mode>3_32bit): Likewise.
2912 (vsx_ior<mode>3_64bit): Likewise.
2913 (vsx_xor<mode>3_32bit): Likewise.
2914 (vsx_xor<mode>3_64bit): Likewise.
2915 (vsx_one_cmpl<mode>2_32bit): Likewise.
2916 (vsx_one_cmpl<mode>2_64bit): Likewise.
2917 (vsx_nor<mode>3_32bit): Likewise.
2918 (vsx_nor<mode>3_64bit): Likewise.
2919 (vsx_andc<mode>3_32bit): Likewise.
2920 (vsx_andc<mode>3_64bit): Likewise.
2921 (vsx_eqv<mode>3_32bit): Likewise.
2922 (vsx_eqv<mode>3_64bit): Likewise.
2923 (vsx_nand<mode>3_32bit): Likewise.
2924 (vsx_nand<mode>3_64bit): Likewise.
2925 (vsx_orc<mode>3_32bit): Likewise.
2926 (vsx_orc<mode>3_64bit): Likewise.
2928 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
2929 logical types in GPRs.
2931 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
2932 logical insns to rs6000.md, and allow TImode operations in
2934 (altivec_ior<mode>3): Likewise.
2935 (altivec_xor<mode>3): Likewise.
2936 (altivec_one_cmpl<mode>2): Likewise.
2937 (altivec_nor<mode>3): Likewise.
2938 (altivec_andc<mode>3): Likewise.
2940 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
2941 attributes for moving the 128-bit logical operations into
2943 (BOOL_REGS_OUTPUT): Likewise.
2944 (BOOL_REGS_OP1): Likewise.
2945 (BOOL_REGS_OP2): Likewise.
2946 (BOOL_REGS_UNARY): Likewise.
2947 (BOOL_REGS_AND_CR0): Likewise.
2948 (one_cmpl<mode>2): Add support for DI logical operations on
2949 32-bit, splitting the operations to 32-bit.
2953 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
2954 changes to combine the 32/64-bit code, allow logical operations on
2955 TI mode in 32-bit, and to use similar match_operator patterns like
2956 scalar mode uses. Combine the Altivec and VSX code for logical
2957 operations, and move it here.
2958 (ior<mode>3, 128-bit types): Likewise.
2959 (xor<mode>3, 128-bit types): Likewise.
2960 (one_cmpl<mode>3, 128-bit types): Likewise.
2961 (nor<mode>3, 128-bit types): Likewise.
2962 (andc<mode>3, 128-bit types): Likewise.
2963 (eqv<mode>3, 128-bit types): Likewise.
2964 (nand<mode>3, 128-bit types): Likewise.
2965 (orc<mode>3, 128-bit types): Likewise.
2966 (and<mode>3_internal): Likewise.
2967 (bool<mode>3_internal): Likewise.
2968 (boolc<mode>3_internal1): Likewise.
2969 (boolc<mode>3_internal2): Likewise.
2970 (boolcc<mode>3_internal1): Likewise.
2971 (boolcc<mode>3_internal2): Likewise.
2972 (eqv<mode>3_internal1): Likewise.
2973 (eqv<mode>3_internal2): Likewise.
2974 (one_cmpl1<mode>3_internal): Likewise.
2976 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2978 * config/microblaze/microblaze.c (microblaze_expand_prologue):
2979 Rename flag_stack_usage to flag_stack_usage_info.
2981 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
2983 * config/microblaze/sync.md: New file.
2984 * config/microblaze/microblaze.md: Include sync.md
2985 * config/microblaze/microblaze.c: Add print_operand 'y'.
2986 * config/microblaze/constraints.md: Add memory_contraint
2987 'Q' which is a single register.
2989 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
2991 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
2993 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
2995 * reload.c (find_reloads): Exit loop once we find this operand
2996 cannot be reloaded somehow for this alternative.
2998 * reload.c (find_reloads): Exit loop once we find a hard register.
3000 * rtlanal.c (computed_jump_p): Exit loop once we find label
3003 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3005 * cfgloopmanip.c (remove_path): Exit loop after setting
3008 * gensupport.c (subst_dup): Avoid loop if code is not
3009 MATCH_DUP nor MATCH_OP_DUP.
3011 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3013 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3015 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3017 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3018 true for SP_REGNUM if mode == ptr_mode.
3019 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3020 with value R0_REGNUM + 31.
3022 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3024 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3025 pad pointer-typed argument downward.
3027 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3029 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3030 and __ILP32__ when the ILP32 model is in use.
3032 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3034 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3035 (aarch64_load_symref_appropriately): In the case of
3036 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3037 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3038 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3039 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3040 if 'mode' doesn't equal to 'ptr_mode'.
3041 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3042 'vcall_offset'; change to call aarch64_emit_move differently depending
3043 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3044 to calculate the upper bound of 'vcall_offset'.
3045 (aarch64_cannot_force_const_mem): Change to also return true if
3047 (aarch64_legitimize_reload_address): In the case of large
3048 displacements, add new local variable 'xmode' and an assertion
3049 based on it; change to use 'xmode' to generate the new rtx and
3051 (aarch64_asm_trampoline_template): Change to generate the template
3052 differently depending on TARGET_ILP32 or not; change to use
3053 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3054 (aarch64_trampoline_size): Removed.
3055 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3056 and replace immediate literals with it. Change to use 'ptr_mode'
3057 instead of 'DImode' and call convert_memory_address if the mode
3058 of 'fnaddr' doesn't equal to 'ptr_mode'.
3059 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3061 (aarch64_elf_asm_destructor): Likewise.
3062 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3063 on TARGET_ILP32 instead of aarch64_trampoline_size.
3064 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3065 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3066 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3067 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3068 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3069 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3070 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3071 depending on the value of 'mode'.
3072 (add_losym_<mode>): New.
3073 (ldr_got_small_<mode>): New, based on ldr_got_small.
3074 (ldr_got_small): Remove.
3075 (ldr_got_small_sidi): New.
3076 * config/aarch64/iterators.md (P): New.
3077 (PTR): Change to 'ptr_mode' in the condition.
3079 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3081 * config.gcc (aarch64*-*-*): Support --with-abi.
3082 (aarch64*-*-elf): Support --with-multilib-list.
3083 (aarch64*-*-linux*): Likewise.
3084 (supported_defaults): Add abi to aarch64*-*-*.
3085 * configure.ac: Mention AArch64 for --with-multilib-list.
3086 * configure: Re-generated.
3087 * config/aarch64/biarchilp32.h: New file.
3088 * config/aarch64/biarchlp64.h: New file.
3089 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3091 (MULTILIB_DEFAULTS): Ditto.
3092 (DRIVER_SELF_SPECS): Ditto.
3093 (ASM_SPEC): Update to also substitute -mabi.
3094 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3095 file whose name depends on -mabi= and -mbig-endian.
3096 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3098 (POINTER_SIZE): New define.
3099 (POINTERS_EXTEND_UNSIGNED): Ditto.
3100 (enum aarch64_abi_type): New enumeration tag.
3101 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3102 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3103 (TARGET_ILP32): New define.
3104 * config/aarch64/aarch64.opt (mabi): New.
3106 (ilp32, lp64): New values for -mabi.
3107 * config/aarch64/t-aarch64 (comma): New define.
3108 (MULTILIB_OPTIONS): Ditto.
3109 (MULTILIB_DIRNAMES): Ditto.
3110 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3111 * doc/invoke.texi: Document -mabi for AArch64.
3113 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3115 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3117 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3118 Anton Blanchard <anton@au1.ibm.com>
3120 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3121 endianness when selecting field to splat.
3123 2013-07-22 Eric Christopher <echristo@gmail.com>
3125 * dwarf2out.c (die_odr_checksum): New function to use
3126 CHECKSUM_ macros and ULEB128 for DIE tag.
3127 (generate_type_signature): Use.
3129 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3131 * config.gcc (sparc*-*-*): Accept leon3 processor.
3132 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3133 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3134 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3135 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3136 (mfix-ut699): Adjust comment.
3137 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3138 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3139 (CPP_CPU_SPEC): Likewise.
3140 (ASM_CPU_SPEC): Likewise.
3141 * config/sparc/sparc.c (leon3_cost): New constant.
3142 (sparc_option_override): Add leon3 support.
3143 (mem_ref): New function.
3144 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3145 (sparc_do_work_around_errata): Look into the instruction in the delay
3146 slot and adjust accordingly. Add fix for the data cache nullify issues
3147 of the UT699. Change insertion position for the NOP.
3148 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3149 (leon3_load): New reservation.
3150 (leon_store): Bump latency to 2.
3151 (grfpu): New automaton.
3152 (grfpu_alu): New unit.
3153 (grfpu_ds): Likewise.
3154 (leon_fp_alu): Adjust.
3155 (leon_fp_mult): Delete.
3156 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3157 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3158 * config/sparc/sparc.md (cpu): Add leon3.
3159 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3161 (atomic_test_and_set): Likewise.
3164 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3166 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3167 default for R5900 targets.
3168 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3169 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3170 * config/mips/mips.c (mips_option_override): Report an error for
3171 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3172 for -march=r5900 -mhard-float.
3174 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3176 * df-problems.c (can_move_insns_across): Exit loop once we
3177 find a non-fixed, non-global register.
3179 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3182 * omega.c (omega_eliminate_red): Break after setting red_found.
3183 (omega_problem_has_red_equations): Similarly after setting found.
3184 (omega_query_variable): Similarly after setting coupled.
3186 2013-07-22 Marek Polacek <polacek@redhat.com>
3188 * gimplify.c: Don't include gimple.h twice.
3190 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3192 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3193 instead of TARGET_THUMB1.
3194 (Pz): New constraint.
3195 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3197 (compare_negsi_si): Likewise.
3198 (compare_addsi2_op0): Likewise.
3199 (compare_addsi2_op1): Likewise.
3200 (addsi3_carryin_<optab>): Likewise.
3201 (addsi3_carryin_alt2_<optab>): Likewise.
3202 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3203 for arm_restrict_it.
3204 (subsi3_carryin): Likewise.
3205 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3206 (minmax_arithsi): Disable for arm_restrict_it.
3207 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3208 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3209 (satsi_<SAT:code>_shift): Likewise.
3210 (arm_shiftsi3): Add alternative for 16-bit encoding.
3211 (arm32_movhf): Disable for arm_restrict_it.
3212 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3213 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3215 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3217 * config/arm/arm.md (attribute "insn"): Delete.
3218 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3219 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3220 (not_shiftsi): Update for attribute change.
3221 (not_shiftsi_compare0): Likewise.
3222 (not_shiftsi_compare0_scratch): Likewise.
3223 (arm_one_cmplsi2): Likewise.
3224 (thumb1_one_cmplsi2): Likewise.
3225 (notsi_compare0): Likewise.
3226 (notsi_compare0_scratch): Likewise.
3227 (thumb1_movdi_insn): Likewise.
3228 (arm_movsi_insn): Likewise.
3229 (movhi_insn_arch4): Likewise.
3230 (movhi_bytes): Likewise.
3231 (arm_movqi_insn): Likewise.
3232 (thumb1_movqi_insn): Likewise.
3233 (arm32_movhf): Likewise.
3234 (thumb1_movhf): Likewise.
3235 (arm_movsf_soft_insn): Likewise.
3236 (thumb1_movsf_insn): Likewise.
3237 (thumb_movdf_insn): Likewise.
3238 (movsicc_insn): Likewise.
3239 (movsfcc_soft_insn): Likewise.
3240 (and_scc): Likewise.
3241 (cond_move): Likewise.
3242 (if_move_not): Likewise.
3243 (if_not_move): Likewise.
3244 (if_shift_move): Likewise.
3245 (if_move_shift): Likewise.
3246 (if_shift_shift): Likewise.
3247 (if_not_arith): Likewise.
3248 (if_arith_not): Likewise.
3249 (cond_move_not): Likewise.
3250 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3251 (neon_mov<mode>): Likewise.
3252 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3253 (thumb2_movsi_vfp): Likewise.
3254 (movsf_vfp): Likewise.
3255 (thumb2_movsf_vfp): Likewise.
3256 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3258 (cortexa7_older_only): Likewise.
3259 (cortexa7_younger): Likewise.
3260 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3261 (1020alu_shift_op): Likewise.
3262 (1020alu_shift_reg_op): Likewise.
3263 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3264 (alu_shift_op): Likewise.
3265 (alu_shift_reg_op): Likewise.
3266 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3267 (11_alu_shift_op): Likewise.
3268 (11_alu_shift_reg_op): Likewise.
3269 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3270 (9_alu_shift_reg_op): Likewise.
3271 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3273 (cortex_a15_alu_shift): Likewise.
3274 (cortex_a15_alu_shift_reg): Likewise.
3275 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3276 (cortex_a5_alu_shift): Likewise.
3277 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3279 (cortex_a53_alu_shift): Likewise.
3280 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3282 (cortex_a7_alu_reg): Likewise.
3283 (cortex_a7_alu_shift): Likewise.
3284 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3285 (cortex_a8_alu_shift): Likewise.
3286 (cortex_a8_alu_shift_reg): Likewise.
3287 (cortex_a8_mov): Likewise.
3288 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3289 (cortex_a9_dp_shift): Likewise.
3290 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3291 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3292 (cortex_r4_mov): Likewise.
3293 (cortex_r4_alu_shift): Likewise.
3294 (cortex_r4_alu_shift_reg): Likewise.
3295 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3296 (526_alu_shift_op): Likewise.
3297 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3298 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3299 (626te_alu_shift_op): Likewise.
3300 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3301 (726te_alu_op): Likewise.
3302 (726te_alu_shift_op): Likewise.
3303 (726te_alu_shift_reg_op): Likewise.
3304 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3305 (mp626_alu_shift_op): Likewise.
3306 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3307 (pj4_alu_e1_conds): Likewise.
3308 (pj4_alu): Likewise.
3309 (pj4_alu_conds): Likewise.
3310 (pj4_shift): Likewise.
3311 (pj4_shift_conds): Likewise.
3312 (pj4_alu_shift): Likewise.
3313 (pj4_alu_shift_conds): Likewise.
3315 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3317 * config/arm/predicates.md (shiftable_operator_strict_it):
3319 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3320 Disable cond_exec version for arm_restrict_it.
3321 (thumb2_smaxsi3): Convert to generate cond_exec.
3322 (thumb2_sminsi3): Likewise.
3323 (thumb32_umaxsi3): Likewise.
3324 (thumb2_uminsi3): Likewise.
3325 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3326 (thumb2_neg_abssi2): Likewise.
3327 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3328 (thumb2_movsicc_insn): Adjust alternatives.
3329 (thumb2_mov_negscc): Disable for arm_restrict_it.
3330 (thumb2_mov_negscc_strict_it): New pattern.
3331 (thumb2_mov_notscc_strict_it): New pattern.
3332 (thumb2_mov_notscc): Disable for arm_restrict_it.
3333 (thumb2_ior_scc): Likewise.
3334 (thumb2_ior_scc_strict_it): New pattern.
3335 (thumb2_cond_move): Adjust for arm_restrict_it.
3336 (thumb2_cond_arith): Disable for arm_restrict_it.
3337 (thumb2_cond_arith_strict_it): New pattern.
3338 (thumb2_cond_sub): Adjust for arm_restrict_it.
3339 (thumb2_movcond): Likewise.
3340 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3341 (thumb2_zero_extendhisi2_v6): Likewise.
3342 (thumb2_zero_extendqisi2_v6): Likewise.
3343 (orsi_notsi_si): Likewise.
3344 (orsi_not_shiftsi_si): Likewise.
3346 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3348 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3349 instruction sequence is 1 byte shorter.
3351 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3353 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3354 it is not needed after split.
3356 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3359 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3360 second label for nonlocal goto receivers. Don't output pic base labels
3361 unless we're producing PIC; mark that action unreachable().
3362 (ix86_save_reg): If the function contains a nonlocal label, save the
3364 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3365 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3366 (update_pic_label_number_if_needed): New.
3367 (machopic_output_function_base_name): Adjust for nonlocal receiver
3369 (machopic_should_output_picbase_label): New.
3370 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3371 (nonlocal_goto_receiver): New insn and split.
3373 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3375 * config/aarch64/aarch64-builtins.c
3376 (aarch64_fold_builtin): Fold abs in all modes.
3377 * config/aarch64/aarch64-simd-builtins.def
3378 (abs): Enable for all modes.
3379 * config/aarch64/arm_neon.h
3380 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
3381 (vabs_f64): Add missing intrinsic.
3383 2013-07-19 Ian Bolton <ian.bolton@arm.com>
3385 * config/aarch64/arm_neon.h (vabs_s64): New function
3387 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
3390 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
3391 * config/avr/avr.md (adjust_len): Add `round'.
3392 * config/avr/avr-protos.h (avr_out_round): New prototype.
3393 (avr_out_plus): Add `out_label' argument.
3394 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
3395 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
3396 Handle the case where `insn' is just a pattern.
3397 (avr_out_bitop): Handle the case where `insn' is just a pattern.
3398 (avr_out_round): New function.
3399 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
3401 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
3403 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3404 Add check for flag_stack_usage to handle -fstack-usage support
3406 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
3408 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
3409 interaction for new Power8 flags and VSX.
3411 2013-07-18 Sriraman Tallam <tmsriram@google.com>
3414 * tree-inline.c (expand_call_inline): Emit errors during
3415 early_inlining only if optimization is not turned on.
3417 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3421 * passes.c (init_optimization_passes): Move the construction of
3422 the pass hierarchy into a new passes.def file.
3424 * Makefile.in (passes.o): Add dependency on passes.def.
3426 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3428 * passes.c (init_optimization_passes): Introduce macros for
3429 constructing the tree of passes (INSERT_PASSES_AFTER,
3430 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
3431 TERMINATE_PASS_LIST).
3433 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
3434 Wei Mi <wmi@google.com>
3436 PR rtl-optimization/57878
3437 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
3439 (reload_pseudo_compare_func): Check nregs first for reload
3442 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3444 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
3446 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
3448 * read-rtl.c (validate_const_int): Once an invalid character is
3449 seen, quit the loop.
3451 * gengtype.c (write_roots): Similarly once we find the "deletable"
3452 or "if_marked" option.
3454 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3456 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
3457 "xtab" and "sat". Move value "clz" from here to ...
3458 (attriubte "type"): ... here.
3459 (satsi_<SAT:code>): Delete "insn" attribute.
3460 (satsi_<SAT:code>_shift): Likewise.
3461 (arm_zero_extendqisi2addsi): Likewise.
3462 (arm_extendqisi2addsi): Likewise.
3463 (clzsi2): Update for attribute changes.
3464 (rbitsi2): Likewise.
3465 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
3467 (arm_usatsihi): Likewise.
3468 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3470 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
3472 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
3473 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
3474 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
3475 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
3476 in alphabetical order.
3477 (attribute "core_cycles"): Update for attribute changes.
3478 (arm_addsi3): Likewise.
3479 (addsi3_compare0): Likewise.
3480 (addsi3_compare0_scratch): Likewise.
3481 (addsi3_compare_op1): Likewise.
3482 (addsi3_compare_op2): Likewise.
3483 (compare_addsi2_op0): Likewise.
3484 (compare_addsi2_op1): Likewise.
3485 (addsi3_carryin_shift_<optab>): Likewise.
3486 (subsi3_carryin_shift): Likewise.
3487 (rsbsi3_carryin_shift): Likewise.
3488 (arm_subsi3_insn): Likewise.
3489 (subsi3_compare0): Likewise.
3490 (subsi3_compare): Likewise.
3491 (arm_andsi3_insn): Likewise.
3492 (thumb1_andsi3_insn): Likewise.
3493 (andsi3_compare0): Likewise.
3494 (andsi3_compare0_scratch): Likewise.
3495 (zeroextractsi_compare0_scratch
3496 (andsi_not_shiftsi_si): Likewise.
3497 (iorsi3_insn): Likewise.
3498 (iorsi3_compare0): Likewise.
3499 (iorsi3_compare0_scratch): Likewise.
3500 (arm_xorsi3): Likewise.
3501 (thumb1_xorsi3_insn): Likewise.
3502 (xorsi3_compare0): Likewise.
3503 (xorsi3_compare0_scratch): Likewise.
3504 (satsi_<SAT:code>_shift): Likewise.
3506 (arm_shiftsi3): Likewise.
3507 (shiftsi3_compare0): Likewise.
3508 (not_shiftsi): Likewise.
3509 (not_shiftsi_compare0): Likewise.
3510 (not_shiftsi_compare0_scratch): Likewise.
3511 (arm_one_cmplsi2): Likewise.
3512 (thumb_one_complsi2): Likewise.
3513 (notsi_compare0): Likewise.
3514 (notsi_compare0_scratch): Likewise.
3515 (thumb1_zero_extendhisi2): Likewise.
3516 (arm_zero_extendhisi2): Likewise.
3517 (arm_zero_extendhisi2_v6): Likewise.
3518 (arm_zero_extendhisi2addsi): Likewise.
3519 (thumb1_zero_extendqisi2): Likewise.
3520 (thumb1_zero_extendqisi2_v6): Likewise.
3521 (arm_zero_extendqisi2): Likewise.
3522 (arm_zero_extendqisi2_v6): Likewise.
3523 (arm_zero_extendqisi2addsi): Likewise.
3524 (thumb1_extendhisi2): Likewise.
3525 (arm_extendhisi2): Likewise.
3526 (arm_extendhisi2_v6): Likewise.
3527 (arm_extendqisi): Likewise.
3528 (arm_extendqisi_v6): Likewise.
3529 (arm_extendqisi2addsi): Likewise.
3530 (thumb1_extendqisi2): Likewise.
3531 (thumb1_movdi_insn): Likewise.
3532 (arm_movsi_insn): Likewise.
3533 (movsi_compare0): Likewise.
3534 (movhi_insn_arch4): Likewise.
3535 (movhi_bytes): Likewise.
3536 (arm_movqi_insn): Likewise.
3537 (thumb1_movqi_insn): Likewise.
3538 (arm32_movhf): Likewise.
3539 (thumb1_movhf): Likewise.
3540 (arm_movsf_soft_insn): Likewise.
3541 (thumb1_movsf_insn): Likewise.
3542 (movdf_soft_insn): Likewise.
3543 (thumb_movdf_insn): Likewise.
3544 (arm_cmpsi_insn): Likewise.
3545 (cmpsi_shiftsi): Likewise.
3546 (cmpsi_shiftsi_swp): Likewise.
3547 (arm_cmpsi_negshiftsi_si): Likewise.
3548 (movsicc_insn): Likewise.
3549 (movsfcc_soft_insn): Likewise.
3550 (arith_shiftsi): Likewise.
3551 (arith_shiftsi_compare0
3552 (arith_shiftsi_compare0_scratch
3553 (sub_shiftsi): Likewise.
3554 (sub_shiftsi_compare0
3555 (sub_shiftsi_compare0_scratch
3556 (and_scc): Likewise.
3557 (cond_move): Likewise.
3558 (if_plus_move): Likewise.
3559 (if_move_plus): Likewise.
3560 (if_move_not): Likewise.
3561 (if_not_move): Likewise.
3562 (if_shift_move): Likewise.
3563 (if_move_shift): Likewise.
3564 (if_shift_shift): Likewise.
3565 (if_not_arith): Likewise.
3566 (if_arith_not): Likewise.
3567 (cond_move_not): Likewise.
3568 (thumb1_ashlsi3): Set type attribute.
3569 (thumb1_ashrsi3): Likewise.
3570 (thumb1_lshrsi3): Likewise.
3571 (thumb1_rotrsi3): Likewise.
3572 (shiftsi3_compare0_scratch): Likewise.
3573 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
3574 (neon_mov<mode>): Likewise.
3575 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
3577 (thumb2_movsi_insn): Likewise.
3578 (thumb2_cmpsi_neg_shiftsi): Likewise.
3579 (thumb2_extendqisi_v6): Likewise.
3580 (thumb2_zero_extendhisi2_v6): Likewise.
3581 (thumb2_zero_extendqisi2_v6): Likewise.
3582 (thumb2_shiftsi3_short): Likewise.
3583 (thumb2_addsi3_compare0_scratch): Likewise.
3584 (orsi_not_shiftsi_si): Likewise.
3585 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3586 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
3588 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3589 (1020alu_shift_op): Likewise.
3590 (1020alu_shift_reg_op): Likewise.
3591 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3592 (alu_shift_op): Likewise.
3593 (alu_shift_reg_op): Likewise.
3594 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3595 (11_alu_shift_op): Likewise.
3596 (11_alu_shift_reg_op): Likewise.
3597 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3598 (9_alu_shift_reg_op): Likewise.
3599 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3601 (cortex_a15_alu_shift): Likewise.
3602 (cortex_a15_alu_shift_reg): Likewise.
3603 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
3605 (cortex_a5_alu_shift): Likewise.
3606 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
3608 (cortex_a53_alu_shift): Likewise.
3609 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3611 (cortex_a7_alu_reg): Likewise.
3612 (cortex_a7_alu_shift): Likewise.
3613 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
3615 (cortex_a8_alu_shift): Likewise.
3616 (cortex_a8_alu_shift_reg): Likewise.
3617 (cortex_a8_mov): Likewise.
3618 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3619 (cortex_a9_dp_shift): Likewise.
3620 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
3622 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
3624 (cortex_r4_mov): Likewise.
3625 (cortex_r4_alu_shift): Likewise.
3626 (cortex_r4_alu_shift_reg): Likewise.
3627 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3628 (526_alu_shift_op): Likewise.
3629 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3630 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3631 (626te_alu_shift_op): Likewise.
3632 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
3633 (726te_alu_op): Likewise.
3634 (726te_alu_shift_op): Likewise.
3635 (726te_alu_shift_reg_op): Likewise.
3636 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3637 (mp626_alu_shift_op): Likewise.
3638 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
3639 (pj4_alu_e1_conds): Likewise.
3640 (pj4_alu): Likewise.
3641 (pj4_alu_conds): Likewise.
3642 (pj4_shift): Likewise.
3643 (pj4_shift_conds): Likewise.
3644 (pj4_alu_shift): Likewise.
3645 (pj4_alu_shift_conds): Likewise.
3646 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3648 (cortexa7_older_only): Likewise.
3649 (cortexa7_younger): Likewise.
3651 2013-07-18 David Malcolm <dmalcolm@redhat.com>
3653 * ipa-pure-const.c (generate_summary): Rename to...
3654 (pure_const_generate_summary): ... this.
3656 2013-07-17 Iain Sandoe <iain@codesourcery.com>
3658 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
3660 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
3663 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
3666 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3668 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
3669 enabled without -march=zEC12.
3670 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
3673 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3675 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
3676 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
3677 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
3678 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
3679 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
3680 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
3681 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
3683 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
3684 and ISA_HAS_NMADD3_NMSUB3.
3685 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
3686 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
3687 (nmsub4<mode>, nmsub3<mode>): Likewise.
3688 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
3690 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3692 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
3693 TARGET_MIPS5400 checking.
3695 2013-07-16 Jakub Jelinek <jakub@redhat.com>
3696 Peter Bergner <bergner@vnet.ibm.com>
3698 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
3699 registers in the comment.
3700 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
3701 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
3702 rather than FIRST_PSEUDO_REGISTERS.
3704 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
3706 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
3707 enable extra ISA flags with TARGET_HTM.
3709 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
3711 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
3714 2013-07-15 Cong Hou <congh@google.com>
3716 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
3717 in compare function for sorting.
3719 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
3721 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
3722 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
3723 * config/rs6000/rs6000.opt: Add -mhtm option.
3724 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
3725 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
3726 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
3727 __HTM__ if the HTM instructions are available.
3728 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
3729 htm_spr_reg_operand): New define_predicates.
3730 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
3731 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
3733 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
3734 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
3735 HTM builtin functions.
3736 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
3737 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
3738 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
3739 (rs6000_builtin_mask_calculate): Likewise.
3740 (rs6000_option_override_internal): Likewise.
3741 (bdesc_htm): Add new HTM builtin support.
3742 (htm_spr_num): New function.
3743 (htm_spr_regno): Likewise.
3744 (rs6000_htm_spr_icode): Likewise.
3745 (htm_expand_builtin): Likewise.
3746 (htm_init_builtins): Likewise.
3747 (rs6000_expand_builtin): Add support for HTM builtin functions.
3748 (rs6000_init_builtins): Likewise.
3749 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
3751 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
3752 (TARGET_HTM, MASK_HTM): Define macros.
3753 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
3754 (FIXED_REGISTERS): Likewise.
3755 (CALL_USED_REGISTERS): Likewise.
3756 (CALL_REALLY_USED_REGISTERS): Likewise.
3757 (REG_ALLOC_ORDER): Likewise.
3758 (enum reg_class): Likewise.
3759 (REG_CLASS_NAMES): Likewise.
3760 (REG_CLASS_CONTENTS): Likewise.
3761 (REGISTER_NAMES): Likewise.
3762 (ADDITIONAL_REGISTER_NAMES): Likewise.
3763 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
3764 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
3765 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
3766 * config/rs6000/htm.md: New file.
3767 * config/rs6000/htmintrin.h: New file.
3768 * config/rs6000/htmxlintrin.h: New file.
3770 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
3772 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
3773 Define SYMBOL_TINY_GOT, update comment.
3774 * config/aarch64/aarch64.c
3775 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
3776 (aarch64_expand_mov_immediate): Likewise.
3777 (aarch64_print_operand): Likewise.
3778 (aarch64_classify_symbol): Likewise.
3779 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
3780 (ldr_got_tiny): Define.
3782 2013-07-13 Tobias Grosser <tobias@grosser.es>
3784 PR tree-optimization/54094
3785 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
3786 scheduling dimension for the parallelism check from the polyhedral
3787 information in the AST.
3788 * graphite-dependences.c (carries_deps): Do not assume the schedule is
3791 2013-07-13 Jason Merrill <jason@redhat.com>
3793 * print-tree.c (debug_vec_tree): Use debug_raw.
3794 (debug_raw (vec<tree, va_gc> &)): New.
3795 (debug_raw (vec<tree, va_gc> *)): New.
3796 * tree.h: Declare them.
3798 2013-07-13 Bin Cheng <bin.cheng@arm.com>
3800 * ifcvt.c (ifcvt_after_combine): New static variable.
3801 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
3803 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
3804 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
3805 rest_of_handle_if_after_reload): Pass new argument for if_convert.
3807 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
3809 * config/mips/mips.c (mips_expand_call): Remove empty statement.
3811 2013-07-12 Michael Matz <matz@suse.de>
3814 * convert.c (convert_to_real): Reject non-float inner types.
3816 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
3818 * config/aarch64/aarch64-protos.h
3819 (aarch64_simd_immediate_valid_for_move): Remove.
3820 * config/aarch64/aarch64.c (simd_immediate_info): New member.
3821 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
3823 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
3825 2013-07-11 Steve Ellcey <sellcey@mips.com>
3827 * config/mips/mips.c (mips_conditional_register_usage): Do not
3828 use t[0-7] registers in MIPS16 mode when optimizing for size.
3830 2013-07-11 Sriraman Tallam <tmsriram@google.com>
3832 * config/i386/i386.c (dispatch_function_versions): Fix array
3833 indexing of function_version_info to match actual_versions.
3835 2013-07-11 Teresa Johnson <tejohnson@google.com>
3837 * vec.h (struct va_gc): Move release out-of-line.
3838 (va_gc::release): Call ggc_free on released vec.
3840 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3842 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
3843 Require GOT register as additional operand in UNSPEC.
3844 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
3845 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3846 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
3847 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
3848 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
3849 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
3850 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
3852 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
3855 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
3856 name seen by assembler/linker rather if available.
3858 2013-07-11 Andreas Schwab <schwab@suse.de>
3860 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
3862 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
3864 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
3866 2013-07-10 Joseph Myers <joseph@codesourcery.com>
3868 * doc/tm.texi.in: Move hook documentation to ....
3869 * target.def: ... here.
3871 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
3873 * doc/tm.texi: Regenerate.
3875 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
3878 * doc/invoke.texi: Document -Wconditionally-supported.
3880 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3883 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
3886 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3889 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
3890 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
3891 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
3892 Remove duplicate devices.
3893 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
3894 * config/avr/t-multilib: Regenerate.
3895 * config/avr/avr-tables.opt: Regenerate.
3896 * doc/avr-mmcu.texi: Regenerate.
3898 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
3901 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
3903 2013-07-10 Graham Stott <graham.stott@btinternet.com>
3905 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
3906 the cost of MULT when optimizing for size.
3908 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3910 * config/cr16/cr16-protos.h: Don't include target.h.
3912 2013-07-09 Joseph Myers <joseph@codesourcery.com>
3914 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
3915 adjust register size for TDmode and TFmode for VSX registers.
3917 2013-07-08 Kai Tietz <ktietz@redhat.com>
3920 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
3921 hook_bool_const_tree_true.
3923 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3925 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
3926 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
3927 * config/s390/s390.md: Define FPR*_REGNUM constants.
3928 Fix FPR2_REGNUM constant (18 -> 17).
3929 ("*trunc<BFP:mode><DFP_ALL:mode>2")
3930 ("*trunc<DFP_ALL:mode><BFP:mode>2")
3931 ("trunc<BFP:mode><DFP_ALL:mode>2")
3932 ("trunc<DFP_ALL:mode><BFP:mode>2")
3933 ("*extend<BFP:mode><DFP_ALL:mode>2")
3934 ("*extend<DFP_ALL:mode><BFP:mode>2")
3935 ("extend<BFP:mode><DFP_ALL:mode>2")
3936 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
3939 2013-07-08 Graham Stott <graham.stott@btinternet.com>
3941 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
3943 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3945 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
3946 and cfun_fpr_bit_p to cfun_fpr_save_p.
3947 (s390_frame_area, s390_register_info, s390_frame_info)
3948 (s390_emit_prologue, s390_emit_epilogue)
3949 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
3951 * config/s390/s390.h: Define *_REGNUM macros for floating point
3954 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
3956 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
3958 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
3960 PR rtl-optimization/57786
3961 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
3962 and break out of the loop when it is set to false.
3964 2013-07-08 Jakub Jelinek <jakub@redhat.com>
3967 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
3968 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
3969 (const_int 63)) 0)).
3970 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
3971 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
3972 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
3974 PR rtl-optimization/57829
3975 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
3976 mask bits outside of mode are just sign-extension from mode to HWI.
3978 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3980 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
3981 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
3982 adjust_address instead of change_address to keep info about alignment.
3983 (emit_strmov): Remove.
3984 (emit_memmov): New function.
3985 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
3986 (expand_movmem_epilogue): Likewise and return updated rtx for
3988 (expand_constant_movmem_prologue): Likewise and return updated rtx for
3989 destination and source.
3990 (decide_alignment): Refactor, handle vector_loop.
3991 (ix86_expand_movmem): Likewise.
3992 (ix86_expand_setmem): Likewise.
3993 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
3995 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
3997 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
3998 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4000 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4002 * config/i386/sse.md (sse_movlhps): Change alternative 3
4003 of operand 2 to "m".
4005 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4008 * config/i386/sse.md (iptr): New mode attribute.
4009 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4010 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4011 (<sse>_vmmul<mode>3): Ditto.
4012 (<sse>_vmdiv<mode>3): Ditto.
4013 (sse_vmrcpv4sf2): Ditto.
4014 (<sse>_vmsqrt<mode>2): Ditto.
4015 (sse_vmrsqrtv4sf2): Ditto.
4016 (<sse>_vm<code><mode>3): Ditto.
4017 (avx_vmcmp<mode>3): Ditto.
4018 (<sse>_vmmaskcmp<mode>3): Ditto.
4019 (<sse>_comi): Ditto.
4020 (<sse>_ucomi): Ditto.
4021 (*xop_vmfrcz_<mode>): Ditto.
4022 (*fmai_fmadd_<mode>): Ditto.
4023 (*fmai_fmsub_<mode>): Ditto.
4024 (*fmai_fnmadd_<mode>): Ditto.
4025 (*fmai_fnmsub_<mode>): Ditto.
4026 (*fma4i_vmfmadd_<mode>): Ditto.
4027 (*fma4i_vmfmsub_<mode>): Ditto.
4028 (*fma4i_vmfnmadd_<mode>): Ditto.
4029 (*fma4i_vmfnmsub_<mode>): Ditto.
4030 (*xop_vmfrcz_<mode>): Ditto.
4031 (sse_cvtps2pi): Ditto.
4032 (sse_cvttps2pi): Ditto.
4033 (sse_cvtss2si): Ditto.
4034 (sse_cvtss2si_2): Ditto.
4035 (sse_cvtss2siq_2): Ditto.
4036 (sse_cvttss2si): Ditto.
4037 (sse_cvttss2siq): Ditto.
4038 (sse_cvtsd2si): Ditto.
4039 (sse_cvtsd2si_2): Ditto.
4040 (sse_cvtsd2siq_2): Ditto.
4041 (sse_cvttsd2si): Ditto.
4042 (sse_cvttsd2siq): Ditto.
4043 (sse_cvtsd2ss): Ditto.
4044 (sse_cvtss2sd): Ditto.
4045 (avx2_pbroadcast<mode>): Ditto.
4046 (avx2_pbroadcast<mode>_1): Ditto.
4047 (*avx_vperm_broadcast_v4sf): Ditto.
4049 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4050 (sse_movlhps): Ditto.
4051 (sse_storehps): Ditto.
4052 (sse_loadhps): Ditto.
4053 (sse_storelps): Ditto.
4054 (sse_loadlps): Ditto.
4055 (*vec_concatv4sf): Ditto.
4056 (*vec_interleave_highv2df): Ditto.
4057 (*vec_interleave_lowv2df): Ditto.
4058 (*vec_extractv2df_1_sse): Ditto.
4059 (*vec_extractv2df_0_sse): Ditto.
4060 (sse2_storelpd): Ditto.
4061 (sse2_loadlpd): Ditto.
4062 (sse2_movsd): Ditto.
4063 (*vec_concatv4si): Ditto.
4064 (vec_concatv2di): Ditto.
4066 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4067 for Intel asm dialect.
4068 (mmx_punpcklwd): Ditto.
4069 (mmx_punpckldq): Ditto.
4071 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4072 for intel assembler dialect.
4074 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4077 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4078 for BUILT_IN_C{LZ,LRSB}*.
4079 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4080 * tree-vrp.c (extract_range_basic): Handle
4081 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4082 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4083 fall thru to code calling set_value*.
4084 * builtins.c (expand_builtin): Remove *IMAX cases.
4085 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4086 if width is bigger than 2*HWI.
4088 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4090 PR rtl-optimization/55342
4091 * lra-int.h (lra_subreg_reload_pseudos): New.
4092 * lra.c: Add undoing optional reloads to the block diagram.
4093 (lra_subreg_reload_pseudos): New.
4094 (lra_optional_reload_pseudos): Change comments.
4095 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4096 lra_optional_reload_pseudos after undo transformations.
4097 * lra-assigns.c (pseudo_prefix_title): New.
4098 (lra_setup_reg_renumber): Use it.
4099 (spill_for): Ditto. Check subreg reload pseudos too.
4100 (assign_by_spills): Consider subreg reload pseudos too.
4101 * lra-constraints.c (simplify_operand_subreg): Use
4102 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4103 (curr_insn_transform): Recognize and do optional reloads.
4104 (undo_optional_reloads): New.
4105 (lra_undo_inheritance): Call undo_optional_reloads.
4107 2013-07-05 Thomas Quinot <quinot@adacore.com>
4109 * tree-complex.c (expand_complex_operations_1): Fix typo.
4111 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4113 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4114 (tune_params): New member 'const vec_costs'.
4115 * config/aarch64/aarch64.c (generic_vector_cost): New.
4116 (generic_tunings): New member 'generic_vector_cost'.
4117 (aarch64_builtin_vectorization_cost): New.
4118 (aarch64_add_stmt_cost): New.
4119 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4120 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4122 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4125 * config/i386/predicates.md (vsib_address_operand): Disallow
4126 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4128 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4131 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4132 expand_builtin_setjmp_receiver.
4133 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4134 with NULL for the label parameter.
4135 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4136 the frame-pointer. Adjust comments.
4137 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4138 only if LABEL is non-NULL.
4140 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4142 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4143 (ARM_ABI_AAPCS64): Ditto.
4145 (ARM_DEFAULT_ABI): Ditto.
4147 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4149 * config/aarch64/aarch64-builtins.c
4150 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4151 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4153 * config/aarch64/aarch64-simd.md
4154 (aarch64_ld1<VALL:mode>): New.
4155 (aarch64_st1<VALL:mode>): Likewise.
4156 * config/aarch64/arm_neon.h
4157 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4159 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4161 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4163 (ix86_option_override_internal):Turn on all -mavx target flags by
4164 default as they are dependent on AVX anyway.
4166 2013-07-02 Cary Coutant <ccoutant@google.com>
4168 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4170 (loc_checksum_ordered): Likewise.
4171 (hash_loc_operands): Remove inline keyword.
4173 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4175 PR tree-optimization/57741
4176 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4177 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4178 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4179 Allow REAL_CST step_exprs if flag_associative_math.
4180 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4182 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4184 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4186 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4188 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4190 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4192 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4194 (iorsi3_insn): Likewise.
4195 (arm_xorsi3): Likewise.
4197 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4199 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4201 (attribute "type"): ... here, and prefix with "wmmx_".
4202 (attribute "core_cycles"): Update for attribute changes.
4203 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4204 (tbcstv4hi): Likewise.
4205 (tbcstv2si): Likewise.
4206 (iwmmxt_iordi3): Likewise.
4207 (iwmmxt_xordi3): Likewise.
4208 (iwmmxt_anddi3): Likewise.
4209 (iwmmxt_nanddi3): Likewise.
4210 (iwmmxt_arm_movdi): Likewise.
4211 (iwmmxt_movsi_insn): Likewise.
4212 (mov<mode>_internal): Likewise.
4213 (and<mode>3_iwmmxt): Likewise.
4214 (ior<mode>3_iwmmxt): Likewise.
4215 (xor<mode>3_iwmmxt): Likewise.
4216 (add<mode>3_iwmmxt): Likewise.
4217 (ssaddv8qi3): Likewise.
4218 (ssaddv4hi3): Likewise.
4219 (ssaddv2si3): Likewise.
4220 (usaddv8qi3): Likewise.
4221 (usaddv4hi3): Likewise.
4222 (usaddv2si3): Likewise.
4223 (sub<mode>3_iwmmxt): Likewise.
4224 (sssubv8qi3): Likewise.
4225 (sssubv4hi3): Likewise.
4226 (sssubv2si3): Likewise.
4227 (ussubv8qi3): Likewise.
4228 (ussubv4hi3): Likewise.
4229 (ussubv2si3): Likewise.
4230 (mulv4hi3_iwmmxt): Likewise.
4231 (smulv4hi3_highpart): Likewise.
4232 (umulv4hi3_highpart): Likewise.
4233 (iwmmxt_wmacs): Likewise.
4234 (iwmmxt_wmacsz): Likewise.
4235 (iwmmxt_wmacu): Likewise.
4236 (iwmmxt_wmacuz): Likewise.
4237 (iwmmxt_clrdi): Likewise.
4238 (iwmmxt_clrv8qi): Likewise.
4239 (iwmmxt_clr4hi): Likewise.
4240 (iwmmxt_clr2si): Likewise.
4241 (iwmmxt_uavgrndv8qi3): Likewise.
4242 (iwmmxt_uavgrndv4hi3): Likewise.
4243 (iwmmxt_uavgv8qi3): Likewise.
4244 (iwmmxt_uavgv4hi3): Likewise.
4245 (iwmmxt_tinsrb): Likewise.
4246 (iwmmxt_tinsrh): Likewise.
4247 (iwmmxt_tinsrw): Likewise.
4248 (iwmmxt_textrmub): Likewise.
4249 (iwmmxt_textrmsb): Likewise.
4250 (iwmmxt_textrmuh): Likewise.
4251 (iwmmxt_textrmsh): Likewise.
4252 (iwmmxt_textrmw): Likewise.
4253 (iwmxxt_wshufh): Likewise.
4254 (eqv8qi3): Likewise.
4255 (eqv4hi3): Likewise.
4256 (eqv2si3): Likewise.
4257 (gtuv8qi3): Likewise.
4258 (gtuv4hi3): Likewise.
4259 (gtuv2si3): Likewise.
4260 (gtv8qi3): Likewise.
4261 (gtv4hi3): Likewise.
4262 (gtv2si3): Likewise.
4263 (smax<mode>3_iwmmxt): Likewise.
4264 (umax<mode>3_iwmmxt): Likewise.
4265 (smin<mode>3_iwmmxt): Likewise.
4266 (umin<mode>3_iwmmxt): Likewise.
4267 (iwmmxt_wpackhss): Likewise.
4268 (iwmmxt_wpackwss): Likewise.
4269 (iwmmxt_wpackdss): Likewise.
4270 (iwmmxt_wpackhus): Likewise.
4271 (iwmmxt_wpackwus): Likewise.
4272 (iwmmxt_wpackdus): Likewise.
4273 (iwmmxt_wunpckihb): Likewise.
4274 (iwmmxt_wunpckihh): Likewise.
4275 (iwmmxt_wunpckihw): Likewise.
4276 (iwmmxt_wunpckilb): Likewise.
4277 (iwmmxt_wunpckilh): Likewise.
4278 (iwmmxt_wunpckilw): Likewise.
4279 (iwmmxt_wunpckehub): Likewise.
4280 (iwmmxt_wunpckehuh): Likewise.
4281 (iwmmxt_wunpckehuw): Likewise.
4282 (iwmmxt_wunpckehsb): Likewise.
4283 (iwmmxt_wunpckehsh): Likewise.
4284 (iwmmxt_wunpckehsw): Likewise.
4285 (iwmmxt_wunpckelub): Likewise.
4286 (iwmmxt_wunpckeluh): Likewise.
4287 (iwmmxt_wunpckeluw): Likewise.
4288 (iwmmxt_wunpckelsb): Likewise.
4289 (iwmmxt_wunpckelsh): Likewise.
4290 (iwmmxt_wunpckelsw): Likewise.
4291 (ror<mode>3): Likewise.
4292 (ashr<mode>3_iwmmxt): Likewise.
4293 (lshr<mode>3_iwmmxt): Likewise.
4294 (ashl<mode>3_iwmmxt): Likewise.
4295 (ror<mode>3_di): Likewise.
4296 (ashr<mode>3_di): Likewise.
4297 (lshr<mode>3_di): Likewise.
4298 (ashl<mode>3_di): Likewise.
4299 (iwmmxt_wmadds): Likewise.
4300 (iwmmxt_wmaddu): Likewise.
4301 (iwmmxt_tmia): Likewise.
4302 (iwmmxt_tmiaph): Likewise.
4303 (iwmmxt_tmiabb): Likewise.
4304 (iwmmxt_tmiatb): Likewise.
4305 (iwmmxt_tmiabt): Likewise.
4306 (iwmmxt_tmiatt): Likewise.
4307 (iwmmxt_tmovmskb): Likewise.
4308 (iwmmxt_tmovmskh): Likewise.
4309 (iwmmxt_tmovmskw): Likewise.
4310 (iwmmxt_waccb): Likewise.
4311 (iwmmxt_wacch): Likewise.
4312 (iwmmxt_waccw): Likewise.
4313 (iwmmxt_waligni): Likewise.
4314 (iwmmxt_walignr): Likewise.
4315 (iwmmxt_walignr0): Likewise.
4316 (iwmmxt_walignr1): Likewise.
4317 (iwmmxt_walignr2): Likewise.
4318 (iwmmxt_walignr3): Likewise.
4319 (iwmmxt_wsadb): Likewise.
4320 (iwmmxt_wsadh): Likewise.
4321 (iwmmxt_wsadbz): Likewise.
4322 (iwmmxt_wsadhz): Likewise.
4323 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4324 (iwmmxt_wabsdiffb): Likewise.
4325 (iwmmxt_wabsdiffh): Likewise.
4326 (iwmmxt_wabsdiffw): Likewise.
4327 (iwmmxt_waddsubhx): Likewise
4328 (iwmmxt_wsubaddhx): Likewise.
4329 (addc<mode>3): Likewise.
4330 (iwmmxt_avg4): Likewise.
4331 (iwmmxt_avg4r): Likewise.
4332 (iwmmxt_wmaddsx): Likewise.
4333 (iwmmxt_wmaddux): Likewise.
4334 (iwmmxt_wmaddsn): Likewise.
4335 (iwmmxt_wmaddun): Likewise.
4336 (iwmmxt_wmulwsm): Likewise.
4337 (iwmmxt_wmulwum): Likewise.
4338 (iwmmxt_wmulsmr): Likewise.
4339 (iwmmxt_wmulumr): Likewise.
4340 (iwmmxt_wmulwsmr): Likewise.
4341 (iwmmxt_wmulwumr): Likewise.
4342 (iwmmxt_wmulwl): Likewise.
4343 (iwmmxt_wqmulm): Likewise.
4344 (iwmmxt_wqmulwm): Likewise.
4345 (iwmmxt_wqmulmr): Likewise.
4346 (iwmmxt_wqmulwmr): Likewise.
4347 (iwmmxt_waddbhusm): Likewise.
4348 (iwmmxt_waddbhusl): Likewise.
4349 (iwmmxt_wqmiabb): Likewise.
4350 (iwmmxt_wqmiabt): Likewise.
4351 (iwmmxt_wqmiatb): Likewise.
4352 (iwmmxt_wqmiatt): Likewise.
4353 (iwmmxt_wqmiabbn): Likewise.
4354 (iwmmxt_wqmiabtn): Likewise.
4355 (iwmmxt_wqmiatbn): Likewise.
4356 (iwmmxt_wqmiattn): Likewise.
4357 (iwmmxt_wmiabb): Likewise.
4358 (iwmmxt_wmiabt): Likewise.
4359 (iwmmxt_wmiatb): Likewise.
4360 (iwmmxt_wmiatt): Likewise.
4361 (iwmmxt_wmiabbn): Likewise.
4362 (iwmmxt_wmiabtn): Likewise.
4363 (iwmmxt_wmiatbn): Likewise.
4364 (iwmmxt_wmiattn): Likewise.
4365 (iwmmxt_wmiawbb): Likewise.
4366 (iwmmxt_wmiawbt): Likewise.
4367 (iwmmxt_wmiawtb): Likewise.
4368 (iwmmxt_wmiawtt): Likewise.
4369 (iwmmxt_wmiawbbn): Likewise.
4370 (iwmmxt_wmiawbtn): Likewise.
4371 (iwmmxt_wmiawtbn): Likewise.
4372 (iwmmxt_wmiawttn): Likewise.
4373 (iwmmxt_wmerge): Likewise.
4374 (iwmmxt_tandc<mode>3): Likewise.
4375 (iwmmxt_torc<mode>3): Likewise.
4376 (iwmmxt_torvsc<mode>3): Likewise.
4377 (iwmmxt_textrc<mode>3): Likewise.
4378 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
4379 (wmmxt_pack): Likewise.
4380 (wmmxt_mult_c1): Likewise.
4381 (wmmxt_mult_c2): Likewise.
4382 (wmmxt_alu_c1): Likewise.
4383 (wmmxt_alu_c2): Likewise.
4384 (wmmxt_alu_c3): Likewise.
4385 (wmmxt_transfer_c1): Likewise.
4386 (wmmxt_transfer_c2): Likewise.
4387 (wmmxt_transfer_c3): Likewise.
4388 (marvell_f_iwmmxt_wstr): Likewise.
4389 (marvell_f_iwmmxt_wldr): Likewise.
4391 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
4393 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
4395 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4398 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4399 * lra-constraints.c (need_for_split_p): Check call used hard regs
4400 living through calls.
4402 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
4403 call used regs for call insn.
4405 2013-06-28 Jakub Jelinek <jakub@redhat.com>
4408 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
4409 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
4411 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
4413 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
4415 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
4417 * lra-constraints.c (need_for_split_p): Check call used hard regs
4418 living through calls.
4420 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
4423 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
4424 to tie with any other modes. Eliminate Altivec vector mode tests,
4425 since these are a subset of ALTIVEC or VSX vector modes. Simplify
4426 code, to return 0 if testing MODE2 for a condition, if we've
4427 already tested MODE1 for the same condition.
4429 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4431 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
4434 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4436 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4437 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
4439 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
4441 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
4443 (aarch64_symbolic_constant_p): Remove.
4444 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
4445 static. Fix line length and white space.
4446 (aarch64_symbolic_constant_p): Remove.
4447 * config/aarch64/predicates.md (aarch64_valid_symref):
4448 Use aarch64_classify_symbol_expression.
4450 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4452 * config/arm/constraints.md (Ts): New constraint.
4453 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
4455 (compare_scc): Use "Ts" constraint for operand 0.
4456 (ior_scc_scc): Likewise.
4457 (and_scc_scc): Likewise.
4458 (and_scc_scc_nodom): Likewise.
4459 (ior_scc_scc_cmp): Likewise for operand 7.
4460 (and_scc_scc_cmp): Likewise.
4461 * config/arm/thumb2.md (thumb2_movsi_insn):
4462 Add alternatives for 16-bit encodings.
4463 (thumb2_movhi_insn): Likewise.
4464 (thumb2_movsicc_insn): Likewise.
4465 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
4466 (thumb2_negscc): Use "Ts" constraint.
4467 Move mvn instruction outside cond_exec block.
4468 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
4469 for 16-bit encodings.
4471 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4473 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
4475 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
4476 (mulsi3subsi): Likewise.
4477 (mulsidi3adddi): Likewise.
4478 (mulsidi3_v6): Likewise.
4479 (umulsidi3_v6): Likewise.
4480 (umulsidi3adddi_v6): Likewise.
4481 (smulsi3_highpart_v6): Likewise.
4482 (umulsi3_highpart_v6): Likewise.
4483 (mulhisi3tb): Likewise.
4484 (mulhisi3bt): Likewise.
4485 (mulhisi3tt): Likewise.
4486 (maddhisi4): Likewise.
4487 (maddhisi4tb): Likewise.
4488 (maddhisi4tt): Likewise.
4489 (maddhidi4): Likewise.
4490 (maddhidi4tb): Likewise.
4491 (maddhidi4tt): Likewise.
4492 (zeroextractsi_compare0_scratch): Likewise.
4493 (insv_zero): Likewise.
4494 (insv_t2): Likewise.
4495 (anddi_notzesidi_di): Likewise.
4496 (anddi_notsesidi_di): Likewise.
4497 (andsi_notsi_si): Likewise.
4498 (iordi_zesidi_di): Likewise.
4499 (xordi_zesidi_di): Likewise.
4500 (andsi_iorsi3_notsi): Likewise.
4502 (smax_m1): Likewise.
4504 (not_shiftsi): Likewise.
4505 (unaligned_loadsi): Likewise.
4506 (unaligned_loadhis): Likewise.
4507 (unaligned_loadhiu): Likewise.
4508 (unaligned_storesi): Likewise.
4509 (unaligned_storehi): Likewise.
4510 (extv_reg): Likewise.
4511 (extzv_t2): Likewise.
4513 (udivsi3): Likewise.
4514 (arm_zero_extendhisi2addsi): Likewise.
4515 (arm_zero_extendqisi2addsi): Likewise.
4516 (compareqi_eq0): Likewise.
4517 (arm_extendhisi2_v6): Likewise.
4518 (arm_extendqisi2addsi): Likewise.
4519 (arm_movt): Likewise.
4520 (thumb2_ldrd): Likewise.
4521 (thumb2_ldrd_base): Likewise.
4522 (thumb2_ldrd_base_neg): Likewise.
4523 (thumb2_strd): Likewise.
4524 (thumb2_strd_base): Likewise.
4525 (thumb2_strd_base_neg): Likewise.
4526 (arm_negsi2): Add alternative for 16-bit encoding.
4527 (arm_one_cmplsi2): Likewise.
4529 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4531 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
4532 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
4533 (movdfcc): Likewise.
4534 * config/arm/vfp.md (*thumb2_movsf_vfp):
4535 Disable predication for arm_restrict_it.
4536 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
4537 (*thumb2_movdfcc_vfp): Likewise.
4538 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
4539 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
4540 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
4541 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
4542 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
4543 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
4544 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
4545 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
4546 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
4547 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
4548 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
4549 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
4550 Disable predication for arm_restrict_it.
4552 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
4554 * config/i386/bmiintrin.h (_bextr_u32): New.
4555 (_bextr_u64): Ditto.
4557 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
4559 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
4560 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
4561 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
4562 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
4563 * config/mips/n32-elf.h: ...this new file.
4565 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
4568 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
4569 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
4571 2013-06-27 Catherine Moore <clm@codesourcery.com>
4573 * config/mips/mips-tables.opt: Regenerate.
4574 * config/mips/mips-cpus.def: Add m14ke and m14kec.
4575 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
4576 * doc/invoke.texi: Add -m14kc.
4578 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4581 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
4582 constraints of operand 1 and 2.
4585 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
4586 to match RTL canonicalization. Swap predicates and
4587 constraints of operand 1 and 2.
4589 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
4591 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
4592 Process OP_INOUT regs for splitting too.
4594 2013-06-27 Jakub Jelinek <jakub@redhat.com>
4596 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
4597 decl before the loop, initialize to NULL.
4598 (vectorizable_load): Initialize ptr_incr to NULL.
4600 2013-06-27 Martin Jambor <mjambor@suse.cz>
4603 * ipa-ref.h (ipa_maybe_record_reference): Declare.
4604 * ipa-ref.c (ipa_maybe_record_reference): New function.
4605 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
4606 * ipa-cp.c (create_specialized_node): Record potential references from
4608 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
4610 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4612 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
4613 parameter 'mode' of type 'enum machine_mode mode'; change to pass
4614 'mode' to force_reg.
4615 (aarch64_add_offset): Update calls to aarch64_force_temporary.
4616 (aarch64_expand_mov_immediate): Likewise.
4618 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
4620 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
4621 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
4623 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4625 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
4626 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4627 (struct machine_function): Add tbegin_p.
4628 (s390_canonicalize_comparison): Fold CC mode compares to
4629 conditional jump if possible.
4630 (s390_emit_jump): Return the emitted jump.
4631 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
4632 Handle CCRAWmode compares.
4633 (s390_option_override): Default to -mhtm if available.
4634 (s390_reg_clobbered_rtx): Handle floating point regs as well.
4635 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
4636 FPRs instead of df_regs_ever_live_p.
4637 (s390_optimize_nonescaping_tx): New function.
4638 (s390_init_frame_layout): Extend clobbered_regs array to cover
4640 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
4641 (s390_expand_tbegin): New function.
4642 (enum s390_builtin): New enum definition.
4643 (code_for_builtin): New array definition.
4644 (s390_init_builtins): New function.
4645 (s390_expand_builtin): New function.
4646 (TARGET_INIT_BUILTINS): Define.
4647 (TARGET_EXPAND_BUILTIN): Define.
4648 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
4649 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
4650 (s390_alc_comparison): Likewise.
4651 * config/s390/s390-modes.def: Add CCRAWmode.
4652 * config/s390/s390.h (processor_flags): Add PF_TX.
4653 (TARGET_CPU_HTM): Define macro.
4654 (TARGET_HTM): Define macro.
4655 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
4656 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
4657 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
4658 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
4659 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
4660 (TBEGIN_MASK, TBEGINC_MASK): New constants.
4661 ("*cc_to_int"): Move up.
4662 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
4663 constants other than 0.
4664 ("*ccraw_to_int"): New insn and splitter definition.
4665 ("tbegin", "tbegin_nofloat", "tbegin_retry")
4666 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
4667 ("tx_assist"): New expander.
4668 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
4669 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
4670 * config/s390/s390.opt: Add -mhtm option.
4671 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
4672 * config/s390/htmxlintrin.h: New file.
4673 * config/s390/htmintrin.h: New file.
4674 * config/s390/s390intrin.h: New file.
4675 * doc/extend.texi: Document htm builtins.
4676 * config.gcc: Add the new header files to extra_headers.
4678 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
4680 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
4681 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
4683 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
4684 Pat Haugen <pthaugen@us.ibm.com>
4685 Peter Bergner <bergner@vnet.ibm.com>
4687 * config/rs6000/power8.md: New.
4688 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
4689 setting for power8 entry.
4690 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
4691 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
4692 test for Power4/Power5 only.
4693 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
4695 (force_new_group): Adjust comment.
4696 * config/rs6000/rs6000.md: Include power8.md.
4698 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
4700 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
4701 * config/arm/arm-protos.h (arm_max_conditional_execute): New
4703 (tune_params): Update comment.
4704 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
4705 (arm_max_conditional_execute): New function.
4706 (thumb2_final_prescan_insn): Use max_insn_skipped and
4707 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
4709 2013-06-25 Jakub Jelinek <jakub@redhat.com>
4711 PR tree-optimization/57705
4712 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
4713 SSA_NAME step, provided that it is not defined inside the loop.
4714 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
4715 (get_initial_def_for_induction): Handle SSA_NAME IV step.
4717 2013-06-25 Martin Jambor <mjambor@suse.cz>
4720 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
4721 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
4723 (ipa_note_param_call): Initialize member_ptr flag.
4724 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
4725 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
4726 (ipa_write_indirect_edge_info): Stream member_ptr flag.
4727 (ipa_read_indirect_edge_info): Likewise.
4729 2013-06-25 Richard Biener <rguenther@suse.de>
4732 * passes.c (init_optimization_passes): Move pass_fold_builtins
4733 and pass_dce earlier with -Og.
4735 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
4737 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
4738 <BIT_FIELD_REF>: Remove trailing TAB.
4739 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
4742 2013-06-24 Martin Jambor <mjambor@suse.cz>
4744 PR tree-optimization/57358
4745 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
4746 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
4747 (ipa_analyze_params_uses): Generate pessimistic info when true.
4749 2013-06-24 Martin Jambor <mjambor@suse.cz>
4751 PR tree-optimization/57539
4752 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
4753 global.inlined_to of the new node to it. All callers changed.
4754 * ipa-inline-transform.c (clone_inlined_nodes): New variable
4755 inlining_into, pass it to cgraph_clone_node.
4756 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
4757 ipa_free_edge_args_substructures.
4758 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
4759 rdesc linked list. Do not assert rdesc edges have inlined caller.
4760 Assert we have found an rdesc in the rdesc list.
4762 2013-06-24 Richard Biener <rguenther@suse.de>
4764 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
4765 (pointer_set_lookup): Declare.
4766 (class pointer_map): New template class implementing a
4767 generic pointer to T map.
4768 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
4769 pointer_map<T>::contains, pointer_map<T>::insert,
4770 pointer_map<T>::traverse): New functions.
4771 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
4772 (pointer_set_lookup): New function.
4773 (pointer_set_contains): Use pointer_set_lookup.
4774 (pointer_set_insert): Likewise.
4775 (insert_aux): Remove.
4776 (struct pointer_map_t): Embed a pointer_set_t.
4777 (pointer_map_create): Adjust.
4778 (pointer_map_destroy): Likewise.
4779 (pointer_map_contains): Likewise.
4780 (pointer_map_insert): Likewise.
4781 (pointer_map_traverse): Likewise.
4782 * tree-streamer.h (struct streamer_tree_cache_d): Use a
4783 pointer_map<unsigned> instead of a pointer_map_t.
4784 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
4785 (streamer_tree_cache_lookup): Likewise.
4786 (streamer_tree_cache_create): Likewise.
4787 (streamer_tree_cache_delete): Likewise.
4788 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
4789 pointer_map<unsigned> instead of a pointer_map_t.
4790 (lto_init_tree_ref_encoder): Adjust.
4791 (lto_destroy_tree_ref_encoder): Likewise.
4792 * lto-section-out.c (lto_output_decl_index): Likewise.
4793 (lto_record_function_out_decl_state): Likewise.
4794 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
4796 2013-06-24 Richard Biener <rguenther@suse.de>
4798 PR tree-optimization/57488
4799 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
4801 2013-06-24 Alan Modra <amodra@gmail.com>
4803 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
4804 (gen_easy_altivec_constant): Likewise.
4805 * config/rs6000/predicates.md (easy_vector_constant_add_self,
4806 easy_vector_constant_msb): Likewise.
4808 2013-06-23 Jakub Jelinek <jakub@redhat.com>
4811 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
4812 add missing return true.
4814 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
4817 * config/sh/predicates.md (general_extend_operand): Invoke
4818 general_movsrc_operand for memory operands.
4819 (general_movsrc_operand): Allow reg+reg addressing, do not use
4820 general_operand for memory operands.
4822 2013-06-23 Sriraman Tallam <tmsriram@google.com>
4824 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
4825 when current target options does not apply.
4826 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
4827 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
4828 * config/i386/bmiintrin.h: Pass appropriate target
4829 attributes to header.
4830 * config/i386/mmintrin.h: Ditto.
4831 * config/i386/nmmintrin.h: Ditto.
4832 * config/i386/avx2intrin.h: Ditto.
4833 * config/i386/fxsrintrin.h: Ditto.
4834 * config/i386/tbmintrin.h: Ditto.
4835 * config/i386/xsaveintrin.h: Ditto.
4836 * config/i386/f16cintrin.h: Ditto.
4837 * config/i386/xtestintrin.h: Ditto.
4838 * config/i386/xsaveoptintrin.h: Ditto.
4839 * config/i386/bmi2intrin.h: Ditto.
4840 * config/i386/lzcntintrin.h: Ditto.
4841 * config/i386/smmintrin.h: Ditto.
4842 * config/i386/wmmintrin.h: Ditto.
4843 * config/i386/x86intrin.h: Remove all header include guards.
4844 * config/i386/prfchwintrin.h: Ditto.
4845 * config/i386/pmmintrin.h: Ditto.
4846 * config/i386/tmmintrin.h: Ditto.
4847 * config/i386/xmmintrin.h: Ditto.
4848 * config/i386/popcntintrin.h: Ditto.
4849 * config/i386/rdseedintrin.h: Ditto.
4850 * config/i386/ammintrin.h: Ditto.
4851 * config/i386/emmintrin.h: Ditto.
4852 * config/i386/immintrin.h: Remove all header include guards.
4853 * config/i386/fma4intrin.h: Ditto.
4854 * config/i386/lwpintrin.h: Ditto.
4855 * config/i386/xopintrin.h: Ditto.
4856 * config/i386/ia32intrin.h: Ditto.
4857 * config/i386/avxintrin.h: Ditto.
4858 * config/i386/rtmintrin.h: Ditto.
4859 * config/i386/fmaintrin.h: Ditto.
4860 * config/i386/mm3dnow.h: Ditto.
4862 2013-06-22 Sriraman Tallam <tmsriram@google.com>
4864 * common/config/i386/i386-common.c: Handle LZCNT.
4866 2013-06-22 Andi Kleen <ak@linux.intel.com>
4868 * doc/extend.texi: Use __atomic_store_n instead of
4869 __atomic_store in HLE example.
4871 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
4873 * config/sh/sh.c: Remove <cstdlib> workaround.
4875 2013-06-21 Andi Kleen <ak@linux.intel.com>
4877 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
4879 2013-06-21 Andi Kleen <ak@linux.intel.com>
4881 * doc/extend.texi: Document that __atomic_clear and
4882 __atomic_test_and_set should only be used with bool.
4884 2013-06-20 Jan Hubicka <jh@suse.cz>
4886 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
4888 * tree.c (decls_same_for_odr): New function.
4889 (same_for_edr): New function.
4890 (types_same_for_odr): New function.
4891 (get_binfo_at_offset): Use it.
4892 * tree.h (types_same_for_odr): Declare.
4894 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
4895 Jason Merrill <jason@redhat.com>
4897 * system.h: Include <cstdlib> as well as <stdlib.h>.
4899 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
4902 * config/i386/i386.c (construct_container): Report error if
4903 long double is used with disabled x87 float returns.
4905 2013-06-20 Jan Hubicka <jh@suse.cz>
4907 * lto-cgraph.c (input_symtab): Do not set cgraph state.
4909 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
4911 PR rtl-optimization/57425
4912 PR rtl-optimization/57569
4913 * alias.c (write_dependence_p): Remove parameters mem_mode and
4914 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
4915 Changed all callers.
4916 (canon_anti_dependence): Get comments and semantics in sync.
4917 Add parameter mem_canonicalized. Changed all callers.
4918 * rtl.h (canon_anti_dependence): Update prototype.
4920 2013-06-20 Richard Biener <rguenther@suse.de>
4922 * data-streamer-in.c (streamer_read_uhwi): Optimize single
4923 byte case, inline streamer_read_uchar and defer section
4926 2013-06-20 Richard Biener <rguenther@suse.de>
4928 PR tree-optimization/57584
4929 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
4930 SSA names into the expanded expression that take part in
4931 abnormal coalescing.
4933 2013-06-19 Sharad Singhai <singhai@google.com>
4935 * gcov.c (print_usage): Handle new option.
4936 (process_args): Ditto.
4937 (get_gcov_intermediate_filename): New function.
4938 (output_intermediate_file): New function.
4939 (output_gcov_file): New function
4940 (generate_results): Handle new option.
4941 (release_function): Relase demangled name.
4942 (read_graph_file): Handle demangled name.
4943 (output_lines): Ditto.
4944 * doc/gcov.texi: Document gcov intermediate format.
4946 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
4949 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
4950 (lra_emit_add): Use the functions. Add comment about Y as an
4953 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
4956 * collect2.c (collect_atexit): New.
4957 (collect_exit): Delete.
4958 (main): Register collect_atexit with atexit.
4959 (collect_wait): Change collect_exit to exit.
4961 * collect2.h (collect_exit): Delete.
4962 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
4964 2013-06-19 Wei Mi <wmi@google.com>
4966 PR rtl-optimization/57518
4967 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
4968 if regno is used in paradoxical subreg.
4969 (update_equiv_regs): Check pdx_subregs[regno] before
4970 set a reg to be equivalent with a mem.
4972 2013-06-19 Matthias Klose <doko@ubuntu.com>
4975 * file-find.h (find_a_file): Add a mode parameter.
4976 * file-find.c (find_a_file): Likewise.
4977 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
4978 with X_OK for the executables.
4979 * collect2.c (main): Call find_a_file with X_OK.
4981 2013-06-19 Steve Ellcey <sellcey@mips.com>
4984 * config/mips/mips.md (casesi_internal_mips16_<mode>):
4985 Use NEXT_INSN instead of next_real_insn.
4987 2013-06-19 Jan Hubicka <jh@suse.cz>
4989 * cgraph.h (const_value_known_p): Replace by ...
4990 (ctor_for_folding): .. this one.
4991 * cgraphunit.c (process_function_and_variable_attributes): Use it.
4992 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
4993 * expr.c (expand_expr_real_1): Likewise.
4994 (string_constant): Likewise.
4995 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
4996 * ipa.c (process_references): Likewise.
4997 (symtab_remove_unreachable_nodes): Likewise.
4998 * ipa-inline-analysis.c (param_change_prob): Likewise.
4999 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5000 (get_base_constructor): Likwise.
5001 * varpool.c (varpool_remove_node): Likewise.
5002 (varpool_remove_initializer): LIkewise.
5003 (dump_varpool_node): LIkwise.
5004 (const_value_known_p): Rewrite to ...
5005 (ctor_for_folding): ... this one.
5007 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5010 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5011 PERSONALITY in $PATH derived prefixes.
5013 2013-06-19 Jeff Law <law@redhat.com>
5015 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5018 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5019 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5022 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5024 * config/arm/vfp.md: Move VFP instruction classification documentation
5026 * config/arm/arm.md: ... here. Update instruction classification
5029 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5031 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5032 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5033 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5035 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5038 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5039 with NEXT_INSN. Use tablejump_p to check for jump table data
5042 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5045 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5046 that now in C++ the value is correct per the C++ standards.
5048 2013-06-19 Richard Biener <rguenther@suse.de>
5050 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5053 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5056 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5059 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5060 with next_active_insn.
5062 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5064 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5065 functions are inlined during failures to flag an error.
5066 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5067 in early inline pass.
5069 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5071 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5074 2013-06-18 Julian Brown <julian@codesourcery.com>
5076 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5077 Permit virtual register pre-reload if !strict.
5078 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5080 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5082 * config/arm/neon.md (movmisalign<mode>): Use
5083 neon_perm_struct_or_reg_operand instead of
5084 neon_struct_or_register_operand.
5085 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5086 neon_permissive_struct_operand instead of neon_struct_operand.
5087 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5088 neon_vector_mem_operand.
5089 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5090 neon_vector_mem_operand.
5091 (neon_permissive_struct_operand): New.
5092 (neon_struct_or_register_operand): Rename to...
5093 (neon_perm_struct_or_reg_operand): This. Adjust call to
5094 neon_vector_mem_operand.
5096 2013-06-18 Richard Biener <rguenther@suse.de>
5098 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5099 * lto-streamer.h: Include pointer-set.h.
5100 (struct lto_decl_slot): Remove.
5101 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5102 Remove next_index entry.
5103 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5104 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5105 (lto_init_tree_ref_encoder): Adjust.
5106 (lto_destroy_tree_ref_encoder): Likewise.
5107 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5108 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5109 (lto_output_decl_index): Adjust.
5110 (lto_new_out_decl_state): Likewise.
5111 (lto_record_function_out_decl_state): Likewise.
5112 * lto-streamer-out.c (copy_function): Likewise.
5114 2013-06-18 Richard Biener <rguenther@suse.de>
5116 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5117 * cgraphunit.c: Include cfgloop.h.
5118 (init_lowered_empty_function): Initialize the loop tree.
5119 (assemble_thunk): Insert new BBs into loops.
5121 2013-06-18 Richard Biener <rguenther@suse.de>
5123 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5124 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5125 the map from cache entry to cache index optional.
5126 (streamer_tree_cache_replace_tree): Adjust accordingly.
5127 (streamer_tree_cache_append): Likewise.
5128 (streamer_tree_cache_delete): Likewise.
5129 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5130 streamer cache map from cache entry to cache index.
5131 * lto-streamer-out.c (create_output_block): Adjust.
5133 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5135 * config/arm/arm.md (attribute "insn"): Move multiplication and
5136 division attributes to...
5137 (attribute "type"): ... here. Remove mult.
5138 (attribute "mul32"): New attribute.
5139 (attribute "mul64"): Add umaal.
5140 (*arm_mulsi3): Update attributes.
5141 (*arm_mulsi3_v6): Likewise.
5142 (*thumb_mulsi3): Likewise.
5143 (*thumb_mulsi3_v6): Likewise.
5144 (*mulsi3_compare0): Likewise.
5145 (*mulsi3_compare0_v6): Likewise.
5146 (*mulsi_compare0_scratch): Likewise.
5147 (*mulsi_compare0_scratch_v6): Likewise.
5148 (*mulsi3addsi): Likewise.
5149 (*mulsi3addsi_v6): Likewise.
5150 (*mulsi3addsi_compare0): Likewise.
5151 (*mulsi3addsi_compare0_v6): Likewise.
5152 (*mulsi3addsi_compare0_scratch): Likewise.
5153 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5154 (*mulsi3subsi): Likewise.
5155 (*mulsidi3adddi): Likewise.
5156 (*mulsi3addsi_v6): Likewise.
5157 (*mulsidi3adddi_v6): Likewise.
5158 (*mulsidi3_nov6): Likewise.
5159 (*mulsidi3_v6): Likewise.
5160 (*umulsidi3_nov6): Likewise.
5161 (*umulsidi3_v6): Likewise.
5162 (*umulsidi3adddi): Likewise.
5163 (*umulsidi3adddi_v6): Likewise.
5164 (*smulsi3_highpart_nov6): Likewise.
5165 (*smulsi3_highpart_v6): Likewise.
5166 (*umulsi3_highpart_nov6): Likewise.
5167 (*umulsi3_highpart_v6): Likewise.
5168 (mulhisi3): Likewise.
5169 (*mulhisi3tb): Likewise.
5170 (*mulhisi3bt): Likewise.
5171 (*mulhisi3tt): Likewise.
5172 (maddhisi4): Likewise.
5173 (*maddhisi4tb): Likewise.
5174 (*maddhisi4tt): Likewise.
5175 (maddhidi4): Likewise.
5176 (*maddhidi4tb): Likewise.
5177 (*maddhidi4tt): Likewise.
5179 (udivsi3): Likewise.
5180 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5181 (thumb2_mulsi_short_compare0): Likewise.
5182 (thumb2_mulsi_short_compare0_scratch): Likewise.
5183 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5184 (1020mult2): Likewise.
5185 (1020mult3): Likewise.
5186 (1020mult4): Likewise.
5187 (1020mult5): Likewise.
5188 (1020mult6): Likewise.
5189 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5191 (cortex_a15_mult64): Likewise.
5192 (cortex_a15_sdiv): Likewise.
5193 (cortex_a15_udiv): Likewise.
5194 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5200 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5201 (pj4_ir_div): Likewise.
5202 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5203 (11_mult2): Likewise.
5204 (11_mult3): Likewise.
5205 (11_mult4): Likewise.
5206 (11_mult5): Likewise.
5207 (11_mult6): Likewise.
5208 (11_mult7): Likewise.
5209 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5210 (cortex_a8_mla): Likewise.
5211 (cortex_a8_mull): Likewise.
5212 (cortex_a8_smulwy): Likewise.
5213 (cortex_a8_smlald): Likewise.
5214 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5215 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5216 (cortex_r4_mul_3): Likewise.
5217 (cortex_r4_mla_4): Likewise.
5218 (cortex_r4_mla_3): Likewise.
5219 (cortex_r4_smlald): Likewise.
5220 (cortex_r4_mull): Likewise.
5221 (cortex_r4_sdiv): Likewise.
5222 (cortex_r4_udiv): Likewise.
5223 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5224 (cortex_a7_idiv): Likewise.
5225 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5226 (9_mult2): Likewise.
5227 (9_mult3): Likewise.
5228 (9_mult4): Likewise.
5229 (9_mult5): Likewise.
5230 (9_mult6): Likewise.
5231 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5232 (cortex_a53_sdiv): Likewise.
5233 (cortex_a53_udiv): Likewise.
5234 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5235 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5236 (mp626_mult2): Likewise.
5237 (mp626_mult3): Likewise.
5238 (mp626_mult4): Likewise.
5239 * config/arm/fa526.md (526_mult1): Update attribute change.
5240 (526_mult2): Likewise.
5241 * config/arm/arm-generic.md (mult): Update attribute change.
5242 (mult_ldsched_strongarm): Likewise.
5243 (mult_ldsched): Likewise.
5244 (multi_cycle): Likewise.
5245 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5246 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5247 (606te_mult2): Likewise.
5248 (606te_mult3): Likewise.
5249 (606te_mult4): Likewise.
5250 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5251 (cortex_a9_mac16): Likewise.
5252 (cortex_a9_multiply): Likewise.
5253 (cortex_a9_mac): Likewise.
5254 (cortex_a9_multiply_long): Likewise.
5255 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5256 (626te_mult2): Likewise.
5257 (626te_mult3): Likewise.
5258 (626te_mult4): Likewise.
5260 2013-06-18 Richard Biener <rguenther@suse.de>
5263 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5265 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5268 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5269 with next_active_insn.
5271 2013-06-18 Alan Modra <amodra@gmail.com>
5273 * config/rs6000/rs6000.h (enum data_align): New.
5274 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5275 (DATA_ABI_ALIGNMENT): Define.
5276 (CONSTANT_ALIGNMENT): Correct comment.
5277 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5278 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5280 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5282 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5283 ATTRIBUTE_UNUSED marking.
5285 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5287 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5288 alternative and update.
5289 (aarch64_dup_lanedi): Delete.
5290 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5291 * config/aarch64/aarch64-simd-builtins.def: Update.
5293 2013-06-17 Richard Biener <rguenther@suse.de>
5295 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5296 (lto_input_scc): Declare.
5297 (lto_input_tree_1): Likewise.
5298 (struct lto_stats_d): Add num_tree_bodies_output and
5299 num_pickle_refs_output.
5300 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5301 (lto_read_tree_1): Split out from ...
5302 (lto_read_tree): ... this.
5303 (lto_input_scc): New function.
5304 (lto_input_tree_1): Split out from ...
5305 (lto_input_tree): ... this. Handle LTO_tree_scc.
5306 (lto_data_in_create): Create the streamer cache without hashes.
5307 * lto-streamer-out.c (create_output_block): Create the streamer
5308 cache with hashes when not doing WPA.
5309 (lto_write_tree_1): Split out from ...
5310 (lto_write_tree): ... this.
5311 (get_symbol_initial_value): New function.
5312 (lto_output_tree_1): Split out from ...
5313 (lto_output_tree): ... this. Write trees as series of SCCs
5314 using a DFS walk via DFS_write_tree.
5315 (struct sccs, struct scc_entry): New types.
5316 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5317 (DFS_write_tree_body): New function.
5318 (DFS_write_tree): Likewise.
5319 (hash_tree): Likewise.
5320 (scc_entry_compare): Likewise.
5321 (hash_scc): Likewise.
5322 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5323 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5324 TREE_CHAIN as regular reference.
5325 (streamer_read_integer_cst): Remove.
5326 (streamer_get_pickled_tree): Adjust.
5327 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5328 of DECL_EXTERNALs in BLOCK_VARS for now.
5329 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5331 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5332 Add hash value argument and record that if hashes are recorded
5334 (streamer_tree_cache_insert_1): Adjust.
5335 (streamer_tree_cache_insert): Likewise.
5336 (streamer_tree_cache_insert_at): Rename to ...
5337 (streamer_tree_cache_replace_tree): ... this and adjust.
5338 (streamer_tree_cache_append): Adjust.
5339 (record_common_node): Likewise.
5340 (streamer_tree_cache_create): Add argument whether to
5341 record hash values together with trees.
5342 (streamer_tree_cache_delete): Adjust.
5343 * tree-streamer.h (struct streamer_tree_cache_d): Add
5345 (streamer_read_integer_cst): Remove.
5346 (streamer_tree_cache_insert): Adjust.
5347 (streamer_tree_cache_append): Likewise.
5348 (streamer_tree_cache_insert_at): Rename to ...
5349 (streamer_tree_cache_replace_tree): ... this and adjust.
5350 (streamer_tree_cache_create): Add argument whether to record hashes.
5351 (streamer_tree_cache_get): Rename to ...
5352 (streamer_tree_cache_get_tree): ... this.
5353 (streamer_tree_cache_get_hash): New function.
5354 * tree.c (cache_integer_cst): New function.
5355 * tree.h (cache_integer_cst): Declare.
5356 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5357 * lto-symtab.c (lto_varpool_replace_node): Only release
5358 DECL_INITIAL of non-prevailing decls.
5359 * varpool.c (varpool_remove_initializer): Do not release
5360 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5362 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5364 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5365 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5366 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5367 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5368 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5369 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5370 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5371 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5372 instead of TARGET_64BIT.
5373 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5374 Require ISA_HAS_<D>DIV.
5376 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
5378 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
5379 (mips*-*-linux*): Move default with_llsc setting to where other
5381 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
5383 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
5384 Likewise. Remove default with_tune setting. Move default float
5385 setting to its own block. Handle with_llsc in the same block as above.
5387 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
5389 PR rtl-optimization/57425
5390 PR rtl-optimization/57569
5391 * alias.c (write_dependence_p): Add new parameters mem_mode,
5392 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
5393 Changed all callers.
5394 (canon_anti_dependence): New function.
5395 * cse.c (check_dependence): Use canon_anti_dependence.
5396 * cselib.c (cselib_invalidate_mem): Likewise.
5397 * rtl.h (canon_anti_dependence): Declare.
5399 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5401 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
5402 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
5403 ".set mips3" for 64-bit targets.
5405 2013-06-15 Dehao Chen <dehao@google.com>
5407 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
5408 * gimple-low.c (gimple_check_call_matching_types): Likewise.
5409 (gimple_check_call_args): Likewise.
5410 * value-prof.c (check_ic_target): Likewise.
5411 * ipa-inline.c (early_inliner): Likewise.
5412 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
5413 * cgraph.c (cgraph_create_edge_1): Likewise.
5414 (cgraph_make_edge_direct): Likewise.
5416 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
5419 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
5420 rs6000_output_move_128bit to handle emitting quad memory
5421 operations. Set attribute length to 8 bytes.
5423 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
5425 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
5427 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
5428 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
5429 (aarch64_<su>mlsl<mode>): Likewise.
5431 2013-06-14 Mike Stump <mikestump@comcast.net>
5433 * Makefile.in (TARGET_H): Add insn-codes.h.
5435 2013-06-14 Alan Modra <amodra@gmail.com>
5439 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
5440 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
5441 bitfield expansion when EXPAND_MEMORY.
5442 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
5444 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
5446 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
5447 test for clearing quad memory on 32-bit later.
5449 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5451 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
5452 (fold_negate_expr): Likewise.
5453 (fold_real_zero_addition_p): Handle vectors.
5454 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
5456 2013-06-14 Alan Modra <amodra@gmail.com>
5458 * varasm.c (force_const_mem): Revert 2013-06-07 change.
5460 2013-06-13 Jan Hubicka <jh@suse.cz>
5462 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
5463 Local comdats are not externally visible.
5464 * symtab.c (dump_symtab_base): Dump externally visible.
5465 (verify_symtab_base): Verify back links in the symtab hash.
5467 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5469 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
5470 CONVERT_EXPR as equal nodes.
5472 2013-06-13 Bin Cheng <bin.cheng@arm.com>
5474 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
5476 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
5478 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
5479 Generalize to complex and vector.
5480 * tree.c (build_all_ones_cst): New function.
5481 * tree.h (build_all_ones_cst): Declare it.
5483 2013-06-13 Alan Modra <amodra@gmail.com>
5485 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
5486 * config/rs6000/rs6000.md (signbittf2): New insn.
5487 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
5488 (abstf2_internal, cmptf_internal2): Likewise.
5489 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
5491 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5492 Pat Haugen <pthaugen@us.ibm.com>
5493 Peter Bergner <bergner@vnet.ibm.com>
5495 * config/rs6000/rs6000.c (emit_load_locked): Add support for
5496 power8 byte, half-word, and quad-word atomic instructions.
5497 (emit_store_conditional): Likewise.
5498 (rs6000_expand_atomic_compare_and_swap): Likewise.
5499 (rs6000_expand_atomic_op): Likewise.
5501 * config/rs6000/sync.md (larx): Add new modes for power8.
5503 (AINT): New mode iterator to include TImode as well as normal
5504 integer modes on power8.
5505 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
5506 that VSX registers are not considered. Use AINT mode iterator
5507 instead of INT1 to allow inclusion of quad word atomic operations
5509 (load_locked<mode>): Likewise.
5510 (store_conditional<mode>): Likewise.
5511 (atomic_compare_and_swap<mode>): Likewise.
5512 (atomic_exchange<mode>): Likewise.
5513 (atomic_nand<mode>): Likewise.
5514 (atomic_fetch_<fetchop_name><mode>): Likewise.
5515 (atomic_nand_fetch<mode>): Likewise.
5516 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
5518 (ATOMIC): On power8, add QImode, HImode modes.
5519 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
5520 modes that promote to SImode.
5521 (load_lockedti): Convert TImode arguments to PTImode, so that we
5522 get a guaranteed even/odd register pair.
5523 (load_lockedpti): Likewise.
5524 (store_conditionalti): Likewise.
5525 (store_conditionalpti): Likewise.
5527 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
5528 atomic load/store instructions.
5531 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
5533 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
5535 (insn_count): New attribute, with most cases extracted from...
5536 (length): ...here. Redefine most cases in terms of insn_count.
5537 (single_insn): Delete.
5538 (can_delay): Use insn_count to check for single instructions.
5539 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
5540 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
5541 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
5542 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
5543 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
5544 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
5545 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
5546 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
5547 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
5548 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
5549 rather than "length".
5550 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
5551 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
5552 Use "insn_count" rather than "length".
5553 * config/mips/mips-dsp.md
5554 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
5555 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
5558 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
5560 PR tree-optimization/57361
5561 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
5563 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
5565 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
5567 (aarch64_simd_combine<mode>): New instruction expansion.
5568 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
5570 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
5571 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
5573 2013-06-12 Jan Hubicka <jh@suse.cz>
5575 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
5576 decl has when in streaming stage.
5577 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
5578 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
5580 2013-06-12 Roland Stigge <stigge@antcom.de>
5583 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
5585 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5587 PR tree-optimization/57537
5588 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
5589 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
5591 2013-06-12 Richard Biener <rguenther@suse.de>
5593 * data-streamer.h (streamer_write_char_stream): CSE
5594 obs->current_pointer.
5595 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
5596 streamer_write_char_stream manually and optimize the resulting loop.
5597 (streamer_write_hwi_stream): Likewise.
5599 2013-06-12 Jan Hubicka <jh@suse.cz>
5601 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
5602 * cgraph.h (varpool_create_empty_node): Declare.
5603 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
5605 * symtab.c (symtab_unregister_node): Be lax about missin entries
5607 (symtab_get_node): Update comment.
5608 * varpool.c (varpool_create_empty_node): Break out from ...
5609 (varpool_node_for_decl): ... here.
5610 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
5612 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
5614 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
5615 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
5616 part. Use straight-line flow at the end.
5617 <COMPONENT_REF>: Remove superfluous else.
5618 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
5620 2013-06-12 Jakub Jelinek <jakub@redhat.com>
5623 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
5624 target hook even for !TREE_PUBLIC decls. If no resolution info
5625 is available, return false for common and external decls.
5627 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
5629 * config/rl78/constraints.md (U): New constraint.
5630 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
5633 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5636 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
5637 to allow returning address to AT_PLATFORM name.
5639 2013-06-11 Jan Hubicka <jh@suse.cz>
5641 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
5642 * cgraph.h (symtab_node_base): Add weakref flag.
5643 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
5644 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
5645 (output_weakrefs): Use weakref flag.
5646 * fold-const.c (simple_operand_p): Handle WEAK.
5647 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
5648 * ipa.c (varpool_externally_visible_p): Drop weakref.
5649 (function_and_variable_visibility): Update comment; fix weakref
5650 sanity checks; do not clear DECL_WEAK on them.
5651 * lto-cgraph.c (lto_output_node): update.
5652 (lto_output_varpool_node): Update.
5653 (input_overwrite_node): Update.
5654 (input_node): Update.
5655 (input_varpool_node): Update.
5656 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
5657 (lto_symtab_merge_symbols): Add sanity check.
5658 (lto_symtab_prevailing_decl): Do not special case weakrefs.
5659 * passes.c (rest_of_decl_compilation): Set static flag, too.
5660 * symtab.c (dump_symtab_base): Dump weakref.
5661 (verify_symtab_base): Sanity check weakrefs.
5662 (symtab_make_decl_local): Remove duplicated code.
5663 (symtab_alias_ultimate_target): Simplify.
5664 * varpool.c (varpool_create_variable_alias): Set weakref flag.
5666 2013-06-11 DJ Delorie <dj@redhat.com>
5668 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
5669 (rl78_unwind_word_mode): New.
5671 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5673 * final.c (debug_prefix_maps): Make static.
5675 2013-06-11 David Malcolm <dmalcolm@redhat.com>
5677 * function.c (initial_trampoline): Remove stray copy.
5679 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
5681 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
5683 2013-06-11 Martin Jambor <mjambor@suse.cz>
5685 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
5686 within bounds at the beginning of the function.
5688 2013-06-11 Alan Modra <amodra@gmail.com>
5690 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
5692 (get_named_section): Don't NULL !DECL_P decl.
5694 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
5696 * doc/invoke.texi (core-avx2): Document.
5698 (atom): Updated with MOVBE.
5700 2013-06-11 Richard Biener <rguenther@suse.de>
5702 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
5704 2013-06-11 Anton Blanchard <anton@samba.org>
5706 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
5707 correct shift value in little-endian mode.
5709 2013-06-11 Jakub Jelinek <jakub@redhat.com>
5712 * varasm.c (get_variable_align): Move #endif to the right place.
5714 2013-06-10 Cary Coutant <ccoutant@google.com>
5716 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
5717 for hash so that hash table traversal order is deterministic.
5719 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
5720 Pat Haugen <pthaugen@us.ibm.com>
5721 Peter Bergner <bergner@vnet.ibm.com>
5723 * config/rs6000/vector.md (GPR move splitter): Do not split moves
5724 of vectors in GPRS if they are direct moves or quad word load or
5727 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
5729 (direct_move_p): Likewise.
5730 (quad_load_store_p): Likewise.
5732 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
5733 classes into bins based on the physical register type.
5734 (reg_class_to_reg_type): Likewise.
5735 (IS_STD_REG_TYPE): Likewise.
5736 (IS_FP_VECT_REG_TYPE): Likewise.
5737 (reload_fpr_gpr): Arrays to determine what insn to use if we can
5738 use direct move instructions.
5739 (reload_gpr_vsx): Likewise.
5740 (reload_vsx_gpr): Likewise.
5741 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
5742 information that is a simplification of register classes. Also
5743 precalculate direct move reload helpers.
5744 (direct_move_p): New function to return true if the operation can
5745 be done as a direct move instruciton.
5746 (quad_load_store_p): New function to return true if the operation
5747 is a quad memory operation.
5748 (rs6000_legitimize_address): If quad memory, only allow register
5749 indirect for TImode addresses.
5750 (rs6000_legitimate_address_p): Likewise.
5751 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
5752 (rs6000_reload_register_type): Likewise.
5753 (register_to_reg_type): Return register type.
5754 (rs6000_secondary_reload_simple_move): New helper function for
5755 secondary reload and secondary memory needed to identify anything
5756 that is a simple move, and does not need reloading.
5757 (rs6000_secondary_reload_direct_move): New helper function for
5758 secondary reload to identify cases that can be done with several
5759 instructions via the direct move instructions.
5760 (rs6000_secondary_reload_move): New helper function for secondary
5761 reload to identify moves between register types that can be done.
5762 (rs6000_secondary_reload): Add support for quad memory operations
5763 and for direct move.
5764 (rs6000_secondary_memory_needed): Likewise.
5765 (rs6000_debug_secondary_memory_needed): Change argument names.
5766 (rs6000_output_move_128bit): New function to return the move to
5767 use for 128-bit moves, including knowing about the various
5768 limitations of quad memory operations.
5770 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
5771 memory operations. call rs6000_output_move_128bit for the actual
5772 instruciton(s) to generate.
5773 (vsx_movti_64bit): Likewise.
5775 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
5776 (UNSPEC_P8V_MTVSRWZ): Likewise.
5777 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
5778 (UNSPEC_P8V_MTVSRD): Likewise.
5779 (UNSPEC_P8V_XXPERMDI): Likewise.
5780 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
5781 (UNSPEC_FUSION_GPR): Likewise.
5782 (FMOVE128_GPR): New iterator for direct move.
5783 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
5786 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
5787 loads and direct move instructions.
5788 (zero_extendsidi2_lfiwzx): Likewise.
5789 (extendsidi2_lfiwax): Likewise.
5790 (extendsidi2_nocell): Likewise.
5791 (floatsi<mode>2_lfiwax): Likewise.
5793 (floatunssi<mode>2_lfiwzx): Likewise.
5795 (fix_trunc<mode>_stfiwx): Likewise.
5796 (fixuns_trunc<mode>_stfiwx): Likewise.
5797 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
5798 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
5799 (parity<mode>2_cmpb): Set length/type attr.
5800 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
5801 for 'mr.' to fast_compare.
5802 (bpermd_<mode>): Change type attr to popcnt.
5803 (p8_fmrgow_<mode>): New insns for power8 direct move support.
5804 (p8_mtvsrwz_1): Likewise.
5805 (p8_mtvsrwz_2): Likewise.
5806 (reload_fpr_from_gpr<mode>): Likewise.
5807 (p8_mtvsrd_1): Likewise.
5808 (p8_mtvsrd_2): Likewise.
5809 (p8_xxpermdi_<mode>): Likewise.
5810 (reload_vsx_from_gpr<mode>): Likewise.
5811 (reload_vsx_from_gprsf): Likewise.
5812 (p8_mfvsrd_3_<mode>): LIkewise.
5813 (reload_gpr_from_vsx<mode>): Likewise.
5814 (reload_gpr_from_vsxsf): Likewise.
5815 (p8_mfvsrd_4_disf): Likewise.
5816 (multi-word GPR splits): Do not split direct moves or quad memory
5819 2013-06-10 David Malcolm <dmalcolm@redhat.com>
5821 * tree-into-ssa.c (interesting_blocks): Make static.
5823 2013-06-10 Jakub Jelinek <jakub@redhat.com>
5826 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
5827 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
5828 Use DATA_ABI_ALIGNMENT for that case instead if defined.
5829 (get_variable_align): New function.
5830 (get_variable_section, emit_bss, emit_common,
5831 assemble_variable_contents, place_block_symbol): Use
5832 get_variable_align instead of DECL_ALIGN.
5833 (assemble_noswitch_variable): Add align argument, use it
5834 instead of DECL_ALIGN.
5835 (assemble_variable): Adjust caller. Use get_variable_align
5836 instead of DECL_ALIGN.
5837 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
5839 (DATA_ABI_ALIGNMENT): Define.
5840 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
5841 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
5842 opt is false, only return the psABI mandated alignment increase.
5843 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
5844 (DATA_ABI_ALIGNMENT): ... this.
5845 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
5846 (DATA_ABI_ALIGNMENT): ... this.
5847 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
5848 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
5849 (DATA_ABI_ALIGNMENT): ... this.
5850 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
5851 * doc/tm.texi: Regenerated.
5853 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
5855 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
5856 cmp_code to construct REG_EQUAL note.
5858 2013-06-09 Jakub Jelinek <jakub@redhat.com>
5861 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
5862 that operands[2] doesn't overlap with operands[0].
5864 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
5865 Jan Hubicka <jh@suse.cz>
5867 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
5868 hack to mark symbols as used.
5870 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
5872 PR rtl-optimization/57559
5873 * lra-constraints.c (process_alt_operands): Don't discourage
5874 memory with known offset for offsetable memory constraint.
5875 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
5877 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
5879 * varasm.c (struct oc_local_state): Reorder fields.
5880 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
5881 and adjust accordingly.
5882 (output_constructor): Reorder initialization code and adjust call to
5883 output_constructor_bitfield.
5885 2013-06-07 Jan Hubicka <jh@suse.cz>
5887 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
5889 2013-06-07 David Malcolm <dmalcolm@redhat.com>
5891 * tree-object-size.c (unknown): Make const.
5893 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5895 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
5896 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
5897 for last alternative in the cpu_facility attribute.
5899 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5902 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
5903 (xordi3): Change operand 2 constraint to arm_xordi_operand.
5904 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
5905 * config/arm/constraints.md (Dg): New constraint.
5906 * config/arm/neon.md (xordi3_neon): Remove.
5907 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
5908 * config/arm/predicates.md (arm_xordi_operand): New predicate.
5910 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5912 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
5913 Clean up alternatives.
5915 2013-06-07 Alan Modra <amodra@gmail.com>
5917 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
5920 2013-06-07 Alan Modra <amodra@gmail.com>
5922 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
5924 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5926 * config/arm/constraints.md (Df): New constraint.
5927 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
5928 Correct length attribute for last two alternatives.
5930 2013-06-07 Alan Modra <amodra@gmail.com>
5932 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
5933 override user -mfp-in-toc.
5934 (offsettable_ok_by_alignment): Consider just the current access
5935 rather than the whole object, unless BLKmode. Handle
5936 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
5937 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
5938 for -mcmodel=medium.
5939 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
5940 override user -mfp-in-toc or -msum-in-toc. Default to
5941 -mno-fp-in-toc for -mcmodel=medium.
5943 2013-06-06 DJ Delorie <dj@redhat.com>
5945 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
5946 TARGET_VALID_POINTER_MODE.
5948 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
5949 Pat Haugen <pthaugen@us.ibm.com>
5950 Peter Bergner <bergner@vnet.ibm.com>
5952 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5953 Document new power8 builtins.
5955 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
5956 condition code register, to allow 128-bit logical operations to be
5957 done in the VSX or GPR registers.
5958 (nor<mode>3): Use the canonical form for nor.
5959 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
5960 vclz*, and vpopcnt* vector instructions.
5961 (nand<mode>3): Likewise.
5962 (orc<mode>3): Likewise.
5963 (clz<mode>2): LIkewise.
5964 (popcount<mode>2): Likewise.
5966 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
5967 that only the GPRs are recognized.
5969 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5970 support for new power8 builtins.
5972 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
5974 (xscvdpspn): Likewise.
5980 (vpopcnt): Likewise.
5981 (vpopcntb): Likewise.
5982 (vpopcnth): Likewise.
5983 (vpopcntw): Likewise.
5984 (vpopcntd): Likewise.
5989 (eqv_v16qi3): Likewise.
5990 (eqv_v8hi3): Likewise.
5991 (eqv_v4si3): Likewise.
5992 (eqv_v2di3): Likewise.
5993 (eqv_v4sf3): Likewise.
5994 (eqv_v2df3): Likewise.
5996 (nand_v16qi3): Likewise.
5997 (nand_v8hi3): Likewise.
5998 (nand_v4si3): Likewise.
5999 (nand_v2di3): Likewise.
6000 (nand_v4sf3): Likewise.
6001 (nand_v2df3): Likewise.
6003 (orc_v16qi3): Likewise.
6004 (orc_v8hi3): Likewise.
6005 (orc_v4si3): Likewise.
6006 (orc_v2di3): Likewise.
6007 (orc_v4sf3): Likewise.
6008 (orc_v2df3): Likewise.
6010 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6011 allow power8 quad mode in 64-bit.
6012 (rs6000_builtin_vectorized_function): Add support to vectorize
6013 ISA 2.07 count leading zeros, population count builtins.
6014 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6015 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6016 (builtin_function_type): Add vgbbd builtin function which takes an
6018 (altivec_expand_vec_perm_const): Add support for new power8 merge
6021 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6022 that does not include TImdoe for use with 32-bit.
6023 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6025 (UNSPEC_VSX_CVDPSPN): Likewise.
6026 (vsx_xscvdpspn): Likewise.
6027 (vsx_xscvspdpn): Likewise.
6028 (vsx_xscvdpspn_scalar): Likewise.
6029 (vsx_xscvspdpn_directmove): Likewise.
6030 (vsx_and<mode>3): Split logical operations into 32-bit and
6031 64-bit. Add support to do logical operations on TImode as well as
6032 VSX vector types. Allow logical operations to be done in either
6033 VSX registers or in general purpose registers in 64-bit mode. Add
6034 splitters if GPRs were used. For AND, add clobber of CCmode to
6035 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6037 (vsx_and<mode>3_32bit): Likewise.
6038 (vsx_and<mode>3_64bit): Likewise.
6039 (vsx_ior<mode>3): Likewise.
6040 (vsx_ior<mode>3_32bit): Likewise.
6041 (vsx_ior<mode>3_64bit): Likewise.
6042 (vsx_xor<mode>3): Likewise.
6043 (vsx_xor<mode>3_32bit): Likewise.
6044 (vsx_xor<mode>3_64bit): Likewise.
6045 (vsx_one_cmpl<mode>2): Likewise.
6046 (vsx_one_cmpl<mode>2_32bit): Likewise.
6047 (vsx_one_cmpl<mode>2_64bit): Likewise.
6048 (vsx_nor<mode>3): Likewise.
6049 (vsx_nor<mode>3_32bit): Likewise.
6050 (vsx_nor<mode>3_64bit): Likewise.
6051 (vsx_andc<mode>3): Likewise.
6052 (vsx_andc<mode>3_32bit): Likewise.
6053 (vsx_andc<mode>3_64bit): Likewise.
6054 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6055 and xxlorc instructions.
6056 (vsx_eqv<mode>3_64bit): Likewise.
6057 (vsx_nand<mode>3_32bit): Likewise.
6058 (vsx_nand<mode>3_64bit): Likewise.
6059 (vsx_orc<mode>3_32bit): Likewise.
6060 (vsx_orc<mode>3_64bit): Likewise.
6062 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6064 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6066 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6067 (p8_vmrgow): Likewise.
6068 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6069 GPRs to be split under VSX.
6070 (p8v_clz<mode>2): Add power8 count leading zero support.
6071 (p8v_popcount<mode>2): Add power8 population count support.
6072 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6075 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6078 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6080 (vec_nand): Likewise.
6081 (vec_vclz): Likewise.
6082 (vec_vclzb): Likewise.
6083 (vec_vclzd): Likewise.
6084 (vec_vclzh): Likewise.
6085 (vec_vclzw): Likewise.
6086 (vec_vgbbd): Likewise.
6087 (vec_vmrgew): Likewise.
6088 (vec_vmrgow): Likewise.
6089 (vec_vpopcnt): Likewise.
6090 (vec_vpopcntb): Likewise.
6091 (vec_vpopcntd): Likewise.
6092 (vec_vpopcnth): Likewise.
6093 (vec_vpopcntw): Likewise.
6095 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6097 PR rtl-optimization/57468
6098 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6101 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6103 PR rtl-optimization/57459
6104 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6105 type when setting live regs.
6107 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6109 * config/s390/s390.opt (mlra): New option.
6110 * config/s390/s390.c (s390_decompose_address): Check displacement
6111 for all registers for LRA.
6112 (s390_secondary_reload): Don't used secondary reloads for LRA.
6113 (s390_lra_p): New function.
6114 (TARGET_LRA_P): Define.
6115 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6116 of attribute cpu_facility to zarch for the last alternative.
6117 (*cmpmem_short): Ditto.
6119 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6121 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6122 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6123 (arm_expand_prologue): Likewise.
6125 2013-06-06 Teresa Johnson <tejohnson@google.com>
6128 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6129 as this is now done by redirect_edge_and_branch_force.
6130 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6131 barriers, and fix interaction with splitting.
6132 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6133 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6134 reflect changes made in the routine.
6135 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6136 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6137 since this is called in cfglayout mode, and replace partition fixup
6138 with assert as that is now done by force_nonfallthru_and_redirect.
6139 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6140 already be marked with region crossing note.
6141 (insert_section_boundary_note): Make non-static, gate on flag
6142 has_bb_partition, rewrite to also check for multiple partitions.
6143 (rest_of_handle_reorder_blocks): Remove call to
6144 insert_section_boundary_note, now done later during free_cfg.
6145 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6146 * bb-reorder.h (insert_section_boundary_note): Declare.
6147 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6148 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6149 invoke insert_section_boundary_note.
6150 (try_redirect_by_replacing_jump): Remove unnecessary
6151 check for region crossing note.
6152 (fixup_partition_crossing): New function.
6153 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6154 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6155 in non-cfglayout mode.
6156 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6157 remove old code that tried to do this. Emit barrier correctly
6158 when we are in cfglayout mode.
6159 (last_bb_in_partition): New function.
6160 (rtl_split_edge): Correctly fixup partition boundaries.
6161 (commit_one_edge_insertion): Remove old code that tried to
6162 fixup region crossing edge since this is now handled in
6163 split_block, and set up insertion point correctly since
6164 block may now end in a jump.
6165 (verify_hot_cold_block_grouping): Guard against checking when not in
6166 linearized RTL mode.
6167 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6169 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6170 rtl_verify_flow_info, so not called in cfglayout mode.
6171 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6172 (fixup_reorder_chain): Remove old code that attempted to fixup region
6173 crossing note as this is now handled in force_nonfallthru_and_redirect.
6174 (duplicate_insn_chain): Don't duplicate switch section notes.
6175 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6177 * basic-block.h (emit_barrier_after_bb): Declare.
6179 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6181 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6182 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6183 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6185 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6187 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6189 * config/arm/ldmstm.md: Regenerate.
6191 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6193 * config/arm/sync.md (atomic_loaddi_1):
6194 Disable predication for arm_restrict_it.
6195 (arm_load_exclusive<mode>): Likewise.
6196 (arm_load_exclusivesi): Likewise.
6197 (arm_load_exclusivedi): Likewise.
6198 (arm_load_acquire_exclusive<mode>): Likewise.
6199 (arm_load_acquire_exclusivesi): Likewise.
6200 (arm_load_acquire_exclusivedi): Likewise.
6201 (arm_store_exclusive<mode>): Likewise.
6202 (arm_store_exclusive<mode>): Likewise.
6203 (arm_store_release_exclusivedi): Likewise.
6204 (arm_store_release_exclusive<mode>): Likewise.
6206 2013-06-06 Richard Biener <rguenther@suse.de>
6208 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6210 (lto_tag_is_tree_code_p): Adjust.
6211 (lto_tag_is_gimple_code_p): Likewise.
6212 (lto_gimple_code_to_tag): Likewise.
6213 (lto_tag_to_gimple_code): Likewise.
6214 (lto_tree_code_to_tag): Likewise.
6215 (lto_tag_to_tree_code): Likewise.
6216 * data-streamer.h (streamer_write_hwi_in_range): Use
6217 uhwi streaming to stream the normalized range.
6218 (streamer_read_hwi_in_range): Likewise.
6220 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6222 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6223 (predicable_short_it): Likewise.
6224 (predicated): Likewise.
6225 (enabled): Handle above.
6226 (define_cond_exec): Set predicated attribute to yes.
6228 2013-06-05 Mike Stump <mikestump@comcast.net>
6230 * gdbinit.in (__FUNCTION__): Add.
6232 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6234 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6235 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6237 2013-06-05 Jan Hubicka <jh@suse.cz>
6239 * varasm.c (mark_decl_referenced): Revert the removal until targets
6242 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6244 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6245 instead of mark_decl_referenced.
6247 2013-06-05 Jan Hubicka <jh@suse.cz>
6249 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6250 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6251 and symtab_used_from_object_file_p.
6252 (cgraph_make_node_local_1): Clear forced_by_abi.
6253 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6254 * cgraph.h (symtab_node_base): Add forced_by_abi.
6255 (decide_is_variable_needed): Remove.
6256 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6257 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6258 (decide_is_symbol_needed): ... this one; handle symbols in general;
6259 always analyze virtuals; honnor forced_by_abi.
6260 (cgraph_finalize_function): Update.
6261 (varpool_finalize_decl): Update.
6262 (symbol_defined_and_needed): Remove.
6263 (analyze_functions): Update.
6264 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6265 output_refs, input_overwrite_node): Handle forced_by_abi.
6266 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6267 (address_taken_from_non_vtable_p): ... this one.
6268 (comdat_can_be_unshared_p_1): New function.
6269 (cgraph_comdat_can_be_unshared_p): Rename to ...
6270 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6271 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6272 (function_and_variable_visibility): Clear forced_by_abi as needed.
6273 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6274 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6275 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6276 * varpool.c (decide_is_variable_needed): Remove.
6278 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6280 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6281 (arm_option_override): Override arm_restrict_it where appropriate.
6282 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6283 * config/arm/arm.opt (mrestrict-it): New command-line option.
6284 * doc/invoke.texi: Document -mrestrict-it.
6286 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6288 * tsan.c (tsan_atomic_table): Make const.
6290 2013-06-05 Richard Biener <rguenther@suse.de>
6292 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6293 index associated with the tree we are supposed to replace.
6294 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6295 TREE_ASM_WRITTEN as zero for everything but SSA names.
6297 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6299 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6301 2013-06-04 Jan Hubicka <jh@suse.cz>
6303 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6304 (input_node, input_varpool_node): Handle correctly external same
6306 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6307 nodes at ltrans stage.
6309 2013-06-04 Jan Hubicka <jh@suse.cz>
6311 * ipa-inline.c (update_caller_keys): Fix availability test.
6312 (update_callee_keys): Likewise.
6313 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6314 to follow ELF standard.
6316 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6318 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6319 (mips64r5900el-*-elf*): New configurations.
6320 * config/mips/mips-cpus.def (r5900): New processor.
6321 * config/mips/mips-tables.opt: Regenerate.
6322 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6323 (mips_issue_rate): Handle PROCESSOR_R5900.
6324 (mips_reorg_process_insns): Force reorder mode for the R5900.
6325 * config/mips/mips.h (TARGET_MIPS5900): Define.
6326 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6328 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6330 * config/mips/mips.md (processor): Add r5900.
6331 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6333 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6335 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6336 into function to generate MOVI instruction.
6337 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6338 (aarch64_preferred_simd_mode): Turn into wrapper.
6339 (aarch64_output_scalar_simd_mov_immediate): New function.
6340 * config/aarch64/aarch64-protos.h: Add prototype for above.
6342 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6344 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6345 element_char member.
6346 (sizetochar): Return signed char.
6347 (aarch64_simd_valid_immediate): Remove elchar and other
6348 unnecessary variables.
6349 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6350 Calculate element_char as required.
6351 * config/aarch64/aarch64-protos.h: Update and move prototype
6352 for aarch64_output_simd_mov_immediate.
6353 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6356 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6358 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6359 information completed by aarch64_simd_valid_immediate.
6360 (aarch64_legitimate_constant_p): Update arguments.
6361 (aarch64_simd_valid_immediate): Work with struct rather than many
6363 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6364 (aarch64_simd_make_constant): Update arguments.
6365 (aarch64_output_simd_mov_immediate): Work with struct rather than
6366 many pointers. Output immediate directly rather than as operand.
6367 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6369 * config/aarch64/constraints.md (Dn): Update arguments.
6371 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6373 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6375 (aarch64_simd_immediate_valid_for_move): Remove.
6376 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
6377 (aarch64_simd_make_constant): Update call.
6378 (aarch64_output_simd_mov_immediate): Update call.
6379 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6381 * config/aarch64/constraints.md (Dn): Update call.
6383 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6385 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
6386 return type to bool for prototype.
6387 (aarch64_legitimate_constant_p): Check for true instead of not -1.
6388 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
6389 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
6391 2013-06-04 Catherine Moore <clm@codesourcery.com>
6393 * config/mips/mips.opt (meva): New.
6394 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
6395 (ASM_SPEC): Handle -meva.
6396 * doc/invoke.texi (meva): Document.
6398 2013-06-04 Alan Modra <amodra@gmail.com>
6400 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
6403 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6405 * rtl.def: Add extra fourth optional field to define_cond_exec.
6406 * gensupport.c (process_one_cond_exec): Process attributes from
6408 * doc/md.texi: Document fourth field in define_cond_exec.
6410 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
6412 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
6413 out the processing order as in store_bit_field_1.
6415 2013-06-04 Jan Hubicka <jh@suse.cz>
6418 * cgraphunit.c (cgraph_process_same_body_aliases): Create
6419 non-VAR_DECL node if it does not exist yet.
6421 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
6423 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
6424 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
6425 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
6426 target_cpu_default setting.
6428 2013-06-03 Teresa Johnson <tejohnson@google.com>
6430 * dumpfile.c (opt_info_switch_p): Change -fopt-info
6431 default to -fopt-info=optimized instead of all.
6432 * doc/invoke.texi: Ditto.
6433 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
6434 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
6435 (execute_vect_slp): Emit BB vectorization success under
6436 MSG_OPTIMIZED_LOCATIONS.
6437 * tree-vect-slp.c (vect_slp_transform_bb): Change
6438 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
6439 * tree-vect-loop.c (vect_transform_loop): Ditto.
6441 2013-06-03 Jason Merrill <jason@redhat.com>
6444 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
6445 Use TARGET_EXPR for C++.
6447 2013-06-03 Jakub Jelinek <jakub@redhat.com>
6449 PR rtl-optimization/57268
6450 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
6451 if DEBUG_INSN_P (insn).
6454 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6456 PR rtl-optimization/57268
6457 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6458 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6460 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6462 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
6463 (ix86_avoid_lea_for_addr): Likewise.
6464 (exact_dependency_1): Likewise.
6465 (ix86_adjust_cost): Likewise.
6466 (swap_top_of_ready_list): Fix formatting and !reload_completed check
6468 (do_reorder_for_imul): Fix typo, formatting and
6469 !reload_completed check removed.
6470 (ix86_sched_reorder): Fix typo and formatting.
6471 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
6474 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
6476 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
6478 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
6480 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
6481 <STRING_CST>: Likewise.
6482 <VECTOR_CST>: Likewise.
6484 2013-06-01 Janus Weil <janus@gcc.gnu.org>
6485 Mikael Morin <mikael@gcc.gnu.org>
6487 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
6488 * config.in: Regenerated.
6489 * configure: Regenerated.
6491 2013-06-01 Jan Hubicka <jh@suse.cz>
6494 * cgraphunit.c (compile): When weakref is not supported,
6495 set up transparent aliases before final output pass.
6496 * varasm.c (assemble_alias): Do not try to do it here.
6498 2013-06-01 Jan Hubicka <jh@suse.cz>
6501 * passes.c (for_per_function): Skip unanalyzed functions.
6503 2013-06-01 Jan Hubicka <jh@suse.cz>
6505 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
6506 (lto_symtab_merge_symbols_1): ... this one.
6507 (lto_symtab_merge_cgraph_nodes): Rename to ...
6508 (lto_symtab_merge_symbols): ... this one; simplify.
6509 * cgraph.c (same_body_aliases_done): Rename to ...
6510 (cpp_implicit_aliases_done): ... this one.
6511 (cgraph_create_function_alias): Update.
6512 (cgraph_same_body_alias): Update.
6513 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
6514 (verify_edge_corresponds_to_fndecl): Simplify.
6515 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
6516 (cgraph_node): Remove same_body_alias.
6517 (varpool_node): Remove alias_of and extra_name_alias.
6518 (same_body_aliases_done): Rename to ..
6519 (cpp_implicit_aliases_done): ... this one.
6520 (symtab_alias_ultimate_target): Add default parameter.
6521 (symtab_resolve_alias): New function.
6522 (fixup_same_cpp_alias_visibility): Declare.
6523 (cgraph_function_node): Add default parameter.
6524 (cgraph_node_asm_name): Likewise.
6525 (cgraph_function_or_thunk_node): Add default parameter; do
6526 not ICE when it is NULL.
6527 (varpool_variable_node): Likewise.
6528 * tree-emutls.c (create_emultls_var): Update.
6529 (ipa_lower_emutls): Update.
6530 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
6531 (cgraph_reset_node): Reset alias info.
6532 (cgraph_finalize_function): Update.
6533 (fixup_same_cpp_alias_visibility): Move to symtab.c.
6534 (analyze_function): Simplify.
6535 (cgraph_process_same_body_aliases): Simplify.
6536 (analyze_functions): Fixup same body aliases.
6537 (handle_alias_pairs): Simplify.
6538 (assemble_thunk): Update.
6539 (assemble_thunks_and_aliases): Update.
6540 (output_weakrefs): Rewrite.
6541 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
6542 (lto_output_varpool_node): Likewise.
6543 (compute_ltrans_boundary): Remve assert.
6544 (get_alias_symbol): New functoin.
6545 (input_node): Rewrite alias handling.
6546 (input_varpool_node): Likewise.
6547 * ipa-pure-const.c (propagate_pure_const): Fix formating.
6548 * ipa.c (process_references): Handle weakrefs correctly.
6549 (symtab_remove_unreachable_nodes): Likewise.
6550 * trans-mem.c (get_cg_data): Update.
6551 (ipa_tm_create_version_alias): Update.
6552 (ipa_tm_execute): Update.
6553 * symtab.c (dump_symtab_base): Dump aliases.
6554 (verify_symtab_base): Verify aliases.
6555 (symtab_node_availability): New function.
6556 (symtab_alias_ultimate_target): Simplify.
6557 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
6558 handle all the fixup cases.
6559 (symtab_resolve_alias): New function.
6560 * passes.c (ipa_write_summaries): Handle weakrefs.
6561 * varpool.c (varpool_analyze_node): Simplify.
6562 (assemble_aliases): Update.
6563 (varpool_create_variable_alias): Simplify.
6564 (varpool_extra_name_alias): Simplify.
6565 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
6566 (lto_symtab_merge_symbols): ... this one.
6568 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
6571 PR rtl-optimization/57268
6572 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6573 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6575 2013-06-01 Tobias Burnus <burnus@net-b.de>
6578 2013-05-31 Tobias Burnus <burnus@net-b.de>
6581 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
6584 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
6586 PR rtl-optimization/57268
6587 * sched-deps.c (sched_analyze_2): Flush dependence lists if
6588 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
6590 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6592 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
6593 unordered comparison operators when -fno-trapping-math is in effect
6595 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
6596 and implement unordered comparison operators properly on the e500.
6598 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
6600 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
6601 for constant scalar integers.
6602 (simplify_relational_operation_1): Likewise.
6604 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
6606 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
6607 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
6610 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6611 Igor Zamyatin <igor.zamyatin@intel.com>
6613 Silvermont (SLM) architecture performance tuning.
6614 * config/i386/i386.h (enum ix86_tune_indices): Add
6615 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
6616 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
6618 * config/i386/i386.c (initial_ix86_tune_features)
6619 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
6620 (ix86_lea_outperforms): Handle Silvermont tuning.
6621 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
6623 (ix86_use_lea_for_mov): Likewise.
6624 (ix86_avoid_lea_for_addr): Likewise.
6625 (ix86_lea_for_add_ok): Likewise.
6626 (exact_dependency_1): New function.
6627 (exact_store_load_dependency): Likewise.
6628 (ix86_adjust_cost): Handle Silvermont tuning.
6629 (do_reoder_for_imul): Likewise.
6630 (swap_top_of_ready_list): New function.
6631 (ix86_sched_reorder): Changed to handle Silvermont tuning.
6633 * config/i386/i386.md (peepholes that split memory operand in fp
6636 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6638 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6639 Remove un-necessary braces.
6641 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
6643 * config/aarch64/aarch64.c (aarch64_classify_symbol):
6644 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
6646 2013-05-31 Tobias Burnus <burnus@net-b.de>
6649 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
6651 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6654 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
6655 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
6656 * config/arm/neon.md (iordi3_neon): Remove.
6657 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
6658 * config/arm/predicates.md (imm_for_neon_logic_operand):
6659 Move to earlier in the file.
6660 (neon_logic_op2): Likewise.
6661 (arm_iordi_operand_neon): New predicate.
6663 2013-05-31 Richard Biener <rguenther@suse.de>
6665 PR tree-optimization/57478
6666 PR tree-optimization/57453
6667 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
6670 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6672 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
6673 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
6675 2013-05-30 Tobias Burnus <burnus@net-b.de>
6676 Thomas Koenig <tkoenig@gcc.gnu.org>
6679 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
6680 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
6682 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
6684 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
6686 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
6688 * target.def (register_usage_leveling_p): New hook.
6689 * targhooks.c (default_register_usage_leveling_p): New.
6690 * targhooks.h (default_register_usage_leveling_p): New prototype.
6691 * lra-assigns.c (register_usage_leveling_p): Use the hook.
6692 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
6693 * doc/tm.texi: Update.
6694 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
6696 2013-05-30 Ian Bolton <ian.bolton@arm.com>
6698 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
6699 (*insv_reg<mode>): New define_insn.
6701 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
6703 PR rtl-optimization/57439
6704 * postreload.c (move2add_valid_value_p): Check that we have
6705 a zero subreg_regno_offset when accessing the register in
6708 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
6709 Igor Zamyatin <igor.zamyatin@intel.com>
6711 Silvermont (SLM) architecture pipeline model, tuning and
6713 * config.gcc: Add slm config options and target.
6715 * config/i386/slm.md: New.
6717 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
6719 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
6721 (ix86_target_macros_internal): Likewise.
6723 * gcc/config/i386/i386.c (slm_cost): New cost.
6724 (m_SLM): New macro flag.
6725 (initial_ix86_tune_features): Set m_SLM.
6726 (x86_accumulate_outgoing_args): Likewise.
6727 (x86_arch_always_fancy_math_387): Likewise.
6728 (processor_target_table): Add slm cost.
6729 (cpu_names): Add slm cpu name.
6730 (x86_option_override_internal): Set SLM ISA.
6731 (ix86_issue_rate): New case PROCESSOR_SLM.
6732 (ia32_multipass_dfa_lookahead): Likewise.
6733 (fold_builtin_cpu): Add slm.
6735 * config/i386/i386.h (TARGET_SLM): New target macro.
6736 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
6737 (processor_type): Add PROCESSOR_SLM.
6739 * config/i386/i386.md (cpu): Add new value "slm".
6740 (slm.md): Include slm.md.
6742 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
6743 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6745 * config/arm/arm-protos.h: Add and update function protos.
6746 * config/arm/arm.c (use_simple_return_p): New added.
6747 (thumb2_expand_return): Check simple_return flag.
6748 * config/arm/arm.md: Add simple_return and conditional simple_return.
6749 * config/arm/iterators.md: Add iterator for return and simple_return.
6751 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6753 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
6754 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
6755 (arm_emit_vfp_multi_reg_pop): Likewise.
6756 (thumb2_emit_ldrd_pop): Likewise.
6757 (arm_expand_epilogue): Add misc REG_CFA notes.
6758 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
6760 2013-05-29 Lawrence Crowl <crowl@google.com>
6762 * config/arm/t-arm: Update for below.
6764 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
6765 Change type to hash_table. Update dependent calls and types.
6767 * config/i386/t-cygming: Update for below.
6769 * config/i386/t-interix: Update for below.
6771 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
6772 Change type to hash_table. Update dependent calls and types.
6773 (i386_find_on_wrapper_list::wrappers): Likewise.
6775 * config/ia64/t-ia64: Update for below.
6777 * config/ia64/ia64.c (bundle_state_table):
6778 Change type to hash_table. Update dependent calls and types.
6780 * config/mips/mips.c (mips_reorg_process_insns::htab):
6781 Change type to hash_table. Update dependent calls and types.
6783 * config/sol2.c (solaris_comdat_htab):
6784 Change type to hash_table. Update dependent calls and types.
6786 * config/t-sol2: Update for above.
6788 2013-05-29 Teresa Johnson <tejohnson@google.com>
6790 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
6791 functions are not yet marked as defined.
6793 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
6794 Pat Haugen <pthaugen@us.ibm.com>
6795 Peter Bergner <bergner@vnet.ibm.com>
6797 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
6801 (vrotl<mode>3): Likewise.
6802 (vashl<mode>3): Likewise.
6803 (vlshr<mode>3): Likewise.
6804 (vashr<mode>3): Likewise.
6806 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6807 support for power8 V2DI builtins.
6809 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
6810 power8 V2DI builtins.
6811 (vupkhsw): Likewise.
6812 (vupklsw): Likewise.
6813 (vaddudm): Likewise.
6818 (vpkudum): Likewise.
6819 (vpksdss): Likewise.
6820 (vpkudus): Likewise.
6821 (vpksdus): Likewise.
6826 (vsubudm): Likewise.
6827 (vcmpequd): Likewise.
6828 (vcmpgtsd): Likewise.
6829 (vcmpgtud): Likewise.
6830 (vcmpequd_p): Likewise.
6831 (vcmpgtsd_p): Likewise.
6832 (vcmpgtud_p): Likewise.
6833 (vupkhsw): Likewise.
6834 (vupklsw): Likewise.
6835 (vaddudm): Likewise.
6840 (vpksdss): Likewise.
6841 (vpksdus): Likewise.
6842 (vpkudum): Likewise.
6843 (vpkudus): Likewise.
6848 (vsubudm): Likewise.
6850 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
6851 support for power8 V2DI instructions.
6853 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
6854 power8 V2DI instructions. Combine pack and unpack insns to use an
6855 iterator for each mode. Check whether a particular mode supports
6856 Altivec instructions instead of just checking TARGET_ALTIVEC.
6857 (UNSPEC_VPKUWUM): Likewise.
6858 (UNSPEC_VPKSHSS): Likewise.
6859 (UNSPEC_VPKSWSS): Likewise.
6860 (UNSPEC_VPKUHUS): Likewise.
6861 (UNSPEC_VPKSHUS): Likewise.
6862 (UNSPEC_VPKUWUS): Likewise.
6863 (UNSPEC_VPKSWUS): Likewise.
6864 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
6865 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
6866 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
6867 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
6868 (UNSPEC_VUPKHSB): Likewise.
6869 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
6870 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
6871 (UNSPEC_VUPKHSH): Likewise.
6872 (UNSPEC_VUPKLSB): Likewise.
6873 (UNSPEC_VUPKLSH): Likewise.
6875 (VI_char): Likewise.
6876 (VI_scalar): Likewise.
6877 (VI_unit): Likewise.
6879 (VP_small): Likewise.
6880 (VP_small_lc): Likewise.
6881 (VU_char): Likewise.
6882 (add<mode>3): Likewise.
6883 (altivec_vaddcuw): Likewise.
6884 (altivec_vaddu<VI_char>s): Likewise.
6885 (altivec_vadds<VI_char>s): Likewise.
6886 (sub<mode>3): Likewise.
6887 (altivec_vsubcuw): Likewise.
6888 (altivec_vsubu<VI_char>s): Likewise.
6889 (altivec_vsubs<VI_char>s): Likewise.
6890 (altivec_vavgs<VI_char>): Likewise.
6891 (altivec_vcmpbfp): Likewise.
6892 (altivec_eq<mode>): Likewise.
6893 (altivec_gt<mode>): Likewise.
6894 (altivec_gtu<mode>): Likewise.
6895 (umax<mode>3): Likewise.
6896 (smax<mode>3): Likewise.
6897 (umin<mode>3): Likewise.
6898 (smin<mode>3): Likewise.
6899 (altivec_vpkuhum): Likewise.
6900 (altivec_vpkuwum): Likewise.
6901 (altivec_vpkshss): Likewise.
6902 (altivec_vpkswss): Likewise.
6903 (altivec_vpkuhus): Likewise.
6904 (altivec_vpkshus): Likewise.
6905 (altivec_vpkuwus): Likewise.
6906 (altivec_vpkswus): Likewise.
6907 (altivec_vpks<VI_char>ss): Likewise.
6908 (altivec_vpks<VI_char>us): Likewise.
6909 (altivec_vpku<VI_char>us): Likewise.
6910 (altivec_vpku<VI_char>um): Likewise.
6911 (altivec_vrl<VI_char>): Likewise.
6912 (altivec_vsl<VI_char>): Likewise.
6913 (altivec_vsr<VI_char>): Likewise.
6914 (altivec_vsra<VI_char>): Likewise.
6915 (altivec_vsldoi_<mode>): Likewise.
6916 (altivec_vupkhsb): Likewise.
6917 (altivec_vupkhs<VU_char>): Likewise.
6918 (altivec_vupkls<VU_char>): Likewise.
6919 (altivec_vupkhsh): Likewise.
6920 (altivec_vupklsb): Likewise.
6921 (altivec_vupklsh): Likewise.
6922 (altivec_vcmpequ<VI_char>_p): Likewise.
6923 (altivec_vcmpgts<VI_char>_p): Likewise.
6924 (altivec_vcmpgtu<VI_char>_p): Likewise.
6925 (abs<mode>2): Likewise.
6926 (vec_unpacks_hi_v16qi): Likewise.
6927 (vec_unpacks_hi_v8hi): Likewise.
6928 (vec_unpacks_lo_v16qi): Likewise.
6929 (vec_unpacks_hi_<VP_small_lc>): Likewise.
6930 (vec_unpacks_lo_v8hi): Likewise.
6931 (vec_unpacks_lo_<VP_small_lc>): Likewise.
6932 (vec_pack_trunc_v8h): Likewise.
6933 (vec_pack_trunc_v4si): Likewise.
6934 (vec_pack_trunc_<mode>): Likewise.
6936 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
6938 (vec_vmaxsd): Likewise.
6939 (vec_vmaxud): Likewise.
6940 (vec_vminsd): Likewise.
6941 (vec_vminud): Likewise.
6942 (vec_vpksdss): Likewise.
6943 (vec_vpksdus): Likewise.
6944 (vec_vpkudum): Likewise.
6945 (vec_vpkudus): Likewise.
6946 (vec_vrld): Likewise.
6947 (vec_vsld): Likewise.
6948 (vec_vsrad): Likewise.
6949 (vec_vsrd): Likewise.
6950 (vec_vsubudm): Likewise.
6951 (vec_vupkhsw): Likewise.
6952 (vec_vupklsw): Likewise.
6954 2013-05-29 Jan Hubicka <jh@suse.cz>
6956 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
6957 flags; reorder rest of fields in more consistent way.
6958 (varpool_node): Remove analyzed, finalized and alias.
6959 (cgraph_ndoe): Likewise.
6960 (symtab_alias_ultimate_target): New function.
6961 (cgraph_function_node): Move offline.
6962 (cgraph_reset_node): Declare.
6963 (cgraph_comdat_can_be_unshared_p): Remove.
6964 (varpool_remove_initializer): Declare.
6965 (varpool_first_defined_variable, varpool_next_defined_variable
6966 cgraph_first_defined_function, cgraph_next_defined_function): Update.
6967 (cgraph_function_with_gimple_body_p): Update.
6968 (varpool_all_refs_explicit_p): Update.
6969 (symtab_alias_target): New function.
6970 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
6971 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
6972 (cgraph_function_or_thunk_node): Simplify using
6973 symtab_alias_ultimate_target.
6974 (varpool_variable_node): Likewise.
6975 * cgraph.c (cgraph_create_function_alias): Update.
6976 (cgraph_add_thunk): Update.
6977 (cgraph_remove_node): Update.
6978 (dump_cgraph_node): Do not dump removed flags.
6979 (cgraph_function_body_availability): Update.
6980 (cgraph_propagate_frequency): Update.
6981 (verify_cgraph_node): Check sanity of local flag.
6982 (cgraph_function_node): Move here from cgraph.h; revamp for
6983 cgraph_function_or_thunk_node.
6984 * lto-symtab.c (lto_varpool_replace_node): Update.
6985 (lto_symtab_resolve_can_prevail_p): Update.
6986 (lto_symtab_merge_cgraph_nodes): Update.
6987 * ipa-cp.c (determine_versionability, initialize_node_lattices,
6988 propagate_constants_accross_call, devirtualization_time_bonus,
6989 ipcp_propagate_stage): Update.
6990 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
6991 * ipa-inline-transform.c (clone_inlined_nodes,
6992 preserve_function_body_p): Update.
6993 * ipa-reference.c (propagate): Update.
6994 (write_node_summary_p): Update.
6995 * toplev.c (wrapup_global_declaration_2): Update.
6996 * cgraphunit.c (cgraph_analyze_function): Rename to ...
6997 (analyze_function) ... this one.
6998 (cgraph_process_new_functions): Update.
6999 (cgraph_reset_node): Export.
7000 (cgraph_finalize_function): Update.
7001 (cgraph_add_new_function): Update.
7002 (process_function_and_variable_attributes): Update.
7003 (varpool_finalize_decl): Update.
7004 (symbol_finalized): Remove.
7005 (symbol_finalized_and_needed): Rename to ...
7006 (symbol_defined_and_needed): ... update.
7007 (cgraph_analyze_functions): Update.
7008 (handle_alias_pairs): Update.
7009 (mark_functions_to_output): Update.
7010 (assemble_thunk): Update.
7011 (output_in_order): Update.
7012 (output_weakrefs): Update.
7013 (finalize_compilation_unit): Update.
7014 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7015 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7016 input_node, input_varpool_node): Update.
7017 * dbxout.c (dbxout_expand_expr): Update.
7018 * cgraphclones.c (cgraph_clone_node): Update.
7019 (cgraph_copy_node_for_versioning): Update.
7020 (cgraph_materialize_clone): Update.
7021 (cgraph_materialize_all_clones): Update.
7022 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7023 propagate_pure_const, propagate_nothrow): Update.
7024 * lto-streamer-out.c (lto_output, write_symbol): Update.
7025 * ipa-utils.c (ipa_reverse_postorder): Update.
7026 * ipa-inline.c (can_inline_edge_p): Update.
7027 (update_caller_keys, ipa_inline): Update.
7028 * dwarf2out.c (reference_to_unused,
7029 premark_types_used_by_global_vars_helper): Update.
7030 * tree-eh.c (tree_could_trap_p): Update.
7031 * ipa-split.c (consider_split, execute_split_functions): Update.
7032 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7033 has_addr_references_p): Update; move ahead in file for better
7035 (process_references): Simplify.
7036 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7038 (cgraph_comdat_can_be_unshared_p): Make static.
7039 (cgraph_externally_visible_p): Update.
7040 (varpool_externally_visible_p): Update.
7041 (function_and_variable_visibility): Update.
7042 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7043 ipa_tm_mark_force_output_node): Update.
7044 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7045 estimate_edge_devirt_benefit, inline_generate_summary,
7046 inline_write_summary): Update.
7047 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7048 * ipa-prop.c (ipa_compute_jump_functions): Update.
7049 (ipa_print_node_params, ipa_prop_read_section,
7050 ipa_update_after_lto_read, read_replacements_section): Update.
7051 * varasm.c (mark_decl_referenced): Update.
7052 (assemble_alias, dump_tm_clone_pairs): Update.
7053 * tree-inline.c (copy_bb): Update.
7054 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7056 * symtab.c (dump_symtab_base): Print new flags.
7057 (verify_symtab_base): Verify new flags.
7058 (symtab_alias_ultimate_target): New function.
7059 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7060 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7062 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7064 * i386.c (ix86_get_function_versions_dispatcher,
7065 ix86_generate_version_dispatcher_body): Update.
7066 (fold_builtin_cpu): Use varpool_add_new_variable.
7067 * varpool.c (varpool_remove_initializer): Break out from ...
7068 (varpool_remove_node): ... this one.
7069 (dump_varpool_node, varpool_node_for_asm,
7070 cgraph_variable_initializer_availability, varpool_analyze_node,
7071 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7072 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7075 2013-05-29 Jan Hubicka <jh@suse.cz>
7077 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7079 2013-05-29 Easwaran Raman <eraman@google.com>
7081 PR tree-optimization/57442
7082 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7083 when control exits the main loop.
7085 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7087 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7089 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7090 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7091 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7094 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7096 PR tree-optimization/57441
7097 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7098 Don't limit size of incr_vec to number of candidates.
7100 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7102 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7103 and mips16 directories.
7104 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7105 (MULTILIB_DIRNAMES): Ditto.
7106 (MULTILIB_EXCEPTIONS): Add new exceptions.
7107 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7108 (MULTILIB_DIRNAMES): Ditto.
7109 (MULTILIB_EXCEPTIONS): Add new exceptions.
7111 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7112 Marcus Shawcroft <marcus.shawcroft@arm.com>
7114 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7115 SYMBOL_TINY_ABSOLUTE.
7116 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7117 SYMBOL_TINY_ABSOLUTE.
7118 (aarch64_expand_mov_immediate): Likewise.
7119 (aarch64_classify_symbol): Likewise.
7120 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7121 Permit SYMBOL_TINY_ABSOLUTE.
7122 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7124 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7125 Marcus Shawcroft <marcus.shawcroft@arm.com>
7127 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7128 Refactor if/switch. Replace gcc_assert with if.
7130 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7132 * config/i386/i386.c (initial_ix86_tune_features): Enable
7133 FP Reassociation for AMD bdver1 and bdver2.
7135 2013-05-29 Martin Jambor <mjambor@suse.cz>
7137 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7138 and IMAGPART_EXPR do not occur within other handled_components.
7140 2013-05-29 Richard Biener <rguenther@suse.de>
7142 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7143 access on whether the use is in the BB we currently try to
7145 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7146 vectorize to vect_bb_slp_scalar_cost.
7148 2013-05-29 Richard Biener <rguenther@suse.de>
7150 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7151 computing scalar cost offsetted by stmts that are kept live
7153 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7154 for computation of scalar cost.
7156 2013-05-28 Steve Ellcey <sellcey@mips.com>
7158 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7160 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7162 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7163 array notation built-in reduction functions.
7164 * doc/passes.texi (Passes): Added documentation about changes done
7166 * doc/invoke.texi (C Dialect Options): Added documentation about
7167 the -fcilkplus flag.
7168 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7169 (BUILTINS_DEF): Depend on cilkplus.def.
7170 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7171 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7172 * cilkplus.def: New file.
7174 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7176 PR rtl-optimization/57439
7177 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7179 2013-05-28 Easwaran Raman <eraman@google.com>
7181 PR tree-optimization/57337
7182 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7183 (find_insert_point): Correctly identify the insertion point
7184 when two statements with the same UID is compared.
7186 2013-05-28 Richard Biener <rguenther@suse.de>
7188 PR tree-optimization/56787
7189 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7190 from the list of data references.
7191 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7193 (vect_analyze_loop_operations): Likewise.
7194 (vect_transform_loop): Remove clobbers.
7196 2013-05-28 Martin Jambor <mjambor@suse.cz>
7198 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7199 and REALPART_EXPRs have scalar type.
7201 2013-05-28 Richard Biener <rguenther@suse.de>
7203 PR tree-optimization/57411
7204 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7206 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7207 virtual operand propagation.
7209 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7211 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7212 destination register for bmasksi_vis.
7213 (vector_init_bshuffle): Likewise.
7214 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7216 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7218 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7219 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7220 mode if the instruction isn't available in the original mode.
7221 * config/sparc/sparc.opt (mfix-ut699): New option.
7222 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7223 (divdf3): Turn into expander.
7224 (divdf3_nofix): New insn.
7225 (divdf3_fix): Likewise.
7226 (divsf3): Disable if -mfix-ut699.
7227 (sqrtdf2): Turn into expander.
7228 (sqrtdf2_nofix): New insn.
7229 (sqrtdf2_fix): Likewise.
7230 (sqrtsf2): Disable if -mfix-ut699.
7232 2013-05-27 Richard Biener <rguenther@suse.de>
7235 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7236 block for the new loop.
7238 2013-05-27 Richard Biener <rguenther@suse.de>
7240 PR tree-optimization/57343
7241 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7242 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7243 (number_of_iterations_cond): Do not build the folded tree.
7245 2013-05-27 Richard Biener <rguenther@suse.de>
7249 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7250 OEP_CONSTANT_ADDRESS_OF retained.
7252 PR tree-optimization/57417
7253 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7255 (set_ssa_val_to): Compare addresses using
7256 get_addr_base_and_unit_offset.
7258 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7260 PR rtl-optimization/56833
7261 * postreload.c (move2add_record_mode): New function.
7262 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7263 (move2add_use_add2_insn): Use move2add_record_sym_value.
7264 (move2add_use_add3_insn): Likewise.
7265 (reload_cse_move2add): Use move2add_valid_value_p and
7266 move2add_record_mode. Invalidate call-clobbered and REG_INC
7267 affected regs by setting reg_mode to VOIDmode.
7268 (move2add_note_store): Don't pretend the inside of a SUBREG is
7269 the actual destination. Invalidate single/leading registers by
7270 setting reg_mode to VOIDmode.
7271 Use move2add_record_sym_value, move2add_valid_value_p and
7272 move2add_record_mode.
7274 2013-05-27 Richard Biener <rguenther@suse.de>
7276 PR tree-optimization/57396
7277 * tree-affine.c (double_int_constant_multiple_p): Properly
7278 return false for val == 0 and div != 0.
7280 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7282 * config/mips/mips.h: Use #elif in preprocessor conditions.
7284 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7287 * config/mips/constraints.md (kl): New constraint.
7288 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7289 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7290 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7292 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7294 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7297 * config/mips/mips.c (mips_can_inline_p): New function.
7298 (TARGET_CAN_INLINE_P): Define.
7300 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7302 * sched-int.h (ds_t, dw_t): Make unsigned int.
7303 Fix documentation that describes how all the ds_t bits are used.
7304 Reserve the last bit for delayed-branch scheduling.
7305 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7306 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7307 (gen_dep_weak_1): Remove prototype.
7308 * sched-deps.c (get_dep_weak_1): Make static.
7309 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7310 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7311 * doc/tm.texi: Regenerate.
7312 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7314 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7317 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7319 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7320 Sandra Loosemore <sandra@codesourcery.com>
7322 * config.gcc (powerpc-*): Allow native for with-cpu.
7324 2013-05-24 Jeff Law <law@redhat.com>
7326 PR tree-optimization/57124
7327 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7328 conversion feeding a condition if the range has an overflow
7329 if -fstrict-overflow. Add warnings for when we do make the
7332 2013-05-24 Dehao Chen <dehao@google.com>
7334 * tree-cfg.c (locus_discrim_map): Fix the typo.
7335 (locus_discrim_hasher): Likewise.
7336 (locus_discrim_hasher::hash): Likewise.
7337 (locus_discrim_hasher::equal): Likewise.
7339 2013-05-24 Martin Jambor <mjambor@suse.cz>
7341 PR tree-optimization/57294
7342 * cgraph.h (ipa_record_stmt_references): Declare.
7343 * cgraphbuild.c (ipa_record_stmt_references): New function.
7344 (build_cgraph_edges): Use ipa_record_stmt_references.
7345 (rebuild_cgraph_edges): Likewise.
7346 (cgraph_rebuild_references): Likewise.
7347 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7348 associated with the old statement and build references from the
7349 newly built statements.
7350 * ipa-ref.c (ipa_remove_stmt_references): New function.
7351 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7353 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7355 * lra-constraints.c (emit_spill_move): Use smaller mode for
7357 (check_and_process_move): Consider mem-reg moves for secondary
7359 (curr_insn_transform): Don't lose insns emitted before for
7360 secondary memory moves.
7361 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7362 reg set up in the current insn.
7364 2013-05-24 Dehao Chen <dehao@google.com>
7366 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7368 (locus_descrim_hasher::equal): Likewise.
7369 (build_gimple_cfg): New discriminator assignment algorithm.
7370 (make_edges): Likewise.
7371 (next_discriminator_for_locus): Likewise.
7372 (same_line_p): Likewise.
7373 (assign_discriminators): Likewise.
7374 (make_cond_expr_edges): Likewise.
7375 (make_gimple_switch_edges): Likewise.
7376 (make_goto_expr_edges): Likewise.
7377 (make_gimple_asm_edges): Likewise.
7379 2013-05-24 Ian Bolton <ian.bolton@arm.com>
7381 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
7382 X format specifier to only display bottom 16 bits.
7383 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
7384 immediate to match for operand 2, since it will be masked.
7386 2013-05-24 Richard Biener <rguenther@suse.de>
7388 PR tree-optimization/57287
7389 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
7390 all SSA names that occur in abnormal PHIs.
7392 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
7395 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
7396 that index is not negative.
7398 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
7400 PR rtl-optimization/55177
7401 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
7402 (simplify_byte_swapping_operation): New.
7403 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
7404 (simplify_relational_operation_1): Deal with BSWAP.
7406 2013-05-23 Richard Henderson <rth@redhat.com>
7409 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
7410 (ix86_reorg): Call it.
7412 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
7415 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
7416 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
7417 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
7419 2013-05-23 Christian Bruel <christian.bruel@st.com>
7422 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
7424 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7425 Marcus Shawcroft <marcus.shawcroft@arm.com>
7427 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
7428 * config/aarch64/constraints.md (Usa): Remove.
7429 * doc/md.texi (AArch64 Usa): Remove.
7431 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7432 Marcus Shawcroft <marcus.shawcroft@arm.com>
7434 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
7435 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
7436 * config/aarch64/predicates.md (aarch64_const_address): Remove.
7437 (aarch64_mov_operand): Use aarch64_mov_operand_p.
7439 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
7441 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
7442 instruction (AdvSIMD).
7443 * config/aarch64/aarch64-builtins.c
7444 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
7445 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
7447 2013-05-23 Martin Jambor <mjambor@suse.cz>
7450 * tree.h (contains_bitfld_component_ref_p): Declare.
7451 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
7452 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
7454 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
7455 not access a bit-field. Assert all final offsets are byte-aligned.
7457 2013-05-23 Richard Biener <rguenther@suse.de>
7459 PR tree-optimization/57380
7460 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
7461 least one invariant or re-used load.
7462 * passes.c (init_optimization_passes): Move pass_phiprop before
7465 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
7467 * config/aarch64/aarch64-simd.md
7468 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
7470 2013-05-23 Richard Biener <rguenther@suse.de>
7473 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7474 OEP_CONSTANT_ADDRESS_OF retained.
7476 2013-05-23 Jakub Jelinek <jakub@redhat.com>
7479 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
7480 don't lower unit. Handle unit not being always BITS_PER_WORD.
7482 2013-05-23 Richard Biener <rguenther@suse.de>
7484 PR rtl-optimization/57341
7485 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
7486 instead of true_dependence.
7488 2013-05-22 David Malcolm <dmalcolm@redhat.com>
7490 * bb-reorder.c (branch_threshold): Make const.
7491 (exec_threshold): Ditto.
7493 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7494 Pat Haugen <pthaugen@us.ibm.com>
7495 Peter Bergner <bergner@vnet.ibm.com>
7497 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
7498 documentation for the power8 crypto builtins.
7500 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
7502 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
7503 macros for defining power8 builtin functions.
7504 (BU_P8V_AV_2): Likewise.
7505 (BU_P8V_AV_P): Likewise.
7506 (BU_P8V_VSX_1): Likewise.
7507 (BU_P8V_OVERLOAD_1): Likewise.
7508 (BU_P8V_OVERLOAD_2): Likewise.
7509 (BU_CRYPTO_1): Likewise.
7510 (BU_CRYPTO_2): Likewise.
7511 (BU_CRYPTO_3): Likewise.
7512 (BU_CRYPTO_OVERLOAD_1): Likewise.
7513 (BU_CRYPTO_OVERLOAD_2): Likewise.
7514 (XSCVSPDP): Fix typo, point to the correct instruction.
7515 (VCIPHER): Add power8 crypto builtins.
7516 (VCIPHERLAST): Likewise.
7517 (VNCIPHER): Likewise.
7518 (VNCIPHERLAST): Likewise.
7519 (VPMSUMB): Likewise.
7520 (VPMSUMH): Likewise.
7521 (VPMSUMW): Likewise.
7522 (VPERMXOR_V2DI): Likewise.
7523 (VPERMXOR_V4SI: Likewise.
7524 (VPERMXOR_V8HI: Likewise.
7525 (VPERMXOR_V16QI: Likewise.
7526 (VSHASIGMAW): Likewise.
7527 (VSHASIGMAD): Likewise.
7529 (VPERMXOR): Likewise.
7530 (VSHASIGMA): Likewise.
7532 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7533 __CRYPTO__ if the crypto instructions are available.
7534 (altivec_overloaded_builtins): Add support for overloaded power8
7537 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
7538 support for power8 crypto builtins.
7539 (builtin_function_type): Likewise.
7540 (altivec_init_builtins): Add support for builtins that take vector
7541 long long (V2DI) arguments.
7543 * config/rs6000/crypto.md: New file, define power8 crypto
7546 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
7547 Pat Haugen <pthaugen@us.ibm.com>
7548 Peter Bergner <bergner@vnet.ibm.com>
7550 * doc/invoke.texi (Option Summary): Add power8 options.
7551 (RS/6000 and PowerPC Options): Likewise.
7553 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
7554 constraints.md instead of rs6000.h. Reorder w* constraints. Add
7555 wm, wn, wr documentation.
7557 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
7558 registers if direct move instructions are enabled.
7559 (wn): New constraint for no registers.
7560 (wq): New constraint for quad word even GPR registers.
7561 (wr): New constraint if 64-bit instructions are enabled.
7562 (wv): New constraint if power8 vector instructions are enabled.
7563 (wQ): New constraint for quad word memory locations.
7565 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
7566 constraint for 0..15 for crypto instructions.
7567 (gpc_reg_operand): If VSX allow registers in VSX registers as well
7568 as GPR and floating point registers.
7569 (int_reg_operand): New predicate to match only GPR registers.
7570 (base_reg_operand): New predicate to match base registers.
7571 (quad_int_reg_operand): New predicate to match even GPR registers
7572 for quad memory operations.
7573 (vsx_reg_or_cint_operand): New predicate to allow vector logical
7574 operations in both GPR and VSX registers.
7575 (quad_memory_operand): New predicate for quad memory operations.
7576 (reg_or_indexed_operand): New predicate for direct move support.
7578 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
7579 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
7580 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
7581 (POWERPC_MASKS): Add power8 options.
7582 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
7585 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7586 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
7588 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
7589 (-mpower8-fusion): New power8 options.
7590 (-mpower8-fusion-sign): Likewise.
7591 (-mpower8-vector): Likewise.
7592 (-mcrypto): Likewise.
7593 (-mdirect-move): Likewise.
7594 (-mquad-memory): Likewise.
7596 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
7598 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
7600 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
7601 (rs6000_debug_vector_unit): Add p8_vector.
7602 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
7603 definitions. Also print fusion state.
7604 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
7605 (rs6000_builtin_mask_calculate): Add power8 builtin support.
7606 (rs6000_option_override_internal): Add support for power8.
7607 (rs6000_common_init_builtins): Add debugging for skipped builtins
7609 (rs6000_adjust_cost): Add power8 support.
7610 (rs6000_issue_rate): Likewise.
7611 (insn_must_be_first_in_group): Likewise.
7612 (insn_must_be_last_in_group): Likewise.
7613 (force_new_group): Likewise.
7614 (rs6000_register_move_cost): Likewise.
7615 (rs6000_opt_masks): Likewise.
7617 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
7618 power8 capable assembler, default to power7 options.
7619 (TARGET_DIRECT_MOVE): Likewise.
7620 (TARGET_CRYPTO): Likewise.
7621 (TARGET_P8_VECTOR): Likewise.
7622 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
7623 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
7624 (VECTOR_MEM_P8_VECTOR_P): Likewise.
7625 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
7626 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
7627 (TARGET_XSCVDPSPN): Likewise.
7628 (TARGET_XSCVSPDPN): Likewsie.
7629 (TARGET_SYNC_HI_QI): Likewise.
7630 (TARGET_SYNC_TI): Likewise.
7631 (MASK_CRYPTO): Likewise.
7632 (MASK_DIRECT_MOVE): Likewise.
7633 (MASK_P8_FUSION): Likewise.
7634 (MASK_P8_VECTOR): Likewise.
7635 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
7636 temporary used by some of the direct move instructions to get two FP
7637 temporary registers does not force creation of a stack frame.
7638 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
7639 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
7640 that any VSX registers are tieable, even if they are also an
7641 Altivec vector mode.
7642 (r6000_reg_class_enum): Add wm, wr, wv constraints.
7643 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
7644 (RS6000_BTM_CRYPTO): Likewise.
7645 (RS6000_BTM_COMMON): Likewise.
7647 * config/rs6000/rs6000.md (cpu attribute): Add power8.
7648 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
7649 (enum rs6000_vector): Add power8 vector support.
7651 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7655 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
7656 (any_sibcall_could_use_r3): this and handle indirect calls.
7657 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
7659 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7661 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
7663 2013-05-22 Richard Biener <rguenther@suse.de>
7666 * profile.c (branch_prob): Do not split blocks that are
7667 abnormally receiving from ECF_RETURNS_TWICE functions.
7669 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7671 * recog.c (offsettable_address_addr_space_p): Fix calculation of
7672 address mode. Move pointer mode initialization to the same place.
7674 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7676 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
7677 while it has any effect.
7679 2013-05-21 Easwaran Raman <eraman@google.com>
7681 PR tree-optimization/57322
7682 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
7683 UID of the statement added to the BB to be 1.
7685 2013-05-21 Jakub Jelinek <jakub@redhat.com>
7687 PR tree-optimization/57331
7688 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
7689 of conversion from pointer type to integral type with integer.
7691 2013-05-21 Martin Jambor <mjambor@suse.cz>
7694 * ipa-prop.c (ipa_read_node_info): Process param_used and
7695 controlled_uses in the same order as when writing.
7697 2013-05-21 Magnus Granberg <baldrick@free.fr>
7700 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
7702 2013-05-21 Richard Biener <rguenther@suse.de>
7704 PR tree-optimization/57318
7705 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
7706 estimate stmts with side-effects as likely eliminated.
7708 2013-05-21 Richard Biener <rguenther@suse.de>
7710 PR tree-optimization/57330
7711 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
7712 preserve the call stmts fntype.
7714 2013-05-21 Richard Biener <rguenther@suse.de>
7716 PR tree-optimization/57303
7717 * tree-ssa-sink.c (statement_sink_location): Improve killing
7718 stmt detection and properly handle self-assignments.
7720 2013-05-21 Christian Bruel <christian.bruel@st.com>
7722 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
7723 spanning registers. LEAF_REG_REMAP is supported only for contiguous
7724 registers. Set register size out of the PARALLEL loop.
7726 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
7729 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
7730 (*fmasf4, *fmasf4_media): New insns.
7732 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7734 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
7735 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
7736 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
7737 (mips_idiv_insns): Update the comments to say that the returned
7738 instruction counts are in units of BASE_INSN_LENGTH.
7739 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
7740 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
7741 using 2 rather than 4 as the length of indirect MIPS16 and
7742 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
7743 length of a NOP. Don't divide MIPS16 lengths by 2.
7744 (mips16_split_long_branches): Assume a branch is long if the
7745 length is greater than 4 rather than 8.
7746 * config/mips/mips.md (length): Give MIPS16 lengths directly,
7747 rather than multiplying them by 2. Multiply instruction counts
7748 by BASE_INSN_LENGTH rather than 4.
7749 (*jump_mips16, tls_get_tp_mips16_<mode>)
7750 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
7752 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7754 * config/mips/mips.md (extended_mips16): Remove branch case.
7755 (length): Remove duplicated extended_mips16 test.
7757 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
7759 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
7761 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
7763 * recog.h (Recog_data): Rename to...
7764 (recog_data_d): ...this.
7765 (recog_data): Update accordingly.
7766 * recog.c (recog_data): Likewise.
7767 * reload.c (save_recog_data): Likewise.
7768 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
7769 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
7771 2013-05-17 Julian Brown <julian@codesourcery.com>
7773 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
7774 found in a REG_EQUAL note, invalidate it.
7776 2013-05-17 Easwaran Raman <eraman@google.com>
7778 * tree-ssa-reassoc.c (find_insert_point): New function.
7779 (insert_stmt_after): Likewise.
7780 (get_def_stmt): Likewise.
7781 (ensure_ops_are_available): Likewise.
7782 (not_dominated_by): Likewise.
7783 (rewrite_expr_tree): Do not move statements beyond what is
7784 necessary. Remove call to swap_ops_for_binary_stmt...
7785 (reassociate_bb): ... and move it here.
7786 (build_and_add_sum): Assign UIDs for new statements.
7787 (linearize_expr): Likewise.
7788 (do_reassoc): Renumber gimple statement UIDs.
7790 2013-05-17 Jan Hubicka <jh@suse.cz>
7792 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
7794 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
7795 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
7797 (output_weakrefs): Update.
7799 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
7800 Martin Jambor <mjambor@suse.cz>
7803 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
7804 value that corresponds to the given aggval is found in values vector.
7806 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
7808 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
7809 sse, sse2, sse3, ssse3 and sse4a flags to options.
7811 2013-05-17 David Malcolm <dmalcolm@redhat.com>
7813 * gengtype-state.c: (s_expr_writer): New class, to handle
7814 prettifying of output layout of s-expressions.
7815 (state_writer): New class, to write out gtype.state.
7816 (state_written_type_count): Move this variable into member data of
7818 (s_expr_writer::s_expr_writer): New code: constructor for new class
7819 (state_writer::state_writer(): ditto
7820 (s_expr_writer::write_new_line): New function
7821 (s_expr_writer::write_any_indent): ditto
7822 (s_expr_writer::begin_s_expr): ditto
7823 (s_expr_writer::end_s_expr): ditto
7824 (write_state_fileloc): convert to method of state_writer...
7825 (state_writer:: write_state_fileloc): ...and use methods of
7826 s_expr_writer to write indentation into the gtype.state output file
7827 to visually represent the hierarchical structure of the list
7829 (write_state_fields): ditto, renaming to...
7830 (state_writer::write_state_fields)
7831 (write_state_a_string): ditto, renaming to...
7832 (state_writer::write_state_a_string)
7833 (write_state_string_option): ditto, renaming to...
7834 (state_writer::write_state_string_option)
7835 (write_state_type_option): ditto, renaming to...
7836 (state_writer::write_state_type_option)
7837 (write_state_nested_option): ditto, renaming to...
7838 (state_writer::write_state_nested_option)
7839 (write_state_option): ditto, renaming to...
7840 (state_writer::write_state_option)
7841 (write_state_options): ditto, renaming to...
7842 (state_writer::write_state_options)
7843 (write_state_lang_bitmap): ditto, renaming to...
7844 (state_writer::write_state_lang_bitmap)
7845 (write_state_version): ditto, renaming to...
7846 (state_writer::write_state_version)
7847 (write_state_scalar_type): ditto, renaming to...
7848 (state_writer::write_state_scalar_type)
7849 (write_state_string_type): ditto, renaming to...
7850 (state_writer::write_state_string_type)
7851 (write_state_undefined_type): ditto, renaming to...
7852 (state_writer::write_state_undefined_type)
7853 (write_state_struct_union_type): ditto, renaming to...
7854 (state_writer::write_state_struct_union_type)
7855 (write_state_struct_type): ditto, renaming to...
7856 (state_writer::write_state_struct_type)
7857 (write_state_user_struct_type): ditto, renaming to...
7858 (state_writer::write_state_user_struct_type)
7859 (write_state_lang_struct_type): ditto, renaming to...
7860 (state_writer::write_state_lang_struct_type)
7861 (write_state_param_struct_type): ditto, renaming to...
7862 (state_writer::write_state_param_struct_type)
7863 (write_state_pointer_type): ditto, renaming to...
7864 (state_writer::write_state_pointer_type)
7865 (write_state_array_type): ditto, renaming to...
7866 (state_writer::write_state_array_type)
7867 (write_state_gc_used): ditto, renaming to...
7868 (state_writer::write_state_gc_used)
7869 (write_state_common_type_content): ditto, renaming to...
7870 (state_writer::write_state_common_type_content)
7871 (write_state_type): ditto, renaming to...
7872 (state_writer::write_state_type)
7873 (write_state_pair_list): ditto, renaming to...
7874 (state_writer::write_state_pair_list)
7875 (write_state_pair): ditto, renaming to...
7876 (state_writer::write_state_pair)
7877 (write_state_typedefs): ditto, renaming to...
7878 (state_writer::write_state_typedefs)
7879 (write_state_structures): ditto, renaming to...
7880 (state_writer::write_state_structures)
7881 (write_state_param_structs): ditto, renaming to...
7882 (state_writer::write_state_param_structs)
7883 (write_state_variables): ditto, renaming to...
7884 (state_writer::write_state_variables)
7885 (write_state_srcdir): ditto, renaming to...
7886 (state_writer::write_state_srcdir)
7887 (write_state_files_list): ditto, renaming to...
7888 (state_writer::write_state_files_list)
7889 (write_state_languages): ditto, renaming to...
7890 (state_writer::write_state_languages)
7891 (write_state): create a state_writer instance and use it when
7892 writing out the state file
7894 2013-05-17 Mike Stump <mikestump@comcast.net>
7896 PR rtl-optimization/57304
7897 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
7898 accessing DF_REF_REAL_LOC.
7900 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7902 PR rtl-optimization/57281
7903 PR rtl-optimization/57300
7904 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
7905 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
7906 what the other splitter did if the registers are dead.
7908 2013-05-17 Richard Biener <rguenther@suse.de>
7910 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
7913 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7915 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
7918 2013-05-17 Marek Polacek <polacek@redhat.com>
7920 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
7921 length when doing non-zero store of storing '\0' to '\0'.
7923 2013-05-17 Jakub Jelinek <jakub@redhat.com>
7925 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
7926 vect_external_def oprnd1 with loop_vinfo, try to emit
7927 optional cast, negation and and stmts on the loop preheader
7928 edge instead of into the pattern def seq.
7930 PR tree-optimization/57051
7931 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
7932 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
7934 2013-05-16 Nick Clifton <nickc@redhat.com>
7936 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
7937 (rl78_is_naked_func): New function.
7938 (rl78_expand_prologue): Skip prologue generation for naked functions.
7939 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
7940 * doc/extend.texi (naked): Add RL78 to the list of processors
7941 that supports this attribute.
7943 2013-05-16 Jeff Law <law@redhat.com>
7945 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
7947 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7949 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
7950 cache parameters using detect_caches_amd also for CYRIX,
7951 NSC and TM2 signatures.
7953 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
7954 Dzianis Kahanovich <mahatma@eu.by>
7958 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7959 VIA/Centaur processors and determine their cache parameters
7960 using detect_caches_amd.
7962 2013-05-16 Teresa Johnson <tejohnson@google.com>
7964 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
7965 (rtl_verify_edges): New function.
7966 (rtl_verify_bb_insns): Ditto.
7967 (rtl_verify_bb_pointers): Ditto.
7968 (rtl_verify_bb_insn_chain): Ditto.
7969 (rtl_verify_fallthru): Ditto.
7970 (rtl_verify_bb_layout): Ditto.
7971 (rtl_verify_flow_info_1): Outline checks into new functions.
7972 (rtl_verify_flow_info): Ditto.
7974 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
7976 * cfghooks.c (copy_bbs): Add update_dominance argument.
7977 * cfghooks.h (copy_bbs): Update prototype.
7978 * tree-cfg.c (gimple_duplicate_sese_region):
7979 Add update_dominance argument.
7980 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
7981 * tree-ssa-loop-ch.c (copy_loop_headers): Update
7982 gimple_duplicate_sese_region call.
7983 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
7984 Update copy_bbs call.
7985 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
7986 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
7988 2013-05-16 Jakub Jelinek <jakub@redhat.com>
7990 * tree-vectorizer.h (NUM_PATTERNS): Increment.
7991 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
7992 vect_recog_rotate_pattern.
7993 (vect_recog_rotate_pattern): New function.
7995 2013-05-16 Jason Merrill <jason@redhat.com>
7997 * Makefile.in (LLINKER): New variable.
7998 (mostlyclean): Remove link mutex.
7999 * configure.ac: Handle --enable-link-mutex.
8000 * lock-and-run.sh: New script.
8002 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8005 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8008 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8010 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8012 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8014 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8015 * config/arm/arm.c (next_consecutive_mem): New function.
8016 (gen_movmem_ldrd_strd): Likewise.
8017 * config/arm/arm.md (movmemqi): Update condition and code.
8018 (unaligned_loaddi, unaligned_storedi): New patterns.
8020 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8022 * config.gcc: Obsolete *-*-solaris2.9*.
8023 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8025 2013-05-16 Richard Biener <rguenther@suse.de>
8027 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8028 earlier, after GRAPHITE transforms and IV canonicalization.
8030 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8032 * omp-low.c (extract_omp_for_data): For collapsed loops,
8033 if at least one of the loops is known at compile time to
8034 iterate zero times, set count to 0.
8035 (expand_omp_regimplify_p): New function.
8036 (expand_omp_for_generic): For collapsed loops, if at least
8037 one of the loops isn't known to iterate at least once,
8038 add runtime check with setting count to 0.
8039 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8040 For unsigned types if it isn't known at compile time that
8041 the loop will iterate at least once, add runtime check to bypass
8042 the whole loop if initial condition isn't true.
8044 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8046 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8048 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8051 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8052 transformations to avoid an infinite loop.
8054 2013-05-16 Marek Polacek <polacek@redhat.com>
8056 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8058 2013-05-15 Leif Ekblad <leif@rdos.net>
8060 * config/i386/i386.c (ix86_decompose_address): Use
8061 DEFAULT_TLS_SEG_REG to access TLS segment register.
8062 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8063 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8064 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8066 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8069 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8070 sibling calls to functions that would normally be lazily bound,
8071 unless $gp is call-clobbered.
8073 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8075 * config/i386/i386.c (ix86_option_override_internal): Update
8076 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8077 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8078 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8079 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8081 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8083 2013-05-15 Andreas Schwab <schwab@suse.de>
8085 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8086 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8089 2013-05-15 Teresa Johnson <tejohnson@google.com>
8091 * loop-unroll.c (report_unroll_peel): Check decision before
8092 emitting unroll/peel message.
8094 2013-05-15 Teresa Johnson <tejohnson@google.com>
8096 * function.h (has_bb_partition): New rtl_data flag.
8097 (bb_reorder_complete): Ditto.
8098 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8099 instead of flag_reorder_blocks_and_partition.
8100 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8101 with some enhancements.
8102 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8103 * bb-reorder.c (connect_traces): Check for has_bb_partition
8104 instead of flag_reorder_blocks_and_partition.
8105 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8106 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8107 verify_hot_cold_block_grouping.
8108 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8110 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8113 * config/arm/predicates.md (call_insn_operand): New predicate.
8114 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8115 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8116 if insn is not a tail call.
8117 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8119 * config/arm/arm.h (enum reg_class): New caller save register class.
8120 (REG_CLASS_NAMES): Likewise.
8121 (REG_CLASS_CONTENTS): Likewise.
8122 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8125 2013-05-15 Richard Biener <rguenther@suse.de>
8127 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8128 of MSG_OPTIMIZED_LOCATIONS.
8129 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8130 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8132 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8133 of MSG_OPTIMIZED_LOCATIONS.
8134 (execute_vect_slp): Likewise.
8135 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8136 (vect_create_cond_for_alias_checks): Likewise.
8137 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8138 (vect_recog_widen_mult_pattern): Likewise.
8139 (vect_recog_widen_sum_pattern): Likewise.
8140 (vect_recog_over_widening_pattern): Likewise.
8141 (vect_recog_widen_shift_pattern): Likewise.
8142 (vect_recog_vector_vector_shift_pattern): Likewise.
8143 (vect_recog_divmod_pattern): Likewise.
8144 (vect_recog_mixed_size_cond_pattern): Likewise.
8145 (vect_recog_bool_pattern): Likewise.
8146 (vect_pattern_recog_1): Likewise.
8148 2013-05-15 Martin Jambor <mjambor@suse.cz>
8150 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8151 non-functions to builtin_unreachable.
8152 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8153 correct when new direct edges were discovered.
8155 2013-05-15 Martin Jambor <mjambor@suse.cz>
8157 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8158 header, print symbol order instead of node uid, print more information
8159 about indirect edge targets.
8160 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8162 (ipa_make_edge_direct_to_target): Likewise.
8163 (remove_described_reference): Likewise.
8164 (propagate_controlled_uses): Likewise.
8165 (ipa_print_node_params): Also print symbol order.
8166 (ipcp_transform_function): Print symbol order instead of node uids.
8167 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8168 (cgraph_get_create_real_symbol_node): Likewise.
8169 * ipa-cp.c (print_lattice): Likewise.
8170 (print_all_lattices): Likewise.
8171 (determine_versionability): Likewise.
8172 (initialize_node_lattices): Likewise.
8173 (estimate_local_effects): Likewise.
8174 (update_profiling_info): Likewise.
8175 (create_specialized_node): Likewise.
8176 (perhaps_add_new_callers): Likewise.
8177 (decide_about_value): Likewise.
8178 (decide_whether_version_node): Likewise.
8179 (identify_dead_nodes): Likewise.
8180 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8181 (dump_inline_summary): Likewise.
8182 (estimate_node_size_and_time): Likewise.
8183 (inline_analyze_function): Likewise.
8184 * ipa-inline.c (report_inline_failed_reason): Likewise.
8185 (want_early_inline_function_p): Likewise.
8186 (edge_badness): Likewise.
8187 (update_edge_key): Likewise.
8188 (inline_small_functions): Likewise. Add dumping of order to two other
8190 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8191 instead of node uids.
8192 (propagate_pure_const): Likewise.
8193 (propagate_pure_const): Likewise.
8194 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8195 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8196 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8198 * tree-pretty-print.c (dump_function_header): Likewise.
8199 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8200 Print symbol order instead of node uids.
8202 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8204 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8205 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8207 2013-05-15 Richard Biener <rguenther@suse.de>
8209 PR tree-optimization/57275
8210 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8211 return value for fail to do runtime alias checks for gather loads.
8213 2013-05-15 Jan Hubicka <jh@suse.cz>
8217 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8218 weakrefs are not external.
8219 (lto_symtab_merge_decls): Fix thinko when dealing with
8220 non-lto_symtab decls.
8221 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8222 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8223 * varpool.c (dump_varpool_node): Dump more flags.
8225 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8227 * config/i386/i386.c (processor_alias_table): Add instruction
8228 FSGSBASE for AMD bdver3 architecture.
8230 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8232 * tree.c (warn_deprecated_use): Print file:line using locus color.
8233 * diagnostic.c (diagnostic_report_current_module): Print file:line
8234 and file:line:column using locus color.
8236 2013-05-14 Mike Stump <mikestump@comcast.net>
8238 * gdbinit.in: Add __null.
8240 2013-05-14 Mike Stump <mikestump@comcast.net>
8242 * recog.h: Rename struct recog_data to Recog_data.
8243 * recog.c: Likewise.
8244 * reload.c (can_reload_into): Likewise.
8245 * config/picochip/picochip.c: Likewise.
8247 2013-05-14 Mike Stump <mikestump@comcast.net>
8249 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8251 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8253 * resource.h (struct resources): Remove unch_memory member.
8254 (CLEAR_RESOURCE): Don't clear unch_memory.
8255 * resource.c (mark_referenced_resources): Don't set it.
8256 (mark_set_resources): Likewise.
8257 (mark_target_live_regs): Don't clear it.
8258 (init_resource_info): Likewise.
8259 * reorg.c (resource_conflicts_p): Don't compare it.
8260 (redundant_insn): Don't set it.
8262 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8264 * emit-rtl.c (next_label): Remove unused function.
8265 (skip_consecutive_labels, link_cc0_insns): Move to ...
8266 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8267 only place where these functions are used, and make them static.
8269 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8271 * fold-const.c (fold_negate_expr): Handle vectors.
8272 (fold_truth_not_expr): Make it static.
8273 (fold_invert_truthvalue): New static function.
8274 (invert_truthvalue_loc): Handle vectors. Do not call
8275 fold_truth_not_expr directly.
8276 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8277 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8278 (fold_comparison): Handle vector constants.
8279 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8280 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8281 * tree.h (fold_truth_not_expr): Remove declaration.
8283 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8285 * config/aarch64/aarch64-simd.md
8286 (aarch64_vcond_internal<mode>): Rename to...
8287 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8288 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8289 float modes. Clarify all iterator modes.
8290 (vcond<mode><mode>): Use new name for vcond expanders.
8291 (vcond<v_cmp_result><mode>): Likewise.
8292 (vcondu<mode><mode>: Likewise.
8293 * config/aarch64/iterators.md (VDQF_COND): New.
8295 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8298 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8299 variable for the shift amount. Check that we shift by non-negative
8302 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8305 * config/arm/arm.h (EPILOGUE_USES): Only return true
8306 for LR_REGNUM after epilogue_completed.
8308 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8310 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8313 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8316 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8317 and Solaris 11+/x86 with gld.
8318 * configure: Regenerate.
8320 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8322 * expmed.c (expand_shift_1): Canonicalize rotates by
8323 constant bitsize / 2 to bitsize - 1.
8324 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8325 case ROTATERT>: Likewise.
8328 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8330 * config/i386/i386.md (rotateinv): New code attr.
8331 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8332 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8333 roll $31, %eax, etc.
8335 2013-05-14 Richard Biener <rguenther@suse.de>
8338 * tree-eh.c (sink_clobbers): Give up for successors with
8339 multiple predecessors and no virtual uses.
8341 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8343 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8344 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8346 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8349 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8350 the case when both op0 and op1 have VOIDmode.
8352 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8354 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8355 in multiply-accumulate mode.
8357 2013-05-13 Guozhi Wei <carrot@google.com>
8359 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8361 2013-05-13 Kai Tietz <ktietz@redhat.com>
8364 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8365 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8366 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8367 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8368 (ix86_expand_prologue): Likewise.
8369 (ix86_expand_split_stack_prologue): Likewise.
8370 (legitimate_pic_address_disp_p): Likewise.
8371 (legitimize_pic_address): Likewise.
8372 (legitimize_tls_address): Likewise.
8373 (legitimize_pe_coff_symbol): Likewise.
8374 (output_pic_addr_const): Likewise.
8375 (construct_plt_address): Likewise.
8376 (ix86_expand_call): Likewise.
8377 (x86_output_mi_thunk): Likewise.
8378 (x86_function_profiler): Likewise.
8380 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
8382 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
8383 similar switch cases.
8384 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
8385 (aarch64_simd_mov_to_<mode>low): Delete.
8386 (aarch64_simd_mov_to_<mode>high): Delete.
8387 (move_lo_quad_<mode>): Add w<-r alternative.
8388 (aarch64_simd_move_hi_quad_<mode>): Likewise.
8389 (aarch64_simd_mov_from_*): Update type attribute.
8390 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
8393 2013-05-13 Jan Hubicka <jh@suse.cz>
8395 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
8396 * config/i386/i386.c (ix86_compute_frame_layout,
8397 ix86_expand_epilogue, emit_i387_cw_initialization,
8398 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
8399 ix86_local_alignment): Fix use of size/speed predicates.
8401 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8403 PR tree-optimization/45216
8404 PR tree-optimization/57157
8405 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
8406 the (-Y) & (B - 1) variant if OP is |.
8407 * expmed.c (expand_shift_1): For rotations by const0_rtx just
8408 return shifted. Use (-op1) & (prec - 1) as other_amount
8409 instead of prec - op1.
8411 2013-05-13 Martin Jambor <mjambor@suse.cz>
8414 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
8415 (ipa_constant_data): New type.
8416 (ipa_jump_func): Use ipa_constant_data to hold information about
8417 constant jump functions.
8418 (ipa_get_jf_constant): Adjust to jump function type changes.
8419 (ipa_get_jf_constant_rdesc): New function.
8420 (ipa_param_descriptor): New field controlled_uses.
8421 (ipa_get_controlled_uses): New function.
8422 (ipa_set_controlled_uses): Likewise.
8423 * ipa-ref.h (ipa_find_reference): Declare.
8424 * ipa-prop.c (ipa_cst_ref_desc): New type.
8425 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
8427 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
8428 New parameter cs. Adjust all callers.
8429 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
8430 (remove_described_reference): New function.
8431 (jfunc_rdesc_usable): Likewise.
8432 (try_make_edge_direct_simple_call): Decrement controlled use count,
8433 attempt to remove reference if it hits zero.
8434 (combine_controlled_uses_counters): New function.
8435 (propagate_controlled_uses): Likewise.
8436 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
8437 (ipa_edge_duplication_hook): Duplicate reference descriptions.
8438 (ipa_print_node_params): Print described use counter.
8439 (ipa_write_jump_function): Adjust to jump function type changes.
8440 (ipa_read_jump_function): New parameter CS, pass it to
8441 ipa_set_jf_constant. Adjust caller.
8442 (ipa_write_node_info): Stream controlled use count
8443 (ipa_read_node_info): Likewise.
8444 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
8446 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
8447 count. Remove cloning-added reference if it reaches zero.
8448 * ipa-ref.c (ipa_find_reference): New function.
8450 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8452 * config/i386/i386.c (processor_target_table): Modified default
8453 alignment values for AMD BD and BT architectures.
8455 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
8457 * tree-vect-generic.c (uniform_vector_p): Move ...
8458 * tree.c (uniform_vector_p): ... here.
8459 * tree.h (uniform_vector_p): Declare it.
8460 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
8463 2013-05-13 Jakub Jelinek <jakub@redhat.com>
8465 PR tree-optimization/57230
8466 * tree-ssa-strlen.c (handle_char_store): Record length for
8467 array store from STRING_CST.
8469 PR tree-optimization/57230
8470 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
8473 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
8475 * config/epiphany/epiphany.c (epiphany_init): Check size of
8476 NUM_MODES_FOR_MODE_SWITCHING.
8477 (epiphany_expand_prologue):
8478 Remove CONFIG_REGNUM initial value handling code.
8479 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
8480 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
8481 (emit_set_fp_mode, epiphany_mode_after): Likewise.
8482 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
8483 Don't return 1 for FP_MODE_NONE.
8484 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
8485 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
8486 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
8487 * config/epiphany/epiphany.md (save_config): New pattern.
8489 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
8491 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
8493 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
8495 * config/i386/i386.md (memory): Handle sseishft1.
8496 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
8497 (*vec_extractv2di_1): Ditto.
8499 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
8501 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
8504 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
8506 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
8507 Add mthumb/march=armv7-a multilib.
8508 Add mthumb/march=armv7-r multilib.
8509 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
8511 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
8513 * config/v850/t-rtems: Add more multilibs.
8515 2013-05-10 Richard Biener <rguenther@suse.de>
8517 PR tree-optimization/57214
8518 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
8519 not propagate from SSA names that occur in abnormal PHI nodes.
8521 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
8523 * stor-layout.c (element_precision): New function.
8524 * machmode.h (element_precision): Declare it.
8525 * tree.c (build_minus_one_cst): New function.
8526 (element_precision): Likewise.
8527 * tree.h (build_minus_one_cst): Declare new function.
8528 (element_precision): Likewise.
8529 * fold-const.c (operand_equal_p): Use element_precision.
8530 (fold_binary_loc): Handle vector types.
8531 * convert.c (convert_to_integer): Use element_precision.
8532 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
8535 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
8537 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
8538 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
8539 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
8540 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8541 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
8542 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
8543 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
8544 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
8545 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
8546 (Uuw8): New constraints.
8547 (Usb4): Move into alphabetical order.
8548 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
8549 (sd8_operand, ub8_operand, uw8_operand): New predicates.
8550 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
8551 previously unnamed patterns.
8552 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
8553 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
8554 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
8555 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
8556 of set_attr_alternative/if_then_else. Use extended_mips16 instead
8557 of specific lengths.
8559 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8561 * config/i386/i386.md (rotateinv): New code attr.
8562 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8563 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8564 roll $31, %eax, etc.
8566 PR tree-optimization/45216
8567 PR tree-optimization/57157
8568 * tree-ssa-forwprop.c (simplify_rotate): New function.
8569 (ssa_forward_propagate_and_combine): Call it.
8571 2013-05-10 Richard Biener <rguenther@suse.de>
8573 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
8574 disable peeling when we version for aliasing.
8575 (vector_alignment_reachable_p): Honor explicit user alignment.
8576 (vect_supportable_dr_alignment): Likewise.
8577 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
8578 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
8579 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
8580 then peeling to arrange for the cost-model check to come first.
8582 2013-05-10 Alan Modra <amodra@gmail.com>
8584 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
8585 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
8586 * configure: Regenerate.
8588 2013-05-10 Alan Modra <amodra@gmail.com>
8591 * varasm.c (default_elf_select_section): Move !DECL_P check..
8592 (get_named_section): ..to here before calling get_section_name.
8594 (default_section_type_flags): Add DECL_P check.
8595 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
8596 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
8598 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
8600 * config/epiphany/epiphany.c (epiphany_expand_prologue):
8601 When using gen_stack_adjust_str with a register offset, add a
8602 REG_FRAME_RELATED_EXPR note.
8604 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
8606 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
8607 (*vec_extractv4si_zext_mem): Ditto.
8608 (*vec_extractv2di): Add 0->x and x->x alternatives.
8609 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
8610 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
8612 2013-05-09 Jason Merrill <jason@redhat.com>
8614 N3639 C++1y VLA support
8615 * gimplify.c (gimplify_vla_decl): Don't touch an existing
8618 * tree.c (build_constructor_va): New.
8619 * tree.h: Declare it.
8621 2013-05-09 Martin Jambor <mjambor@suse.cz>
8624 * gimple-fold.c (canonicalize_constructor_val): Call
8625 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
8627 2013-05-09 Jan Hubicka <jh@suse.cz>
8628 Richard Biener <rguenther@suse.de>
8631 * symtab.c (symtab_make_decl_local): Do not add private names.
8633 2013-05-09 Jan Hubicka <jh@suse.cz>
8636 * symtab.c (insert_to_assembler_name_hash): Handle clones.
8637 (unlink_from_assembler_name_hash): Likewise.
8638 (symtab_prevail_in_asm_name_hash, symtab_register_node,
8639 symtab_unregister_node, symtab_initialize_asm_name_hash,
8640 change_decl_assembler_name): Update.
8642 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
8644 * config/aarch64/aarch64.md: New movtf split.
8645 (*movtf_aarch64): Update.
8646 (aarch64_movdi_tilow): Handle TF modes and rename to
8647 aarch64_movdi_<mode>low.
8648 (aarch64_movdi_tihigh): Handle TF modes and rename to
8649 aarch64_movdi_<mode>high
8650 (aarch64_movtihigh_di): Handle TF modes and rename to
8651 aarch64_mov<mode>high_di
8652 (aarch64_movtilow_di): Handle TF modes and rename to
8653 aarch64_mov<mode>low_di
8654 (aarch64_movtilow_tilow): Remove spurious whitespace.
8655 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
8657 (aarch64_print_operand): Update.
8659 2013-05-09 Alan Modra <amodra@gmail.com>
8661 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
8663 * configure: Regenerate.
8665 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8667 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
8668 splitter preparation statements.
8669 * config/i386/sse.md (*vec_extract* splitters): Ditto.
8670 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
8673 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8675 * gimple-ssa-strength-reduction.c (count_candidates): Change
8676 return value to int.
8677 (analyze_candidates_and_replace): Change type of length to int.
8679 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
8681 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
8682 (*vec_extract<mode>): Use VI12_128 mode iterator.
8683 (*vec_extract<mode>_mem): Ditto.
8684 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
8687 2013-05-08 Diego Novillo <dnovillo@google.com>
8692 2012-08-17 Diego Novillo <dnovillo@google.com>
8695 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
8696 * config.in: Regenerate.
8697 * configure: Regenerate.
8698 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
8700 2013-05-08 Jan Hubicka <jh@suse.cz>
8703 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
8704 * cgraph.h (symtab_node_base): Add unique_name.
8705 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8706 input_overwrite_node, input_varpool_node): Stream unique_name.
8707 * cgraphclones.c (cgraph_create_virtual_clone,
8708 cgraph_function_versioning): Set unique_name.
8709 * ipa.c (function_and_variable_visibility): Set unique_name.
8711 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8713 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
8714 (alloc_cand_and_find_basis): Restrict conditional candidate
8715 processing to CAND_MULTs.
8717 2013-05-08 Jan Hubicka <jh@suse.cz>
8720 lto-symtab.c (lto_symtab_symbol_p): New function.
8721 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
8722 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
8723 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
8724 Skip static symbols.
8726 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
8728 PR tree-optimization/57200
8729 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
8730 Only call inform if the preceding warning_at returns true.
8732 2013-05-07 Han Shen <shenhan@google.com>
8734 * cfgexpand.c (record_or_union_type_has_array_p): New function.
8735 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
8736 * common.opt (fstack-protector-strong): New option.
8737 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
8738 * doc/invoke.texi (Optimization Options): Document
8739 "-fstack-protector-strong".
8740 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
8742 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8744 * config/mips/mips.c (mips_machine_reorg2): Return 0.
8746 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
8748 * ira.c (update_equiv_regs): Add insn having equiv memory even if
8749 it is not lhs of the insn.
8750 (setup_reg_equiv): Remove insn having equiv memory which it is not
8752 * lra-constraints.c (process_address): Try to improve generation
8753 code for address base + disp.
8754 (lra_constraints): Make correct the code for checking insn setting
8755 up backward equivalence. Remove insn only if it is in the init
8757 * lra-eliminations.c (update_reg_eliminate): Change return value.
8758 (lra_eliminate): Use the result.
8760 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
8762 * config/i386/sse.md (ssescalarnummask): New mode attribute.
8763 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
8764 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
8765 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
8766 register target operands.
8767 (*vec_extractv8hi_sse2): New pattern.
8768 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
8769 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
8770 (*vec_extract<mode>_mem): New insn and split pattern.
8772 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
8774 * config/arm/arm.c (arm_asan_shadow_offset): New function.
8775 (TARGET_ASAN_SHADOW_OFFSET): Define.
8776 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
8777 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
8779 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8781 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
8782 (incr_vec_index): Return -1 if increment not found.
8783 (create_add_on_incoming_edge): Assert if increment not found.
8784 (record_increment): Limit number of increments recorded.
8785 (all_phi_incrs_profitable): Return false if an increment not found.
8786 (replace_profitable_candidates): Don't process increments that were
8788 (analyze_candidates_and_replace): Limit size of incr_vec.
8790 2013-05-07 Richard Biener <rguenther@suse.de>
8792 * calls.c (special_function_p): setjmp-like functions are leaf.
8793 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
8794 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
8796 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
8798 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
8799 (aarch64_simd_mov<mode>): New expander.
8800 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
8801 (aarch64_simd_mov_to_<mode>high): Likewise.
8802 (aarch64_simd_mov_from_<mode>low): Likewise.
8803 (aarch64_simd_mov_from_<mode>high): Likewise.
8804 (aarch64_dup_lane<mode>): Update.
8805 (aarch64_dup_lanedi): New instruction pattern.
8806 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
8807 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
8809 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8811 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
8812 (replace_mult_candidate): Remove unnecessary argument; remove
8813 unnecessary parameter from call to introduce_cast_before_cand.
8814 (replace_unconditional_candidate): Remove unnecessary parameter
8815 from call to replace_mult_candidate.
8816 (replace_conditional_candidate): Likewise.
8817 (insert_initializers): Use make_temp_ssa_name.
8818 (introduce_cast_before_cand): Remove unnecessary argument; use
8820 (replace_one_candidate): Remove unnecessary argument; remove
8821 unnecessary parameter from calls to introduce_cast_before_cand.
8822 (replace_profitable_candidates): Remove unnecessary parameters
8823 from calls to replace_one_candidate.
8825 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8827 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
8828 phi def as possibly hiding a basis for a CAND_ADD whose operands
8829 have been commuted in the analysis.
8830 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
8832 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
8834 * config/aarch64/aarch64.md
8835 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
8836 shift value between 0-4.
8838 2013-05-07 Richard Biener <rguenther@suse.de>
8840 * double-int.h (rshift): New overload.
8841 * double-int.c (rshift): New function.
8842 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
8843 (create_reference_ops_from_ref): Remove.
8844 (vn_reference_insert): Use shared ops for constructing the
8845 reference and copy it.
8847 2013-05-07 Richard Biener <rguenther@suse.de>
8850 * tree-eh.c (sink_clobbers): Properly propagate
8851 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
8853 2013-05-07 Jakub Jelinek <jakub@redhat.com>
8855 PR tree-optimization/57149
8856 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
8857 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
8858 collect_phi_def_edges, execute_late_warn_uninitialized): Use
8859 uninit_undefined_value_p instead of ssa_undefined_value_p.
8862 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
8863 for modifier == EXPAND_INITIALIZER.
8865 2013-05-07 Anton Blanchard <anton@samba.org>
8867 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
8868 for powerpc64 little endian.
8869 * configure: Regenerate.
8871 2013-05-06 Graham Stott <grahams@btinternet.com>
8873 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
8874 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
8875 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
8876 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
8878 2013-05-06 Graham Stott <grahams@btinternet.com>
8880 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
8881 codes which allow non-lvalues.
8883 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
8885 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
8886 components are all 1s.
8887 (integer_minus_onep): New function.
8888 * tree.h (integer_minus_onep): Declare it.
8889 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
8890 integer_minus_onep instead of integer_all_onesp.
8892 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8895 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
8896 variations of these patterns.
8898 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8900 * config/i386/i386.md (isa): Add x64_sse4 member.
8901 (enabled): Handle x64_sse4.
8902 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
8903 instruction for 64bit SSE4_1 targets. Update insn attributes.
8904 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
8905 instruction for SSE4_1 targets. Update insn attributes.
8906 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
8907 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
8908 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
8910 (*vec_extractv4si): Rename from *sse4_1_pextrd.
8911 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
8912 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
8914 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8917 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
8919 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
8921 * final.c (do_assembler_dialects): Don't handle curly braces and
8922 vertical bar escaped by % as dialect delimiters.
8923 (output_asm_insn): Print curly braces and vertical bar if escaped
8924 by % and ASSEMBLER_DIALECT defined.
8925 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
8926 * doc/tm.texi: Regenerated.
8928 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
8930 * config/mips/mips.c: Include tree-pass.h.
8931 (mips_reorg): Split in pre- and post-dbr_schedule parts.
8932 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
8933 (pass_mips_machine_reorg2): New machine specific pass.
8934 (insert_pass_mips_machine_reorg2): New pass plugin definition.
8935 (mips_option_override): Register the new pass.
8936 * rtl.h (cleanup_barriers): Remove prototype.
8937 (dbr_schedule): Likewise.
8938 * jump.c (cleanup_barriers): Make static.
8939 * reorg.c (dbr_schedule): Likewise.
8941 2013-05-06 Richard Biener <rguenther@suse.de>
8943 PR tree-optimization/57185
8944 * tree-parloops.c (add_field_for_reduction): Handle anonymous
8947 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
8950 * config/i386/i386.c (add_parameter_dependencies): Add dependence
8951 between "first_arg" and "insn", not "last" and "insn".
8953 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8955 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
8956 (find_candidates_in_block): Re-enable slsr_process_phi.
8957 (create_phi_basis): Fix double counting of candidate adjustment.
8959 2013-05-06 Richard Biener <rguenther@suse.de>
8962 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
8963 the edge is also fallthru, preserve it and just clear the
8965 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
8966 also complex, preserve that and just clear the fallthru flag.
8967 * tree-inline.c (update_ssa_across_abnormal_edges): Also
8968 update virtual operands.
8970 2013-05-06 Alan Modra <amodra@gmail.com>
8972 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
8973 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
8974 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
8975 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
8976 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8977 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
8979 2013-05-06 Alan Modra <amodra@gmail.com>
8981 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
8982 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
8983 (DEFAULT_ASM_ENDIAN): Define.
8984 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
8985 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
8986 Update -K PIC clause from sysv4.h.
8987 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
8988 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
8990 2013-05-06 Alan Modra <amodra@gmail.com>
8992 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
8993 twice for little-endian.
8994 (ashrdi3_no_power, ashrdi3): Support little-endian.
8996 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
8999 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9000 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9002 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9003 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9004 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9007 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9008 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9010 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9011 * config/pa/pa.opt: Make mbig-switch a no-op.
9012 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9013 (CASE_VECTOR_MODE): Always return SImode.
9014 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9015 for the !TARGET_BIG_SWITCH case.
9016 * config/pa/pa-linux.h: Likewise.
9017 * config/pa/pa-openbsd.h: Likewise.
9018 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9019 * config/pa/pa.md (short_jump): Remove define_insn.
9020 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9021 (casesi0): Remove define_insn.
9022 (type): Remove btable_branch.
9023 (pa_combine_type): Likewise.
9024 (in_nullified_branch_delay): Likewise.
9025 (in_call_delay): Likewise.
9026 (define_delay): Likewise.
9027 (define_insn_reservation "Z3"): Likewise.
9028 (define_insn_reservation "Z4"): Likewise.
9029 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9030 (pa_adjust_insn_length): Remove adjustment for btable branches.
9031 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9034 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9036 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9037 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9038 Add m->r,x alternatives.
9039 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9040 splitters using SWI48x mode iterator.
9041 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9042 TARGET_64BIT. Add m->x alternative.
9043 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9044 Add o->x alternative. Enable for TARGET_SSE.
9045 (sse_storeq): Remove expander.
9046 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9047 with memory input operand.
9048 (*vec_extractv2di_1 splitter): New.
9049 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9050 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9052 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9054 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9055 (INT_LOWPART): Delete.
9056 (extract_MB): Adjust.
9057 (extract_ME): Adjust.
9058 (print_operand): Adjust.
9060 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9062 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9063 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9064 (reg_or_logical_cint_operand, easy_fp_constant,
9065 logical_const_operand): Delete "CONST_DOUBLE" case.
9066 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9067 "HOST_BITS_PER_WIDE_INT == 64" test.
9068 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9069 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9070 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9071 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9073 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9074 CONST_DOUBLE DImode/VOIDmode case.
9075 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9076 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9077 CONST_DOUBLE VOIDmode case.
9078 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9079 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9080 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9081 Delete CONST_DOUBLE case.
9082 (splitters for mov FMOVE64 const_double): Delete
9083 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9084 "HOST_BITS_PER_WIDE_INT >= 64" test.
9085 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9087 (mov DI const_double): Delete.
9089 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9091 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9092 on op shows all bits zero in mode of a lowpart subreg, return zero.
9094 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9097 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9098 to save TFmode registers and DImode to save TImode registers for
9099 caller save operations.
9100 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9101 mark being partially clobbered since they only use the first
9104 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9105 and TDmode only use the upper 64-bits of each VSX register.
9107 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9109 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9110 (find_candidates_in_block): Disable slsr_process_phi.
9112 2013-05-03 Guozhi Wei <carrot@google.com>
9114 * coverage.c (coverage_obj_init): Move the construction of gcov
9116 (build_init_ctor): ... here.
9118 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9120 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9121 (slsr_cand_d): Redefine def_phi.
9122 (stride_status, phi_adjust_status, count_phis_status): New enums.
9123 (find_phi_def): New.
9124 (find_basis_for_base_expr): New.
9125 (find_basis_for_candidate): Handle hidden bases.
9126 (alloc_cand_and_find_basis): Handle phi candidates.
9127 (slsr_process_phi): New.
9128 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9129 (create_mul_imm_cand): Likewise.
9130 (create_add_ssa_cand): Exclude phi base candidates.
9131 (create_add_imm_cand): Likewise.
9132 (slsr_process_cast): Likewise.
9133 (slsr_process_copy): Likewise.
9134 (find_candidates_in_block): Handle phi candidates.
9135 (dump_candidate): Likewise.
9136 (unconditional_cands): Delete.
9137 (unconditional_cands_with_known_stride_p): Delete.
9138 (phi_dependent_cand_p): New.
9139 (cand_increment): Handle phi-dependent candidates.
9140 (replace_dependent): Delete.
9141 (replace_mult_candidate): New.
9142 (replace_unconditional_candidate): New.
9143 (incr_vec_index): Move to avoid forward reference.
9144 (create_add_on_incoming_edge): New.
9145 (create_phi_basis): New.
9146 (replace_dependents): Delete.
9147 (replace_conditional_candidate): New.
9148 (phi_add_costs): New.
9149 (replace_uncond_cands_and_profitable_phis): New.
9150 (record_increment): Handle phi adjustments.
9151 (record_phi_increments): New.
9152 (record_increments): Handle phi adjustments.
9153 (phi_incr_cost): New.
9154 (lowest_cost_path): Handle phis.
9155 (total_savings): Likewise.
9156 (analyze_increments): Likewise.
9157 (ncd_with_phi): New.
9158 (ncd_of_cand_and_phis): New.
9159 (nearest_common_dominator_for_cands): Handle phi increments.
9160 (all_phi_incrs_profitable): New.
9161 (replace_profitable_candidates): Handle phi-dependent candidates.
9162 (analyze_candidates_and_replace): Likewise.
9164 2013-05-03 Teresa Johnson <tejohnson@google.com>
9167 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9168 do not exceed REG_BR_PROB_BASE.
9170 2013-05-03 Jeff Law <law@redhat.com>
9172 PR tree-optimization/57144
9173 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9174 operand of the condition will bit into the new type when eliminating
9175 a cast feeding a condition.
9177 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9179 PR rtl-optimization/57130
9180 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9181 of COMPARE as in_code to the recursive call if needed.
9183 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9185 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9186 (enabled): Handle new members.
9187 * config/i386/sse.md (*vec_concatv2si): Merge from
9188 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9189 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9191 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9193 PR tree-optimization/57027
9194 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9195 for fnms opportunity, check we got the prerequisite kind
9196 of tree / gimple before using accessor functions.
9198 2013-05-03 Richard Biener <rguenther@suse.de>
9200 * double-int.h (lshift): New overload without precision
9202 (operator *=, operator +=, operator -=): Move ...
9203 * double-int.c (operator *=, operator +=, operator -=): ... here
9204 and implement more efficiently.
9205 (mul_double_with_sign): Remove.
9206 (lshift_double): Adjust to take unsinged shift argument, push
9207 dispatching code to callers.
9208 (mul_double_wide_with_sign): Add early out for callers that
9209 are not interested in high parts or overflow.
9210 (lshift): New function.
9211 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9213 (lrotate, rrotate): Use logical shifts.
9214 * expr.c (get_inner_reference): Use lshift.
9215 * fixed-value.c (do_fixed_divide): Likewise.
9216 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9217 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9218 (indirect_refs_may_alias_p): Likewise.
9219 (stmt_kills_ref_p_1): Likewise.
9221 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9223 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9225 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9227 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9228 scalar form of FABD instruction.
9230 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9232 * lra-constraints.c (process_alt_operands): Add checking alt
9233 number to choose the best alternative.
9235 2013-05-02 Richard Biener <rguenther@suse.de>
9237 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9238 bitmap and its handling.
9239 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9241 2013-05-02 Richard Biener <rguenther@suse.de>
9244 * tree-inline.c (copy_loops): Properly handle removed loops.
9245 (copy_cfg_body): Mark destination loops for fixup if source
9248 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9251 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9252 generating simple_return for naked functions.
9254 2013-05-02 Martin Jambor <mjambor@suse.cz>
9257 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9258 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9260 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9261 ipa_agg_replacement_value structures.
9262 (known_aggs_to_agg_replacement_list): Likewise.
9263 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9264 (read_agg_replacement_chain): Likewise.
9265 (ipcp_transform_function): Also check that by_ref flags match.
9267 2013-05-02 Richard Biener <rguenther@suse.de>
9269 * graphds.h (struct graph): Add obstack member.
9270 * graphds.c (new_graph): Initialize obstack and allocate
9272 (add_edge): Allocate edge from the obstack.
9273 (free_graph): Free the obstack instead of all edges and vertices.
9275 2013-05-02 Teresa Johnson <tejohnson@google.com>
9277 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9279 * cfg.c (update_bb_profile_for_threading): Ditto.
9280 * tree-inline.c (copy_bb): Ditto.
9281 (copy_edges_for_bb): Ditto.
9282 (initialize_cfun): Ditto.
9283 (copy_cfg_body): Ditto.
9284 (expand_call_inline): Ditto.
9285 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9286 (estimate_node_size_and_time): Ditto.
9287 (inline_merge_summary): Ditto.
9288 * cgraphclones.c (cgraph_clone_edge): Ditto.
9289 (cgraph_clone_node): Ditto.
9290 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9291 (compute_trg_info): Ditto.
9293 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9295 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9296 S reg when fp attribute set.
9297 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9299 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9301 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9303 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9304 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9305 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9307 2013-05-02 Richard Biener <rguenther@suse.de>
9309 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9310 (struct instantiate_cache_entry): New type.
9311 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9312 (struct instantiate_cache_type): New type.
9313 (set_instantiated_value, get_instantiated_value): Remove.
9314 (get_instantiated_value_entry): New function.
9315 (instantiate_scev_name): Use the new cache and adjust.
9316 (instantiate_scev_poly): Adjust.
9317 (instantiate_scev_binary): Likewise.
9318 (instantiate_array_ref): Likewise.
9319 (instantiate_scev_convert): Likewise.
9320 (instantiate_scev_not): Likewise.
9321 (instantiate_scev_3): Likewise.
9322 (instantiate_scev_2): Likewise.
9323 (instantiate_scev_r): Likewise.
9324 (instantiate_scev): Likewise.
9325 (resolve_mixers): Likewise.
9327 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9330 * lra-constraints.c (best_small_class_operands_num): Remove.
9331 (process_alt_operands): Remove small_class_operands_num. Take
9332 small classes operands into losers and only if the operand is not
9333 matched. Modify debugging output.
9334 (curr_insn_transform): Remove best_small_class_operands_num.
9337 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9339 * config/aarch64/aarch64-builtins.c
9340 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9341 * config/aarch64/aarch64-simd-builtins.def
9342 (reduc_splus_): Add new modes.
9343 (reduc_uplus_): New.
9344 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9345 (reduc_uplus_v4sf): Likewise.
9346 (reduc_splus_v4sf): Likewise.
9347 (aarch64_addv<mode>): Likewise.
9348 (reduc_uplus_<mode>): Likewise.
9349 (reduc_splus_<mode>): Likewise.
9350 (aarch64_addvv2di): Likewise.
9351 (reduc_uplus_v2di): Likewise.
9352 (reduc_splus_v2di): Likewise.
9353 (aarch64_addvv2si): Likewise.
9354 (reduc_uplus_v2si): Likewise.
9355 (reduc_splus_v2si): Likewise.
9356 (reduc_<sur>plus_<mode>): New.
9357 (reduc_<sur>plus_v2di): Likewise.
9358 (reduc_<sur>plus_v2si): Likewise.
9359 (reduc_<sur>plus_v4sf): Likewise.
9360 (aarch64_addpv4sf): Likewise.
9361 * config/aarch64/arm_neon.h
9362 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9363 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9364 add UNSPEC_SADDV, UNSPEC_UADDV.
9366 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9368 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9370 * config/aarch64/arm_neon.h
9371 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9373 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9375 * config/aarch64/aarch64-builtins
9376 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
9378 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9380 * config/aarch64/aarch64-simd-builtins.def
9382 (reduc_smin_): Likewise.
9383 (reduc_umax_): Likewise.
9384 (reduc_umin_): Likewise.
9385 (reduc_smax_nan_): Likewise.
9386 (reduc_smin_nan_): Likewise.
9389 (smax): Update for V2SF, V4SF and V2DF modes.
9392 (smin_nan): Likewise.
9393 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
9394 (<su><maxmin><mode>3): ...This, refactor.
9395 (s<maxmin><mode>3): New.
9396 (<maxmin_uns><mode>3): Likewise.
9397 (reduc_<maxmin_uns>_<mode>): Refactor.
9398 (reduc_<maxmin_uns>_v4sf): Likewise.
9399 (reduc_<maxmin_uns>_v2si): Likewise.
9400 (aarch64_<fmaxmin><mode>: Remove.
9401 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
9403 (vmin<q>_f<32,64>): Likewise.
9404 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9406 (su): Add mappings for smax, smin, umax, umin.
9408 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
9409 (FMAXMIN): Rename as...
9410 (FMAXMIN_UNS): ...This.
9412 (fmaxminv): Likewise.
9413 (fmaxmin): Likewise.
9415 (maxmin_uns_op): Likewise.
9417 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9419 * config/aarch64/arm_neon.h
9420 (vac<ge, gt><sd>_f<32, 64>): Rename to...
9421 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
9422 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
9424 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9426 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
9427 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
9429 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9431 * config/aarch64/aarch64-simd.md
9432 (vcond<mode>_internal): Handle special cases for constant masks.
9433 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
9434 (vcondu<mode><mode>): Likewise.
9435 (vcond<v_cmp_result><mode>): New.
9437 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9439 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
9440 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
9441 * config/aarch64/aarch64-simd-builtins.def
9442 (cmeq): Update to BUILTIN_VALLDI.
9447 * config/aarch64/arm_neon.h
9448 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
9449 to builtins or C as appropriate.
9451 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9453 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
9455 (cmhi): Rename to...
9457 * config/aarch64/aarch64-simd.md
9458 (simd_mode): Add SF.
9459 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
9460 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
9461 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
9462 (cstore<mode>_neg): ...This.
9463 * config/aarch64/iterators.md
9465 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
9467 (UCOMPARISONS): Likewise.
9468 (optab): Add missing comparisons.
9476 (V_cmp_result): Add DF, SF modes.
9477 (v_cmp_result): Likewise.
9480 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
9482 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
9484 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
9485 define_insn to define_insn_and_split.
9486 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
9487 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
9488 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
9489 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
9490 (thumb2_negscc): Likewise.
9492 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9494 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
9496 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
9498 * config/arm/thumb2.md: Remove trailing whitespaces.
9500 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9502 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
9503 Use gen_int_mode rather than GEN_INT.
9505 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
9507 * value-prof.c (stream_in_histogram_value): Remove the strayed
9510 2013-04-30 Richard Biener <rguenther@suse.de>
9513 * cfghooks.c (split_edge): Properly check for the loop latch edge.
9515 2013-04-30 Richard Biener <rguenther@suse.de>
9518 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
9520 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
9522 PR rtl-optimization/56957
9523 PR rtl-optimization/57105
9524 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
9525 variable. Use just INSN_UID for determining whether an insn
9526 should be only disconnected from the insn stream.
9527 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
9529 2013-04-30 Jakub Jelinek <jakub@redhat.com>
9531 PR tree-optimization/57104
9532 * tsan.c (instrument_expr): Don't instrument accesses to
9533 DECL_HARD_REGISTER VAR_DECLs.
9535 2013-04-30 Richard Biener <rguenther@suse.de>
9537 * function.h (loops_for_fn): New inline function.
9538 (set_loops_for_fn): Likewise.
9539 * cfgloop.h (place_new_loop): Add struct function parameter.
9540 (get_loop): Likewise.
9541 (get_loops): Likewise.
9542 (number_of_loops): Likewise.
9544 (fel_init): Likewise.
9545 * cfg.c (get_loop_copy): Adjust.
9546 * cfgloop.c (flow_loops_dump): Likewise.
9547 (record_loop_exits): Likewise.
9548 (verify_loop_structure): Likewise.
9549 * cfgloopanal.c (mark_irreducible_loops): Likewise.
9550 (estimate_reg_pressure_cost): Likewise.
9551 (mark_loop_exit_edges): Likewise.
9552 * cfgloopmanip.c (place_new_loop): Likewise.
9553 (add_loop): Likewise.
9554 (duplicate_loop): Likewise.
9555 * graph.c (draw_cfg_nodes): Likewise.
9556 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
9557 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
9558 (extract_affine_chrec): Likewise.
9559 (build_scop_iteration_domain): Likewise.
9560 * graphite.c (graphite_initialize): Likewise.
9561 * ira-build.c (create_loop_tree_nodes): Likewise.
9562 (more_one_region_p): Likewise.
9563 (rebuild_regno_allocno_maps): Likewise.
9564 (mark_loops_for_removal): Likewise.
9565 (mark_all_loops_for_removal): Likewise.
9566 (remove_unnecessary_regions): Likewise.
9567 (ira_build): Likewise.
9568 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
9569 * loop-init.c (fix_loop_structure): Likewise.
9570 (gate_rtl_move_loop_invariants): Likewise.
9571 (gate_rtl_unswitch): Likewise.
9572 (gate_rtl_unroll_and_peel_loops): Likewise.
9573 (rtl_doloop): Likewise.
9574 * lto-streamer-in.c (input_cfg): Likewise.
9575 * lto-streamer-out.c (output_cfg): Likewise.
9576 * modulo-sched.c (sms_schedule): Likewise.
9577 * predict.c (tree_estimate_probability): Likewise.
9578 (tree_estimate_probability_driver): Likewise.
9579 (estimate_loops): Likewise.
9580 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
9581 (move_sese_region_to_fn): Likewise.
9582 (debug_loop_num): Likewise.
9583 * tree-chrec.c (chrec_evaluate): Likewise.
9584 (hide_evolution_in_other_loops_than_loop): Likewise.
9585 (chrec_component_in_loop_num): Likewise.
9586 (reset_evolution_in_loop): Likewise.
9587 (evolution_function_is_invariant_rec_p): Likewise.
9588 * tree-if-conv.c (main_tree_if_conversion): Likewise.
9589 * tree-inline.c (copy_loops): Likewise.
9590 (copy_cfg_body): Likewise.
9591 (tree_function_versioning): Likewise.
9592 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
9593 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
9595 (add_to_evolution_1): Likewise.
9596 (scev_const_prop): Likewise.
9597 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
9598 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
9599 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
9600 (tree_ssa_lim_initialize): Likewise.
9601 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
9602 (verify_loop_closed_ssa): Likewise.
9603 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
9604 (tree_ssa_loop_im): Likewise.
9605 (tree_ssa_loop_unswitch): Likewise.
9606 (tree_vectorize): Likewise.
9607 (check_data_deps): Likewise.
9608 (tree_ssa_loop_ivcanon): Likewise.
9609 (tree_ssa_loop_bounds): Likewise.
9610 (tree_complete_unroll): Likewise.
9611 (tree_complete_unroll_inner): Likewise.
9612 (tree_parallelize_loops): Likewise.
9613 (tree_ssa_loop_prefetch): Likewise.
9614 (tree_ssa_loop_ivopts): Likewise.
9615 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
9616 * tree-vectorizer.c (vectorize_loops): Likewise.
9618 2013-04-29 Mike Frysinger <vapier@gentoo.org>
9620 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
9621 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
9622 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
9623 with EABI_LINK_SPEC.
9625 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9628 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
9631 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
9634 * lra-constraints.c (process_alt_operands): Discourage a bit more
9635 using memory for pseudos. Print cost dump for alternatives.
9636 Modify cost values for conflicts with early clobbers.
9637 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
9639 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9642 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
9644 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9646 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
9648 (movdi_aarch64): Support LDR/STR from/to D register.
9650 2013-04-29 Ian Bolton <ian.bolton@arm.com>
9652 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
9653 or higher by default.
9655 2013-04-29 Richard Biener <rguenther@suse.de>
9658 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
9659 even if not adding abnormal edges for calls that can make
9662 2013-04-29 Richard Biener <rguenther@suse.de>
9665 * tree-cfg.c (move_stmt_op): Fix condition under which to update
9667 (move_stmt_r): Remove redundant checking.
9669 2013-04-29 Teresa Johnson <tejohnson@google.com>
9672 * basic-block.h (apply_scale): New function.
9673 (apply_probability): Use apply_scale.
9674 * gimple-streamer-in.c (input_bb): Ditto.
9675 * lto-streamer-in.c (input_cfg): Ditto.
9676 * lto-cgraph.c (merge_profile_summaries): Ditto.
9677 * tree-optimize.c (execute_fixup_cfg): Ditto.
9678 * tree-inline.c (copy_bb): Update comment to use apply_scale.
9679 (copy_edges_for_bb): Ditto.
9680 (copy_cfg_body): Ditto.
9682 2013-04-29 Tom de Vries <tom@codesourcery.com>
9684 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
9685 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
9686 (tail_merge_optimize): Handle current_loops == NULL.
9688 2013-04-26 Jeff Law <law@redhat.com>
9690 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
9691 (simplify_cond_using_ranges): Generalize code to simplify
9692 COND_EXPRs where one argument is a constant and the other
9693 is an SSA_NAME created by an integral type conversion.
9695 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9697 * config/arm/arm.md (store_minmaxsi): Use only when
9698 optimize_insn_for_size_p.
9700 2013-04-29 Christian Bruel <christian.bruel@st.com>
9703 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
9705 2013-04-29 Richard Biener <rguenther@suse.de>
9708 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
9709 loop tree make sure to schedule a fixup for the child as well.
9710 (expand_omp_for_generic): Properly add loops.
9711 (expand_omp_for_static_nochunk): Likewise.
9712 (expand_omp_for_static_chunk): Likewise.
9713 (expand_omp_for): For the degenerate case fixup loops.
9714 (expand_omp_sections): Fix default bb placement in loops.
9715 (expand_omp_atomic_pipeline): Properly add loops.
9717 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9719 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
9721 2013-04-29 Tom de Vries <tom@codesourcery.com>
9723 * tree-ssa-tail-merge.c: Update header comment.
9725 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9727 * config/aarch64/arm_neon.h
9728 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
9729 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
9730 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
9731 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
9732 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
9733 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
9734 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
9735 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
9737 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9739 * config/aarch64/aarch64-simd.md
9740 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
9741 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
9742 fix_trunc, fixuns_trunc.
9743 (ftrunc<VDQF:mode>2): New.
9744 * config/aarch64/iterators.md (optab): Add fix, fixuns.
9745 (fix_trunc_optab): New.
9747 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9749 * config/aarch64/aarch64-builtins.c
9750 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
9751 iceilf, lround, iroundf.
9753 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
9756 * config/i386/i386.h (enum ix86_tune_indices)
9757 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
9758 New, split from X86_TUNE_INTER_UNIT_MOVES.
9759 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
9760 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
9761 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
9762 (TARGET_INTER_UNIT_MOVES): Remove.
9763 * config/i386/i386.c (initial_ix86_tune_features): Update.
9764 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
9765 (ix86_expand_convert_uns_didf_sse): Use
9766 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9767 (ix86_expand_vector_init_one_nonzero): Ditto.
9768 (ix86_expand_vector_init_interleave): Ditto.
9769 (inline_secondary_memory_needed): Return true for moves from SSE class
9770 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
9771 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
9772 * config/i386/constraints.md (Yi, Ym): Depend on
9773 TARGET_INTER_UNIT_MOVES_TO_VEC.
9774 (Yj, Yn): New constraints.
9775 * config/i386/i386.md (*movdi_internal): Change constraints of
9776 operand 1 from Yi to Yj and from Ym to Yn.
9777 (*movsi_internal): Ditto.
9778 (*movdf_internal): Ditto.
9779 (*movsf_internal): Ditto.
9780 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
9781 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
9782 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
9783 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
9784 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
9785 * config/i386/sse.md (movdi_to_sse): Ditto.
9786 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
9787 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
9788 TARGET_INTER_UNIT_MOVES.
9789 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
9790 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
9791 instead of TARGET_INTER_UNIT_MOVES.
9792 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
9793 operand 1 from Yi to Yj and from Ym to Yn.
9795 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9797 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
9798 (float_truncate_hi_): Likewise.
9799 (float_extend_lo_): Likewise.
9800 (float_truncate_lo_): Likewise.
9801 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
9802 (aarch64_float_extend_lo_v2df): Likewise.
9803 (vec_unpacks_hi_v4sf): Likewise.
9804 (aarch64_float_truncate_lo_v2sf): Likewise.
9805 (aarch64_float_truncate_hi_v4sf): Likewise.
9806 (vec_pack_trunc_v2df): Likewise.
9807 (vec_pack_trunc_df): Likewise.
9809 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9811 * config/aarch64/aarch64-builtins.c
9812 (aarch64_fold_builtin): Fold float conversions.
9813 * config/aarch64/aarch64-simd-builtins.def
9814 (floatv2si, floatv4si, floatv2di): New.
9815 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
9816 * config/aarch64/aarch64-simd.md
9817 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
9818 * config/aarch64/iterators.md (FLOATUORS): New.
9819 (optab): Add float, floatuns.
9820 (su_optab): Likewise.
9822 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9824 * config/aarch64/aarch64-builtins.c
9825 (aarch64_builtin_vectorized_function): Use new names for
9827 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
9828 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
9829 (fcvtzu): Split as...
9830 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
9831 (fcvtas): Split as...
9832 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
9833 (fcvtau): Split as...
9834 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
9835 (fcvtps): Split as...
9836 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
9837 (fcvtpu): Split as...
9838 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
9839 (fcvtms): Split as...
9840 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
9841 (fcvtmu): Split as...
9842 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
9843 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
9844 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
9845 (lfrintnusf, lfrintnudf): Likewise.
9846 * config/aarch64/aarch64-simd.md
9847 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
9849 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
9850 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
9851 (fcvt_pattern): Likewise.
9853 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9855 * config/aarch64/aarch64-simd.md
9856 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
9857 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
9859 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9861 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
9862 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
9863 (vrnd<a,m,n,p>_f32): Implement using builtins.
9864 (vrnd<i,x><q>_f<32, 64>): New.
9866 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
9868 * config/aarch64/aarch64-builtins.c
9869 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
9870 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
9871 (frintz): Rename to...
9873 (frintp): Rename to...
9875 (frintm): Rename to...
9877 (frinti): Rename to...
9878 (nearbyint): ...this.
9879 (frintx): Rename to...
9881 (frinta): Rename to...
9883 * config/aarch64/aarch64-simd.md
9884 (aarch64_frint<frint_suffix><mode>): Delete.
9885 (<frint_pattern><mode>2): Convert to insn.
9886 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
9887 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
9888 (frint_pattern): Likewise.
9889 (frint_suffix): Likewise.
9891 2013-04-29 Richard Biener <rguenther@suse.de>
9893 PR tree-optimization/57081
9894 * loop-init.c: Include tree-flow.h.
9895 (loop_optimizer_finalize): Free number of iteration estimates.
9896 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
9898 2013-04-29 Jakub Jelinek <jakub@redhat.com>
9900 PR tree-optimization/57083
9901 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
9902 non-singleton shift count range, zero extend low_bound for uns case.
9904 * config/i386/predicates.md (general_vector_operand): New predicate.
9905 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
9906 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
9907 if they aren't nonimmediate operands. If their original values
9908 satisfy const_vector_equal_evenodd_p, don't shift them.
9909 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
9910 predicates. For the SSE4.1 case force operands[{1,2}] into registers
9911 if not nonimmediate_operand.
9912 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
9913 instead of register_operand.
9914 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
9916 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
9918 * stor-layout.c (finalize_size_functions): Allocate a structure and
9919 reset cfun before dumping the functions.
9921 2013-04-27 Jakub Jelinek <jakub@redhat.com>
9923 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
9926 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
9927 use xop_pmacsdqh if uns_p.
9928 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
9929 the immediate rotate count.
9931 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
9933 * rtl.h (struct rtx_def): Add comment for field jump.
9934 (LRA_SUBREG_P): New macro.
9935 * recog.c (register_operand): Check LRA_SUBREG_P.
9936 * lra.c (lra): Add note at the end of RTL code. Align non-empty
9938 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
9939 (lra_final_code_change): Skip subreg change for operators.
9940 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
9941 if there are no operand changes.
9942 * lra-constraints.c (curr_insn_set): New.
9943 (match_reload): Set LRA_SUBREG_P.
9944 (emit_spill_move): Ditto.
9945 (check_and_process_move): Use curr_insn_set. Process only single
9946 set insns. Don't initialize sec_mem_p and change_p.
9947 (simplify_operand_subreg): Use LRA_SUBREG_P.
9948 (reg_in_class_p): New function.
9949 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
9950 of #ifdef. Add code to remove cycling.
9951 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
9952 non-null disp. Reload inner instead of disp when base and index
9953 are null. Try to put lo_sum into register.
9954 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
9955 (check_and_process_move): Move code for move cost check to
9956 simple_move_p. Remove equiv_substitution.
9957 (simple_move_p): New function.
9958 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
9959 curr_insn_set. Call check_and_process_move only for single set
9960 insns. Use the new function. Move call of check_and_process_move
9961 after operand equiv substitution and address process.
9963 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9966 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
9967 with nonlocal goto receivers or returns twice calls, ignore
9968 unininitialized values from abnormal edges to nl goto receiver
9969 or returns twice call.
9971 2013-04-26 Jakub Jelinek <jakub@redhat.com>
9973 PR tree-optimization/57051
9974 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
9975 and VEC_RSHIFT_EXPR if shift count is a multiple of element
9978 2013-04-26 Richard Biener <rguenther@suse.de>
9980 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
9981 (expand_omp_taskreg): Likewise. Mark loops for fixup.
9982 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
9983 (fixup_loop_arrays_after_move): New function.
9984 (move_sese_region_to_fn): Properly outline the loop tree parts
9987 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
9989 * config/i386/i386.md (type, unit): Fix long lines.
9991 2013-04-26 Richard Biener <rguenther@suse.de>
9993 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
9994 (lto-streamer-out.o): Likewise.
9995 * cfgloop.c (init_loops_structure): Export, add struct function
9996 argument and adjust.
9997 (flow_loops_find): Adjust.
9998 * cfgloop.h (enum loop_estimation): Add EST_LAST.
9999 (init_loops_structure): Declare.
10000 * lto-streamer-in.c: Include cfgloop.h.
10001 (input_cfg): Input the loop tree.
10002 * lto-streamer-out.c: Include cfgloop.h.
10003 (output_cfg): Output the loop tree.
10004 (output_struct_function_base): Do not drop PROP_loops.
10006 2013-03-26 Richard Biener <rguenther@suse.de>
10008 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10009 (pass_build_cfg): Provide PROP_loops.
10010 (move_sese_region_to_fn): Remove loops that are outlined into fn
10012 * tree-inline.c: Include cfgloop.h.
10013 (initialize_cfun): Do not drop PROP_loops.
10014 (copy_loops): New function.
10015 (copy_cfg_body): Copy loop structure.
10016 (tree_function_versioning): Initialize destination loop tree.
10017 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10018 (pass_parallelize_loops): Do IL verification.
10019 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10020 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10021 the CFG make sure we fixup loops as well.
10022 * tree-ssa-tail-merge.c: Include cfgloop.h.
10023 (replace_block_by): When merging loop latches mark loops for fixup.
10024 * lto-streamer-out.c (output_struct_function_base): Drop
10025 PROP_loops for now.
10026 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10027 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10028 * ipa-split.c: Include cfgloop.h.
10029 (split_function): Add the new return block to the loop tree root.
10030 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10031 whether we have removed the forwarder block.
10032 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10033 * cfgloop.h (place_new_loop): Declare.
10034 * cfgloopmanip.c (place_new_loop): Export.
10035 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10036 (tree-switch-conversion.o): Likewise.
10037 (tree-complex.o): Likewise.
10038 (tree-inline.o): Likewise.
10039 (tree-ssa-tailmerge.o): Likewise.
10040 (ipa-split.o): Likewise.
10041 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10042 (tree-ssa-copy.o): Likewise.
10043 * tree-switch-conversion.c: Include cfgloop.h
10044 (process_switch): If we emit a bit-test cascade, schedule loops
10046 * tree-complex.c: Include cfgloop.h.
10047 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10048 * asan.c: Include cfgloop.h.
10049 (create_cond_insert_point): Properly add new basic-blocks to
10050 loops, schedule loop fixup.
10051 * cfgloop.c (verify_loop_structure): Check that looks are not
10053 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10055 (expand_omp_for_generic): Likewise.
10056 (expand_omp_sections): Likewise.
10057 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10058 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10059 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10060 is initialized, not when loops are present.
10061 * tree-parloops.c (parallelize_loops): Remove checking here.
10062 * passes.c (init_optimization_passes): Schedule a copy-propagation
10063 pass before complete unrolling of inner loops.
10065 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10067 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10068 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10069 (colorize_init): Add argument to _WIN32 version.
10070 * toplev.c: Include diagnostic-color.h.
10071 (process_options): Default to -fdiagnostics-color=auto if
10072 GCC_COLORS env var is in the environment.
10073 * common.opt (fdiagnostics-color=): Add Var and Init.
10074 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10075 env var is in the environment, the default is auto rather than never.
10077 * diagnostic.h (file_name_as_prefix): Add context argument.
10078 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10079 the string as locus.
10080 * langhooks.c (lhd_print_error_function): Adjust caller.
10082 2013-04-25 Lawrence Crowl <crowl@google.com>
10084 * var-tracking.c (shared_hash_def::htab):
10085 Change type to hash_table. Update dependent calls and types.
10087 2013-04-25 Lawrence Crowl <crowl@google.com>
10089 * Makefile.in: Update as needed below.
10091 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10092 Move declaration to after the type's method definitons.
10094 * attribs.c (htab_t scoped_attributes::attribute_hash):
10095 Change type to hash_table. Update dependent calls and types.
10097 * bitmap.c (htab_t bitmap_desc_hash):
10098 Change type to hash_table. Update dependent calls and types.
10100 * cselib.c (htab_t cselib_hash_table):
10101 Change type to hash_table. Update dependent calls and types.
10103 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10104 (hash_string_slot_node): Move implementation into lto-streamer.h
10105 struct string_slot_hasher.
10106 (eq_string_slot_node): Likewise.
10108 * data-streamer-out.c: Update output_block::string_hash_table
10109 dependent calls and types.
10111 * dwarf2cfi.c (htab_t trace_index):
10112 Change type to hash_table. Update dependent calls and types.
10114 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10115 Change type to hash_table. Update dependent calls and types.
10116 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10117 (htab_t optimize_external_refs::map): Likewise.
10118 (htab_t output_comp_unit::extern_map): Likewise.
10119 (htab_t output_comdat_type_unit::extern_map): Likewise.
10120 (htab_t output_macinfo::macinfo_htab): Likewise.
10121 (htab_t optimize_location_lists::htab): Likewise.
10122 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10124 * except.c (htab_t ehspec_hash_type):
10125 Change type to hash_table. Update dependent calls and types.
10126 (assign_filter_values::ttypes): Likewise.
10127 (assign_filter_values::ehspec): Likewise.
10128 (sjlj_assign_call_site_values::ar_hash): Likewise.
10129 (convert_to_eh_region_ranges::ar_hash): Likewise.
10131 * gcse.c (htab_t pre_ldst_table):
10132 Change type to hash_table. Update dependent calls and types.
10134 * ggc-common.c (htab_t saving_htab):
10135 Change type to hash_table. Update dependent calls and types.
10136 (htab_t loc_hash): Likewise.
10137 (htab_t ptr_hash): Likewise.
10138 (call_count): Rename ggc_call_count.
10139 (call_alloc): Rename ggc_call_alloc.
10140 (loc_descriptor): Rename make_loc_descriptor.
10141 (add_statistics): Rename ggc_add_statistics.
10143 * ggc-common.c (saving_htab):
10144 Change type to hash_table. Update dependent calls and types.
10146 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10147 (push_gimplify_context): Likewise.
10148 (pop_gimplify_context): Likewise.
10149 (struct gimple_temp_hash_elt): Added.
10150 (struct gimplify_hasher): Likewise.
10151 (struct gimplify_ctx.temp_htab):
10152 Change type to hash_table. Update dependent calls and types.
10154 * gimple-fold.c: Include gimplify-ctx.h.
10156 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10157 Change type to hash_table. Update dependent calls and types.
10158 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10159 avoid potential global name collision.
10161 * gimplify.c: Include gimplify-ctx.h.
10162 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10163 (htab_t gimplify_ctx::temp_htab):
10164 Update dependent calls and types for new type hash_table.
10165 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10166 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10168 * gimplify-ctx.h: New.
10169 (struct gimple_temp_hash_elt): Move from gimplify.c.
10170 (class gimplify_hasher): New.
10171 (struct gimplify_ctx): Move from gimple.h.
10172 (htab_t gimplify_ctx::temp_htab):
10173 Change type to hash_table. Update dependent calls and types.
10175 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10176 (htab_t ivs_params::newivs_index):
10177 Change type to hash_table. Update dependent calls and types.
10178 (htab_t ivs_params::params_index): Likewise.
10179 (htab_t print_generated_program::params_index): Likewise.
10180 (htab_t gloog::newivs_index): Likewise.
10181 (htab_t gloog::params_index): Likewise.
10183 * graphite.c: Include graphite-htab.h.
10184 4htab_t graphite_transform_loops::bb_pbb_mapping):
10185 Change type to hash_table. Update dependent calls and types.
10187 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10188 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10189 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10191 * graphite-dependences.c: Include graphite-htab.h.
10192 (loop_is_parallel_p): Change hash table type of parameter.
10194 * graphite-htab.h: New.
10195 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10196 (extern find_pbb_via_hash): Move from graphite-poly.h.
10197 (extern loop_is_parallel_p): Move from graphite-poly.h.
10198 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10200 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10201 (extern loop_is_parallel_p): Move to graphite-htab.h.
10202 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10204 * haifa-sched.c (htab_t delay_htab):
10205 Change type to hash_table. Update dependent calls and types.
10206 (htab_t delay_htab_i2): Likewise.
10208 * ira-color.c (htab_t allocno_hard_regs_htab):
10209 Change type to hash_table. Update dependent calls and types.
10211 * ira-costs.c (htab_t cost_classes_htab):
10212 Change type to hash_table. Update dependent calls and types.
10214 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10215 Change type to hash_table. Update dependent calls and types.
10217 * loop-iv.c (htab_t bivs):
10218 Change type to hash_table. Update dependent calls and types.
10220 * loop-unroll.c (htab_t opt_info::insns_to_split):
10221 Change type to hash_table. Update dependent calls and types.
10222 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10224 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10225 (struct string_slot_hasher): New.
10226 (htab_t output_block::string_hash_table):
10227 Change type to hash_table. Update dependent calls and types.
10229 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10230 (htab_t file_name_hash_table):
10231 Change type to hash_table. Update dependent calls and types.
10233 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10236 * lto-streamer.c (htab_t tree_htab):
10237 Change type to hash_table. Update dependent calls and types.
10239 * omp-low.c: Include gimplify-ctx.h.
10241 * passes.c (htab_t name_to_pass_map):
10242 Change type to hash_table. Update dependent calls and types.
10243 (pass_traverse): Rename to passes_pass_traverse.
10245 * plugin.c (htab_t event_tab):
10246 Change type to hash_table. Update dependent calls and types.
10248 * postreload-gcse.c (htab_t expr_table):
10249 Change type to hash_table. Update dependent calls and types.
10250 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10252 * sese.c (debug_rename_map_1): Make extern.
10253 (htab_t copy_bb_and_scalar_dependences::rename_map):
10254 Change type to hash_table. Update dependent calls and types.
10256 * sese.h (extern debug_rename_map): Move to .c file.
10258 * store-motion.c (htab_t store_motion_mems_table):
10259 Change type to hash_table. Update dependent calls and types.
10261 * trans-mem.c (htab_t tm_new_mem_hash):
10262 Change type to hash_table. Update dependent calls and types.
10264 * tree-browser.c (htab_t TB_up_ht):
10265 Change type to hash_table. Update dependent calls and types.
10267 * tree-cfg.c (htab_t discriminator_per_locus):
10268 Change type to hash_table. Update dependent calls and types.
10270 * tree-complex.c: Include tree-hasher.h
10271 (htab_t complex_variable_components):
10272 Change type to hash_table. Update dependent calls and types.
10274 * tree-eh.c (htab_t finally_tree):
10275 Change type to hash_table. Update dependent calls and types.
10277 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10278 struct int_tree_hasher.
10279 (extern int_tree_map_eq): Likewise.
10280 (uid_decl_map_hash): Removed.
10281 (extern decl_tree_map_eq): Likewise.
10283 * tree-hasher.h: New.
10284 (struct int_tree_hasher): New.
10285 (typedef int_tree_htab_type): New.
10287 * tree-inline.c: Include gimplify-ctx.h.
10289 * tree-mudflap.c: Include gimplify-ctx.h.
10291 * tree-parloops.c: Include tree-hasher.h.
10292 (htab_t eliminate_local_variables_stmt::decl_address):
10293 Change type to hash_table. Update dependent calls and types.
10294 (htab_t separate_decls_in_region::decl_copies): Likewise.
10296 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10297 Change type to hash_table. Update dependent calls and types.
10299 * tree-sra.c (candidates):
10300 Change type to hash_table. Update dependent calls and types.
10302 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10304 (int_tree_map_hash): Likewise.
10306 * tree-ssa-dom.c (htab_t avail_exprs):
10307 Change type to hash_table. Update dependent calls and types.
10309 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10310 Change type to hash_table. Update dependent calls and types.
10312 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10313 Change type to hash_table. Update dependent calls and types.
10315 * tree-ssa-phiopt.c (seen_ssa_names):
10316 Change type to hash_table. Update dependent calls and types.
10318 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10319 Change type to hash_table. Update dependent calls and types.
10321 * tree-ssa-uncprop.c (equiv):
10322 Change type to hash_table. Update dependent calls and types.
10324 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10326 PR rtl-optimization/57003
10327 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10328 call note_stores with kill_clobbered_value callback again after
10329 killing regs_invalidated_by_call.
10331 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10333 * config/aarch64/aarch64-simd.md
10334 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10335 (aarch64_simd_bsl<mode>): Likewise.
10336 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10338 2013-04-25 Marek Polacek <polacek@redhat.com>
10340 PR tree-optimization/57066
10341 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10343 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10345 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10347 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10349 * config/aarch64/aarch64-builtins.c
10350 (aarch64_fold_builtin): New.
10351 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10352 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10353 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10354 * config/aarch64/arm_neon.h
10355 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10357 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10358 Tejas Belagod <tejas.belagod@arm.com>
10360 * config/aarch64/aarch64-builtins.c
10361 (aarch64_gimple_fold_builtin): New.
10362 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10363 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10364 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10365 (addvv4sf): Update.
10366 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10368 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10370 * config/aarch64/aarch64.md
10371 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10373 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10375 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10376 (*ngcsi_uxtw): New pattern.
10378 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10379 Julian Brown <julian@codesourcery.com>
10381 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
10382 (TB_DREG): Add T_V4HF.
10383 (v4hf_UP): New macro.
10384 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10385 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10386 Handle initialisation of V4HF. Adjust initialisation of reinterpret
10388 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10389 (arm_vector_mode_supported_p): Handle V4HF.
10390 (arm_mangle_map): Handle V4HFmode.
10391 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
10392 * config/arm/arm_neon_builtins.def: Add entries for
10393 vcvtv4hfv4sf, vcvtv4sfv4hf.
10394 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
10395 (neon_vcvtv4hfv4sf): Likewise.
10396 * config/arm/neon-gen.ml: Handle half-precision floating point
10398 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
10399 * config/arm/arm_neon.h: Regenerate.
10400 * config/arm/neon.ml (type elts): Add F16.
10401 (type vectype): Add T_float16x4, T_floatHF.
10402 (type vecmode): Add V4HF.
10403 (type features): Add Requires_FP_bit feature.
10404 (elt_width): Handle F16.
10405 (elt_class): Likewise.
10406 (elt_of_class_width): Likewise.
10407 (mode_of_elt): Refactor.
10408 (type_for_elt): Handle F16, fix error messages.
10409 (vectype_size): Handle T_float16x4.
10410 (vcvt_sh): New function.
10411 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
10412 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
10413 (string_of_mode): Handle V4HF.
10414 * doc/arm-neon-intrinsics.texi: Regenerate.
10416 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10418 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
10419 format specifier in 'X' case.
10421 2013-04-25 Alan Modra <amodra@gmail.com>
10424 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
10425 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
10426 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
10427 Repeat for many other rotate/shift and mask patterns using subregs.
10428 Name lshiftrt insns.
10429 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
10430 on WORDS_BIG_ENDIAN.
10432 2013-04-25 Alan Modra <amodra@gmail.com>
10434 * config.gcc: Support little-endian powerpc-linux targets.
10435 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
10436 (LINK_OS_LINUX_SPEC): Define.
10437 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
10438 Preserve MASK_LITTLE_ENDIAN.
10439 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
10440 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
10441 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
10442 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
10443 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
10444 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
10445 Correct fp word order for little-endian. Don't shift toc entries
10446 smaller than a word for little-endian.
10447 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
10448 (bswapdi2 splits): Correct low-part subreg for little-endian.
10449 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
10450 low/high where such is correct only for be.
10451 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
10452 little-endian for -mcall-aixdesc.
10454 2013-04-25 Alan Modra <amodra@gmail.com>
10456 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10457 replace_equiv_address_nv.
10459 2013-04-25 Alan Modra <amodra@gmail.com>
10461 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
10463 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10466 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10467 * rtl.h (struct rtx_def): ...
10469 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10471 PR rtl-optimizations/57046
10472 * lra-constraints (split_reg): Set up lra_risky_transformations_p
10473 for multi-reg splits.
10475 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
10477 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
10479 2013-04-24 Sterling Augustine <saugustine@google.com>
10481 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
10482 (comp_dir_string, debug_str_dwo_section): New.
10483 (DEBUG_STR_DWO_SECTION): Rename to ...
10484 (DEBUG_DWO_STR_SECTION): ... this.
10485 (DEBUG_NORM_STR_SECTION): Delete.
10486 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
10487 (DEBUG_STR_DWO_SECTION_FLAGS): New.
10488 (find_AT_string): Move most logic to ...
10489 (find_AT_string_in_table): ... here. New.
10490 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
10491 add_skeleton_AT_string. Delete logic.
10492 (output_skeleton_debug_sections): Remove call to
10493 add_top_level_skeleton_die_attrs.
10494 (add_comp_dir_attribute): Move logic to comp_dir_string.
10495 (dwarf2out_init): Initialize debug_str_dwo_section.
10496 (output_indirect_string): Call find_string_form.
10497 (output_indirect_strings): Rewrite.
10498 (prune_unused_types): Empty skeleton_debug_str_hash.
10499 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
10500 (dwarf2out_finish): Call output_indirect_strings.
10502 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10504 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
10506 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
10508 * rtl.h (struct rtx_def): Add comment for field jump.
10509 (LRA_SUBREG_P): New macro.
10510 * recog.c (register_operand): Check LRA_SUBREG_P.
10511 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10513 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10514 (lra_final_code_change): Skip subreg change for operators.
10515 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10516 if there are no operand changes.
10517 * lra-constraints.c (curr_insn_set): New.
10518 (match_reload): Set LRA_SUBREG_P.
10519 (emit_spill_move): Ditto.
10520 (check_and_process_move): Use curr_insn_set. Process only single
10521 set insns. Don't initialize sec_mem_p and change_p.
10522 (simplify_operand_subreg): Use LRA_SUBREG_P.
10523 (reg_in_class_p): New function.
10524 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10525 of #ifdef. Add code to remove cycling.
10526 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10527 non-null disp. Reload inner instead of disp when base and index
10528 are null. Try to put lo_sum into register.
10529 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10530 (check_and_process_move): Move code for move cost check to
10531 simple_move_p. Remove equiv_substitution.
10532 (simple_move_p): New function.
10533 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10534 curr_insn_set. Call check_and_process_move only for single set
10535 insns. Use the new function. Move call of check_and_process_move
10536 after operand equiv substitution and address process.
10538 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
10540 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
10541 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
10542 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
10544 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
10546 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
10548 2013-04-24 Marek Polacek <polacek@redhat.com>
10550 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
10551 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
10552 (select_loops_exit_conditions): Likewise.
10553 (number_of_iterations_for_all_loops): Likewise.
10554 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
10555 (scev_analysis): Likewise.
10557 2013-04-02 Catherine Moore <clm@codesourcery.com>
10558 Chao-ying Fu <fu@mips.com>
10560 * config/mips/micromips.md (jraddiusp): New pattern.
10561 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
10562 instruction if possible.
10564 2013-04-24 Alan Modra <amodra@gmail.com>
10566 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
10568 2013-04-24 Julian Brown <julian@codesourcery.com>
10569 Chung-Lin Tang <cltang@codesourcery.com>
10571 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
10572 dependency behavior in enumeration type DIE generation. Add TODO note
10573 to comments about future DW_FORM_sdata/udata re-work of related code.
10575 2013-04-23 Lawrence Crowl <crowl@google.com>
10577 * Makefile.in: Update as needed below.
10579 * hash-table.h (class hash_table):
10580 Correct many methods with parameter types compare_type to the correct
10581 value_type. (Correct code was unlikely to notice the change.)
10582 (hash_table::elements_with_deleted) New.
10583 (class hashtable::iterator): New.
10584 (hashtable::begin()): New.
10585 (hashtable::end()): New.
10586 (FOR_EACH_HASH_TABLE_ELEMENT): New.
10588 * statistics.c (statistics_hashes):
10589 Change type to hash_table. Update dependent calls and types.
10591 * tree-into-ssa.c (var_infos):
10592 Change type to hash_table. Update dependent calls and types.
10594 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
10595 Change type to hash_table. Update dependent calls and types.
10597 * tree-ssa-loop-im.c (struct mem_ref.refs):
10598 Change type to hash_table. Update dependent calls and types.
10600 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
10601 Change type to hash_table. Update dependent calls and types.
10603 * tree-ssa-sccvn.c (vn_tables_s::nary):
10604 Change type to hash_table. Update dependent calls and types.
10605 (vn_tables_s::phis): Likewise.
10606 (vn_tables_s::references): Likewise.
10608 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
10609 (vn_reference_eq): Update parameter and return types.
10611 * tree-ssa-structalias.c (pointer_equiv_class_table):
10612 Change type to hash_table. Update dependent calls and types.
10613 (location_equiv_class_table): Likewise.
10615 * tree-vect-data-refs.c: Consequential changes for making
10616 peeling a hash_table.
10618 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
10619 (destroy_loop_vec_info): Dependent hash_table update.
10621 * tree-vectorizer.h (peeling_htab):
10622 Change type to hash_table. Update dependent calls and types.
10624 2013-04-23 Shiva Chen <shiva0217@gmail.com>
10626 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
10627 to check the register content is equal or not.
10628 * lra-constraints.c (match_reload): Use lra_assign_reg_val
10629 to assign register content record.
10630 * lra-eliminations.c (update_reg_eliminate): Use
10631 lra_update_reg_val_offset to update register content offset.
10632 * lra-int.h (struct lra_reg): Add offset member.
10633 (lra_reg_val_equal_p): New static inline function.
10634 (lra_update_reg_val_offset): New static inline function.
10635 (lra_assign_reg_val): New static inline function.
10636 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
10637 to assign register content record.
10638 (initialize_lra_reg_info_element): Initial offset to zero.
10640 2013-04-23 Catherine Moore <clm@codesourcery.com>
10642 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
10643 operands. Record compression.
10645 2013-04-23 Xinliang David Li <davidxl@google.com>
10647 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
10649 2013-04-23 Richard Biener <rguenther@suse.de>
10651 PR middle-end/57036
10652 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
10653 parameter, only add abnormal goto edges from the copied body
10654 if the call could perform abnormal gotos.
10655 (copy_cfg_body): Adjust.
10657 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
10659 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
10661 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
10663 * coretypes.h (gimple_stmt_iterator): Add struct to make
10666 2013-04-23 Richard Biener <rguenther@suse.de>
10668 PR tree-optimization/57026
10669 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
10670 from SSA names occuring in abnormal PHI nodes.
10672 2013-04-22 Andi Kleen <ak@linux.intel.com>
10674 * lto/lto.c (print_lto_report_1): Fix LTO report names.
10676 2013-04-22 Andi Kleen <ak@linux.intel.com>
10678 * lto/lto.c (print_lto_report_1): Declare early.
10679 (read_cgraph_and_symbols): Call print_lto_report_1 early.
10681 2013-04-22 Andi Kleen <ak@linux.intel.com>
10683 * common.opt (-flto-report-wpa): Add.
10684 * doc/invoke.texi (-flto-report-wpa): Add.
10685 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
10688 2013-04-22 Xinliang David Li <davidxl@google.com>
10690 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
10691 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
10692 * Makefile.in: New dependency
10694 David Daney <ddaney.cavm@gmail.com>
10696 * configure.ac (gcc_cv_as_micromips_support): Use the
10697 --fatal-warnings option.
10698 * configure: Regenerate.
10700 2013-04-22 Marek Polacek <polacek@redhat.com>
10703 * tsan.c (instrument_expr): Don't instrument expression
10704 in case its size is zero.
10706 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
10710 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
10712 * config/alpha/alpha.c (TARGET_LRA_P): New define.
10714 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10716 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
10717 (gimple_stmt_iterator): New typedef.
10718 * gimple.h (gimple_stmt_iterator): Rename to...
10719 (gimple_stmt_iterator_d): ... This.
10720 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
10721 trees be valid for GIMPLE and GENERIC.
10722 (TARGET_GIMPLE_FOLD_BUILTIN): New.
10723 * gimple-fold.c (gimple_fold_call): Call target hook
10724 gimple_fold_builtin.
10725 * hooks.c (hook_bool_gsiptr_false): New.
10726 * hooks.h (hook_bool_gsiptr_false): New.
10727 * target.def (fold_stmt): New.
10728 * doc/tm.texi: Regenerate.
10730 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
10733 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
10734 a set sp if no stack realignment.
10736 2013-04-22 Nick Clifton <nickc@redhat.com>
10738 * config.gcc (tilegx-linux): Extend extra_objs rather than
10740 (tilepro-linux): Likewise.
10742 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10744 * config/aarch64/aarch64-builtins.c
10746 (CF0, CF1, CF2, CF3, CF4, CF10): New.
10747 (VAR<1-12>): Add MAP parameter.
10748 (BUILTIN_*): Likewise.
10749 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
10750 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
10751 (aarch64_ushl_n<mode>): Likewise.
10752 (aarch64_sshr_n<mode>): Likewise.
10753 (aarch64_ushr_n<mode>): Likewise.
10754 (aarch64_<maxmin><mode>): Likewise.
10755 (aarch64_sqrt<mode>): Likewise.
10756 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
10757 (vshr<q>_n_*): Likewise.
10759 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
10761 * config/aarch64/aarch64-builtins.c
10762 (aarch64_simd_builtin_type_mode): Handle SF types.
10764 (BUILTIN_GPF): Define.
10765 (aarch64_init_simd_builtins): Handle SF types.
10766 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
10767 (frecps): Likewise.
10768 (frecpx): Likewise.
10769 * config/aarch64/aarch64-simd.md
10770 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
10771 (aarch64_frecpe<mode>): New.
10772 (aarch64_frecps<mode>): Likewise.
10773 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
10774 (v8type): Add frecp<esx>.
10775 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
10776 (aarch64_frecps<mode>): Likewise.
10777 * config/aarch64/iterators.md (FRECP): New.
10778 (frecp_suffix): Likewise.
10779 * config/aarch64/arm_neon.h
10780 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
10782 2013-04-22 Christian Bruel <christian.bruel@st.com>
10785 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
10786 (REG_CLASS_NAMES): Idem.
10787 (REG_CLASS_CONTENTS): Idem.
10788 (REGCLASS_HAS_FP_REG): Idem.
10789 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
10790 (sh_conditional_register_usage): Idem.
10792 2013-04-21 Jeff Law <law@redhat.com>
10794 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
10795 (ssa_forward_propagate_and_combine): Use it.
10797 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10799 * lra.c: Update the flow chart diagram.
10801 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
10803 PR rtl-optimization/56847
10804 * lra-constraints.c (process_alt_operands): Discourage alternative
10805 with non-matche doffsettable memory constraint fro memory with
10808 2013-04-19 Richard Biener <rguenther@suse.de>
10810 PR tree-optimization/56982
10811 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
10813 * gimplify.c (gimplify_call_expr): Notice special calls.
10814 (gimplify_modify_expr): Likewise.
10815 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
10816 abnormal control flow receivers.
10817 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
10818 in the same way as cfun->has_nonlocal_labels.
10819 (gimple_purge_dead_abnormal_call_edges): Likewise.
10820 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
10821 receivers start a basic-block.
10823 2013-04-19 Richard Biener <rguenther@suse.de>
10825 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
10827 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
10828 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
10829 (SLP_TREE_LOAD_PERMUTATION): Add.
10830 (vect_transform_slp_perm_load): Adjust prototype.
10831 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
10832 (vect_free_slp_instance): Likewise.
10833 (vect_create_new_slp_node): Likewise.
10834 (vect_supported_slp_permutation_p): Remove.
10835 (vect_slp_rearrange_stmts): Adjust.
10836 (vect_supported_load_permutation_p): Likewise. Inline
10837 vect_supported_slp_permutation_p here.
10838 (vect_analyze_slp_instance): Compute load permutations per
10839 slp node instead of per instance.
10840 (vect_get_slp_defs): Adjust.
10841 (vect_transform_slp_perm_load): Likewise.
10842 (vect_schedule_slp_instance): Remove redundant code.
10843 (vect_schedule_slp): Remove hack for PR56270, add it ...
10844 * tree-vect-stmts.c (vectorizable_load): ... here, do not
10845 CSE loads for SLP. Adjust.
10847 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10849 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
10850 spelling in two comments.
10852 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
10855 * config/arm/arm.c (load_multiple_sequence): Require SP
10856 as base register for loads if SP is in the register list.
10858 2013-04-19 Martin Jambor <mjambor@suse.cz>
10860 PR tree-optimization/56718
10861 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
10862 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
10863 and made public. Adjusted all callers.
10864 (ipa_intraprocedural_devirtualization): New function.
10865 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
10866 (ipa_intraprocedural_devirtualization): Likewise.
10867 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
10869 2013-04-19 Richard Biener <rguenther@suse.de>
10871 PR tree-optimization/57000
10872 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
10874 2013-04-19 Terry Guo <terry.guo@arm.com>
10876 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
10878 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
10879 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
10880 (cortex_m4_fmacs): Use new reservations.
10881 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
10883 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
10885 PR rtl-optimization/56999
10886 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
10888 (lra_coalesce): Remove split_origin_bitmap and related code.
10889 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
10890 ranges if necessary.
10892 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
10894 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
10896 (ix86_expand_call): Remove clobbered_registers array and use
10897 x86_64_ms_sysv_extra_clobbered_registers instead.
10898 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
10900 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
10902 * config/i386/i386.md (*call_rex64_ms_sysv): Use
10903 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
10904 (*call_value_rex64_ms_sysv): Ditto.
10906 2013-04-18 Cary Coutant <ccoutant@google.com>
10908 * dwarf2out.c (output_pubnames): Check die_perennial_p of
10909 parent instead of die_mark.
10911 2013-04-18 Diego Novillo <dnovillo@google.com>
10913 * gimple.c (create_gimple_tmp): New.
10914 (get_expr_type): New.
10915 (build_assign): New.
10916 (build_type_cast): New.
10917 * gimple.h (enum ssa_mode): Define.
10918 (gimple_seq_set_location): New.
10919 * asan.c (build_check_stmt): Change some gimple_build_* calls
10920 to use build_assign and build_type_cast.
10922 2013-04-18 Richard Biener <rguenther@suse.de>
10924 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
10925 handle negative step. Remove redundant checks.
10926 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
10927 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
10928 for negative step and grouped loads fail to vectorize.
10930 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10932 * emit-rtl.c (reset_insn_used_flags): New function.
10933 (reset_all_used_flags): Use it.
10934 (verify_insn_sharing): New function.
10935 (verify_rtl_sharing): Fix verification for SEQUENCEs.
10937 2013-04-18 Jakub Jelinek <jakub@redhat.com>
10939 PR tree-optimization/56984
10940 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
10941 and (x >> M) >= N don't register any assertion if N << M is the
10944 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
10946 * lower-subreg.c (resolve_simple_move): If called self-recursive,
10947 do not delete_insn insns that have not yet been emitted, only
10948 unlink them with remove_insn.
10949 * df-scan.c (df_insn_delete): Revert r197492.
10951 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
10953 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
10954 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
10956 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10958 * config/arm/arm.md (movsicc_insn): Convert define_insn into
10959 define_insn_and_split.
10960 (and_scc,ior_scc,negscc): Likewise.
10961 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
10963 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
10965 * config/arm/arm.c (use_return_insn): Return 0 for targets that
10966 can benefit from using a sequence of LDRD instructions in epilogue
10967 instead of a single LDM instruction.
10969 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
10972 * doc/extend.texi: Fix typo.
10974 2013-04-17 Richard Biener <rguenther@suse.de>
10976 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
10977 (vect_build_slp_tree): ... here.
10978 (vect_build_slp_tree_1): Compute which stmts of the SLP group
10979 match. Remove special-casing of mismatched complex loads.
10980 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
10981 re-try the match with swapped commutative operands.
10982 (vect_supported_load_permutation_p): Remove special-casing of
10983 mismatched complex loads.
10984 (vect_analyze_slp_instance): Adjust.
10986 2013-04-17 Richard Biener <rguenther@suse.de>
10988 PR rtl-optimization/56921
10989 * cfgloop.h (struct loop): Add simple_loop_desc member.
10990 (struct niter_desc): Mark with GTY(()).
10991 (simple_loop_desc): Do not use aux field but simple_loop_desc.
10992 * loop-iv.c (get_simple_loop_desc): Likewise.
10993 (free_simple_loop_desc): Likewise.
10996 2013-04-16 Richard Biener <rguenther@suse.de>
10998 PR rtl-optimization/56921
10999 * loop-init.c (pass_rtl_move_loop_invariants): Add
11000 TODO_do_not_ggc_collect to todo_flags_finish.
11001 (pass_rtl_unswitch): Same.
11002 (pass_rtl_unroll_and_peel_loops): Same.
11003 (pass_rtl_doloop): Same.
11005 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11007 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11008 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11009 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11011 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11012 * tree-streamer.c (record_common_node): Adjust reference in comment.
11014 2013-04-17 Terry Guo <terry.guo@arm.com>
11016 * config/arm/cortex-m4.md: Add a new bypass.
11018 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11020 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11022 (*subs_<optab><mode>_multp2): New pattern.
11023 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11024 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11026 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11028 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11029 (*subs_mul_imm_<mode>): New pattern.
11031 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11034 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11035 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11036 (vsx_movti_32bit): Change j->wa to O->wa.
11038 2013-04-16 Richard Biener <rguenther@suse.de>
11040 PR rtl-optimization/56921
11041 * loop-init.c (pass_rtl_move_loop_invariants): Add
11042 TODO_do_not_ggc_collect to todo_flags_finish.
11043 (pass_rtl_unswitch): Same.
11044 (pass_rtl_unroll_and_peel_loops): Same.
11045 (pass_rtl_doloop): Same.
11047 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11049 * config/arm/arm.c (emit_multi_reg_push): New declaration
11050 for an existing function.
11051 (arm_emit_strd_push): New function.
11052 (arm_expand_prologue): Used here.
11053 (arm_emit_ldrd_pop): New function.
11054 (arm_expand_epilogue): Used here.
11055 (arm_get_frame_offsets): Update condition.
11056 (arm_emit_multi_reg_pop): Add a special case for load of a single
11057 register with writeback.
11059 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11061 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11064 2013-04-16 Richard Biener <rguenther@suse.de>
11066 PR tree-optimization/56756
11067 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11068 (first_mem_ref_loc): New.
11069 (execute_sm): Place the load temporarily before a previous
11070 access instead of in the latch edge to ensure its SSA dependencies
11071 are defined at points dominating the load.
11073 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11075 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11076 correct fix by moving header and footer insn to the footer of
11077 the merged basic block. Clear BB_END of the merged-away block.
11079 PR middle-end/43631
11080 * emit-rtl.c (make_note_raw): New function.
11081 (link_insn_into_chain): New static inline function.
11082 (add_insn): Use it.
11083 (add_insn_before, add_insn_after): Factor insn chain linking code...
11084 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11085 using link_insn_into_chain.
11086 (note_outside_basic_block_p): New helper function for emit_note_after
11087 and emit_note_before.
11088 (emit_note_after): Use nobb variant of add_insn_after if the note
11089 should not be contained in a basic block.
11090 (emit_note_before): Use nobb variant of add_insn_before if the note
11091 should not be contained in a basic block.
11092 (emit_note_copy): Use make_note_raw.
11093 (emit_note): Likewise.
11094 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11095 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11096 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11097 the moved barrier the tail of the basic block it follows.
11098 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11100 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11102 PR tree-optimization/56962
11103 * gimple-ssa-strength-reduction.c (record_increment): Only set
11104 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11105 either rhs1 or rhs2 is equal to c->base_expr.
11107 2013-04-15 Richard Biener <rguenther@suse.de>
11109 PR tree-optimization/56933
11110 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11112 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11113 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11114 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11115 dependence check ...
11116 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11118 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11119 GROUP_READ_WRITE_DEPENDENCE.
11121 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11123 * emit-rtl.c (reset_all_used_flags): New function.
11124 (verify_rtl_sharing): Call reset_all_used_flags before and after
11125 performing the checks.
11127 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11129 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11130 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11131 * config/arm/constraints.md (De): New constraint.
11132 * config/arm/neon.md (anddi3_neon): Delete.
11133 (neon_vand<mode>): Expand to standard anddi3 pattern.
11134 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11135 Move earlier in the file.
11136 (neon_inv_logic_op2): Likewise.
11137 (arm_anddi_operand_neon): New predicate.
11139 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11141 * configure.ac (gcc_cv_ld_as_needed): Set
11142 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11143 Use -z ignore, -z record on *-*-solaris2*.
11144 (HAVE_LD_AS_NEEDED): Update comment.
11145 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11146 * configure: Regenerate.
11147 * config.in: Regenerate.
11148 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11149 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11150 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11151 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11152 equivalents. Fix markup.
11153 * doc/tm.texi: Regenerate.
11155 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11157 * config/i386/i386.opt: New option mstack-protector-guard=.
11158 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11159 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11160 TARGET_SSP_TLS_GUARD.
11161 * config/i386/i386.c (ix86_option_override_internal): Set
11162 ix86_stack_protector_guard.
11163 * config/i386/i386.md (stack_protect_set): Enable for
11164 TARGET_SSP_TLS_GUARD only.
11165 (stack_protect_set_<mode>): Ditto.
11166 (stack_protect_test): Ditto.
11167 (stack_protect_test_<mode>): Ditto.
11168 * doc/invoke.texi (i386 Option): Document.
11170 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11173 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11174 (S_MODES): Set H_MODE bit.
11175 (SF_MODES): Set only S_MODE and SF_MODE bits.
11176 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11177 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11178 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11179 <MODE_FLOAT>: Likewise.
11181 2013-04-15 Joey Ye <joey.ye@arm.com>
11183 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11185 2013-04-15 Joey Ye <joey.ye@arm.com>
11187 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11189 (thumb_far_jump_used_p): Count instruction size and set
11192 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11194 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11195 * resource.c (mark_target_live_regs): Reformat conditional block.
11197 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11199 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11200 notes, they are emitted only just before final.
11201 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11203 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11205 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11206 * cfgrtl.c (delete_insn): Call it here instead.
11207 * lra-spills.c (lra_final_code_change): Use delete_insn.
11208 * haifa-sched.c (sched_remove_insn): Likewise.
11209 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11210 returning to the nop pool.
11211 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11212 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11214 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11216 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11217 * doc/tm.texi: Regenerated.
11219 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11221 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11224 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11226 * df-core.c (df_find_def): Compare register numbers.
11227 (df_find_use): Likewise.
11229 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11232 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11233 lra_in_progress for return.
11235 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11237 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11238 define_insn into define_insn_and_split and emit movsicc patterns.
11240 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11242 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11244 2013-04-12 Richard Biener <rguenther@suse.de>
11246 * tree-pass.h (TODO_do_not_ggc_collect): New.
11247 * passes.c (execute_one_ipa_transform_pass): Honor
11248 TODO_do_not_ggc_collect.
11249 (execute_one_pass): Likewise.
11252 2013-04-10 Richard Biener <rguenther@suse.de>
11254 * passes.c (init_optimization_passes): Remove reload pass.
11255 * ira.c (do_reload): Merge into ...
11257 (rest_of_handle_reload): Remove.
11258 (pass_reload): Likewise.
11259 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11260 of reload for vzeroupper pass placement.
11262 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11264 PR tree-optimization/56918
11265 PR tree-optimization/56920
11266 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11267 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11268 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11269 use wide_mul_with_sign method.
11271 2013-04-12 Richard Biener <rguenther@suse.de>
11273 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11274 not be considered a gimple constant.
11276 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11278 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11279 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11282 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11283 Jakub Jelinek <jakub@redhat.com>
11285 * opts.c: Include diagnostic-color.h.
11286 (common_handle_option): Handle OPT_fdiagnostics_color_.
11287 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11288 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11289 (diagnostic-color.o): New.
11290 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11291 (diagnostic_color_rule): New enum.
11292 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11293 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11294 the location string.
11295 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11296 either NULL, or color kind.
11297 * diagnostic-color.c: New file.
11298 * diagnostic-color.h: New file.
11299 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11301 * doc/invoke.texi (-fdiagnostics-color): Document.
11302 * pretty-print.h (pp_show_color): Define.
11303 (struct pretty_print_info): Add show_color field.
11304 * diagnostic.c: Include diagnostic-color.h.
11305 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11306 macros. Colorize error:, warning: etc. strings and also the location
11308 (diagnostic_show_locus): Colorize the caret line.
11309 * pretty-print.c: Include diagnostic-color.h.
11310 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11311 inside of %< %> quotes or quoted through q format modifier.
11313 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11315 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11317 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11319 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11320 code in CC_NZ mode.
11321 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11324 2013-04-11 Marek Polacek <polacek@redhat.com>
11326 PR tree-optimization/48184
11327 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11329 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11331 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11332 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11333 (skip_simple_arithmetic): Tidy up.
11334 * tree.h (skip_simple_constant_arithmetic): Declare.
11336 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11338 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11340 2013-04-11 Richard Biener <rguenther@suse.de>
11342 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11343 generate vector constants.
11345 2013-04-11 Richard Biener <rguenther@suse.de>
11347 PR tree-optimization/56878
11348 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11349 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11351 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11352 Prefer to align the DR with the most invariant base address.
11354 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11356 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11359 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11361 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11362 floating-point vector comparisons against 0.
11364 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11366 PR tree-optimization/56899
11367 * fold-const.c (extract_muldiv_1): Apply distributive law
11368 only if TYPE_OVERFLOW_WRAPS (ctype).
11370 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11373 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11374 loading has side effect.
11376 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
11378 * config/sparc/sparc.c: Include tree-pass.h.
11379 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
11380 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
11381 head of file. Change return type. Split off gate function.
11382 (sparc_gate_work_around_errata): New function.
11383 (pass_work_around_errata): New pass definition.
11384 (insert_pass_work_around_errata) New pass insert definition to
11385 insert pass_work_around_errata just after delayed-branch scheduling.
11386 (sparc_option_override): Insert the pass.
11387 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
11389 2013-04-10 David S. Miller <davem@davemloft.net>
11391 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
11392 or -mcpu=hypersparc.
11394 * target.def (cstore_mode): New hook.
11395 * target.h: Include insn-codes.h
11396 * targhooks.c: Likewise.
11397 (default_cstore_mode): New function.
11398 * targhooks.h: Declare it.
11399 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
11400 * doc/tm.texi: Rebuild.
11401 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
11402 target hook, rather than inspecting the insn_data.
11403 * config/sparc/sparc.c (sparc_cstore_mode): New function.
11404 (TARGET_CSTORE_MODE): Redefine.
11405 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
11407 * config/sparc/predicates.md (cstore_result_operand): New special
11409 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
11410 Use it for operand 0.
11411 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
11412 (*snesi_special): Likewise.
11413 (*snesi_zero): Likewise.
11414 (*seqsi_zero): Likewise.
11415 (*sltu_insn): Likewise.
11416 (*sgeu_insn): Likewise.
11417 (*seqdi_special): Make operand 0 and comparison operation be of
11419 (*snedi_special): Likewise.
11420 (*snedi_special_vis3): Likewise.
11421 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
11422 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
11423 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
11424 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
11425 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
11426 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
11427 (*sltu_extend_sp64): Likewise.
11428 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
11429 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
11430 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
11431 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
11432 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
11434 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
11436 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
11437 (aarch64_start_file): Use the new function.
11439 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11440 Jason Merrill <jason@redhat.com>
11442 * common.opt: Add -gdwarf.
11443 * opts.c (common_handle_option): Handle it.
11444 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
11446 2013-04-10 Richard Biener <rguenther@suse.de>
11448 * passes.c (execute_todo): Do not call ggc_collect conditional here.
11449 (execute_one_ipa_transform_pass): But unconditionally here.
11450 (execute_one_pass): And here.
11451 (init_optimization_passes): Remove reload pass.
11452 * tree-pass.h (TODO_ggc_collect): Remove.
11453 (pass_reload): Likewise.
11454 * ira.c (do_reload): Merge into ...
11456 (rest_of_handle_reload): Remove.
11457 (pass_reload): Likewise.
11458 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11459 of reload for vzeroupper pass placement.
11460 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
11461 and todo_flags_finish of all passes.
11463 2013-04-10 Richard Biener <rguenther@suse.de>
11465 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
11466 first_const_oprnd field, rename first_def_type to first_op_type.
11467 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
11468 (vect_get_and_check_slp_defs): Always use the type of the
11469 operand. Allow mixed vect_external_def, vect_constant_def types.
11470 (vect_get_constant_vectors): Handle mixed vect_external_def,
11471 vect_constant_def types.
11473 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
11475 PR tree-optimization/55524
11476 * tree-ssa-math-opts.c
11477 (convert_mult_to_fma): Don't use an fms construct
11478 when we don't have an fms operation, but fnma, and it looks
11479 likely that we'll be able to use the latter.
11481 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11483 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
11485 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
11486 inline fail caused by overwritable functions.
11488 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
11490 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
11491 unnecessary bits in the constant power of two case.
11493 2013-04-10 Richard Biener <rguenther@suse.de>
11495 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
11496 broken code swapping operands.
11497 (vect_build_slp_tree): Do not compute load permutations here.
11498 (vect_analyze_slp_instance): Compute load permutations here,
11499 after building the SLP tree.
11501 2013-04-09 Christian Bruel <christian.bruel@st.com>
11503 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
11504 of next/prev_real_insn.
11506 2013-04-09 Jan Hubicka <jh@suse.cz>
11508 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11509 Drop aliased parameter.
11510 (function_and_variable_visibility): Do not handle alias pairs.
11511 * cgraph.c (varpool_externally_visible_p): Update prototype.
11512 * varpool.c (varpool_add_new_variable): Update.
11514 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11516 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
11518 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
11520 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
11522 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
11524 2013-04-09 Marek Polacek <polacek@redhat.com>
11526 PR tree-optimization/48762
11527 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
11529 2013-04-09 Richard Biener <rguenther@suse.de>
11531 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
11533 (vect_build_slp_tree): Likewise.
11534 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
11535 calculating the cost of a SLP instance.
11536 (vect_analyze_slp_instance): Use it from here, after building
11539 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11541 PR middle-end/56883
11542 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
11543 expand_omp_for_static_chunk): Use simple_p = true in
11544 force_gimple_operand_gsi calls when assigning to addressable decls.
11546 2013-04-09 Jeff Law <law@redhat.com>
11548 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
11549 when the boolean was created by converting a wider object which
11550 had a boolean range.
11552 2013-04-09 Richard Biener <rguenther@suse.de>
11554 * tree-vectorizer.h (slp_void_p): Remove.
11555 (slp_tree): Typedef before _slp_tree declaration.
11556 (struct _slp_tree): Use a vector of slp_tree as children.
11557 (vect_get_place_in_interleaving_chain): Remove.
11558 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
11560 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
11562 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
11563 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
11564 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
11565 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
11566 Use slp_node instead of slp_void_p and adjust.
11568 2013-04-09 Richard Biener <rguenther@suse.de>
11570 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
11571 work that is not necessary.
11573 2013-04-09 Jakub Jelinek <jakub@redhat.com>
11575 PR tree-optimization/56854
11576 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
11577 forward into clobber stmts if it would change MEM_REF lhs into
11580 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
11582 * tree.c (type_hash_lookup, type_hash_add): Make static.
11583 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
11585 2013-04-09 Richard Biener <rguenther@suse.de>
11587 * tree.h (unsave_expr_now): Remove.
11588 * tree-inline.c (mark_local_for_remap_r): Remove.
11589 (unsave_expr_1): Likewise.
11590 (unsave_r): Likewise.
11591 (unsave_expr_now): Likewise.
11592 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
11593 (propagate_tree_value): Likewise.
11595 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
11597 * doc/rtl.texi (sequence): Rewrite documentation to match the
11598 current use of SEQUENCE rtl objects.
11599 * rtl.def (SEQUENCE): Likewise.
11601 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
11602 Update documentation.
11603 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
11604 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
11606 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
11608 2013-04-08 Teresa Johnson <tejohnson@google.com>
11610 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
11611 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
11613 (estimate_edge_size_and_time): Add comment to suggest using rounding
11615 (estimate_node_size_and_time): Ditto.
11616 (remap_edge_change_prob): Use helper rounding divide methods.
11617 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
11618 (gimple_mod_pow2_value_transform): Ditto.
11619 (gimple_mod_subtract_transform): Ditto.
11620 (gimple_ic_transform): Ditto.
11621 (gimple_stringops_transform): Ditto.
11622 * stmt.c (conditional_probability): Ditto.
11623 (emit_case_dispatch_table): Ditto.
11624 * lto-cgraph.c (merge_profile_summaries): Ditto.
11625 * tree-optimize.c (execute_fixup_cfg): Ditto.
11626 * cfgcleanup.c (try_forward_edges): Ditto.
11627 * cfgloopmanip.c (scale_loop_profile): Ditto.
11629 (duplicate_loop_to_header_edge): Ditto.
11630 (lv_adjust_loop_entry_edge): Ditto.
11631 * tree-vect-loop.c (vect_transform_loop): Ditto.
11632 * profile.c (compute_branch_probabilities): Ditto.
11633 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
11634 * lto-streamer-in.c (input_cfg): Ditto.
11635 * gimple-streamer-in.c (input_bb): Ditto.
11636 * ipa-cp.c (update_profiling_info): Ditto.
11637 (update_specialized_profile): Ditto.
11638 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
11639 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
11641 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11642 (compute_trg_info): Ditto.
11643 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
11644 (purge_dead_edges): Ditto.
11645 * loop-unswitch.c (unswitch_loop): Ditto.
11646 * cgraphclones.c (cgraph_clone_edge): Ditto.
11647 (cgraph_clone_node): Ditto.
11648 * tree-inline.c (copy_bb): Ditto.
11649 (copy_edges_for_bb): Ditto.
11650 (initialize_cfun): Ditto.
11651 (copy_cfg_body): Ditto.
11652 (expand_call_inline): Ditto.
11654 2013-04-08 Kai Tietz <ktietz@redhat.com>
11656 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
11657 TARGET_CYGWIN64 by TARGET_64BIT.
11659 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11661 * config/epiphany/epiphany.md (GPR_1): New constant.
11662 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
11663 * config/epiphany/epiphany.c (gen_compare_reg):
11664 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
11665 is already in place.
11666 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
11667 Don't require being called during rtl expansion; If y operlaps r0,
11669 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
11670 (epiphany_expand_epilogue): Likewise.
11672 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
11673 Don't use CC_FPmode for ORDERED / UNORDERED.
11674 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
11676 * config/epiphany/constraints.md (CnL): New constraint.
11677 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
11678 * config/epiphany/predicates.md (add_operand): Allow 1024.
11680 * config/epiphany/epiphany.md (logical_op): New code iterator.
11681 (op_mnc): New code attribute.
11682 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
11683 (mov_f+1, mov_f+2): New peephole2 patterns.
11685 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
11686 (cstoresi4): Also allow re-use of zero result when doing a NE
11687 comparison to a non-zero operand.
11688 Use (clobber (scratch)) for first insn if the gpr output is not needed.
11690 * config/epiphany/epiphany.md (<insn_opname>v2si3):
11691 Use gen_addsi3_i / gen_subsi3_i.
11693 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11697 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
11698 contain anything but clobbers, at most one __builtin_stack_restore,
11699 optionally debug stmts and final resx, and if it has at least one
11700 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
11701 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
11702 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
11703 which isn't defaut definition, remove them.
11704 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
11705 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
11706 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
11707 with MEM_REF LHS with SSA_NAME address.
11709 2013-04-08 Jeff Law <law@redhat.com>
11711 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
11713 2013-04-08 Richard Biener <rguenther@suse.de>
11715 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
11717 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
11718 determined vector type.
11719 (vect_analyze_data_refs): Likewise.
11720 (vect_get_new_vect_var): Adjust.
11721 (vect_create_destination_var): Preserve SSA name versions.
11722 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
11723 not dump anything here.
11725 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
11727 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
11728 Add member lr_slot_known.
11729 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
11731 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11732 Remove code that sets lr_slot_offset according to what a previous
11733 version of epiphany_emit_save_restore used to do.
11734 (epiphany_emit_save_restore): When doing an lr save or restore,
11735 set/verify lr_slot_known and lr_slot_offset.
11737 2013-04-08 Xinyu Qi <xyqi@marvell.com>
11740 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
11743 2013-04-08 Richard Biener <rguenther@suse.de>
11745 * alias.c (find_base_term): Fix thinko in previous change.
11747 2013-04-08 Jakub Jelinek <jakub@redhat.com>
11749 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
11750 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
11751 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
11752 if possible to compute val.
11753 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
11754 For QImode integers don't require anything about precision. Use
11755 const_with_all_bytes_same to find out if the constant doesn't have
11756 repeated bytes in it.
11758 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11760 * config/s390/s390.c (s390_expand_insv): Only accept insertions
11763 2013-04-08 Marek Polacek <polacek@redhat.com>
11765 PR rtl-optimization/48182
11766 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
11769 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11772 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
11773 nuses, make sure we have a label.
11775 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11778 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
11779 (rs6000_emit_swdiv_low_precision): Remove.
11780 (rs6000_emit_swdiv): Rewrite to handle between one and four
11781 iterations of Newton-Raphson generally; modify required number of
11782 iterations for some cases.
11783 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
11785 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
11787 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
11788 set-but-unused variable.
11790 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
11791 basic blocks of released function bodies garbage-collectable.
11793 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
11794 (struct rtl_opt_pass): Add TODO_df_finish.
11796 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
11798 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11800 * config/arm/constraints.md (q): New constraint.
11801 * config/arm/ldrdstrd.md: New file.
11802 * config/arm/arm.md (ldrdstrd.md) New include.
11803 (arm_movdi): Use "q" instead of "r" constraint
11804 for double-word memory access.
11805 (movdf_soft_insn): Likewise.
11806 * config/arm/vfp.md (movdi_vfp): Likewise.
11807 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
11808 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
11809 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
11810 (mem_ok_for_ldrd_strd): Likewise.
11811 (output_move_double): Update assertion.
11813 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11815 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
11817 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11819 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
11820 define_insn_and_split.
11821 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
11823 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11825 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
11826 define_insn_and_split.
11827 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
11828 (shiftsi3_compare): New pattern.
11829 (rrx): New pattern.
11830 * config/arm/unspecs.md (UNSPEC_RRX): New.
11832 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11834 * config/arm/arm.md (negdi_extendsidi): New pattern.
11835 (negdi_zero_extendsidi): Likewise.
11837 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11839 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
11840 define_insn_and_split.
11841 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
11842 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
11844 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11846 * config/arm/arm.md (arm_subdi3): Convert define_insn into
11847 define_insn_and_split.
11848 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
11849 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
11851 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11853 * config/arm/arm.md (subsi3_carryin): New pattern.
11854 (subsi3_carryin_const): Likewise.
11855 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
11856 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
11858 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11860 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
11862 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
11864 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
11865 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
11867 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11869 * config/arm/arm.c (arm_expand_builtin): Change fcode
11870 type to unsigned int.
11872 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11874 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
11876 2013-04-04 Ian Lance Taylor <iant@google.com>
11878 * doc/standards.texi (Standards): The Go frontend supports the Go 1
11881 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
11883 PR middle-end/56729
11884 * df-scan.c (df_insn_delete): Disable failing assert.
11886 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11888 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
11889 New function prototype.
11890 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
11891 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
11892 (arm_builtin_vectorized_function): New function.
11894 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11896 * config/arm/arm_neon_builtins.def: New file.
11897 * config/arm/arm.c (neon_builtin_data): Move contents to
11898 arm_neon_builtins.def.
11899 (enum arm_builtins): Include neon builtin definitions.
11900 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
11901 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
11903 2013-04-04 Marek Polacek <polacek@redhat.com>
11905 PR tree-optimization/48186
11906 * predict.c (maybe_hot_frequency_p): Return false if
11907 HOT_BB_FREQUENCY_FRACTION is 0.
11908 (cgraph_maybe_hot_edge_p): Likewise.
11910 2013-04-04 Richard Biener <rguenther@suse.de>
11912 PR tree-optimization/56826
11913 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
11916 2013-04-04 Richard Biener <rguenther@suse.de>
11918 PR tree-optimization/56213
11919 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
11920 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
11922 2013-04-04 Richard Biener <rguenther@suse.de>
11924 PR tree-optimization/56837
11925 * tree-loop-distribution.c (classify_partition): For non-zero
11926 values require that the value has the same precision as its
11927 mode to be useful as memset value.
11929 2013-04-03 Nick Clifton <nickc@redhat.com>
11931 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
11932 (fmssf4): Use fmsf.s on E3V5 architectures.
11933 (fnmasf4): Use fnmaf.s on E3V5 architectures.
11934 (fnmssf4): Use fnmsf.s on E3V5 architectures.
11936 2013-04-03 Jeff Law <law@redhat.com>
11938 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
11939 (lra-eliminations.o): Likewise.
11941 2013-04-03 Teresa Johnson <tejohnson@google.com>
11943 * gcov-io.c (compute_working_sets): Moved most of body of old
11944 compute_working_sets here from profile.c.
11945 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
11946 (gcov_working_set_t): Moved typedef here from basic-block.h
11947 (compute_working_set): Declare.
11948 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
11949 (get_working_sets): Renamed from compute_working_set,
11950 replace most of body with call to new compute_working_sets.
11951 (get_exec_counts): Replace call to compute_working_sets
11952 to get_working_sets.
11953 * profile.h (get_working_sets): Renamed from compute_working_set.
11954 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
11955 to get_working_sets.
11956 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
11957 * gcov-dump.c (dump_working_sets): New function.
11959 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
11961 * hwint.c (sext_hwi, zext_hwi): New functions.
11962 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
11963 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
11964 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
11965 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
11966 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
11967 (sext_hwi, zext_hwi): New functions.
11969 2013-04-03 Jeff Law <law@redhat.com>
11971 PR tree-optimization/56799
11972 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
11973 back test for widening conversion erroneously dropped in prior change.
11975 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11978 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
11979 instead of next_real_insn.
11981 2013-04-03 Marek Polacek <polacek@redhat.com>
11984 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
11986 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11989 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
11991 (thumb1_output_casesi): Likewise.
11992 (thumb2_output_casesi): Likewise.
11994 2013-04-03 Richard Biener <rguenther@suse.de>
11996 PR tree-optimization/56817
11997 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
11999 (tree_unroll_loops_completely_1): ... new function to manually
12000 walk the loop tree, properly defering outer loops of unrolled
12001 loops to later iterations.
12003 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12005 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12006 (vectorizable_load): Likewise.
12007 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12008 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12010 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12012 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12015 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12017 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12019 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12021 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12023 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12025 PR tree-optimization/56790
12026 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12029 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12031 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12033 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12034 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12037 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12040 * tree.h (force_folding_builtin_constant_p): New decl.
12041 * builtins.c (force_folding_builtin_constant_p): New variable.
12042 (fold_builtin_constant_p): Fold immediately also if
12043 force_folding_builtin_constant_p.
12045 2013-04-03 Richard Biener <rguenther@suse.de>
12047 PR tree-optimization/56812
12048 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12049 DRs of the same interleaving chain are independent.
12051 2013-04-02 Jason Merrill <jason@redhat.com>
12053 * gdbinit.in (pbb): Use debug fn.
12055 2013-04-02 Lawrence Crowl <crowl@google.com>
12057 * sese.h (struct ivtype_map_elt_s): Remove unused.
12058 (extern debug_ivtype_map): Remove unused.
12059 (extern eq_ivtype_map_elts): Remove unused.
12060 * sese.c (debug_ivtype_map): Removed unused.
12061 (debug_ivtype_map_1): Removed unused.
12062 (debug_ivtype_elt): Remove unused.
12063 (eq_ivtype_map_elts): Remove unused.
12066 2013-04-02 Kai Tietz <ktietz@redhat.com>
12069 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12070 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12071 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12073 (legitimize_pe_coff_symbol): Likewise.
12074 (is_imported_p): New helper-function.
12075 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12076 for Windows x64 targets.
12077 (ix86_expand_prologue): Optimize for pe-coff targets.
12078 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12079 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12080 medium/large code-model.
12081 (legitimize_pic_address): Likewise.
12082 (legitimize_tls_address): Likewise.
12083 (ix86_expand_call): Likewise.
12084 (x86_output_mi_thunk): Likewise.
12085 (get_dllimport_decl): Add new beimport argument.
12086 (construct_plt_address): Don't assert for x64 pe-coff targets.
12087 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12089 (SYMBOL_FLAG_STUBVAR): New macro.
12090 (SYMBOL_REF_STUBVAR_P): Likewise.
12091 * config/i386/winnt.c (stub_list): New structure.
12092 (stub_head): New local variable.
12093 (i386_pe_record_stub): New function.
12094 (i386_pe_file_end): Emit refptr-stubs.
12096 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12098 PR rtl-optimization/56745
12099 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12100 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12103 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12104 and both of them are MEM_REFs, just compare first argument for
12105 equality and attempt to deal even with differing offsets.
12108 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12109 of gimple_clobber_p to be MEM_REF.
12110 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12111 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12112 after gimplification.
12113 * asan.c (get_mem_ref_of_assignment): Don't instrument
12114 gimple_clobber_p stmts.
12115 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12116 gimple_clobber_p stmt if they have MEM_REF lhs and
12117 are dead because of another gimple_clobber_p stmt.
12118 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12119 gimple_clobber_p stmts like debug stmts.
12120 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12121 that refer to unused VAR_DECLs or uninitialized values.
12122 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12123 gimple_clobber_p stmts if they refer to removed parameters.
12124 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12127 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12129 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12130 using SWI48 mode attribute.
12132 2013-04-02 Wei Mi <wmi@google.com>
12134 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12135 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12136 *<rotate_insn><mode>3_mask in i386.md.
12138 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12140 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12142 2013-04-02 Richard Biener <rguenther@suse.de>
12144 PR tree-optimization/56778
12145 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12146 Runtime alias tests are not supported for gather loads.
12147 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12148 stmts referenced from SSA operands before updating SSA form.
12150 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12151 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12153 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12154 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12155 * config/arm/cortex-a53.md: New file.
12156 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12157 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12158 * config/arm/arm.c (arm_issue_rate): Likewise.
12159 * config/arm/arm-tune.md: Regenerate
12160 * config/arm/arm-tables.opt: Regenerate.
12161 * config/arm/arm-cores.def: Add cortex-a53.
12163 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12165 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12168 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12170 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12171 scalar load/store operations using B/H registers.
12172 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12174 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12176 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12178 * config/aarch64/aarch64.c
12179 (aarch64_simd_scalar_immediate_valid_for_move): New.
12180 * config/aarch64/aarch64-protos.h
12181 (aarch64_simd_scalar_immediate_valid_for_move): New.
12182 * config/aarch64/constraints.md (Dh, Dq): New.
12183 * config/aarch64/iterators.md (hq): New.
12185 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12187 * reorg.c (get_branch_condition): Deal with conditional returns.
12188 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12190 2013-04-01 Wei Mi <wmi@google.com>
12192 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12193 Truncate operand 2 using %b asm operand modifier.
12194 (*<shift_insn><mode>3_mask): Ditto.
12195 (*<rotate_insn><mode>3_mask): Ditto.
12197 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12199 PR middle-end/56798
12200 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12202 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12204 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12206 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12208 2013-03-30 Lawrence Crowl <crowl@google.com>
12210 * dse.c (clear_alias_sets): Remove never set.
12211 (disqualified_clear_alias_sets): Remove never set.
12212 (clear_alias_mode_pool): Remove never set.
12213 (dse_step0): Remove condition that is never true.
12214 (canon_address): Remove condition that is never true.
12215 (dse_step7): Remove condition that is never true.
12216 (rest_of_handle_dse): Remove condition that is never true.
12217 (rest_of_handle_dse::did_global): Remove never read from above.
12218 (dse_step2_spill): Remove never called from above.
12219 (dse_step5_spill): Remove never called from above.
12221 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12223 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12224 JUMP_TABLE_DATA changes.
12225 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12226 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12227 (Insns) <jump_table_data>: New entry.
12228 * doc/tm.texi: Regenerate.
12230 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12232 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12233 for table jump at the end of a basic block using tablejump_p.
12234 * targhooks.c (default_invalid_within_doloop): Likewise.
12235 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12236 target hook implementation that is identical to the default hook.
12237 (rs6000_invalid_within_doloop): Remove.
12239 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12240 unused variable from tablejump_p call.
12242 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12243 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12244 (INSN_DELETED_P): Likewise.
12245 (emit_jump_table_data): New prototype.
12246 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12247 after 4th as unused.
12248 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12249 * sched-vis.c (print_insn): Likewise.
12250 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12251 insn for compatibility with back ends that use next_active_insn to
12252 identify jump table data.
12253 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12254 (remove_insn): Likewise.
12255 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12257 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12258 (emit_jump_table_data): New function.
12260 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12261 basic block, a JUMP_TABLE_DATA never is.
12262 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12263 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12264 off from code handling real insns.
12265 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12266 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12267 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12268 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12269 is not a NONDEBUG_INSN_P.
12270 * ira-costs.c (scan_one_insn): Likewise.
12271 * jump.c (mark_all_labels): Likewise.
12272 (mark_jump_label_1): Likewise.
12273 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12274 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12276 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12277 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12278 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12279 JUMP_TABLE_DATA_P insns.
12280 (calculate_elim_costs_all_insns): Likewise.
12281 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12282 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12283 (delete_output_reload): Code style fixups.
12284 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12285 insn flags on this non-insn.
12286 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12287 as scheduling barriers, for pre-change compatibility.
12288 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12289 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12291 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12292 redundant JUMP_TABLE_DATA_P test.
12293 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12294 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12295 (frv_for_each_packet): Likewise.
12296 * config/i386/i386.c (min_insn_size): Likewise.
12297 (ix86_avoid_jump_mispredicts): Likewise.
12298 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12299 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12300 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12301 (mips16_insn_length): Robustify.
12302 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12303 (mips16_split_long_branches): Likewise.
12304 * config/pa/pa.c (pa_combine_instructions): Likewise.
12305 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12306 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12307 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12308 as contributing to pool range lengths.
12309 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12310 Remove redundant JUMP_TABLE_DATA_P test.
12311 (sh_loop_align): Likewise.
12312 (split_branches): Likewise.
12313 (sh_insn_length_adjustment): Likewise.
12314 * config/spu/spu.c (get_branch_target): Likewise.
12316 2013-03-29 Jan Hubicka <jh@suse.cz>
12318 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12319 gcov streaming; stream hot bb threshold to ltrans.
12320 * predict.c (get_hot_bb_threshold): Break out from ....
12321 (maybe_hot_count_p): ... here.
12322 (set_hot_bb_threshold): New function.
12323 * lto-section-in.c (lto_section_name): Add profile.
12324 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12325 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12326 and data-streamer.h
12327 (histogram_entry): New structure.
12328 (histogram, histogram_pool): New global vars.
12329 (histogram_hash): New structure.
12330 (histogram_hash::hash): New method.
12331 (histogram_hash::equal): Likewise.
12332 (account_time_size): New function.
12333 (cmp_counts): New function.
12334 (dump_histogram): New function.
12335 (ipa_profile_generate_summary): New function.
12336 (ipa_profile_write_summary): New function.
12337 (ipa_profile_read_summary): New function.
12338 (ipa_profile): Decide on threshold.
12339 (pass_ipa_profile): Add ipa_profile_write_summary and
12340 ipa_profile_read_summary.
12341 * Makefile.in (ipa.o): Update dependencies.
12342 * lto-streamer.h (LTO_section_ipa_profile): New section.
12344 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12346 * tree.h (VAR_P): New.
12348 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12351 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12353 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12355 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12356 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12357 (control_flow_insn_p): Likewise.
12358 * cfgrtl.c (duplicate_insn_chain): Likewise.
12359 * final.c (get_attr_length_1): Likewise.
12360 (shorten_branches): Likewise.
12361 (final_scan_insn): Likewise.
12362 * function.c (instantiate_virtual_regs): Likewise.
12363 * gcse.c (insert_insn_end_basic_block): Likewise.
12364 * ira-costs.c (scan_one_insn): Likewise.
12365 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12366 * lra.c (check_rtl): Likewise.
12367 * reload1.c (elimination_costs_in_insn): Likewise.
12368 * reorg.c (follow_jumps): Likewise.
12370 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12371 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12372 (thumb_far_jump_used_p): Likewise.
12373 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12374 (workaround_speculation): Likewise.
12375 (add_sched_insns_for_speculation): Likewise.
12376 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
12377 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12378 (frv_for_each_packet): Likewise.
12379 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
12380 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
12381 (final_emit_insn_group_barriers): Likewise.
12382 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12383 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12384 (mips16_insn_length): Likewise.
12385 * config/pa/pa.c (pa_reorg): Likewise.
12386 (pa_combine_instructions): Likewise.
12387 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
12388 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
12389 (sh_reorg): Likewise.
12390 (split_branches): Likewise.
12391 * config/spu/spu.c (get_branch_target): Likewise.
12393 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
12396 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
12398 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
12399 Fix declaration name.
12401 2013-03-28 Lawrence Crowl <crowl@google.com>
12403 * graphds.h (struct graph.indicies): Remove unused.
12404 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
12405 (SCOP_ORIGINAL_PDDRS): Remove unused.
12406 * sese.h (extern insert_loop_close_phis): Removed unused.
12407 (extern insert_guard_phis): Removed unused.
12408 (extern ivtype_map_elt_info): Removed unused.
12409 (new_ivtype_map_elt): Removed unused.
12410 * sese.c (ivtype_map_elt_info): Removed unused.
12412 2013-03-28 Lawrence Crowl <crowl@google.com>
12414 * Makefile.in: Add several missing include dependences.
12416 (test-dump.o): New. This object is not added to any executable,
12417 but is present for ad-hoc testing.
12419 (debug (const bitmap_head_def &)): New.
12420 (debug (const bitmap_head_def *)): New.
12422 (extern debug (const bitmap_head_def &)): New.
12423 (extern debug (const bitmap_head_def *)): New.
12425 (debug (edge_def &)): New.
12426 (debug (edge_def *)): New.
12428 (debug (basic_block_def &)): New.
12429 (debug (basic_block_def *)): New.
12431 (dump_node (const_tree, int, FILE *)): Correct source file.
12433 (debug (die_struct &)): New.
12434 (debug (die_struct *)): New.
12436 (extern debug (die_struct &)): New.
12437 (extern debug (die_struct *)): New.
12438 * gimple-pretty-print.c
12439 (debug (gimple_statement_d &)): New.
12440 (debug (gimple_statement_d *)): New.
12441 * gimple-pretty-print.h
12442 (extern debug (gimple_statement_d &)): New.
12443 (extern debug (gimple_statement_d *)): New.
12445 (debug (ira_allocno_copy &)): New.
12446 (debug (ira_allocno_copy *)): New.
12447 (debug (ira_allocno &)): New.
12448 (debug (ira_allocno *)): New.
12450 (extern debug (ira_allocno_copy &)): New.
12451 (extern debug (ira_allocno_copy *)): New.
12452 (extern debug (ira_allocno &)): New.
12453 (extern debug (ira_allocno *)): New.
12455 (debug (live_range &)): New.
12456 (debug (live_range *)): New.
12458 (debug (lra_live_range &)): New.
12459 (debug (lra_live_range *)): New.
12461 (debug (lra_live_range &)): New.
12462 (debug (lra_live_range *)): New.
12464 (debug (omega_pb_d &)): New.
12465 (debug (omega_pb_d *)): New.
12467 (extern debug (omega_pb_d &)): New.
12468 (extern debug (omega_pb_d *)): New.
12470 (debug (const rtx_def &)): New.
12471 (debug (const rtx_def *)): New.
12473 (debug_tree (tree): Move within file.
12474 (debug_raw (const tree_node &)): New.
12475 (debug_raw (const tree_node *)): New.
12476 (dump_tree_via_hooks (const tree_node *, int)): New.
12477 (debug (const tree_node &)): New.
12478 (debug (const tree_node *)): New.
12479 (debug_verbose (const tree_node &)): New.
12480 (debug_verbose (const tree_node *)): New.
12481 (debug_head (const tree_node &)): New.
12482 (debug_head (const tree_node *)): New.
12483 (debug_body (const tree_node &)): New.
12484 (debug_body (const tree_node *)): New.
12485 (debug_vec_tree (tree): Move and reimplement in terms of dump.
12486 (debug (vec<tree, va_gc> &)): New.
12487 (debug (vec<tree, va_gc> *)): New.
12489 (extern debug (const rtx_def &)): New.
12490 (extern debug (const rtx_def *)): New.
12492 (debug_raw (simple_bitmap_def &)): New.
12493 (debug_raw (simple_bitmap_def *)): New.
12494 (debug (simple_bitmap_def &)): New.
12495 (debug (simple_bitmap_def *)): New.
12497 (extern debug (simple_bitmap_def &)): New.
12498 (extern debug (simple_bitmap_def *)): New.
12499 (extern debug_raw (simple_bitmap_def &)): New.
12500 (extern debug_raw (simple_bitmap_def *)): New.
12502 (debug (vinsn_def &)): New.
12503 (debug (vinsn_def *)): New.
12504 (debug_verbose (vinsn_def &)): New.
12505 (debug_verbose (vinsn_def *)): New.
12506 (debug (expr_def &)): New.
12507 (debug (expr_def *)): New.
12508 (debug_verbose (expr_def &)): New.
12509 (debug_verbose (expr_def *)): New.
12510 (debug (vec<rtx> &)): New.
12511 (debug (vec<rtx> *)): New.
12513 (extern debug (vinsn_def &)): New.
12514 (extern debug (vinsn_def *)): New.
12515 (extern debug_verbose (vinsn_def &)): New.
12516 (extern debug_verbose (vinsn_def *)): New.
12517 (extern debug (expr_def &)): New.
12518 (extern debug (expr_def *)): New.
12519 (extern debug_verbose (expr_def &)): New.
12520 (extern debug_verbose (expr_def *)): New.
12521 (extern debug (vec<rtx> &)): New.
12522 (extern debug (vec<rtx> *)): New.
12524 (_list_iter_cond_expr): Make inline instead of static.
12526 (debug (sreal &)): New.
12527 (debug (sreal *)): New.
12529 (extern debug (sreal &)): New.
12530 (extern debug (sreal *)): New.
12532 (extern debug_raw (const tree_node &)): New.
12533 (extern debug_raw (const tree_node *)): New.
12534 (extern debug (const tree_node &)): New.
12535 (extern debug (const tree_node *)): New.
12536 (extern debug_verbose (const tree_node &)): New.
12537 (extern debug_verbose (const tree_node *)): New.
12538 (extern debug_head (const tree_node &)): New.
12539 (extern debug_head (const tree_node *)): New.
12540 (extern debug_body (const tree_node &)): New.
12541 (extern debug_body (const tree_node *)): New.
12542 (extern debug (vec<tree, va_gc> &)): New.
12543 (extern debug (vec<tree, va_gc> *)): New.
12545 (debug (struct loop &)): New.
12546 (debug (struct loop *)): New.
12547 (debug_verbose (struct loop &)): New.
12548 (debug_verbose (struct loop *)): New.
12549 * tree-dump.c: Add header dependence.
12551 (extern debug (struct loop &)): New.
12552 (extern debug (struct loop *)): New.
12553 (extern debug_verbose (struct loop &)): New.
12554 (extern debug_verbose (struct loop *)): New.
12556 (debug (data_reference &)): New.
12557 (debug (data_reference *)): New.
12558 (debug (vec<data_reference_p> &)): New.
12559 (debug (vec<data_reference_p> *)): New.
12560 (debug (vec<ddr_p> &)): New.
12561 (debug (vec<ddr_p> *)): New.
12563 (extern debug (data_reference &)): New.
12564 (extern debug (data_reference *)): New.
12565 (extern debug (vec<data_reference_p> &)): New.
12566 (extern debug (vec<data_reference_p> *)): New.
12567 (extern debug (vec<ddr_p> &)): New.
12568 (extern debug (vec<ddr_p> *)): New.
12570 (debug (pt_solution &)): New.
12571 (debug (pt_solution *)): New.
12573 (extern debug (pt_solution &)): New.
12574 (extern debug (pt_solution *)): New.
12576 (debug (_var_map &)): New.
12577 (debug (_var_map *)): New.
12578 (debug (tree_live_info_d &)): New.
12579 (debug (tree_live_info_d *)): New.
12581 (extern debug (_var_map &)): New.
12582 (extern debug (_var_map *)): New.
12583 (extern debug (tree_live_info_d &)): New.
12584 (extern debug (tree_live_info_d *)): New.
12586 2013-03-28 Jan Hubicka <jh@suse.cz>
12588 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
12590 2013-03-28 Ian Bolton <ian.bolton@arm.com>
12592 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
12593 record only when desired or required.
12595 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
12597 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
12598 *vec_extractv2di_1_rex64. Use x64 isa attribute.
12600 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12602 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
12603 (*andsi3_compare0_uxtw): New pattern.
12604 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
12605 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
12607 2013-03-28 Jan Hubicka <jh@suse.cz>
12609 * data-streamer-in.c (streamer_read_gcov_count): New function.
12610 * gimple-streamer-out.c: Include value-prof.h.
12611 (output_gimple_stmt): Output histogram.
12612 (output_bb): Use streamer_write_gcov_count.
12613 * value-prof.c: Include data-streamer.h
12614 (dump_histogram_value): Add HIST_TYPE_MAX.
12615 (stream_out_histogram_value): New function.
12616 (stream_in_histogram_value): New function.
12617 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
12618 (stream_out_histogram_value, stream_in_histogram_value): Declare.
12619 * data-streamer-out.c (streamer_write_gcov_count): New function.
12620 (streamer_write_gcov_count_stream): New function.
12621 * lto-cgraph.c (lto_output_edge): Update counter streaming.
12622 (lto_output_node): Likewise.
12623 (input_node, input_edge): Likewise.
12624 * lto-streamer-out.c (output_cfg): Update streaming.
12625 * lto-streamer-in.c (input_cfg): Likewise.
12626 * data-streamer.h (streamer_write_gcov_count,
12627 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
12628 * gimple-streamer-in.c: Include value-prof.h
12629 (input_gimple_stmt): Input histograms.
12630 (input_bb): Update profile streaming.
12632 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12634 * genmodes.c (emit_max_int): New function.
12635 (emit_insn_modes_h): Added call to emit_max_function.
12636 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
12638 * machmode.def: Fixed comment.
12640 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
12642 * combine.c (try_combine): Removed useless assert.
12643 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
12645 2013-03-28 Marek Polacek <polacek@redhat.com>
12646 Richard Biener <rguenther@suse.de>
12648 PR tree-optimization/56695
12649 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
12650 build signed result of a vector comparison.
12651 * tree-cfg.c (verify_gimple_comparison): Check that a result
12652 of a vector comparison has signed type.
12654 2013-03-28 Richard Biener <rguenther@suse.de>
12656 PR tree-optimization/37021
12657 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
12658 do not restrict gaps between groups.
12659 * tree-vect-stmts.c (vectorizable_load): Properly account for
12660 a gap between groups.
12662 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
12664 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
12665 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
12668 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
12670 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
12671 * doc/extend.texi (Named Address Spaces): Ditto.
12672 (Variable Attributes): Ditto.
12674 2013-03-27 Kai Tietz <ktietz@redhat.com>
12676 * config.build: Add support for cygwin x64 target.
12677 * config.gcc: Likewise.
12678 * config.host: Likewise.
12679 * configure.ac: Likewise
12680 * configure: Regenerated.
12682 2013-03-27 Kai Tietz <ktietz@redhat.com>
12684 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
12685 * config/i386/t-cygwin-w64: New file.
12686 * config/i386/cygwin-w64.h: New file.
12687 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
12688 and add support for x64-cygwin target.
12689 (CPP_SPEC): Likewise.
12690 (CXX_WRAP_SPEC_LIST): Undefine before define.
12691 (LIBGCJ_SONAME): Use 15 as version.
12693 2013-03-27 Richard Biener <rguenther@suse.de>
12695 PR tree-optimization/56716
12696 * tree-ssa-structalias.c (perform_var_substitution): Adjust
12697 dumping for ref nodes.
12699 2013-03-27 Martin Jambor <mjambor@suse.cz>
12701 PR tree-optimization/55334
12702 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
12703 restricted pointers to arrays.
12705 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
12707 * Makefile.in (.SUFFIXES): Add .cc.
12708 (.c.o): Apply same recipe for implicit rule .cc.o.
12710 2013-03-27 Richard Biener <rguenther@suse.de>
12712 PR tree-optimization/37021
12713 * tree-vect-data-refs.c (vect_check_strided_load): Allow
12714 REALPART/IMAGPART_EXPRs around the supported refs.
12715 * tree-ssa-structalias.c (find_func_aliases): Assume that
12716 floating-point values are not used to transfer pointers.
12718 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
12720 * target.def (TARGET_HAS_IFUNC_P): New target hook.
12721 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
12722 * doc/tm.texi: Regenerate.
12723 * targhooks.h (default_has_ifunc_p): New.
12724 * targhooks.c (default_has_ifunc_p): Ditto.
12725 * config/linux-protos.h: New file.
12726 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
12727 this hook for linux which disables support of indirect functions in
12729 * config/linux-android.c: New file.
12730 * config/t-linux-android.c: Ditto.
12731 * config.gcc: Added new object file linux-android.o.
12732 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
12733 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
12734 * varasm.c (do_assemble_alias): Likewise.
12735 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
12736 doesn't support indirect functions.
12737 * configure: Regenerate.
12739 2013-03-27 Bin Cheng <bin.cheng@arm.com>
12742 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
12743 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
12746 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12748 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
12750 2013-03-27 Terry Guo <terry.guo@arm.com>
12752 * config/arm/arm-cores.def: Added core cortex-r7.
12753 * config/arm/arm-tune.md: Regenerated.
12754 * config/arm/arm-tables.opt: Regenerated.
12755 * doc/invoke.texi: Added entry for core cortex-r7.
12757 2013-03-27 Walter Lee <walt@tilera.com>
12759 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
12760 double-decrement of next_scratch_regno.
12762 2013-03-27 Walter Lee <walt@tilera.com>
12764 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
12766 (insn_v1mulus): Ditto.
12767 (insn_v2muls): Ditto.
12769 2013-03-27 Walter Lee <walt@tilera.com>
12771 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
12772 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
12774 2013-03-27 Walter Lee <walt@tilera.com>
12776 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
12777 (*sibcall_value): Ditto.
12779 2013-03-27 Walter Lee <walt@tilera.com>
12781 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
12782 (insn_mnz_v8qi): ... this ...
12783 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
12785 (insn_v<n>mnz): Replaced by ...
12786 (insn_v1mnz): ... this ...
12787 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
12789 (insn_mz_<mode>): Replaced by ...
12790 (insn_mz_v8qi): ... this ...
12791 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
12793 (insn_v<n>mz): Replaced by ...
12794 (insn_v1mz): ... this ...
12795 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
12798 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
12800 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
12802 2013-03-26 Roland McGrath <mcgrathr@google.com>
12804 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
12805 than fprintf with a non-constant, non-format string.
12807 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12809 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
12810 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
12811 operand 0 predicate.
12812 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
12813 attribute. Use general_x64nomem_operand as operand 1 predicate.
12814 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
12815 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12816 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
12817 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
12818 (mov<mode>_insv_1): Remove expander. Merge insn with
12819 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
12820 Use general_x64nomem_operand as operand 1 predicate.
12821 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
12822 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
12823 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12824 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
12825 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
12826 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
12827 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
12828 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
12829 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
12830 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
12831 (general_x64nomem_operand): Ditto.
12833 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
12835 * config/rtems.opt: Add -pthread option.
12837 2013-03-26 Richard Biener <rguenther@suse.de>
12839 * alias.c (find_base_term): Avoid redundant and not used recursion.
12840 (base_alias_check): Get the initial base term from the caller.
12841 (true_dependence_1): Compute and pass base terms to base_alias_check.
12842 (write_dependence_p): Likewise.
12843 (may_alias_p): Likewise.
12845 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
12847 * config/aarch64/aarch64.c (aarch64_classify_address): Support
12848 PC-relative load in SI modes and above only.
12850 2013-03-26 Xinyu Qi <xyqi@marvell.com>
12852 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
12853 * config/arm/iwmmxt.md (WCGR0): Update.
12854 (WCGR1, WCGR2, WCGR3): Likewise.
12856 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
12858 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
12859 Use x64 and nox64 isa attributes.
12861 2013-03-26 Richard Biener <rguenther@suse.de>
12863 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
12864 alignment computations and rely on get_object_alignment_1
12865 for the !TYPE_P case.
12866 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
12868 2013-03-26 Walter Lee <walt@tilera.com>
12870 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
12871 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
12873 2013-03-25 Jeff Law <law@redhat.com>
12875 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
12876 check for INTEGRAL_TYPE_P that was missing due to checking in
12877 wrong version of prior patch.
12879 2013-03-25 Walter Lee <walt@tilera.com>
12881 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
12882 TILEGX_INSN_SHUFFLEBYTES1.
12883 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
12885 (tilegx_builtins): Ditto.
12886 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
12888 2013-03-25 Walter Lee <walt@tilera.com>
12890 * config/tilegx/tilegx.md (floatsisf2): New pattern.
12891 (floatunssisf2): New pattern.
12892 (floatsidf2): New pattern.
12893 (floatunssidf2): New pattern.
12895 2013-03-25 Walter Lee <walt@tilera.com>
12897 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
12898 tests for constraint J, K, N, P.
12900 2013-03-25 Walter Lee <walt@tilera.com>
12902 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
12903 Use indirect/pcrel encoding.
12904 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
12907 2013-03-25 Steve Ellcey <sellcey@mips.com>
12909 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
12910 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
12911 * config/mips/mips.c (mips_option_override): Set IMADD default.
12912 * config/mips/mips.h (PTF_AVOID_IMADD): New.
12913 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
12914 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
12915 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
12916 * doc/invoke.texi (-mimadd/-mno-imadd): New.
12918 2013-03-25 Jeff Law <law@redhat.com>
12920 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
12921 slightly to avoid creating and folding useless trees. Simplify
12922 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
12924 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
12926 * config/i386/i386.md (*zero_extendsidi2): Merge with
12927 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
12928 * config/i386/predicates.md (x86_64_zext_operand): Rename from
12929 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
12930 targets. Clarify comment.
12932 2013-03-25 Martin Jambor <mjambor@suse.cz>
12934 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
12935 pass-through jump functions differently.
12936 (ipa_read_jump_function): Likewise. Also use setter functions to set
12939 2013-03-25 Martin Jambor <mjambor@suse.cz>
12941 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
12942 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
12944 (ipa_get_indirect_edge_target): New function.
12945 (devirtualization_time_bonus): New parameter known_aggs, pass it to
12946 ipa_get_indirect_edge_target. Update all callers.
12947 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
12948 ipa_get_indirect_edge_target_1 instead of calling
12949 ipa_get_indirect_edge_target.
12950 (create_specialized_node): Pass aggvlas to
12951 ipcp_discover_new_direct_edges.
12953 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12955 * config/arm/arm.md (f_sels, f_seld): New types.
12956 (*cmov<mode>): New pattern.
12957 * config/arm/predicates.md (arm_vsel_comparison_operator): New
12960 2013-03-25 Kai Tietz <ktietz@redhat.com>
12962 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
12963 POSIX-printf for mingw-hosted builds.
12965 2013-03-25 Richard Biener <rguenther@suse.de>
12967 PR middle-end/56694
12968 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
12969 must-not-throw stmt location.
12971 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12973 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
12974 Emit load-acquire versions when acq is true.
12975 (arm_emit_store_exclusive): Add rel parameter.
12976 Emit store-release versions when rel is true.
12977 (arm_split_compare_and_swap): Use acquire-release instructions
12979 of barriers when appropriate.
12980 (arm_split_atomic_op): Likewise.
12981 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
12982 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
12983 (VUNSPEC_SLX): Likewise.
12984 (VUNSPEC_LDA): Likewise.
12985 (VUNSPEC_STL): Likewise.
12986 * config/arm/sync.md (atomic_load<mode>): New pattern.
12987 (atomic_store<mode>): Likewise.
12988 (arm_load_acquire_exclusive<mode>): Likewise.
12989 (arm_load_acquire_exclusivesi): Likewise.
12990 (arm_load_acquire_exclusivedi): Likewise.
12991 (arm_store_release_exclusive<mode>): Likewise.
12993 2013-03-25 Catherine Moore <clm@codesourcery.com>
12995 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
12996 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
12997 * config/mip/predicates.md (lwsp_swsp_operand,
12998 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
12999 sb16_operand, db4_operand, db7_operand, ib3_operand,
13000 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13001 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13002 andi16_operand): New predicates.
13003 * config/mips/mips.md (compression): New attribute.
13004 (enabled): New attribute.
13005 (length): Consider compression in computing length.
13006 (shift_compression): New code attribute.
13007 (*add<mode>3): New operands. Record compression.
13008 (sub<mode>3): Likewise.
13009 (one_cmpl<mode>2): Likewise.
13010 (*and<mode>3): Likewise.
13011 (*ior<mode>3): Likewise.
13012 (unnamed pattern for xor): Likewise.
13013 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13014 (*<optab><mode>3): Likewise.
13015 (*mov<mode>_internal: Likewise.
13016 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13017 (mips_unsigned_immediate_p): New.
13018 (umips_lwsp_swsp_address_p): New.
13019 (m16_based_address_p): New.
13020 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13021 (mips_unsigned_immediate_p): New prototype.
13022 (lwsp_swsp_address_p): New prototype.
13023 (m16_based_address_p): New prototype.
13024 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13025 (mips_signed_immediate_p): New function.
13026 (m16_based_address_p): New function.
13027 (lwsp_swsp_address_p): New function.
13028 (mips_print_operand_punctuation): Recognize short delay slot insns
13029 for microMIPS.add<mode>3"
13031 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13034 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13035 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13037 2013-03-25 Richard Biener <rguenther@suse.de>
13039 PR tree-optimization/56689
13040 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13043 2013-03-25 Richard Biener <rguenther@suse.de>
13045 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13047 (memory_references): Likewise.
13048 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13049 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13050 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13051 (gather_mem_refs_in_loops): Fold into ...
13052 (analyze_memory_references): ... this. Move initialization
13053 to tree_ssa_lim_initialize.
13054 (fill_always_executed_in): Rename to ...
13055 (fill_always_executed_in_1): ... this.
13056 (fill_always_executed_in): Move contains_call computation to
13057 this new function from ...
13058 (tree_ssa_lim_initialize): ... here.
13059 (tree_ssa_lim): Call fill_always_executed_in.
13061 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13063 * postreload.c (reload_combine): Fix code detecting returns.
13065 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13067 * function.c (emit_use_return_register_into_block): On cc0 targets,
13068 do not emit the sequence between cc0 setter and user.
13070 2013-03-25 Kai Tietz <ktietz@redhat.com>
13072 * config/i386/predicates.md (local_symbolic_operand): Interpret
13073 dll-imported symbols as none-local.
13075 2013-03-25 Richard Biener <rguenther@suse.de>
13077 * tree-ssa-loop-im.c (struct depend): Remove.
13078 (struct lim_aux_data): Make depends a vec of gimples.
13079 (free_lim_aux_data): Adjust.
13080 (add_dependency): Likewise.
13081 (set_level): Likewise.
13083 2013-03-25 Richard Biener <rguenther@suse.de>
13085 PR middle-end/56434
13086 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13087 the pointer returned by calls with ECF_MALLOC set.
13089 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13091 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13093 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13095 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13096 using MMXMODE mode iterator.
13097 (*move<mode>_internal): Merge with *movv2sf_internal and
13098 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13100 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13102 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13103 (record_last_mem_set_info): Likewise.
13105 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13106 of XNEWVEC followed by memset.
13107 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13109 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13111 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13112 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13113 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13114 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13115 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13116 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13117 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13118 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13119 BARRIER_P instead of GET_CODE.
13121 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13123 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13124 inaccuracy in the probing code.
13126 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13127 (ctrapdi4): Likewise.
13129 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13131 * calls.c (expand_call): Add missing guard to code handling return
13132 of non-BLKmode structures in MSB.
13133 * function.c (expand_function_end): Likewise.
13135 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13137 * combine.c (try_combine): Adjust comment. Do not add the set of
13138 insn #0 if the destination indirectly is set or dies in insn #2.
13139 Tidy up code to distribute a new note.
13141 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13143 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13144 also for alternatives 16 and 17.
13146 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13148 * config/i386/sse.md (*mov<mode>_internal): Merge with
13149 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13150 Emit insn template depending on type attribute. Use
13151 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13152 movd instead of movq mnemonic for interunit moves. Rewrite mode
13153 attribute calculation. Remove unit attribute calculation.
13154 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13155 Set prefix_data16 attribute for DImode ssemov types.
13156 Use Ym instead of y for SSE-MMX conversion alternatives.
13157 Reorder operand constraints.
13159 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13161 * df.h (df_insn_delete): Adjust prototype.
13162 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13163 and let it decide whether mark the basic block dirty.
13164 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13165 * df-scan.c (df_insn_info_delete): New helper function, split
13166 off from df_insn_delete.
13167 (df_scan_free_bb_info): Use it.
13168 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13170 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13171 that the insn is actually an insn and it has a non-NULL basic block.
13172 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13174 2013-03-22 Richard Biener <rguenther@suse.de>
13176 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13178 (mem_ref_alloc): Do not allocate them.
13179 (refs_independent_p): Do not query or maintain a cache.
13181 2013-03-22 Richard Biener <rguenther@suse.de>
13183 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13184 (gather_mem_refs_in_loops): Do not compute it.
13185 (analyze_memory_references): Do not allocate it.
13186 (tree_ssa_lim_finalize): Do not free it.
13187 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13189 2013-03-22 Richard Biener <rguenther@suse.de>
13191 * is-a.h (as_a): Use gcc_checking_assert.
13193 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13195 * config/aarch64/aarch64.c (aarch64_print_operand): New
13196 format specifier for printing a constant in hex.
13197 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13198 format specifier for printing second operand.
13200 2013-03-22 Richard Biener <rguenther@suse.de>
13202 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13204 (gather_mem_refs_in_loops): Perform store accumulation here.
13205 (create_vop_ref_mapping_loop): Remove.
13206 (create_vop_ref_mapping): Likewise.
13207 (analyze_memory_references): Initialize refs_stored_in_loop.
13208 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13209 (record_indep_loop): Remove.
13210 (record_dep_loop): New function.
13211 (ref_indep_loop_p_1): Adjust to only walk over references
13212 in the loop, not its subloops.
13213 (ref_indep_loop_p): Rename to ...
13214 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13215 maintaining a more fine-grained cache.
13216 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13217 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13219 2013-03-22 Richard Biener <rguenther@suse.de>
13221 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13222 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13223 aggregate mem_ref_loc.
13224 (free_mem_ref_locs): Inline into ...
13225 (memref_free): ... this and adjust.
13226 (mem_ref_alloc): Adjust.
13227 (mem_ref_locs_alloc): Remove.
13228 (record_mem_ref_loc): Adjust.
13229 (get_all_locs_in_loop): Rewrite into ...
13230 (for_all_locs_in_loop): ... this iterator.
13231 (rewrite_mem_ref_loc): New functor.
13232 (rewrite_mem_refs): Use for_all_locs_in_loop.
13233 (sm_set_flag_if_changed): New functor.
13234 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13235 (ref_always_accessed): New functor.
13236 (ref_always_accessed_p): Use for_all_locs_in_loop.
13238 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13240 * tree-pass.h (PROP_gimple_lvec): New.
13241 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13242 (init_optimization_passes): Move pass_lower_vector.
13243 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13245 (pass_lower_vector): Provide PROP_gimple_lvec.
13246 (pass_lower_vector_ssa): Likewise.
13247 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13249 2013-03-21 Mark Wielaard <mjw@redhat.com>
13251 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13253 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13255 * config/i386/i386.md (*movdi_internal): Disparage slightly
13256 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13257 conversion alternatives.
13259 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13261 PR middle-end/48087
13262 * diagnostic.def (DK_WERROR): New kind.
13263 * diagnostic.h (werrorcount): Define.
13264 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13265 promoted to DK_ERROR, increment DK_WERROR counter instead of
13267 * toplev.c (toplev_main): Call print_ignored_options even if
13268 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13269 even if just werrorcount is non-zero.
13272 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13274 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13275 (string_cst_pool_decl): New function.
13276 (optimize_one_addr_into_implicit_ptr): New function.
13277 (resolve_addr_in_expr): Optimize DWARF location expression
13278 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13279 which doesn't live in memory, but has DW_AT_location or
13280 DW_AT_const_value, or refers to a string literal, into
13281 DW_OP_GNU_implicit_pointer.
13282 (optimize_location_into_implicit_ptr): New function.
13283 (resolve_addr): If removing DW_AT_location of a variable because
13284 it was DW_OP_addr of address of the variable, but the variable doesn't
13285 live in memory, try to emit const value attribute for the initializer.
13287 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13289 * tree.h (VECTOR_TYPE_P): New macro.
13290 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13291 TYPE_MODE): Use it.
13292 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13293 VEC_COND_EXPR cannot be lvalues.
13294 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13296 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13298 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13299 Restrict the transformation to equal modes.
13301 2013-03-21 Richard Biener <rguenther@suse.de>
13303 PR tree-optimization/39326
13304 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13305 (MEM_ANALYZABLE): Adjust.
13306 (record_mem_ref_loc): Move bitmap ops ...
13307 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13308 unanalyzable refs, do not record locations for it.
13309 (analyze_memory_references): Allocate ref zero as shared
13311 (refs_independent_p): Do not test for unanalyzed mems here.
13312 (ref_indep_loop_p_1): Special-case disambiguation against
13313 the unanalyzed ref.
13314 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13316 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13318 * config/arm/arm-protos.h (tune_params): Add
13319 prefer_neon_for_64bits field.
13320 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13321 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13322 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13323 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13324 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13325 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13326 (arm_option_override): Handle -mneon-for-64bits new option.
13327 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13328 (prefer_neon_for_64bits): Declare new variable.
13329 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13330 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13331 (arch_enabled): Handle new arch types. Remove support for onlya8
13333 (one_cmpldi2): Use new arch names.
13334 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13335 * config/arm/arm.opt (mneon-for-64bits): Add option.
13336 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13337 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13338 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13340 * doc/invoke.texi (-mneon-for-64bits): Document.
13342 2013-03-21 Richard Biener <rguenther@suse.de>
13344 PR tree-optimization/39326
13345 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13346 (sort_bbs_in_loop_postorder_cmp): New function.
13347 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13349 2013-03-21 Richard Biener <rguenther@suse.de>
13351 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13352 (vect_insert_into_interleaving_chain): Likewise.
13353 (vect_drs_dependent_in_basic_block): Inline ...
13354 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13356 (vect_analyze_data_ref_dependence): ... here. Simplify.
13357 (vect_check_interleaving): Simplify.
13358 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13359 (vect_slp_analyze_data_ref_dependences): ... this new function.
13360 (dr_group_sort_cmp): New function.
13361 (vect_analyze_data_ref_accesses): Compute data-reference groups
13362 here instead of in vect_analyze_data_ref_dependence. Use
13363 a more efficient algorithm.
13364 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13365 vect_slp_analyze_data_ref_dependences. Call
13366 vect_analyze_data_ref_accesses earlier.
13367 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13368 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13369 (vect_slp_analyze_data_ref_dependences): New prototype.
13371 2013-03-21 Richard Biener <rguenther@suse.de>
13373 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13374 ref is stored in the loop.
13375 (find_refs_for_sm): Walk only over all stores.
13376 (store_motion_loop): Allocate from lim_bitmap_obstack.
13377 (store_motion): Likewise.
13379 2013-03-21 Richard Biener <rguenther@suse.de>
13381 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
13382 Update virtual SSA form.
13384 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13386 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
13387 * configure: Regenerate.
13388 * config.in: Regenerate.
13389 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
13390 if !HAVE_LD_EH_FRAME_CIEV3.
13392 2013-03-21 Richard Biener <rguenther@suse.de>
13394 * tree-cfg.c (verify_expr_no_block): New function.
13395 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
13396 nor DECL_VALUE_EXPR have locations with associated blocks.
13397 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
13398 (clear_unused_block_pointer): Remove code dealing with
13399 blocks in DECL_DEBUG_EXPR locations.
13401 2013-03-21 Richard Biener <rguenther@suse.de>
13403 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
13404 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
13405 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
13406 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
13407 instead of DECL_DEBUG_EXPR_IS_FROM.
13408 * gimplify.c (gimplify_modify_expr): Likewise.
13409 * tree-cfg.c (verify_expr_location_1): Likewise.
13410 * tree-complex.c (create_one_component_var): Likewise.
13411 * tree-sra.c (create_access_replacement): Likewise.
13412 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
13413 (clear_unused_block_pointer): Likewise.
13414 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
13415 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
13416 * var-tracking.c (var_debug_decl): Likewise.
13417 (track_expr_p): Likewise.
13418 * tree-inline.c (add_local_variables): Likewise. Set
13419 DECL_HAS_DEBUG_EXPR_P after copying it.
13420 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
13421 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
13423 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13426 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
13427 * configure: Regenerate.
13428 * config.in: Regenerate.
13429 * config/i386/i386.md (*movdf_internal): Use
13430 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13431 movd instead of movq mnemonic for interunit moves.
13432 (*movdi_internal): Ditto.
13434 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13436 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
13437 (abd<mode>_3): New pattern.
13438 (aba<mode>_3): New pattern.
13439 (fabd<mode>_3): New pattern.
13441 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13443 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
13444 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
13445 occurrence of REGISTER_PREFIX as its empty string.
13447 2013-03-20 Jeff Law <law@redhat.com>
13449 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
13450 addititional equivalences for equality comparisons between an SSA_NAME
13451 and a constant where the SSA_NAME was set from a widening conversion.
13453 2013-03-20 Walter Lee <walt@tilera.com>
13455 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
13457 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13459 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
13460 depending on type attribute.
13461 (*movti_internal): Ditto.
13462 (*movtf_internal): Ditto.
13463 (*movxf_internal): Ditto.
13464 (*movdf_internal): Ditto.
13465 (*movsf_internal): Ditto.
13467 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13469 * config/i386/i386.md (*movti_internal): Set prefix attribute to
13470 maybe_vex for sselog1 and ssemov types.
13471 (*movdi_internal): Reorder operand constraints.
13472 (*movsi_internal): Ditto. Set prefix attribute to
13473 maybe_vex for sselog1 and ssemov types.
13474 (*movtf_internal): Set prefix attribute to maybe_vex
13475 for sselog1 and ssemov types.
13476 (*movdf_internal): Ditto. Set prefix_data16 attribute for
13477 DImode ssemov types. Reorder operand constraints.
13478 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
13479 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
13480 attribute for SImode ssemov types. Reorder operand constraints.
13482 2013-03-20 Martin Jambor <mjambor@suse.cz>
13484 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
13485 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
13487 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
13489 * config/rs6000/predicates.md (indexed_address, update_address_mem
13490 update_indexed_address_mem): New predicates.
13491 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
13492 attribute for load/store instructions.
13493 * config/rs6000/dfp.md (movsd_store): Likewise.
13494 (movsd_load): Likewise.
13495 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
13496 (unnamed HI->DI extend define_insn): Likewise.
13497 (unnamed SI->DI extend define_insn): Likewise.
13498 (unnamed QI->SI extend define_insn): Likewise.
13499 (unnamed QI->HI extend define_insn): Likewise.
13500 (unnamed HI->SI extend define_insn): Likewise.
13501 (unnamed HI->SI extend define_insn): Likewise.
13502 (extendsfdf2_fpr): Likewise.
13503 (movsi_internal1): Likewise.
13504 (movsi_internal1_single): Likewise.
13505 (movhi_internal): Likewise.
13506 (movqi_internal): Likewise.
13507 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
13508 attribute for load/store instructions.
13509 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
13511 (mov<mode>_softfloat): Likewise.
13512 (mov<mode>_hardfloat32): Likewise.
13513 (mov<mode>_hardfloat64): Likewise.
13514 (mov<mode>_softfloat64): Likewise.
13515 (movdi_internal32): Likewise.
13516 (movdi_internal64): Likewise.
13517 (probe_stack_<mode>): Likewise.
13519 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
13521 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
13522 floating point, and decimal floating point to reload iterator.
13524 * config/rs6000/constraints.md (wl constraint): New constraints to
13525 return FLOAT_REGS if certain options are used to reduce the number
13526 of separate patterns that exist in the file.
13527 (wx constraint): Likewise.
13528 (wz constraint): Likewise.
13530 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13531 -mdebug=reg, print wg, wl, wx, and wz constraints.
13532 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13533 Initialize the reload functions for 64-bit binary/decimal floating
13535 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13536 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13537 create the buffer on the stack to overcome not having a 32-bit
13539 (rs6000_emit_move): Likewise.
13540 (rs6000_secondary_memory_needed_rtx): Likewise.
13541 (rs6000_alloc_sdmode_stack_slot): Likewise.
13542 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13543 via xxlxor, just like DFmode 0.0.
13545 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
13546 define as 1 if we are running on a power7 or newer.
13547 (enum r6000_reg_class_enum): Add new constraints.
13549 * config/rs6000/dfp.md (movsd): Delete, combine with binary
13550 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13551 with other moves by using conditional constraits (wg). Use LFIWZX
13552 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
13553 (movsd splitter): Likewise.
13554 (movsd_hardfloat): Likewise.
13555 (movsd_softfloat): Likewise.
13557 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
13558 binary and decimal floating point moves.
13559 (fmove_ok): New attributes to combine binary and decimal floating
13560 point moves, and to combine power6x (mfpgpr) moves along normal
13562 (real_value_to_target): Likewise.
13563 (f32_lr): Likewise.
13564 (f32_lm): Likewise.
13565 (f32_li): Likewise.
13566 (f32_sr): Likewise.
13567 (f32_sm): Likewise.
13568 (f32_si): Likewise.
13569 (movsf): Combine binary and decimal floating point moves. Combine
13570 power6x (mfpgpr) moves with other moves by using conditional
13571 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
13572 (mov<mode> for SFmode/SDmode); Likewise.
13573 (SFmode/SDmode splitters): Likewise.
13574 (movsf_hardfloat): Likewise.
13575 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
13576 (movsf_softfloat): Likewise.
13577 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
13579 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
13580 wx and wz constraints.
13582 * config/rs6000/constraints.md (wg constraint): New constraint to
13583 return FLOAT_REGS if -mmfpgpr (power6x) was used.
13585 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
13588 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
13589 -mdebug=reg, print wg, wl, wx, and wz constraints.
13590 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
13591 Initialize the reload functions for 64-bit binary/decimal floating
13593 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
13594 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
13595 create the buffer on the stack to overcome not having a 32-bit
13597 (rs6000_emit_move): Likewise.
13598 (rs6000_secondary_memory_needed_rtx): Likewise.
13599 (rs6000_alloc_sdmode_stack_slot): Likewise.
13600 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
13601 via xxlxor, just like DFmode 0.0.
13603 * config/rs6000/dfp.md (movdd): Delete, combine with binary
13604 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
13605 with other moves by using conditional constraits (wg). Use LFIWZX
13606 and STFIWX for loading SDmode on power7.
13607 (movdd splitters): Likewise.
13608 (movdd_hardfloat32): Likewise.
13609 (movdd_softfloat32): Likewise.
13610 (movdd_hardfloat64_mfpgpr): Likewise.
13611 (movdd_hardfloat64): Likewise.
13612 (movdd_softfloat64): Likewise.
13614 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
13615 64-bit binary and decimal floating point moves.
13616 (FMOVE64X): Likewise.
13617 (movdf): Combine 64-bit binary and decimal floating point moves.
13618 Combine power6x (mfpgpr) moves with other moves by using
13619 conditional constraits (wg).
13620 (mov<mode> for DFmode/DDmode): Likewise.
13621 (DFmode/DDmode splitters): Likewise.
13622 (movdf_hardfloat32): Likewise.
13623 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
13624 (movdf_softfloat32): Likewise.
13625 (movdf_hardfloat64_mfpgpr): Likewise.
13626 (movdf_hardfloat64): Likewise.
13627 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
13628 (movdf_softfloat64): Likewise.
13629 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
13630 (reload_<mode>_load): Move to later in the file so they aren't in
13631 the middle of the floating point move insns.
13632 (reload_<mode>_store): Likewise.
13634 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
13637 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
13638 constraint if -mdebug=reg.
13639 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
13640 Enable using dd reload support if needed.
13642 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
13643 binary and decimal floating point moves in rs6000.md.
13644 (movtd_internal): Likewise.
13646 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
13647 decimal floating point moves.
13649 (movtf_internal): Likewise.
13650 (mov<mode>_internal, TDmode/TFmode): Likewise.
13651 (movtf_softfloat): Likewise.
13652 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
13654 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
13655 movdi_internal64, using wg constraint for move direct operations.
13656 (movdi_internal64): Likewise.
13658 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
13659 MODES_TIEABLE_P for selected modes. Print the numerical value of
13660 the various virtual registers. Use GPR/FPR first/last values,
13661 instead of hard coding the register numbers. Print which modes
13662 have reload functions registered.
13663 (rs6000_option_override_internal): If -mdebug=reg, trace the options
13664 settings before/after setting cpu, target and subtarget settings.
13665 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
13666 and for secondary reload failures in rs6000_secondary_reload_inner.
13667 (rs6000_secondary_reload_fail): Likewise.
13668 (rs6000_secondary_reload_inner): Likewise.
13670 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
13671 macros for first/last GPR and FPR registers.
13672 (LAST_GPR_REGNO): Likewise.
13673 (FIRST_FPR_REGNO): Likewise.
13674 (LAST_FPR_REGNO): Likewise.
13676 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
13677 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
13678 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
13679 (vcond<mode><mode>): Likewise.
13680 (vcondu<mode><mode>): Likewise.
13681 (vector_gtu<mode>): Likewise.
13682 (vector_gte<mode>): Likewise.
13683 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
13684 to prevent the compiler from converting DImode operations to TImode.
13685 (ior<mode>3): Likewise.
13686 (and<mode>3): Likewise.
13687 (one_cmpl<mode>2): Likewise.
13688 (nor<mode>3): Likewise.
13689 (andc<mode>3): Likewise.
13691 * config/rs6000/constraints.md (wt constraint): New constraint
13692 that returns VSX_REGS if TImode is allowed in VSX registers.
13694 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
13695 constant under VSX.
13697 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
13698 similar to TImode, but it is restricted to being in the GPRs.
13700 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
13701 TImode to occupy a single VSX register.
13703 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
13704 -mvsx-timode for power7/power8.
13705 (power7 cpu): Likewise.
13706 (power8 cpu): Likewise.
13708 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
13709 sure that TFmode/TDmode take up two registers if they are ever
13710 allowed in the upper VSX registers.
13711 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
13713 (rs6000_init_hard_regno_mode_ok): Likewise.
13714 (rs6000_debug_reg_global): Add debugging for PTImode and wt
13715 constraint. Print if LRA is turned on.
13716 (rs6000_option_override_internal): Give an error if -mvsx-timode
13717 and VSX is not enabled.
13718 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
13719 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
13720 to reg+offset addressing. Use PTImode when checking offset
13721 addresses for validity.
13722 (reg_offset_addressing_ok_p): Likewise.
13723 (rs6000_legitimate_offset_address_p): Likewise.
13724 (rs6000_legitimize_address): Likewise.
13725 (rs6000_legitimize_reload_address): Likewise.
13726 (rs6000_legitimate_address_p): Likewise.
13727 (rs6000_eliminate_indexed_memrefs): Likewise.
13728 (rs6000_emit_move): Likewise.
13729 (rs6000_secondary_reload): Likewise.
13730 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
13731 reloads to fpr registers to continue to use reg+offset addressing,
13732 but 64-bit reloads to altivec registers need reg+reg addressing.
13733 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
13734 it. Treat LO_SUM like a PLUS operation.
13735 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
13736 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
13737 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
13738 registers to share a register with a smaller sized type, since VSX
13739 puts scalars in the upper 64-bits.
13740 (print_operand): Add support for PTImode.
13741 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
13742 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
13743 registers, but don't have arithmetic support.
13744 (rs6000_memory_move_cost): Add test for VSX.
13745 (rs6000_opt_masks): Add -mvsx-timode.
13747 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
13750 (VSr): Use wt constraint for TImode.
13751 (VSv): Drop TImode support.
13752 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
13753 (vsx_movti_64bit): Likewise.
13754 (vsx_movti_32bit): Likewise.
13755 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
13756 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
13757 one '?' on the appropriate output constraint. Do not allow TImode
13758 logical operations on 32-bit systems.
13759 (vsx_ior<mode>3): Likewise.
13760 (vsx_xor<mode>3): Likewise.
13761 (vsx_one_cmpl<mode>2): Likewise.
13762 (vsx_nor<mode>3): Likewise.
13763 (vsx_andc<mode>3): Likewise.
13764 (vsx_concat_<mode>): Likewise.
13765 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
13767 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
13768 OPTION_MASK_VSX_TIMODE.
13769 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
13770 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
13772 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
13773 (TI2 iterator): New iterator for TImode, PTImode.
13774 (wd mode attribute): Add values for vector types.
13775 (movti_string): Replace TI move operations with operations for TImode
13776 and PTImode. Add support for TImode being allowed in VSX registers.
13777 (mov<mode>_string, TImode/PTImode): Likewise.
13778 (movti_ppc64): Likewise.
13779 (mov<mode>_ppc64, TImode/PTImode): Likewise.
13780 (TI mode splitters): Likewise.
13782 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
13785 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
13787 PR tree-optimization/56355
13788 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
13789 Also handle integers with undefined overflow.
13791 2013-03-20 Catherine Moore <clm@codesourcery.com>
13792 Maciej W. Rozycki <macro@codesourcery.com>
13793 Tom de Vries <tom@codesourcery.com>
13794 Nathan Sidwell <nathan@codesourcery.com>
13795 Iain Sandoe <iain@codesourcery.com>
13796 Nathan Froyd <froydnj@codesourcery.com>
13797 Chao-ying Fu <fu@mips.com>
13799 * doc/extend.texi: (micromips, nomicromips, nocompression):
13800 Document new function attributes.
13801 * doc/invoke.texi (minterlink-compressed, mmicromips,
13802 m14k, m14ke, m14kec): Document new options.
13803 (minterlink-mips16): Update documentation.
13804 * doc/md.texi (ZC, ZD): Document new constraints.
13805 * configure.ac (gcc_cv_as_micromips): Check if linker
13806 supports the .set micromips directive.
13807 * configure: Regenerate.
13808 * config.in: Regenerate.
13809 * config/mips/mips-tables.opt: Regenerate.
13810 * config/mips/micromips.md: New file.
13811 * constraints.md (ZC, ZD): New constraints.
13812 * config/mips/predicates.md (movep_src_register): New predicate.
13813 (movep_src_operand): New predicate.
13814 (non_volatile_mem_operand): New predicate.
13815 * config/mips/mips.md (multimem): New type.
13816 (length): Differentiate between 17-bit and 18-bit branch offsets.
13817 (MOVEP1, MOVEP2): New mode iterator.
13818 (mov_<load>l): Use ZC constraint.
13819 (mov_<load>r): Likewise.
13820 (mov_<store>l): Likewise.
13821 (mov_<store>r): Likewise.
13822 (*branch_equality<mode>_inverted): Add microMIPS support.
13823 (*branch_equality<mode>): Likewise.
13824 (*jump_absolute): Likewise.
13825 (indirect_jump_<mode>): Likewise.
13826 (tablejump_<mode>): Likewise.
13827 (<optab>_internal): Likewise.
13828 (sibcall_internal): Likewise.
13829 (sibcall_value_internal): Likewise.
13830 (prefetch): Use constraint ZD.
13831 * config/mips/mips.opt (minterlink-compressed): New option.
13832 (minterlink-mips16): Now an alias for minterlink-compressed.
13833 (mmicromips): New option.
13834 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
13835 (compare_and_swap_12): Likewise.
13836 (sync_add<mode>): Likewise.
13837 (sync_<optab>_12): Likewise.
13838 (sync_old_<optab>_12): Likewise.
13839 (sync_new_<optab>_12): Likewise.
13840 (sync_nand_12): Likewise.
13841 (sync_old_nand_12): Likewise.
13842 (sync_new_nand_12): Likewise.
13843 (sync_sub<mode>): Likewise.
13844 (sync_old_add<mode>): Likewise.
13845 (sync_old_sub<mode>): Likewise.
13846 (sync_new_add<mode>): Likewise.
13847 (sync_new_sub<mode>): Likewise.
13848 (sync_<optab><mode>): Likewise.
13849 (sync_old_<optab><mode>): Likewise.
13850 (sync_new_<optab><mode>): Likewise.
13851 (sync_nand<mode>): Likewise.
13852 (sync_old_nand<mode>): Likewise.
13853 (sync_new_nand<mode>): Likewise.
13854 (sync_lock_test_and_set<mode>): Likewise.
13855 (test_and_set_12): Likewise.
13856 (atomic_compare_and_swap<mode>): Likewise.
13857 (atomic_exchange<mode>_llsc): Likewise.
13858 (atomic_fetch_add<mode>_llsc): Likewise.
13859 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
13860 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
13861 (umips_save_restore_pattern_p): Likewise.
13862 (umips_load_store_pair_p): Likewise.
13863 (umips_output_load_store_pair): Likewise.
13864 (umips_movep_target_p): Likewise.
13865 (umips_12bit_offset_address_p): Likewise.
13866 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
13867 (mips_base_mips16): Rename this...
13868 (mips_base_compression_flags): ...to this. Update all uses.
13869 (mips_attribute_table): Add micromips, nomicromips and nocompression.
13870 (mips_mips16_decl_p): Delete.
13871 (mips_nomips16_decl_p): Delete.
13872 (mips_get_compress_on_flags): New function.
13873 (mips_get_compress_off_flags): New function.
13874 (mips_get_compress_mode): New function.
13875 (mips_get_compress_on_name): New function.
13876 (mips_get_compress_off_name): New function.
13877 (mips_insert_attributes): Support multiple compression types.
13878 (mips_merge_decl_attributes): Likewise.
13879 (umips_12bit_offset_address_p): New function.
13880 (mips_start_function_definition): Emit .set micromips directive.
13881 (mips_call_may_need_jalx_p): New function.
13882 (mips_function_ok_for_sibcall): Add microMIPS support.
13883 (mips_print_operand_punctuation): Support short delay slots and
13885 (umips_swm_mask, umips_swm_encoding): New.
13886 (umips_build_save_restore): New function.
13887 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
13888 (was_mips16_p): Remove.
13889 (old_compression_mode): New.
13890 (mips_set_compression_mode): New function.
13891 (mips_set_current_function): Add microMIPS support.
13892 (mips_option_override): Likewise.
13893 (umips_save_restore_pattern_p): New function.
13894 (umips_output_save_restore): New function.
13895 (umips_load_store_pair_p_1): New function.
13896 (umips_load_store_pair_p): New function.
13897 (umips_output_load_store_pair_1): New function.
13898 (umips_output_load_store_pair): New function.
13899 (umips_movep_target_p) New function.
13900 (mips_prepare_pch_save): Add microMIPS support.
13901 * config/mips/mips.h (TARGET_COMPRESSION): New.
13902 (TARGET_CPU_CPP_BUILTINS): Update macro
13903 to use new compression flags and to support microMIPS.
13904 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
13905 (MIPS_ARCH_FLOAT_SPEC): Likewise.
13906 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
13907 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
13908 (ASM_SPEC): Support mmicromips and mno-micromips.
13909 (M16STORE_REG_P): New macro.
13910 (MIPS_CALL): Support TARGET_MICROMIPS.
13911 (MICROMIPS_J): New macro.
13912 (mips_base_mips16): Rename this...
13913 (mips_base_compression_flags): ...to this.
13914 (UMIPS_12BIT_OFFSET_P): New macro.
13915 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
13916 (MULTILIB_DIRNAMES): Likewise.
13917 2013-03-20 Richard Biener <rguenther@suse.de>
13919 PR tree-optimization/56661
13920 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
13921 the result does not have to be distinct.
13923 2013-03-20 Richard Biener <rguenther@suse.de>
13925 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
13928 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13929 Steven Bosscher <steven@gcc.gnu.org>
13931 PR rtl-optimization/56605
13932 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
13934 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
13937 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
13938 that require movd instead of movq.
13940 2013-03-20 Richard Biener <rguenther@suse.de>
13942 * tree-ssa-structalias.c (struct variable_info): Add pointer
13943 to the first field of an aggregate with sub-vars. Make
13944 this and the pointer to the next subfield its ID.
13945 (vi_next): New function.
13946 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
13947 storedanything_id, integer_id): Increment by one.
13948 (new_var_info, get_call_vi, lookup_call_clobber_vi,
13949 get_call_clobber_vi): Adjust.
13950 (solution_set_expand): Simplify and speedup.
13951 (solution_set_add): Inline into ...
13952 (set_union_with_increment): ... this. Adjust accordingly.
13953 (do_sd_constraint): Likewise.
13954 (do_ds_constraint): Likewise.
13955 (do_complex_constraint): Simplify.
13956 (build_pred_graph): Adjust.
13957 (solve_graph): Likewise. Simplify and speedup.
13958 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
13959 get_constraint_for_component_ref, get_constraint_for_1,
13960 first_vi_for_offset, first_or_preceding_vi_for_offset,
13961 create_function_info_for, create_variable_info_for_1,
13962 create_variable_info_for, intra_create_variable_infos): Adjust.
13963 (init_base_vars): Push NULL for ID zero.
13964 (compute_points_to_sets): Adjust.
13966 2013-03-20 Richard Biener <rguenther@suse.de>
13968 * cfgloop.c (verify_loop_structure): Streamline and avoid
13969 ICEing on corrupt loop tree.
13970 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
13973 2013-03-20 Richard Biener <rguenther@suse.de>
13975 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
13976 check whether an SSA update is needed.
13978 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
13980 * config/mips/constraints.md (T): Rename to...
13984 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
13985 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
13987 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13989 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
13990 (*subsi3_carryin_uxtw): Likewise.
13992 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13994 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
13995 (*rorsi3_insn_uxtw): Likewise.
13997 2013-03-19 Ian Bolton <ian.bolton@arm.com>
13999 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14000 (*extrsi5_insn_uxtw): Likewise.
14002 2013-03-19 Richard Biener <rguenther@suse.de>
14004 PR tree-optimization/56273
14005 * passes.c (init_optimization_passes): Move second VRP after DOM.
14007 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14009 * config/i386/i386.md (*movti_internal): Merge from
14010 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14011 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14012 nox64 isa attributes.
14014 2013-03-18 Richard Biener <rguenther@suse.de>
14016 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14018 (merge_node_constraints): Likewise.
14019 (build_succ_graph): Likewise.
14020 (valid_graph_edge): Inline into single caller.
14021 (unify_nodes): Likewise. Use bitmap_set_bit return value
14023 (scc_visit): Fix formatting and variable use.
14024 (do_sd_constraint): Use gcc_checking_assert.
14025 (do_ds_constraint): Likewise.
14026 (do_complex_constraint): Likewise.
14027 (condense_visit): Likewise. Cleanup.
14028 (dump_pred_graph): New function.
14029 (perform_var_substitution): Dump the pred-graph before
14030 variable substitution.
14031 (find_equivalent_node): Use gcc_checking_assert.
14032 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14034 2013-03-18 Richard Biener <rguenther@suse.de>
14036 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14037 Remove cond_expr_stmt_list argument and do not gimplify the
14039 (vect_loop_versioning): Adjust.
14040 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14041 Cleanup to use less temporaries.
14042 (vect_create_data_ref_ptr): Cleanup.
14044 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14046 PR tree-optimization/56635
14047 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14048 require types_compatible_p types.
14050 2013-03-18 Nick Clifton <nickc@redhat.com>
14052 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14053 spurious backslash.
14055 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14056 Add missing line to comment describing function.
14058 2013-03-18 Richard Biener <rguenther@suse.de>
14060 PR tree-optimization/56210
14061 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14062 Handle string / character search functions.
14063 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14065 2013-03-18 Richard Biener <rguenther@suse.de>
14067 PR middle-end/56483
14068 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14069 and implement properly.
14070 * gimple.h (gimple_cond_single_var_p): Remove.
14072 2013-03-18 Richard Biener <rguenther@suse.de>
14074 * tree-data-ref.h (find_data_references_in_loop): Declare.
14075 * tree-data-ref.c (get_references_in_stmt): Use a stack
14076 vector pre-allocated in the callers.
14077 (find_data_references_in_stmt): Adjust.
14078 (graphite_find_data_references_in_stmt): Likewise.
14079 (create_rdg_vertices): Likewise.
14080 (find_data_references_in_loop): Export.
14081 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14082 Compute dependences here...
14083 (vect_analyze_data_refs): ...not here. When we encounter
14084 a non-vectorizable data reference in basic-block vectorization
14085 truncate the data reference vector. Do not bother to
14086 fixup data-dependence information for gather loads.
14087 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14088 of data references, as reported.
14090 2013-03-18 Richard Biener <rguenther@suse.de>
14092 PR tree-optimization/3713
14093 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14094 has_constants and expr.
14095 (stmt_has_constants): Properly valueize SSA names when deciding
14096 whether the stmt has constants.
14098 2013-03-18 Richard Biener <rguenther@suse.de>
14100 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14101 whole function when there is nothing to do.
14102 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14103 * tree-vectorizer.c (vectorize_loops): Update virtual and
14104 loop-closed SSA once.
14105 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14107 2013-03-18 Richard Biener <rguenther@suse.de>
14109 PR middle-end/56113
14110 * domwalk.c (bb_postorder): New global static.
14111 (cmp_bb_postorder): New function.
14112 (walk_dominator_tree): Replace scheme imposing an order for
14113 visiting dominator sons by one sorting them at the time they
14114 are pushed on the stack.
14116 2013-03-18 Richard Biener <rguenther@suse.de>
14118 PR tree-optimization/39326
14119 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14120 (struct mem_ref): Replace mem member with ao_ref typed member.
14121 (MEM_ANALYZABLE): Adjust.
14122 (memref_eq): Likewise.
14123 (mem_ref_alloc): Likewise.
14124 (gather_mem_refs_stmt): Likewise.
14125 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14126 (execute_sm_if_changed_flag_set): Adjust.
14127 (execute_sm): Likewise.
14128 (ref_always_accessed_p): Likewise.
14129 (refs_independent_p): Likewise.
14130 (can_sm_ref_p): Likewise.
14132 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14135 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14136 return 1 even for !unsignedp.
14138 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14140 * config/i386/i386.md (isa): Add x64 and nox64.
14141 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14142 (*pushtf): Enable *roF alternative for x64 isa only.
14143 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14144 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14145 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14146 (*movtf_internal): Merge from *movtf_internal_rex64 and
14147 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14148 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14149 nox64 isa attributes.
14150 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14151 nox64 isa attributes.
14152 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14154 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14156 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14158 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14161 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14162 class names. Remove trailing comma after "ALL_REGS".
14164 2013-03-16 Jan Hubicka <jh@suse.cz>
14166 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14167 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14168 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14169 of cgraph_get_create_node.
14170 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14172 2013-03-16 Jason Merrill <jason@redhat.com>
14175 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14176 with DW_AT_default_value.
14178 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14180 * BASE-VER: Set to 4.9.0.
14182 2013-03-14 Andi Kleen <ak@linux.intel.com>
14185 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14186 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14187 Document _x* TSX intrinsics.
14189 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14190 David Holsgrove <david.holsgrove@xilinx.com>
14192 * configure.ac: Add MicroBlaze TLS support detection.
14193 * configure: Regenerate.
14194 * config/microblaze/microblaze-protos.h
14195 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14196 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14197 * config/microblaze/microblaze.c (microblaze_address_type): Add
14198 ADDRESS_TLS and tls_reloc address types.
14199 (microblaze_address_info): Add tls_reloc.
14200 (TARGET_HAVE_TLS): Define.
14201 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14202 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14203 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14204 load_tls_operand, microblaze_call_tls_get_addr,
14205 microblaze_legitimize_tls_address): New functions.
14206 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14207 (get_base_reg): Use microblaze_tls_symbol_p.
14208 (microblaze_classify_address): Handle TLS.
14209 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14210 label_mentioned_p and microblaze_tls_referenced_p.
14211 (microblaze_legitimize_address): Handle TLS.
14212 (microblaze_address_insns): Handle ADDRESS_TLS.
14213 (pic_address_needs_scratch): Handle TLS.
14214 (print_operand_address): Handle TLS.
14215 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14216 (microblaze_expand_move): Handle TLS.
14217 (microblaze_legitimate_constant_p): Check
14218 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14219 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14220 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14221 (PIC_OFFSET_TABLE_REGNUM): Set.
14222 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14223 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14224 (addsi3, movsi_internal2, movdf_internal): Update constraints
14225 * config/microblaze/predicates.md (arith_plus_operand): Define
14226 (move_operand): Redefine as move_src_operand,
14227 check microblaze_tls_referenced_p.
14229 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14231 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14232 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14234 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14236 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14239 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14241 PR tree-optimization/53265
14242 * common.opt (Waggressive-loop-optimizations): New option.
14243 * tree-ssa-loop-niter.c: Include tree-pass.h.
14244 (do_warn_aggressive_loop_optimizations): New function.
14245 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14246 if number_of_latch_executions returned constant.
14247 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14248 early. If number_of_latch_executions returned constant, set
14249 nb_iterations_upper_bound back to it.
14250 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14252 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14253 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14255 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14256 (MULTILIB_OSDIRNAMES): Set.
14257 * genmultilib: If defaultosdirname doesn't start with :: , set
14258 defaultosdirname2 instead, clear it and emit two . multilib_raw
14259 entries instead of just one.
14261 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14263 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14264 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14265 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14266 (SUBTARGET_OVERRIDE_OPTIONS): New.
14268 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14271 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14272 (musermode): Convert to Var(TARGET_USERMODE).
14273 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14274 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14275 * config/sh/sh.c (sh_option_override): Use
14276 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14277 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14279 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14281 (udivsi3_i4_single, divsi3_i4_single): Use
14282 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14284 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14286 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14289 2013-03-13 Richard Biener <rguenther@suse.de>
14291 PR tree-optimization/56608
14292 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14293 calls when vectorizing basic-blocks.
14295 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14298 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14301 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14303 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14305 2013-03-11 Jan Hubicka <jh@suse.cz>
14308 * lto-streamer-out.c (output_symbol_p): Skip references from
14309 constructors of external variables.
14311 2013-03-11 Jan Hubicka <jh@suse.cz>
14313 PR middle-end/56571
14314 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14316 * emit-rtl.c (verify_rtx_sharing): Likewise.
14317 (copy_insn_1): Likewise.
14318 * rtl.c (copy_rtx): Likewise.
14320 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14323 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14324 output_operand_lossage message.
14326 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14329 * arm.c (shift_op): Validate RTL pattern on the fly.
14330 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14333 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14336 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14337 new call patterns that clobber %fr12.
14338 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14339 split and postreload patterns.
14340 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14341 registers %fr12 and %fr12R as call used.
14343 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14345 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14346 (canon_address, record_store, replace_read, check_mem_read_rtx,
14347 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14348 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14349 rest_of_handle_dse): Likewise.
14351 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14353 PR middle-end/56524
14354 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14356 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14357 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14358 (save_optabs_if_changed): Replace with...
14359 (init_tree_optimization_optabs): ...this.
14360 * optabs.c (save_optabs_if_changed): Rename to...
14361 (init_tree_optimization_optabs): ...this. Take the optimization node
14362 as argument. Do nothing if the base optabs are already correct.
14363 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14364 to recompute optabs.
14365 * function.h (function): Remove optabs field.
14366 * function.c (invoke_set_current_function_hook): Call
14367 init_tree_optimization_optabs. Use the result to initialize
14370 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14372 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14373 if GTMA_HAS_NO_INSTRUMENTATION.
14374 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14375 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14376 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
14377 * gimple-pretty-print.c (dump_gimple_transaction): Handle
14378 GTMA_HAS_NO_INSTRUMENTATION.
14380 2013-03-08 Jakub Jelinek <jakub@redhat.com>
14382 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
14385 2013-03-08 Marek Polacek <polacek@redhat.com>
14386 Jakub Jelinek <jakub@redhat.com>
14388 PR tree-optimization/56478
14389 * predict.c (is_comparison_with_loop_invariant_p): Change the
14390 type of loop_step to tree.
14391 (predict_loops): Adjust.
14392 (predict_iv_comparison): Perform the computations on double_ints.
14394 2013-03-08 Richard Biener <rguenther@suse.de>
14396 PR tree-optimization/56570
14397 * tree-cfg.c (verify_expr_location_1): Verify locations for
14399 * tree-sra.c (create_access_replacement): Strip locations
14400 from DECL_DEBUG_EXPRs.
14402 2013-03-08 Richard Biener <rguenther@suse.de>
14404 * tree-inline.c (expand_call_inline): Do not associate
14405 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
14406 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
14408 2013-03-08 Richard Biener <rguenther@suse.de>
14410 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
14411 or block changes with -Og. Fix for location / block encoding
14412 changes and PHI arguments with locations.
14414 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
14416 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
14418 (struct output_info): Likewise.
14419 (register_overhead): Remove bad gcc_assert.
14420 (bitmap_find_bit): If there is only a single bitmap element, do not
14421 count a miss as a search.
14422 (print_statistics): Update for counter type changes.
14423 (dump_bitmap_statistics): Likewise. Print headers such that they
14424 are properly lined up with the printed counters.
14426 2013-03-07 Jakub Jelinek <jakub@redhat.com>
14428 PR tree-optimization/56559
14429 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
14430 check that it has only a single use.
14432 2013-03-07 Richard Biener <rguenther@suse.de>
14434 * doc/invoke.texi (fwhole-program): Discourage use in combination
14437 2013-03-06 Jakub Jelinek <jakub@redhat.com>
14439 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
14441 PR tree-optimization/56539
14442 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
14443 instead of GSI_CONTINUE_LINKING as last argument to
14444 force_gimple_operand_gsi. Adjust function comment.
14446 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
14449 PR middle-end/56548
14450 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
14451 promoted mode, convert the result back to the original mode.
14453 2013-03-06 Richard Biener <rguenther@suse.de>
14455 PR middle-end/56294
14456 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
14457 (insert_updated_phi_nodes_compare_uids): New function.
14458 (update_ssa): Sort symbols_to_rename after UID before
14459 traversing it to insert PHI nodes.
14461 2013-03-06 Richard Biener <rguenther@suse.de>
14463 PR middle-end/50494
14464 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
14465 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
14468 2013-02-13 Richard Biener <rguenther@suse.de>
14471 * varasm.c (output_constant_def_1): Get the decl representing
14472 the constant as argument.
14473 (output_constant_def): Wrap output_constant_def_1.
14474 (make_decl_rtl): Use output_constant_def_1 with the decl
14475 representing the constant.
14476 (build_constant_desc): Optionally re-use a decl already
14477 representing the constant.
14478 (tree_output_constant_def): Adjust.
14480 2013-03-06 Joey Ye <joey.ye@arm.com>
14483 * gcc.c (convert_white_space): New function.
14484 (main): Handles white space in function name.
14486 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
14489 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
14490 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
14491 to SH_DIV_CALL_TABLE for TARGET_SH2.
14492 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
14494 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
14495 call-table options.
14497 2013-03-05 Sterling Augustine <saugustine@google.com>
14498 Cary Coutant <ccoutant@google.com>
14501 * dwarf2out.c (resolve_addr): Don't call
14502 remove_loc_list_addr_table_entries a second time for the same
14505 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14508 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
14509 (avoid_complex_debug_insns): New function.
14510 (expand_debug_locations): Call it.
14512 PR rtl-optimization/56484
14513 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
14514 lifetimes of hard registers on small register class machines.
14516 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14518 * config/microblaze/microblaze-protos.h: Rename
14519 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
14520 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
14522 (microblaze_fast_interrupt_function_p): New function.
14523 (microblaze_is_interrupt_handler): Rename to
14524 microblaze_is_interrupt_variant and add fast_interrupt check.
14525 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
14526 (save_restore_insns): Likewise.
14527 (compute_frame_size): Likewise.
14528 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
14529 (microblaze_globalize_label): Likewise.
14530 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
14531 * config/microblaze/microblaze.md: Use wrapper
14532 microblaze_is_interrupt_variant.
14534 2013-03-05 Kai Tietz <ktietz@redhat.com>
14536 * sdbout.c (sdbout_one_type): Switch to current function's section
14537 supporting cold/hot.
14539 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
14541 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
14544 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14546 PR middle-end/56461
14547 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
14548 if VALGRIND_GET_VBITS is defined, temporarily make object
14549 memory all defined, and restore previous valgrind addressability
14550 and definability afterwards. Free this_object at the end.
14552 PR middle-end/56461
14553 * lra.c (lra): Call lra_clear_live_ranges if live_p,
14554 right before calling lra_create_live_ranges, also call it
14555 when clearing live_p. Only call lra_clear_live_ranges
14556 at the end if live_p.
14558 PR middle-end/56461
14559 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
14561 2013-03-05 Richard Biener <rguenther@suse.de>
14563 PR tree-optimization/56521
14564 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
14567 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
14570 * except.h (remove_unreachable_eh_regions): New prototype.
14571 * except.c (remove_eh_handler_splicer): New function, split out
14572 of remove_eh_handler.
14573 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
14574 warning about running it on many EH regions one at a time.
14575 (remove_unreachable_eh_regions_worker): New function, walk the
14576 EH tree in depth-first order and remove non-marked regions.
14577 (remove_unreachable_eh_regions): New function.
14578 * tree-eh.c (mark_reachable_handlers): New function, split out
14579 from remove_unreachable_handlers.
14580 (remove_unreachable_handlers): Use mark_reachable_handlers and
14581 remove_unreachable_eh_regions.
14582 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
14583 and remove_unreachable_eh_regions.
14585 2013-03-05 Richard Biener <rguenther@suse.de>
14587 PR middle-end/56525
14588 * loop-init.c (fix_loop_structure): Remove loops in two stages,
14589 not freeing them until the end.
14591 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14593 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
14595 2013-03-05 Richard Biener <rguenther@suse.de>
14597 PR tree-optimization/56270
14598 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
14599 of loads after scheduling an SLP instance.
14601 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14603 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
14605 (check_gcc_parallelize): Run guality.exp as a separate job from
14606 vect.exp with unsorted.exp and $(dg_target_exps) separately from
14607 struct-layout-1.exp with stackalign.exp.
14609 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
14611 PR middle-end/56461
14612 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
14613 load_index sbitmap even if some bit in it isn't set.
14615 PR middle-end/56461
14616 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
14617 (discover_iteration_bound_by_body_walk): Change queues to
14618 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
14619 spelling in comment. Call safe_push on queues[bound_index] directly.
14620 Release queues[queue_index] in every iteration unconditionally.
14621 Release bounds vector.
14623 PR middle-end/56461
14624 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
14625 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
14626 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
14629 2013-03-05 Richard Biener <rguenther@suse.de>
14632 * tree-inline.c (remap_blocks_to_null): New function.
14633 (expand_call_inline): When expanding a call stmt without
14634 an associated block inline remap all callee blocks to NULL.
14636 2013-03-05 Jakub Jelinek <jakub@redhat.com>
14638 PR rtl-optimization/56494
14639 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
14640 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
14641 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
14643 PR middle-end/56461
14644 * sel-sched-ir.c (free_sched_pools): Release
14645 succs_info_pool.stack[succs_info_pool.max_top] vectors too
14646 if succs_info_pool.max_top isn't -1.
14649 * opts.c (opts_obstack, opts_concat): Moved to...
14650 * opts-common.c (opts_obstack, opts_concat): ... here.
14652 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14654 PR middle-end/56461
14655 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
14657 2013-03-04 Martin Jambor <mjambor@suse.cz>
14659 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
14660 all appropriate places.
14662 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
14664 PR tree-optimization/56424
14665 * ipa-split.c (split_function): Do not set the RSO flag if result is
14666 not by reference and its type is a register type.
14668 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
14670 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
14671 (microblaze_legitimate_pic_operand): Likewise
14672 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
14673 new function microblaze_legitimate_pic_operand
14674 * config/microblaze/microblaze-protos.h
14675 (microblaze_legitimate_pic_operand): Declare.
14677 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14679 * config/microblaze/predicates.md (call_insn_simple_operand):
14680 New predicate for supported rtx code types.
14681 * config/microblaze/microblaze.md (call_internal1): Use
14682 call_insn_simple_operand predicate.
14684 2013-03-04 Jakub Jelinek <jakub@redhat.com>
14686 PR middle-end/56461
14687 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
14688 partitions.ordered_remove.
14690 PR middle-end/56461
14691 * tree-vect-stmts.c (vectorizable_conversion): Don't call
14692 vec_oprnds0.create (1) for modifier == NONE.
14694 PR middle-end/56461
14695 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
14696 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
14697 vec_oprnds1 right before pushing anything to it for
14700 PR middle-end/56461
14701 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
14702 set nbbs to 0 instead of having separate code path.
14703 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
14704 instead of false as last argument if returning NULL.
14706 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
14708 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
14709 the attribute is now called "target" instead of "option".
14710 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
14711 * doc/tm.texi.in (Target Attributes): Likewise document the correct
14712 attribute/pragma name for TARGET_OPTION_VALID_P and
14713 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
14714 * doc/tm.texi: Regenerated.
14716 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
14718 * config/microblaze/microblaze.c:
14719 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
14720 * config/microblaze/microblaze.h: Add -mxl-reorder to
14722 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
14723 instructions emitted if TARGET_REORDER.
14724 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
14725 or 0 for -m/-mno case, but initialises as 2 to detect default use case
14728 2013-03-01 Xinliang David Li <davidxl@google.com>
14730 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
14733 2013-03-01 Jakub Jelinek <jakub@redhat.com>
14735 PR middle-end/56461
14736 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
14737 vector even when returning true. Fix up function comment formatting.
14739 PR middle-end/56461
14740 * ira-build.c (ira_loop_nodes_count): New variable.
14741 (create_loop_tree_nodes): Initialize it.
14742 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
14744 PR middle-end/56461
14745 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
14746 method on dr_chain and result_chain.
14747 * tree-vect-stmts.c (vectorizable_store): Only call
14748 result_chain.create if j == 0.
14750 PR middle-end/56461
14751 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
14752 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
14753 before overwriting it.
14755 2013-03-01 Tobias Burnus <burnus@net-b.de>
14757 * doc/extended.texi (C Extensions): Change order in @menu
14759 (Other MIPS Built-in Functions): Move last MIPS entry before
14760 "picoChip Built-in Functions".
14761 (SH Built-in Functions): Move after RX Built-in Functions.
14762 * doc/gcc.texi (Introduction): Change order in @menu
14764 * doc/md.texi (Constraints): Ditto.
14765 * gty.texi (Type Information): Ditto.
14766 (User-provided marking routines for template types): Make
14768 * doc/invoke.texi (AArch64 Options): Move before
14769 "Adapteva Epiphany Options".
14771 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
14772 Jakub Jelinek <jakub@redhat.com>
14775 * asan.c (gate_asan): Lookup no_sanitize_address instead of
14776 no_address_safety_analysis attribute.
14777 * doc/extend.texi (no_address_safety_attribute): Rename to
14778 no_sanitize_address attribute, mention no_address_safety_analysis
14779 attribute as deprecated alias.
14781 2013-02-28 Jakub Jelinek <jakub@redhat.com>
14783 PR middle-end/56461
14784 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
14785 type to vec<vec<tree> > *.
14786 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
14787 to be vec<tree> instead of vec<tree> *, set vec_defs
14788 to vNULL and call vec_defs.create (number_of_vects), adjust other
14790 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
14791 vectorizable_condition): Adjust vect_get_slp_defs callers.
14793 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14795 * config/aarch64/aarch64.c
14796 (aarch64_float_const_representable): Remove unused variable.
14798 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14800 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
14802 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14804 * config/aarch64/aarch64-builtins.c
14805 (aarch64_init_simd_builtins): Make static.
14807 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
14809 * config/aarch64/aarch64.c
14810 (aarch64_simd_make_constant): Make static.
14812 2013-02-28 Martin Jambor <mjambor@suse.cz>
14814 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
14815 with no initialization to the RHS of debug statements.
14817 2013-02-28 Martin Jambor <mjambor@suse.cz>
14819 PR tree-optimization/56294
14820 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
14822 (get_access_replacement): Do not call create_access_replacement.
14823 Assert a replacement exists.
14824 (get_repl_default_def_ssa_name): Create the replacement declaration
14827 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14829 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
14830 final_end_function.
14832 2013-02-28 Marek Polacek <polacek@redhat.com>
14834 PR rtl-optimization/56466
14835 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
14836 if we're changing a loop.
14837 (peel_loops_completely): Likewise.
14839 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
14842 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
14844 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14847 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
14848 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
14849 INTX_FTYPE_FX, FX_FTYPE_INTX.
14850 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14852 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14854 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
14855 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
14856 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
14857 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
14858 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
14859 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
14860 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
14861 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
14862 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
14863 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
14864 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
14865 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
14866 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
14867 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
14868 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
14869 (avrxmega6): Increase max flash segments from 5 to 6.
14870 * config/avr/t-multilib: Regenerate.
14871 * config/avr/avr-tables.opt: Regenerate.
14872 * doc/avr-mmcu.texi: Regenerate.
14874 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
14876 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
14877 (avr_device_to_arch): Rename to avr_device_to_ld.
14878 (avr_device_to_as): New prototype.
14879 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
14880 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
14881 * config/avr/driver-avr.c (avr_device_to_as): New.
14882 (avr_device_to_arch): Rename to avr_device_to_ld.
14884 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14886 PR middle-end/56461
14887 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
14888 method on dr_chain and result_chain.
14890 PR middle-end/56461
14891 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
14892 pointer_set_destroy on not_executed_last_iteration.
14894 PR middle-end/56461
14895 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
14897 PR middle-end/56461
14898 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
14899 FOR_EACH_DEFINED_FUNCTION when freeing state.
14901 PR middle-end/56461
14902 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
14904 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
14907 PR middle-end/56461
14908 * ipa-cp.c (decide_whether_version_node): Call vec_free on
14909 known_aggs[i].items and release known_aggs vector.
14911 PR middle-end/56461
14912 * ipa-reference.c (propagate): Free node_info even for alias nodes.
14914 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
14916 * config/microblaze/microblaze.c (microblaze_emit_compare):
14917 Use xor for EQ/NE comparisions.
14918 * config/microblaze/microblaze.md (cstoresf4): Add constraints
14919 (cbranchsf4): Adjust operator to comparison_operator.
14921 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14923 PR middle-end/56461
14924 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
14926 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
14927 vec_safe_push, always update *slot.
14928 (redirect_edge_var_map_clear): Use vec_free.
14929 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
14930 (free_var_map_entry): Use vec_free.
14931 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
14932 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
14934 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
14936 PR middle-end/45472
14937 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
14938 when the may_trap_p bit of the exprs being merged differs.
14939 Reorder tests for speculativeness in the logical and operator.
14941 2013-02-27 Jakub Jelinek <jakub@redhat.com>
14943 * incpath.c (add_standard_paths): Use reconcat instead of concat
14944 where appropriate and avoid leaking memory.
14946 * opts.h: Include obstack.h.
14947 (opts_concat): New prototype.
14948 (opts_obstack): New declaration.
14949 * opts.c (opts_concat): New function.
14950 (opts_obstack): New variable.
14951 (init_options_struct): Call gcc_init_obstack on opts_obstack.
14952 (finish_options): Use opts_concat instead of concat
14953 and XOBNEWVEC instead of XNEWVEC.
14954 * opts-common.c (generate_canonical_option, decode_cmdline_option,
14955 generate_option): Likewise.
14956 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
14957 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
14960 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
14961 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
14963 2013-02-26 Jakub Jelinek <jakub@redhat.com>
14965 PR middle-end/56461
14966 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
14968 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14970 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
14971 (arm_block_move_unaligned_straight): Likewise.
14972 (arm_adjust_block_mem): Likewise.
14974 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
14977 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
14978 temp, cond and label.
14979 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
14982 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
14983 * config/c6x/c6x.h (dbx_register_map): Update declaration.
14986 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
14987 of prologue/epilogue functions.
14990 * config/tilegx/tilegx.c (tilegx_expand_prologue):
14991 Remove unused variable cfa_offset.
14992 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
14995 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
14996 type promotion to unsigned.
14999 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15000 for HOST_WIDE_INT of 32 bit / same size as int.
15001 (arm_block_move_unaligned_straight): Likewise.
15002 (arm_adjust_block_mem): Likewise.
15005 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15008 2013-02-26 Marek Polacek <polacek@redhat.com>
15010 PR tree-optimization/56426
15011 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15013 2013-02-26 Richard Biener <rguenther@suse.de>
15016 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15017 unused variable loops.
15019 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15021 PR tree-optimization/56448
15022 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15023 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15024 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15025 later operands of the references, or even first operand for
15026 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15028 PR tree-optimization/56443
15029 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15030 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15031 to type_for_mode langhook.
15033 2013-02-25 Matt Turner <mattst88@gmail.com>
15035 * doc/invoke.texi: Document r4700.
15037 2013-02-25 Richard Biener <rguenther@suse.de>
15039 PR tree-optimization/56175
15040 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15042 (simplify_bitwise_binary): ... here. Also guard the conversion
15043 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15045 2013-02-25 Catherine Moore <clm@codesourcery.com>
15048 2013-02-24 Catherine Moore <clm@codesourcery.com>
15049 Maciej W. Rozycki <macro@codesourcery.com>
15050 Tom de Vries <tom@codesourcery.com>
15051 Nathan Sidwell <nathan@codesourcery.com>
15052 Iain Sandoe <iain@codesourcery.com>
15053 Nathan Froyd <froydnj@codesourcery.com>
15054 Chao-ying Fu <fu@mips.com>
15056 * doc/extend.texi: (micromips, nomicromips, nocompression):
15057 Document new function attributes.
15058 * doc/invoke.texi (minterlink-compressed, mmicromips,
15059 m14k, m14ke, m14kec): Document new options.
15060 (minterlink-mips16): Update documentation.
15061 * doc/md.texi (ZC, ZD): Document new constraints.
15062 * configure.ac (gcc_cv_as_micromips): Check if linker
15063 supports the .set micromips directive.
15064 * configure: Regenerate.
15065 * config.in: Regenerate.
15066 * config/mips/mips-tables.opt: Regenerate.
15067 * config/mips/micromips.md: New file.
15068 * constraints.md (ZC, AD): New constraints.
15069 * config/mips/predicates.md (movep_src_register): New predicate.
15070 (movep_src_operand): New predicate.
15071 (non_volatile_mem_operand): New predicate.
15072 * config/mips/mips.md (multimem): New type.
15073 (length): Differentiate between 17-bit and 18-bit branch offsets.
15074 (MOVEP1, MOVEP2): New mode iterator.
15075 (mov_<load>l): Use ZC constraint.
15076 (mov_<load>r): Likewise.
15077 (mov_<store>l): Likewise.
15078 (mov_<store>r): Likewise.
15079 (*branch_equality<mode>_inverted): Add microMIPS support.
15080 (*branch_equality<mode>): Likewise.
15081 (*jump_absolute): Likewise.
15082 (indirect_jump_<mode>): Likewise.
15083 (tablejump_<mode>): Likewise.
15084 (<optab>_internal): Likewise.
15085 (sibcall_internal): Likewise.
15086 (sibcall_value_internal): Likewise.
15087 (prefetch): Use constraint ZD.
15088 * config/mips/mips.opt (minterlink-compressed): New option.
15089 (minterlink-mips16): Now an alias for minterlink-compressed.
15090 (mmicromips): New option.
15091 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15092 (compare_and_swap_12): Likewise.
15093 (sync_add<mode>): Likewise.
15094 (sync_<optab>_12): Likewise.
15095 (sync_old_<optab>_12): Likewise.
15096 (sync_new_<optab>_12): Likewise.
15097 (sync_nand_12): Likewise.
15098 (sync_old_nand_12): Likewise.
15099 (sync_new_nand_12): Likewise.
15100 (sync_sub<mode>): Likewise.
15101 (sync_old_add<mode>): Likewise.
15102 (sync_old_sub<mode>): Likewise.
15103 (sync_new_add<mode>): Likewise.
15104 (sync_new_sub<mode>): Likewise.
15105 (sync_<optab><mode>): Likewise.
15106 (sync_old_<optab><mode>): Likewise.
15107 (sync_new_<optab><mode>): Likewise.
15108 (sync_nand<mode>): Likewise.
15109 (sync_old_nand<mode>): Likewise.
15110 (sync_new_nand<mode>): Likewise.
15111 (sync_lock_test_and_set<mode>): Likewise.
15112 (test_and_set_12): Likewise.
15113 (atomic_compare_and_swap<mode>): Likewise.
15114 (atomic_exchange<mode>_llsc): Likewise.
15115 (atomic_fetch_add<mode>_llsc): Likewise.
15116 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15117 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15118 (umips_save_restore_pattern_p): Likewise.
15119 (umips_load_store_pair_p): Likewise.
15120 (umips_output_load_store_pair): Likewise.
15121 (umips_movep_target_p): Likewise.
15122 (umips_12bit_offset_address_p): Likewise.
15123 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15124 (mips_base_mips16): Rename this...
15125 (mips_base_compression_flags): ...to this. Update all uses.
15126 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15127 (mips_mips16_decl_p): Delete.
15128 (mips_nomips16_decl_p): Delete.
15129 (mips_get_compress_on_flags): New function.
15130 (mips_get_compress_off_flags): New function.
15131 (mips_get_compress_mode): New function.
15132 (mips_get_compress_on_name): New function.
15133 (mips_get_compress_off_name): New function.
15134 (mips_insert_attributes): Support multiple compression types.
15135 (mips_merge_decl_attributes): Likewise.
15136 (umips_12bit_offset_address_p): New function.
15137 (mips_start_function_definition): Emit .set micromips directive.
15138 (mips_call_may_need_jalx_p): New function.
15139 (mips_function_ok_for_sibcall): Add microMIPS support.
15140 (mips_print_operand_punctuation): Support short delay slots and
15142 (umips_swm_mask, umips_swm_encoding): New.
15143 (umips_build_save_restore): New function.
15144 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15145 (was_mips16_p): Remove.
15146 (old_compression_mode): New.
15147 (mips_set_compression_mode): New function.
15148 (mips_set_current_function): Add microMIPS support.
15149 (mips_option_override): Likewise.
15150 (umips_save_restore_pattern_p): New function.
15151 (umips_output_save_restore): New function.
15152 (umips_load_store_pair_p_1): New function.
15153 (umips_load_store_pair_p): New function.
15154 (umips_output_load_store_pair_1): New function.
15155 (umips_output_load_store_pair): New function.
15156 (umips_movep_target_p) New function.
15157 (mips_prepare_pch_save): Add microMIPS support.
15158 * config/mips/mips.h (TARGET_COMPRESSION): New.
15159 (TARGET_CPU_CPP_BUILTINS): Update macro
15160 to use new compression flags and to support microMIPS.
15161 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15162 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15163 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15164 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15165 (ASM_SPEC): Support mmicromips and mno-micromips.
15166 (M16STORE_REG_P): New macro.
15167 (MIPS_CALL): Support TARGET_MICROMIPS.
15168 (MICROMIPS_J): New macro.
15169 (mips_base_mips16): Rename this...
15170 (mips_base_compression_flags): ...to this.
15171 (UMIPS_12BIT_OFFSET_P): New macro.
15172 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15173 (MULTILIB_DIRNAMES): Likewise.
15175 2013-02-25 Tom de Vries <tom@codesourcery.com>
15177 PR rtl-optimization/56131
15178 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15179 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15180 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15182 2013-02-25 Tobias Burnus <burnus@net-b.de>
15184 * doc/invoke.texi (-fsanitize=): Move from optimization
15185 to debugging options.
15187 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15189 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15191 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15192 Alexander Monakov <amonakov@ispras.ru>
15194 PR middle-end/56077
15195 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15196 flush pending lists also on non-jumps. Adjust comment.
15198 2013-02-24 Catherine Moore <clm@codesourcery.com>
15199 Maciej W. Rozycki <macro@codesourcery.com>
15200 Tom de Vries <tom@codesourcery.com>
15201 Nathan Sidwell <nathan@codesourcery.com>
15202 Iain Sandoe <iain@codesourcery.com>
15203 Nathan Froyd <froydnj@codesourcery.com>
15204 Chao-ying Fu <fu@mips.com>
15206 * doc/extend.texi: (micromips, nomicromips, nocompression):
15207 Document new function attributes.
15208 * doc/invoke.texi (minterlink-compressed, mmicromips,
15209 m14k, m14ke, m14kec): Document new options.
15210 (minterlink-mips16): Update documentation.
15211 * doc/md.texi (ZC, ZD): Document new constraints.
15212 * configure.ac (gcc_cv_as_micromips): Check if linker
15213 supports the .set micromips directive.
15214 * configure: Regenerate.
15215 * config.in: Regenerate.
15216 * config/mips/mips-tables.opt: Regenerate.
15217 * config/mips/micromips.md: New file.
15218 * constraints.md (ZC, AD): New constraints.
15219 * config/mips/predicates.md (movep_src_register): New predicate.
15220 (movep_src_operand): New predicate.
15221 (non_volatile_mem_operand): New predicate.
15222 * config/mips/mips.md (multimem): New type.
15223 (length): Differentiate between 17-bit and 18-bit branch offsets.
15224 (MOVEP1, MOVEP2): New mode iterator.
15225 (mov_<load>l): Use ZC constraint.
15226 (mov_<load>r): Likewise.
15227 (mov_<store>l): Likewise.
15228 (mov_<store>r): Likewise.
15229 (*branch_equality<mode>_inverted): Add microMIPS support.
15230 (*branch_equality<mode>): Likewise.
15231 (*jump_absolute): Likewise.
15232 (indirect_jump_<mode>): Likewise.
15233 (tablejump_<mode>): Likewise.
15234 (<optab>_internal): Likewise.
15235 (sibcall_internal): Likewise.
15236 (sibcall_value_internal): Likewise.
15237 (prefetch): Use constraint ZD.
15238 * config/mips/mips.opt (minterlink-compressed): New option.
15239 (minterlink-mips16): Now an alias for minterlink-compressed.
15240 (mmicromips): New option.
15241 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15242 (compare_and_swap_12): Likewise.
15243 (sync_add<mode>): Likewise.
15244 (sync_<optab>_12): Likewise.
15245 (sync_old_<optab>_12): Likewise.
15246 (sync_new_<optab>_12): Likewise.
15247 (sync_nand_12): Likewise.
15248 (sync_old_nand_12): Likewise.
15249 (sync_new_nand_12): Likewise.
15250 (sync_sub<mode>): Likewise.
15251 (sync_old_add<mode>): Likewise.
15252 (sync_old_sub<mode>): Likewise.
15253 (sync_new_add<mode>): Likewise.
15254 (sync_new_sub<mode>): Likewise.
15255 (sync_<optab><mode>): Likewise.
15256 (sync_old_<optab><mode>): Likewise.
15257 (sync_new_<optab><mode>): Likewise.
15258 (sync_nand<mode>): Likewise.
15259 (sync_old_nand<mode>): Likewise.
15260 (sync_new_nand<mode>): Likewise.
15261 (sync_lock_test_and_set<mode>): Likewise.
15262 (test_and_set_12): Likewise.
15263 (atomic_compare_and_swap<mode>): Likewise.
15264 (atomic_exchange<mode>_llsc): Likewise.
15265 (atomic_fetch_add<mode>_llsc): Likewise.
15266 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15267 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15268 (umips_save_restore_pattern_p): Likewise.
15269 (umips_load_store_pair_p): Likewise.
15270 (umips_output_load_store_pair): Likewise.
15271 (umips_movep_target_p): Likewise.
15272 (umips_12bit_offset_address_p): Likewise.
15273 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15274 (mips_base_mips16): Rename this...
15275 (mips_base_compression_flags): ...to this. Update all uses.
15276 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15277 (mips_mips16_decl_p): Delete.
15278 (mips_nomips16_decl_p): Delete.
15279 (mips_get_compress_on_flags): New function.
15280 (mips_get_compress_off_flags): New function.
15281 (mips_get_compress_mode): New function.
15282 (mips_get_compress_on_name): New function.
15283 (mips_get_compress_off_name): New function.
15284 (mips_insert_attributes): Support multiple compression types.
15285 (mips_merge_decl_attributes): Likewise.
15286 (umips_12bit_offset_address_p): New function.
15287 (mips_start_function_definition): Emit .set micromips directive.
15288 (mips_call_may_need_jalx_p): New function.
15289 (mips_function_ok_for_sibcall): Add microMIPS support.
15290 (mips_print_operand_punctuation): Support short delay slots and
15292 (umips_swm_mask, umips_swm_encoding): New.
15293 (umips_build_save_restore): New function.
15294 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15295 (was_mips16_p): Remove.
15296 (old_compression_mode): New.
15297 (mips_set_compression_mode): New function.
15298 (mips_set_current_function): Add microMIPS support.
15299 (mips_option_override): Likewise.
15300 (umips_save_restore_pattern_p): New function.
15301 (umips_output_save_restore): New function.
15302 (umips_load_store_pair_p_1): New function.
15303 (umips_load_store_pair_p): New function.
15304 (umips_output_load_store_pair_1): New function.
15305 (umips_output_load_store_pair): New function.
15306 (umips_movep_target_p) New function.
15307 (mips_prepare_pch_save): Add microMIPS support.
15308 * config/mips/mips.h (TARGET_COMPRESSION): New.
15309 (TARGET_CPU_CPP_BUILTINS): Update macro
15310 to use new compression flags and to support microMIPS.
15311 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15312 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15313 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15314 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15315 (ASM_SPEC): Support mmicromips and mno-micromips.
15316 (M16STORE_REG_P): New macro.
15317 (MIPS_CALL): Support TARGET_MICROMIPS.
15318 (MICROMIPS_J): New macro.
15319 (mips_base_mips16): Rename this...
15320 (mips_base_compression_flags): ...to this.
15321 (UMIPS_12BIT_OFFSET_P): New macro.
15322 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15323 (MULTILIB_DIRNAMES): Likewise.
15325 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15328 * target-globals.c (save_target_globals): For init_reg_sets and
15329 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15331 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15333 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15334 * config/aarch64/t-aarch64
15335 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15337 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15339 PR inline-asm/56148
15340 * lra-constraints.c (process_alt_operands): Reload operand
15341 conflicting with earlier clobber only if no more other conflicting
15344 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15347 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15348 if not linking a shared library.
15350 2013-02-22 Seth LaForge <sethml@google.com>
15352 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15354 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15356 * config/arm/arm.md (split for extendsidi): Update condition.
15357 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15358 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15359 (qhs_zextenddi_cstr): Likewise.
15361 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15363 PR middle-end/56420
15364 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15365 avoid signed wrapping.
15366 (expand_mult): Handle properly multiplication by
15367 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15368 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15369 in the compiler if coeff is HOST_WIDE_INT_MIN.
15370 (expand_divmod): Don't make ext_op1 static, change it's type to
15371 uhwi. Avoid undefined behavior in -INTVAL (op1).
15373 PR rtl-optimization/50339
15374 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15376 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
15377 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
15378 into splitting_ashiftrt field.
15379 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
15381 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
15384 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
15386 PR middle-end/56108
15387 * trans-mem.c (execute_tm_mark): Do not expand transactions that
15388 are sure to go irrevocable.
15390 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
15392 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
15393 scalars are valid operands.
15395 2013-02-21 Martin Jambor <mjambor@suse.cz>
15397 PR tree-optimization/56310
15398 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
15399 only matching indices and non-negative final offsets.
15400 (intersect_aggregates_with_edge): Pass src_idx to
15401 agg_replacements_to_vector. Pass src_idx insstead of index to
15402 intersect_with_agg_replacements.
15404 2013-02-21 Martin Jambor <mjambor@suse.cz>
15406 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
15407 instead of hard-wired defaults.
15409 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
15411 * doc/invoke.texi (MIPS Options): Update documentation of the
15412 floating-point multiply-accumulate instruction restrictions.
15414 2013-02-21 Kostya Serebryany <kcc@google.com>
15416 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
15417 asan_shadow_offset on x86_64 linux.
15419 2013-02-21 Richard Biener <rguenther@suse.de>
15421 PR tree-optimization/56415
15423 2013-02-11 Richard Biener <rguenther@suse.de>
15425 PR tree-optimization/56273
15426 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15429 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15432 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
15435 PR inline-asm/56405
15436 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
15437 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
15439 2013-02-20 Jan Hubicka <jh@suse.cz>
15441 PR tree-optimization/56265
15442 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
15443 when target is referenced for first time.
15445 2013-02-20 Richard Biener <rguenther@suse.de>
15447 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
15448 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
15449 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
15450 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
15451 not return anything.
15452 (rename_ssa_copies): Do not remove unused locals.
15453 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
15454 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
15455 * passes.c (execute_function_todo): Do not schedule unused locals
15456 removal if cleanup_tree_cfg did something.
15457 * tree-ssa-live.c (remove_unused_locals): Dump statistics
15458 about the number of removed locals.
15460 2013-02-20 Richard Biener <rguenther@suse.de>
15462 PR tree-optimization/56398
15463 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
15465 2013-02-20 Martin Jambor <mjambor@suse.cz>
15467 PR tree-optimization/55334
15468 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
15469 restricted pointers to arrays.
15471 2013-02-20 Richard Biener <rguenther@suse.de>
15472 Jakub Jelinek <jakub@redhat.com>
15474 PR tree-optimization/56396
15475 * tree-ssa-ccp.c (n_const_val): New static variable.
15476 (get_value): Return NULL for SSA names we don't have a lattice
15478 (ccp_initialize): Initialize n_const_val.
15479 * tree-ssa-copy.c (n_copy_of): New static variable.
15480 (init_copy_prop): Initialize n_copy_of.
15481 (get_value): Return NULL_TREE for SSA names we don't have a
15484 2013-02-20 Martin Jambor <mjambor@suse.cz>
15486 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
15488 2013-02-20 Richard Biener <rguenther@suse.de>
15490 * genpreds.c (write_lookup_constraint): Do not compare first
15491 letter of the constraint again.
15493 2013-02-20 Richard Biener <rguenther@suse.de>
15495 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
15497 (get_use_iv_cost): Terminate hashtable walk when coming across
15500 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
15502 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
15503 reassociation for avx2 targets.
15505 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15507 * config/microblaze/microblaze.c: microblaze_has_clz = 0
15508 Add version check for v8.10.a to enable microblaze_has_clz
15509 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
15510 version and TARGET_PATTERN_COMPARE check
15511 * config/microblaze/microblaze.md: New clzsi2 instruction
15513 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15515 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
15516 function before branching.
15518 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15520 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
15522 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
15524 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
15526 PR middle-end/55889
15527 * sel-sched.c: Include ira.h.
15528 (implicit_clobber_conflict_p): New function.
15529 (moveup_expr): Use it.
15530 * Makefile.in (sel-sched.o): Depend on ira.h.
15532 2013-02-19 Richard Biener <rguenther@suse.de>
15534 PR tree-optimization/56384
15535 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
15536 (vn_hash_type): Split out from ...
15537 (vn_hash_constant_with_type): ... here.
15538 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
15539 (vn_phi_eq): Compare types from vn_phi_s structure.
15540 (vn_phi_lookup): Populate vn_phi_s type.
15541 (vn_phi_insert): Likewise.
15543 2013-02-19 Jakub Jelinek <jakub@redhat.com>
15545 PR tree-optimization/56350
15546 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
15547 if haven't found reduction or nested cycle operand, rather than
15548 asserting we must find it.
15550 PR tree-optimization/56381
15551 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
15554 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
15555 Jakub Jelinek <jakub@redhat.com>
15558 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
15559 (swap_optab_enable): Same.
15560 (init_all_optabs): Use argument instead of global.
15561 * tree.h (struct tree_optimization_option): New field target_optabs.
15562 * expr.h (init_all_optabs): Add argument to prototype.
15563 (TREE_OPTIMIZATION_OPTABS): New.
15564 (save_optabs_if_changed): Protoize.
15565 * optabs.h: Declare this_fn_optabs.
15566 * optabs.c (save_optabs_if_changed): New.
15567 Declare this_fn_optabs.
15568 (init_optabs): Add argument to init_all_optabs() call.
15569 * function.c (invoke_set_current_function_hook): Handle per
15571 * function.h (struct function): New field optabs.
15572 * config/mips/mips.c (mips_set_mips16_mode): Handle when
15573 optimization_current_node has changed.
15574 * target-globals.h (save_target_globals_default_opts): Protoize.
15575 * target-globals.c (save_target_globals_default_opts): New.
15577 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15580 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
15581 registers %fr12 and %fr12R as call used.
15584 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
15585 and HImode, require all displacements to be an integer multiple of
15587 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
15588 only allow QImode and HImode when reload is in progress and strict is
15589 true. Likewise for symbolic addresses. Use base14_operand to check
15590 displacements in REG+BASE addresses.
15592 2013-02-18 Richard Biener <rguenther@suse.de>
15594 PR tree-optimization/56366
15595 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15596 handle sign-conversion of outer-loop initial induction value.
15598 2013-02-18 Richard Biener <rguenther@suse.de>
15600 PR middle-end/56349
15601 * cfghooks.c (merge_blocks): If we merge a latch into another
15602 block adjust references to it.
15603 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
15604 (verify_loop_structure): Verify that a recorded latch is in fact
15607 2013-02-18 Richard Biener <rguenther@suse.de>
15609 PR tree-optimization/56321
15610 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
15611 order SSA name release and virtual operand unlinking.
15613 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15615 * config/microblaze/microblaze.md (save_stack_block): Define.
15616 (restore_stack_block): Likewise.
15618 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15620 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
15621 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
15622 * config/microblaze/microblaze.c (microblaze_option_override):
15623 Bail out early for PIC modes when target does not support PIC.
15625 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15627 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
15628 Replace with a microblaze version.
15629 (microblaze_trampoline_init): Adapt for microblaze.
15630 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
15633 2013-02-16 Jakub Jelinek <jakub@redhat.com>
15634 Dodji Seketeli <dodji@redhat.com>
15637 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
15638 (instrument_mem_region_access): Do not forget to always put
15639 instrumentation of the of 'base' and 'base + len' in a "if (len !=
15640 0) statement, even for cases where either 'base' or 'base + len'
15641 are not instrumented -- because they have been previously
15642 instrumented. Simplify the logic by putting all the statements
15643 instrument 'base + len' inside a sequence, and then insert that
15644 sequence right before the current insertion point. Then, to
15645 instrument 'base + len', just get an iterator on that statement.
15646 And do not forget to update the pointer to iterator the function
15647 received as argument.
15649 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
15651 PR rtl-optimization/56348
15652 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
15654 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
15656 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
15657 (clean_graph_dump_file): Pass base to start_graph_dump.
15659 2013-02-14 Richard Henderson <rth@redhat.com>
15662 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
15664 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
15666 * collect2-aix.h: Define F_LOADONLY.
15668 2013-02-14 Richard Biener <rguenther@suse.de>
15671 * varasm.c (output_constant_def_1): Get the decl representing
15672 the constant as argument.
15673 (output_constant_def): Wrap output_constant_def_1.
15674 (make_decl_rtl): Use output_constant_def_1 with the decl
15675 representing the constant.
15676 (build_constant_desc): Optionally re-use a decl already
15677 representing the constant.
15678 (tree_output_constant_def): Adjust.
15680 2013-02-14 Dodji Seketeli <dodji@redhat.com>
15683 * asan.c (instrument_builtin_call): Really put the length of the
15684 second source argument into src1_len.
15686 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15688 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
15689 argument. If it is false, don't create edge from then_bb to
15691 (insert_if_then_before_iter): Pass true to it.
15692 (build_check_stmt): Pass false to it.
15693 (transform_statements): Flush hash table only on extended basic
15694 block boundaries, rather than at the beginning of every bb.
15695 Don't flush hash table on nonfreeing_call_p calls.
15696 * tree-flow.h (nonfreeing_call_p): New prototype.
15697 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
15699 2013-02-13 David S. Miller <davem@davemloft.net>
15701 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
15703 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
15706 * ira.c (max_regno_before_ira): Move from ...
15708 (fix_reg_equiv_init): Use max_regno_before_ira instead of
15711 2013-02-13 Jakub Jelinek <jakub@redhat.com>
15713 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
15715 2013-02-13 Richard Biener <rguenther@suse.de>
15718 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
15719 globals in MEM_REFs.
15721 2013-02-13 Richard Biener <rguenther@suse.de>
15723 * loop-init.c (loop_optimizer_init): Clear loop state when
15724 re-initializing preserved loops.
15725 * loop-unswitch.c (unswitch_single_loop): Return whether
15726 we unswitched the loop. Do not verify loop state here.
15727 (unswitch_loops): When we unswitched a loop discover new loops.
15729 2013-02-13 Kostya Serebryany <kcc@google.com>
15731 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
15733 * sanitizer.def: Rename __asan_init to __asan_init_v1.
15735 2013-02-12 Dodji Seketeli <dodji@redhat.com>
15737 Avoid instrumenting duplicated memory access in the same basic block
15738 * Makefile.in (asan.o): Add new dependency on hash-table.h
15739 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
15740 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
15741 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
15742 (free_mem_ref_resources, has_mem_ref_been_instrumented)
15743 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
15744 (get_mem_ref_of_assignment): New functions.
15745 (get_mem_refs_of_builtin_call): Extract from
15746 instrument_builtin_call and tweak a little bit to make it fit with
15748 (instrument_builtin_call): Use the new
15749 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
15750 of is_gimple_builtin_call.
15751 (instrument_derefs, instrument_mem_region_access): Insert the
15752 instrumented memory reference into the hash table.
15753 (maybe_instrument_assignment): Renamed instrument_assignment into
15754 this, and change it to advance the iterator when instrumentation
15755 actually happened and return true in that case. This makes it
15756 homogeneous with maybe_instrument_assignment, and thus give a
15757 chance to callers to be more 'regular'.
15758 (transform_statements): Clear the memory reference hash table
15759 whenever we enter a new BB, when we cross a function call, or when
15760 we are done transforming statements. Use
15761 maybe_instrument_assignment instead of instrumentation. No more
15762 need to special case maybe_instrument_assignment and advance the
15763 iterator after calling it; it's now handled just like
15764 maybe_instrument_call. Update comment.
15766 2013-02-13 Richard Biener <rguenther@suse.de>
15768 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
15769 Fix loop discovery code.
15771 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
15773 PR inline-asm/56148
15774 * lra-constraints.c (process_alt_operands): Match early clobber
15775 operand with itself. Check conflicts with earlyclobber only if
15776 the operand is not reloaded. Prefer to reload conflicting operand
15777 if earlyclobber and matching operands are the same.
15779 2013-02-12 Richard Biener <rguenther@suse.de>
15782 * lto-streamer-out.c (write_symbol): Do not output symbols
15783 for hard register variables.
15785 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
15788 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
15789 (umulsidi3_insn, mulsidi3_insn): New insns.
15791 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
15793 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
15794 (struct tune_params): Add vec_costs field.
15795 * config/arm/arm.c (arm_builtin_vectorization_cost)
15796 (arm_add_stmt_cost): New functions.
15797 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
15798 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
15799 (arm_default_vec_cost): New struct of type cpu_vec_costs.
15800 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
15801 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
15802 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
15803 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
15805 2013-02-12 Richard Biener <rguenther@suse.de>
15808 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
15809 decls again if possible.
15811 2013-02-12 Richard Biener <rguenther@suse.de>
15813 PR middle-end/56288
15814 * tree-ssa.c (verify_ssa_name): Fix check, move
15815 SSA_NAME_IN_FREE_LIST check up.
15817 2013-02-12 Jakub Jelinek <jakub@redhat.com>
15818 Steven Bosscher <steven@gcc.gnu.org>
15820 PR rtl-optimization/56151
15821 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
15822 equal to op0 or op1, and last_insn pattern is CODE operation
15823 with MEM dest and one of the operands matches that MEM.
15825 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
15827 * doc/extend.texi: Document Function Multiversioning and "default"
15828 parameter string to target attribute.
15829 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
15830 target attribute parameter is "default".
15831 (ix86_compare_version_priority): Remove checks for target attribute.
15832 (ix86_mangle_function_version_assembler_name): Change error to sorry.
15833 Remove check for target attribute equal to NULL. Add assert.
15834 (ix86_generate_version_dispatcher_body): Change error to sorry.
15836 2013-02-11 Iain Sandoe <iain@codesourcery.com>
15837 Jack Howarth <howarth@bromo.med.uc.edu>
15838 Patrick Marlier <patrick.marlier@gmail.com>
15841 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
15842 define ENDFILE_SPEC as TM_DESTRUCTOR.
15843 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
15845 2013-02-11 Alexander Potapenko <glider@google.com>
15846 Jack Howarth <howarth@bromo.med.uc.edu>
15847 Jakub Jelinek <jakub@redhat.com>
15850 * config/darwin.c (cdtor_record): Rename ctor_record.
15851 (sort_cdtor_records): Rename sort_ctor_records.
15852 (finalize_dtors): New routine to sort destructors by
15853 priority before use in assemble_integer.
15854 (machopic_asm_out_destructor): Use finalize_dtors if needed.
15856 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
15858 PR rtl-optimization/56275
15859 * simplify-rtx.c (avoid_constant_pool_reference): Check that
15860 offset is non-negative and less than cmode size before
15861 calling simplify_subreg.
15863 2013-02-11 Richard Biener <rguenther@suse.de>
15865 PR tree-optimization/56264
15866 * cfgloop.h (fix_loop_structure): Adjust prototype.
15867 * loop-init.c (fix_loop_structure): Return the number of
15868 newly discovered loops.
15869 * tree-cfgcleanup.c (repair_loop_structures): When new loops
15870 are discovered, do a full loop-closed SSA rewrite.
15872 2013-02-11 Richard Biener <rguenther@suse.de>
15874 PR tree-optimization/56273
15875 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
15877 (check_array_ref): Fix missing newline in dumps.
15878 (search_for_addr_array): Likewise.
15880 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
15882 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
15884 2013-02-09 Jakub Jelinek <jakub@redhat.com>
15887 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
15889 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
15891 PR rtl-optimization/56246
15892 * lra-constraints.c (simplify_operand_subreg): Try to reuse
15894 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
15895 constraints are satisfied.
15897 2013-02-08 Jeff Law <law@redhat.com>
15900 * emit-rtl.c (reg_is_parm_p): New function.
15901 * regs.h (reg_is_parm_p): New prototype.
15902 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
15903 callee-clobbered registers.
15905 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
15908 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
15909 If there is no implicit builtin declaration, just return NULL.
15911 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15913 * config/i386/sse.md (FMAMODEM): New mode iterator.
15914 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
15915 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
15917 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15919 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
15920 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
15921 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
15923 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15925 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
15926 (microblaze*-*-elf): Likewise.
15927 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
15929 * config/microblaze/microblaze-c.c: Add builtin defines for
15930 _LITTLE_ENDIAN and _BIG_ENDIAN.
15931 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
15932 add to TARGET_DEFAULT flags.
15933 Expand ASM_SPEC and LINK_SPEC.
15934 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
15935 * config/microblaze/microblaze.md: Update extendsidi2 and
15936 movdi_internal instructions to use low-order / high-order reg
15938 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
15939 options and inversemask / mask of LITTLE_ENDIAN.
15940 * config/microblaze/t-microblaze: Expand multilib options to
15941 include mlittle-endian (le) and update exceptions patterns.
15943 2013-02-08 Jakub Jelinek <jakub@redhat.com>
15945 PR rtl-optimization/56195
15946 * lra-constraints.c (get_reload_reg): Don't reuse regs
15947 if they have smaller mode than requested, if they have
15948 wider mode than requested, try to return a SUBREG.
15950 PR tree-optimization/56250
15951 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
15952 if type is unsigned and code isn't MULT_EXPR.
15954 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
15956 PR tree-optimization/56064
15957 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
15958 bits according to mode.
15959 * fixed-value.h (fixed_from_double_int)
15960 (const_fixed_from_double_int): Adjust comments.
15962 2013-02-08 Richard Biener <rguenther@suse.de>
15965 * lto-streamer.h (struct data_in): Remove current_file, current_line
15966 and current_col members.
15967 * lto-streamer-out.c (lto_output_location): Stream changed bits
15968 en-block for efficiency.
15969 * lto-streamer-in.c (clear_line_info): Remove.
15970 (lto_input_location): Cache current file, line and column
15971 globally via local statics. Read changed bits en-block.
15972 (input_function): Do not call clear_line_info.
15973 (lto_read_body): Likewise.
15974 (lto_input_toplevel_asms): Likewise.
15976 2013-02-08 Michael Matz <matz@suse.de>
15978 PR tree-optimization/52448
15979 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
15980 (nt_call_phase): New static.
15981 (add_or_mark_expr): Only mark accesses with newer phase than any
15983 (nonfreeing_call_p): New.
15984 (nt_init_block): Update nt_call_phase, mark blocks as visited.
15985 (nt_fini_block): Keep blocks marked as visited.
15986 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
15988 2013-02-08 Richard Biener <rguenther@suse.de>
15990 * ira.c (ira): Free broken dominator information.
15992 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
15994 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
15996 2013-02-08 Marek Polacek <polacek@redhat.com>
15998 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16000 2013-02-08 Richard Biener <rguenther@suse.de>
16002 PR middle-end/56181
16003 * cfgloop.h (flow_loops_find): Adjust.
16004 (bb_loop_header_p): Declare.
16005 * cfgloop.c (bb_loop_header_p): New function split out from ...
16006 (flow_loops_find): ... here. Adjust function signature,
16007 support incremental loop structure update.
16008 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16009 * cfgloopmanip.c (fix_loop_structure): Move ...
16010 * loop-init.c (fix_loop_structure): ... here.
16011 (apply_loop_flags): Split out from ...
16012 (loop_optimizer_init): ... here.
16013 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16014 in incremental mode, only remove dead loops here.
16016 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16019 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16020 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16021 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16022 (*round<mode>3.libgcc): New insns for fixed-modes.
16023 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16024 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16025 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16026 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16027 implementations. Define to __builtin_avr_absFX,
16028 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16029 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16030 __builtin_avr_countlsFX, respectively.
16031 * config/avr/avr-c.c (target.h): Include it.
16032 (enum avr_builtin_id): New enum.
16033 (avr_resolve_overloaded_builtin): New static function.
16034 (avr_register_target_pragmas): Use it to set
16035 targetm.resolve_overloaded_builtin.
16036 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16037 tree nodes used by DEF_BUILTIN.
16038 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16039 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16040 <AVR_BUILTIN_xxBITS>: Same.
16042 2013-02-08 Richard Biener <rguenther@suse.de>
16044 * cfgloop.c (verify_loop_structure): Properly handle
16045 a loop exiting to another loop header.
16046 * ira-int.h (ira_loops): Remove.
16047 * ira.c (ira_loops): Remove.
16048 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16049 (do_reload): Use loop_optimizer_finalize.
16050 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16051 number_of_loops to access the loop tree.
16052 (more_one_region_p): Likewise.
16053 (finish_loop_tree_nodes): Likewise.
16054 (rebuild_regno_allocno_maps): Likewise.
16055 (mark_loops_for_removal): Likewise.
16056 (mark_all_loops_for_removal): Likewise.
16057 (remove_unnecessary_regions): Likewise.
16058 (ira_build): Likewise.
16059 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16061 2013-02-08 Richard Biener <rguenther@suse.de>
16063 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16064 * ipa-pure-const.c (analyze_function): Avoid calling
16065 mark_irreducible_loops twice.
16066 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16068 2013-02-07 David S. Miller <davem@davemloft.net>
16070 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16072 * var-tracking.c (vt_add_function_parameter): Test the presence of
16073 HAVE_window_save properly and do not remap argument registers when
16074 we have a leaf function.
16076 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16079 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16081 * config/i386/i386.c (ix86_print_operand): Ditto.
16083 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16085 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16087 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16089 PR rtl-optimization/56225
16090 * lra-constraints.c (process_alt_operands): Check that reload hard
16091 reg can hold value for strict_low_part.
16093 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16096 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16097 dwarf2out_end_function.
16098 (in_first_function_p, maybe_at_text_label_p,
16099 first_loclabel_num_not_at_text_label): New variables.
16100 (dwarf2out_var_location): In the first function find out
16101 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16102 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16105 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16107 PR rtl-optimization/56178
16108 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16109 SUBREG of a register. Tidy up related block of code.
16110 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16111 note if the source is a register or a SUBREG of a register.
16113 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16116 * config/rs6000/rs6000.md (ptrm): New mode attr.
16117 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16118 call_value_indirect_aix<pttrsize>,
16119 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16122 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16124 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16125 if -bnortl. Convert to strcmp and strncmp.
16127 2013-02-07 Alan Modra <amodra@gmail.com>
16130 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16131 addresses won't wrap when offsetting.
16132 (rs6000_secondary_reload): Provide secondary reloads needed for
16133 wrapping LO_SUM addresses.
16135 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16137 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16138 MACH, just __MACH__.
16140 2013-02-06 Richard Biener <rguenther@suse.de>
16142 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16143 instead of calling fix_loop_structure.
16145 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16147 PR middle-end/56217
16148 * omp-low.c (use_pointer_for_field): Return false if
16149 lower_send_shared_vars doesn't generate any copy-out code.
16151 2013-02-06 Tom de Vries <tom@codesourcery.com>
16153 PR rtl-optimization/56131
16154 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16155 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16156 of the label is NULL. Add comment.
16158 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16160 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16163 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16164 (STATIC_LIBTSAN_LIBS): Likewise.
16165 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16166 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16167 is defined, don't add anything else beyond that.
16168 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16169 (LINK_COMMAND_SPEC): Use them.
16171 PR tree-optimization/56205
16172 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16173 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16174 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16176 2013-02-05 Richard Biener <rguenther@suse.de>
16178 PR tree-optimization/53342
16179 PR tree-optimization/53185
16180 * tree-vectorizer.h (vect_check_strided_load): Remove.
16181 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16182 not disallow peeling for vectorized strided loads.
16183 (vect_check_strided_load): Make static and simplify.
16184 (vect_analyze_data_refs): Adjust.
16185 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16186 correctly when vectorizing strided loads.
16188 2013-02-05 Richard Biener <rguenther@suse.de>
16190 * doc/install.texi: Refer to ISL, not PPL.
16192 2013-02-05 Jan Hubicka <jh@suse.cz>
16194 PR tree-optimization/55789
16195 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16197 2013-02-05 Jan Hubicka <jh@suse.cz>
16199 PR tree-optimization/55789
16200 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16201 the dead call anyway.
16203 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16206 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16208 2013-02-04 Alexander Potapenko <glider@google.com>
16209 Jack Howarth <howarth@bromo.med.uc.edu>
16210 Jakub Jelinek <jakub@redhat.com>
16213 * config/darwin.c (sort_ctor_records): Stabilized qsort
16214 on constructor priority by using original position.
16215 (finalize_ctors): New routine to sort constructors by
16216 priority before use in assemble_integer.
16217 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16219 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16222 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16223 about visibility on artificial decls.
16224 * config/sol2.c (solaris_assemble_visibility): Likewise.
16226 2013-02-04 Kai Tietz <ktietz@redhat.com>
16229 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16230 argument and improve checking of return-argument types for 16-byte
16232 (ix86_function_value_1): Add additional valtype argument on call
16233 of function_value_64.
16234 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16235 handling infunction_value_64 function.
16237 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16239 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16241 2013-02-04 Richard Biener <rguenther@suse.de>
16243 PR tree-optimization/56188
16244 * tree-ssa-structalias.c (label_visit): Consider case with
16245 initially non-empty points-to set.
16246 (perform_var_substitution): Dump node mapping and clean up.
16248 2013-02-04 Richard Guenther <rguenther@suse.de>
16251 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16252 node prevail as last resort.
16253 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16254 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16256 2013-02-04 Richard Biener <rguenther@suse.de>
16258 PR tree-optimization/56113
16259 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16261 (equiv_class_lookup_or_add): ... this.
16262 (label_visit): Adjust and fix error in previous patch.
16263 (perform_var_substitution): Adjust.
16265 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16267 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16268 * config/sh/sh4-300.md: Likewise.
16269 * config/sh/sh4a.md: Likewise.
16270 * config/sh/constraints.md: Likewise.
16271 * config/sh/sh.md: Likewise.
16272 * config/sh/netbsd-elf.h: Likewise.
16273 * config/sh/predicates.md: Likewise.
16274 * config/sh/sh-protos.h: Likewise.
16275 * config/sh/ushmedia.h: Likewise.
16276 * config/sh/linux.h: Likewise.
16277 * config/sh/sh.c: Likewise.
16278 * config/sh/superh.h: Likewise.
16279 * config/sh/elf.h: Likewise.
16280 * config/sh/sh4.md: Likewise.
16281 * config/sh/sh.h: Likewise.
16283 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16285 * config/pa/constraints.md: Adjust unused letters. Change "T"
16286 constraint to match_test floating_point_store_memory_operand().
16287 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16288 (base14_operand): New.
16289 (floating_point_store_memory_operand): New.
16290 (integer_store_memory_operand): Revise to use base14_operand and
16291 reg_plus_base_memory_operand.
16292 (move_dest_operand): Allow symbolic_memory_operands.
16293 (symbolic_memory_operand): Check for LO_SOM.
16294 (symbolic_operand): Change default case to break.
16295 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16296 CONST_DOUBLE values to be reloaded by putting them into memory when
16297 the destination is a floating point register.
16298 (movdf): Remove code to handle CONST_DOUBLE.
16300 (reload_indf_r1): New.
16301 (reload_insf_r1): New.
16302 Consistently use "Q" and "T" constraints with integer and floating
16303 point move instructions, respectively.
16304 (movdi): Remove FAIL.
16305 Change predicate for source operand unamed DImode move from
16306 general_operand to move_src_operand.
16307 (umulsidi3): Change predicate for destination operand to
16309 Likewise for similar unamed patterns.
16310 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16311 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16312 (hppa_legitimize_address): Simplify mask calculation.
16313 (pa_emit_move_sequence): Revised handling of secondary reloads from
16314 REG+D addresses for floating point loads and stores. Directly handle
16315 loading CONST0_RTX (mode) to a floating point register.
16316 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16317 to floating point registers. Don't restrict secondary reloads to
16318 floating point registers to integer modes. Revise some comments and
16320 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16321 (pa_legitimate_address_p): New.
16322 (pa_legitimize_reload_address): New.
16323 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16324 (STRICT_REG_OK_FOR_BASE_P): New.
16325 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16326 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16328 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16329 Andrew Dixie <andrewd@gentrack.com>
16331 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16334 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16336 * expmed.c (extract_bit_field_1): Pass the full width of the
16337 structure to get_best_reg_extraction_insn.
16339 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16342 * configure.ac (use_cxa_atexit): Add AIX.
16343 * configure: Regenerate.
16345 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16347 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16350 * final.c (need_profile_function): New variable.
16351 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16352 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16353 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16354 notes, targetm.asm_out.function_prologue doesn't emit anything,
16355 HAVE_prologue and profiler should be emitted before prologue,
16356 set need_profile_function instead of emitting it.
16357 (final_scan_insn): If need_profile_function, emit
16358 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16359 NOTE_INSN_FUNCTION_BEG note.
16361 2013-02-01 Richard Henderson <rth@redhat.com>
16363 * config/rs6000/rs6000.md (smulditi3): New.
16366 * config/alpha/alpha.md (umulditi3): New.
16368 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16370 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16371 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16373 2013-02-01 Richard Biener <rguenther@suse.de>
16375 PR tree-optimization/56113
16376 * tree-ssa-structalias.c (label_visit): Reduce work for
16377 single-predecessor nodes.
16379 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
16381 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
16382 range isn't testing for zero.
16384 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
16386 PR middle-end/56113
16387 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
16389 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
16390 Nick Clifton <nickc@redhat.com>
16392 * config/v850/constraints.md (Q): Define as a memory constraint.
16393 * config/v850/predicates.md (label_ref_operand): New predicate.
16394 (e3v5_shift_operand): New predicate.
16395 (ior_operator): New predicate.
16396 * config/v850/t-v850: Add e3v5 multilib.
16397 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
16398 (v850_gen_movdi): Prototype.
16399 * config/v850/v850.c: Add support for e3v5 architecture.
16400 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
16402 (construct_save_jarl): Add e3v5 long JARL support.
16403 (v850_adjust_insn_length): New function. Adjust length of call
16404 insns when using e3v5 instructions.
16405 (v850_gen_movdi): New function: Generate instructions to move a
16407 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
16408 (CPP_SPEC): Define __v850e3v5__ as appropriate.
16409 (TARGET_USE_FPU): Enable for e3v5.
16410 (CONST_OK_FOR_W): New macro.
16411 (ADJUST_INSN_LENGTH): Define.
16412 * config/v850/v850.md (UNSPEC_LOOP): Define.
16413 (attr cpu): Add v850e3v5.
16414 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
16415 (movdi): New pattern.
16416 (movdi_internal): New pattern.
16417 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
16418 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
16419 (cstoresf4): Likewise.
16420 (cstoredf4): Likewise.
16421 (insv): New pattern.
16422 (rotlso3_a): New pattern.
16423 (rotlsi3_b): New pattern
16424 (rotlsi3_v850e3v5): New pattern.
16425 (doloop_begin): New pattern.
16426 (fix_loop_counter): New pattern.
16427 (doloop_end): New pattern.
16428 (branch_normal): Add e3v5 long branch support.
16429 (branch_invert): Likewise.
16430 (branch_z_normal): Likewise.
16431 (branch_z_invert): Likewise.
16432 (branch_nz_normal): Likewise.
16433 (branch_nz_invert): Likewise.
16434 (call_internal_short): Add e3v5 register-indirect JARL support.
16435 (call_internal_long): Likewise.
16436 (call_value_internal_short): Likewise.
16437 (call_value_internal_long): Likewise.
16438 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
16439 (mloop): New option.
16440 * config.gcc: Add support for configuring v840e3v5 target.
16441 * doc/invoke.texi: Document new v850 specific command line options.
16443 2013-01-31 Paul Koning <ni1d@arrl.net>
16447 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
16448 children if parent is a class.
16449 (prune_unused_types_prune): Don't add DW_AT_declaration.
16451 2013-01-31 Richard Biener <rguenther@suse.de>
16453 PR tree-optimization/56157
16454 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
16455 match up operand with SLP child.
16457 2013-01-31 Jason Merrill <jason@redhat.com>
16460 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
16461 parameters the first time.
16462 (gen_scheduled_generic_parms_dies): Check completeness here.
16464 2013-01-31 Richard Biener <rguenther@suse.de>
16466 PR middle-end/53073
16467 * common.opt (faggressive-loop-optimizations): New flag,
16468 enabled by default.
16469 * doc/invoke.texi (faggressive-loop-optimizations): Document.
16470 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
16471 infer_loop_bounds_from_undefined by it.
16473 2013-01-31 Richard Biener <rguenther@suse.de>
16475 PR tree-optimization/56150
16476 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
16477 visit virtual operands.
16478 (find_uses_to_rename_bb): Likewise.
16480 2013-01-31 Richard Biener <rguenther@suse.de>
16482 PR tree-optimization/56150
16483 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
16484 mixed store non-store stmts.
16486 2013-01-30 Jakub Jelinek <jakub@redhat.com>
16489 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
16490 LIBASAN_EARLY_SPEC is defined.
16491 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
16492 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
16494 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
16497 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
16498 invalid args instead of ICEing on it.
16499 (ix86_valid_target_attribute_tree): Return error_mark_node if
16500 ix86_valid_target_attribute_inner_p failed.
16501 (ix86_valid_target_attribute_p): Return false only if
16502 ix86_valid_target_attribute_tree returned error_mark_node. Allow
16503 target("default") attribute.
16504 (sorted_attr_string): Change argument from const char * to tree,
16505 merge in all target attribute arguments rather than just one.
16506 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
16507 instead of free. Avoid using strcat.
16508 (ix86_mangle_function_version_assembler_name): Mangle
16509 target("default") as if no target attribute is present. Adjust
16510 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
16511 instead of xmalloc and XDELETEVEC instead of free.
16512 (ix86_function_versions): Don't return true if one of the decls
16513 doesn't have target attribute. If they don't and one of the decls
16514 is DECL_FUNCTION_VERSIONED, report an error. Adjust
16515 sorted_attr_string caller. Use XDELETEVEC instead of free.
16516 (ix86_supports_function_versions): Remove.
16517 (make_name): Fix up formatting.
16518 (make_dispatcher_decl): Remove resolver_name and its initialization.
16519 Avoid leaking memory.
16520 (is_function_default_version): Return true if there is
16521 target("default") attribute rather than no target attribute at all.
16522 (make_resolver_func): Avoid leaking memory.
16523 (ix86_generate_version_dispatcher_body): Likewise.
16524 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
16525 * target.def (supports_function_versions): Remove.
16526 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
16527 * doc/tm.texi: Regenerated.
16529 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
16531 PR rtl-optimization/56144
16532 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
16533 for values with side effects.
16535 2013-01-30 Richard Biener <rguenther@suse.de>
16537 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
16538 (sparseset_pop): Likewise.
16539 * cfganal.c (compute_idf): Likewise. Increase work-stack size
16540 to be able to use quick_push in the worker loop.
16542 2013-01-30 Marek Polacek <polacek@redhat.com>
16544 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
16546 2013-01-30 Richard Biener <rguenther@suse.de>
16549 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
16551 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
16553 PR tree-optimization/56064
16554 * fixed-value.c (fixed_from_double_int): New function.
16555 * fixed-value.h (fixed_from_double_int): New prototype.
16556 (const_fixed_from_double_int): New static inline function.
16557 * fold-const.c (native_interpret_fixed): New static function.
16558 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
16559 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
16560 (native_encode_fixed): New static function.
16561 (native_encode_expr) <FIXED_CST>: Use it.
16562 (native_interpret_int): Move double_int worker code to...
16563 * double-int.c (double_int::from_buffer): ...this new static method.
16564 * double-int.h (double_int::from_buffer): Prototype it.
16566 2013-01-30 Richard Biener <rguenther@suse.de>
16568 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
16569 New pointer-map and obstack.
16570 (init_alias_vars): Allocate pointer-map and obstack.
16571 (delete_points_to_sets): Free them.
16572 (find_what_var_points_to): Cache result.
16573 (find_what_p_points_to): Adjust for changed interface of
16574 find_what_var_points_to.
16575 (compute_points_to_sets): Likewise.
16576 (ipa_pta_execute): Likewise.
16578 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16580 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
16581 * configure: Regenerate.
16582 * config.in: Regenerate.
16583 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
16584 #nobits/#progbits if supported.
16586 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
16589 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
16590 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
16591 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
16593 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16595 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
16596 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
16598 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16600 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
16601 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
16603 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16605 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
16607 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
16608 * config/arm/cortex-a7.md: New bypasses using
16609 arm_mac_accumulator_is_result.
16611 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16613 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
16614 (cortex_a7_neon_mla): Likewise.
16615 (cortex_a7_fpfmad): New reservation.
16616 (cortex_a7_fpmacs): Use ffmas and update required units.
16617 (cortex_a7_fpmuld): Update required units and latency.
16618 (cortex_a7_fpmacd): Likewise.
16619 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
16620 (cortex_a7_neon). Likewise.
16621 (bypass) Update participating units.
16623 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
16625 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
16626 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
16628 * config/arm/vfp11.md (vfp_farith): Use ffmas.
16629 (vfp_fmul): Use ffmad.
16630 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
16631 (cortex_r4_fmacd): Use ffmad.
16632 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
16633 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
16634 (cortex_a9_fmacd): Use ffmad.
16635 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
16636 (cortex_a8_vfp_macd): Use ffmad.
16637 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
16638 (cortex_a5_fpmacd): Use ffmad.
16639 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
16640 (cortex_a15_vfp_macd): Use ffmad.
16641 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
16643 2013-01-29 Jason Merrill <jason@redhat.com>
16646 * varasm.c (default_assemble_visibility): Don't warn about
16647 visibility on artificial decls.
16649 2013-01-29 Richard Biener <rguenther@suse.de>
16651 PR tree-optimization/56113
16652 * tree-ssa-structalias.c (equiv_class_lookup): Also return
16654 (label_visit): Free duplicate bitmaps and record the leader instead.
16655 (perform_var_substitution): Adjust.
16657 2013-01-29 Richard Biener <rguenther@suse.de>
16659 PR tree-optimization/55270
16660 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
16661 the CFG, schedule loops for fixup.
16663 2013-01-29 Nick Clifton <nickc@redhat.com>
16665 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
16668 2013-01-28 Leif Ekblad <leif@rdos.net>
16670 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
16671 * config/i386/i386.h (TARGET_RDOS): New macro.
16672 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
16673 * config/i386/i386.c (ix86_option_override_internal): For 64bit
16674 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
16675 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
16676 DEFAULT_LARGE_SECTION_THRESHOLD.
16677 * config/i386/i386.md (R14_REG, R15_REG): New constants.
16678 * config/i386/rdos.h: New file.
16679 * config/i386/rdos64.h: New file.
16681 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
16684 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
16687 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16689 PR rtl-optimization/56117
16690 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
16691 call cselib_lookup_from_insn on the MEM before calling
16692 add_insn_mem_dependence.
16694 2013-01-28 Richard Biener <rguenther@suse.de>
16696 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
16697 to a stmt that didn't have one.
16698 (copy_phis_for_bb): Likewise for PHI arguments.
16699 (copy_debug_stmt): Likewise for debug stmts.
16701 2013-01-28 Richard Biener <rguenther@suse.de>
16703 PR tree-optimization/56034
16704 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
16705 (partition_builtin_p): Adjust.
16706 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
16707 it is the last partition.
16708 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
16709 up the vertex for the definition.
16710 (classify_partition): Classify whether a partition is a
16711 PKIND_REDUCTION, thus has uses outside of the loop.
16712 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
16713 Merge all PKIND_REDUCTION partitions into the last partition.
16714 (tree_loop_distribution): Seed partitions from reductions as well.
16716 2013-01-28 Jakub Jelinek <jakub@redhat.com>
16718 PR tree-optimization/56125
16719 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
16720 pow(x,c) into sqrt(x) * powi(x, n/2) or
16721 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
16722 optimizing for size.
16723 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
16724 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
16727 PR tree-optimization/56094
16728 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
16729 to UNKNOWN_LOCATION while gimplifying expr.
16731 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
16734 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
16735 operand 0 in movabs insn template for -masm=intel asm alternative.
16736 (*movabs<mode>_2): Ditto for operand 1.
16738 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
16741 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
16744 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16746 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
16749 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16751 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
16752 Undef to avoid warning.
16754 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16756 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
16757 * configure: Regenerate.
16759 2013-01-25 Jakub Jelinek <jakub@redhat.com>
16761 PR tree-optimization/56098
16762 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
16763 for stmts with volatile ops.
16764 (cond_store_replacement): Don't optimize if assign has volatile ops.
16765 (cond_if_else_store_replacement_1): Don't optimize if either
16766 then_assign or else_assign have volatile ops.
16767 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
16770 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16772 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
16774 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16776 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
16777 missing ':' in asm example.
16779 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
16781 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
16782 entries into lane and laneq entries.
16783 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
16784 Remove AdvSIMD scalar modes.
16785 (aarch64_sq<r>dmulh_laneq<mode>): New.
16786 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
16788 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
16789 builtin implementations to relfect changes in RTL in aarch64-simd.md.
16790 * config/aarch64/iterators.md (VCOND): New.
16793 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
16796 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
16797 Add NULL LIBNAME argument to existing definitions.
16798 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
16799 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
16800 * config/avr/avr.c (DEF_BUILTIN): Same.
16801 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
16802 (avr_expand_builtin): Expand to a vanilla call if a libgcc
16803 implementation is available (DECL_ASSEMBLER_NAME is set).
16804 (avr_fold_absfx): New static function.
16805 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
16806 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
16807 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
16808 AVR_BUILTIN_ABSLLK.
16809 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
16810 (abshk, absk, abslk, absllk): Provide as static inline functions.
16812 2013-01-25 Marek Polacek <polacek@redhat.com>
16814 PR tree-optimization/56035
16815 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
16817 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
16819 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
16820 (*movtf_internal_rex64): Add (!o,C) alternative
16821 (*movxf_internal_rex64): Ditto.
16822 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
16824 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
16826 * doc/invoke.texi: fix typo.
16827 * doc/objc.texi: fix typo.
16829 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
16831 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
16832 for the first two alternatives.
16834 2013-01-24 Diego Novillo <dnovillo@google.com>
16836 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
16837 (ggc-zone.o): Remove.
16838 * configure.ac: Remove option --with-gc.
16839 * configure: Re-generate.
16840 * doc/install.texi: Remove documentation for --with-gc.
16841 * gengtype.c (write_enum_defn): Remove. Update all users.
16842 (write_Types_process_field): Remove generation of gt_e_* argument.
16843 (output_type_enum): Remove. Update all users.
16844 (write_enum_defn): Remove. Update all users.
16845 (enum alloc_zone): Remove. Update all users.
16846 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
16847 * ggc-common.c (ggc_splay_alloc): Remove first argument.
16848 Update all callers.
16849 (struct ptr_data): Remove field TYPE. Update all users.
16850 (gt_pch_note_object): Remove argument TYPE. Update all users.
16851 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
16853 * ggc-none.c (ggc_alloc_typed_stat): Remove.
16854 (struct alloc_zone): Remove.
16855 (ggc_internal_alloc_zone_stat): Remove.
16856 (ggc_internal_cleared_alloc_zone_stat): Remove.
16857 * ggc-page.c (ggc_alloc_typed_stat): Remove.
16858 (ggc_pch_count_object): Remove last argument. Update all users.
16859 (ggc_pch_alloc_object): Remove last argument. Update all users.
16860 (struct alloc_zone): Remove.
16861 * ggc-zone.c: Remove.
16862 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
16863 (struct alloc_zone): Remove.
16864 (ggc_alloc_typed_stat): Remove.
16865 (ggc_alloc_typed): Remove.
16866 (ggc_splay_alloc): Remove first argument.
16867 (rtl_zone): Remove. Update all users.
16868 (tree_zone): Remove. Update all users.
16869 (tree_id_zone): Remove. Update all users.
16870 (ggc_internal_zone_alloc_stat): Remove. Update all users.
16871 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
16872 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
16873 * tree-ssanames.c: Remove references to zone allocator in comments.
16875 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
16877 * config/avr/avr.c (avr_out_fract): Make register numbers that
16878 might be outside of source operand signed.
16880 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
16882 * config/i386/constraints.md (Yf): New constraint.
16883 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
16884 of f constraint to conditionaly disable x87 register preferences.
16885 (*movdf_internal): Ditto.
16886 (*movsf_internal): Ditto.
16888 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
16890 PR inline-asm/55934
16891 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
16892 that have operands with impossible constraints.
16893 Add a FIXME for a speed-up opportunity.
16894 * lra-constraints.c (process_alt_operands): Verify that a class
16895 selected from constraints on asms is valid for the operand mode.
16896 (curr_insn_transform): Remove incorrect comment.
16898 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
16900 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
16901 TOC operand is a valid symbol ref in the constant pool.
16903 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16905 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
16907 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
16910 * config/avr/stdfix.h: New file.
16911 * t-avr (stdfix-gcc.h): New rule to build it.
16912 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
16914 2013-01-23 Kostya Serebryany <kcc@google.com>
16916 * config/darwin.h: remove dependency on
16917 CoreFoundation (asan on Mac OS).
16919 2013-01-23 Jakub Jelinek <jakub@redhat.com>
16922 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
16923 instead of cmpdi_operand for first comparison operand.
16924 Don't assert that comparison operands aren't both constants.
16926 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16928 * doc/install.texi (Downloading the Source): Update references to
16929 downloading separate components.
16931 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
16933 * doc/extend.texi (__int128): Improve grammar.
16935 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
16938 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
16939 alternative to (o,r).
16940 (*movdi_internal_rex64): Remove (!o,n) alternative.
16941 (DImode immediate->memory splitter): Remove.
16942 (DImode immediate->memory peephole2): Remove.
16943 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
16944 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
16945 alternative to (!o,*r).
16946 (*movtf_internal_sse): New pattern.
16947 (*movxf_internal_rex64): New pattern.
16948 (*movxf_internal): Disable for TARGET_64BIT.
16949 (*movdf_internal_rex64): Remove (!o,F) alternative.
16951 2013-01-22 Jakub Jelinek <jakub@redhat.com>
16953 PR middle-end/56074
16954 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
16955 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
16956 * tree-vect-loop-manip.c (find_loop_location): Also ignore
16957 stmt locations where LOCATION_LOCUS of the stmt location is
16958 UNKNOWN_LOCATION or BUILTINS_LOCATION.
16961 * config/i386/i386.md (UNSPEC_STOS): New.
16962 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
16963 *strsetqi_1): Add UNSPEC_STOS.
16965 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
16968 * doc/invoke.texi: Remove left over -Wsynth example.
16970 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16972 PR tree-optimization/56051
16973 * fold-const.c (fold_binary_loc): Don't fold
16974 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
16975 a narrowing conversion, or widening conversion from signed
16978 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
16980 PR rtl-optimization/56023
16981 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
16982 dependent on debug instruction.
16984 2013-01-21 Martin Jambor <mjambor@suse.cz>
16986 PR middle-end/56022
16987 * function.c (allocate_struct_function): Call
16988 invoke_set_current_function_hook earlier.
16990 2013-01-21 Jakub Jelinek <jakub@redhat.com>
16992 * reload1.c (init_reload): Only initialize reload_obstack
16993 during the first call.
16995 2013-01-21 Marek Polacek <polacek@redhat.com>
16997 * cfgloop.c (verify_loop_structure): Fix up grammar.
16999 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17001 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17002 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17004 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17007 * config/arm/marvell-pj4.md: Update copyright year.
17008 Fix up use of alu to alu_reg and simple_alu_imm.
17010 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17012 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17014 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17017 * lra-constraints.c (curr_insn_transform): Don't reuse original
17018 insn for secondary memory move when memory mode should be different.
17020 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17022 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17023 atomic_storedi_1): New patterns.
17025 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17027 btver2 pipeline descriptions.
17028 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17030 * config/i386/i386.md (btver2_decode): New type attributes.
17031 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17033 * config/i386/btver2.md: New file describing btver2 pipelines.
17035 2013-01-19 Andrew Pinski <apinski@cavium.com>
17037 PR tree-optimization/52631
17038 * tree-ssa-sccvn (visit_use): Before looking up the original
17039 statement, try looking up the simplified expression.
17041 2013-01-19 Anthony Green <green@moxielogic.com>
17043 * config/moxie/moxie.c (moxie_expand_prologue): Set
17044 current_function_static_stack_size.
17046 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17048 PR tree-optimization/56029
17049 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17050 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17052 2013-01-18 Sharad Singhai <singhai@google.com>
17054 PR tree-optimization/55995
17055 * dumpfile.c (dump_loc): Print location only if available.
17056 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17058 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17061 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17062 secondary memory move.
17063 (inherit_reload_reg): Use rclass instead of cl for
17064 check_secondary_memory_needed_p.
17066 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17068 PR middle-end/56015
17069 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17070 the case where writing real complex part of target modifies op1.
17072 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17074 * config/aarch64/aarch64-simd.md
17075 (aarch64_vcond_internal<mode>): Handle unordered cases.
17076 * config/aarch64/iterators.md (v_cmp_result): New.
17078 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17079 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17081 * config/arm/marvell-pj4.md: New file.
17082 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17083 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17084 (generic_vfp): Likewise.
17085 * config/arm/arm-cores.def: Add marvell-pj4.
17086 * config/arm/arm-tune.md: Regenerate.
17087 * config/arm/arm-tables.opt: Regenerate.
17088 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17089 * doc/invoke.texi: Document marvell-pj4.
17091 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17093 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17095 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17100 * var-tracking.c (negative_power_of_two_p): New.
17101 (global_get_addr_cache, local_get_addr_cache): New.
17102 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17103 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17105 (vt_stack_offset_p): Remove.
17106 (vt_canon_true_dep): Always canonicalize loc's address.
17107 (clobber_overlapping_mems): Make sure we have a MEM.
17108 (local_get_addr_clear_given_value): New.
17109 (val_reset): Clear local cached entries.
17110 (compute_bb_dataflow): Create and release the local cache.
17111 Disable duplicate MEMs clobbering.
17112 (emit_notes_in_bb): Clobber MEMs likewise.
17113 (vt_emit_notes): Create and release the local cache.
17114 (vt_initialize, vt_finalize): Create and release the global
17115 cache, respectively.
17116 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17118 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17120 PR libmudflap/53359
17121 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17122 not found in the symtab.
17124 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17127 PR rtl-optimization/55547
17128 PR rtl-optimization/53827
17131 * alias.c (offset_overlap_p): New, factored out of...
17132 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17133 the conservative special case for symbolic constants. Don't
17134 adjust zero sizes on alignment.
17136 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17138 PR rtl-optimization/52573
17139 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17140 REG_UNUSED for the same register.
17142 2013-01-17 Richard Biener <rguenther@suse.de>
17143 Marek Polacek <polacek@redhat.com>
17145 PR rtl-optimization/55833
17146 * loop-unswitch.c (unswitch_loops): Move loop verification...
17147 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17148 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17149 Set it to true when we're removing a loop from hierarchy tree in
17150 an irreducible region.
17151 (fix_bb_placements): Adjust caller.
17152 (fix_loop_placements): Likewise.
17154 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17156 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17157 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17158 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17159 Remove ID. Adjust comments.
17160 * config/avr/avr-c.c (avr_builtin_name): Remove.
17161 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17162 * config/avr/avr.c (avr_tolower): New static function.
17163 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17164 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17165 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17168 2013-01-17 Jan Hubicka <jh@suse.cz>
17170 PR tree-optimization/55273
17171 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17173 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17176 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17177 store through atomic_store<mode>_1.
17178 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17180 2013-01-17 Martin Jambor <mjambor@suse.cz>
17182 PR tree-optimizations/55264
17183 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17184 for virtual methods.
17185 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17186 virtual methods before inlining is over.
17187 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17189 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17192 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17194 PR rtl-optimization/56005
17195 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17196 pending reads for prefetch.
17198 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17200 * config/aarch64/aarch64.md
17201 (*cstoresi_neg_uxtw): New pattern.
17202 (*cmovsi_insn_uxtw): New pattern.
17203 (*<optab>si3_uxtw): New pattern.
17204 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17205 (*<optab>si3_insn_uxtw): New pattern.
17206 (*bswapsi2_uxtw): New pattern.
17208 2013-01-16 Richard Biener <rguenther@suse.de>
17210 * tree-inline.c (tree_function_versioning): Remove set but
17211 never used variable.
17213 2013-01-16 Richard Biener <rguenther@suse.de>
17215 PR tree-optimization/55964
17216 * tree-flow.h (rename_variables_in_loop): Remove.
17217 (rename_variables_in_bb): Likewise.
17218 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17219 (copy_loop_before): Adjust and delete update-ssa status.
17220 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17221 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17222 (rename_variables_in_loop): Remove.
17223 (slpeel_update_phis_for_duplicate_loop): Likewise.
17224 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17225 use available cfg machinery instead of duplicating it.
17226 Update PHI nodes and perform poor-mans SSA update here.
17227 (slpeel_tree_peel_loop_to_edge): Adjust.
17229 2013-01-16 Richard Biener <rguenther@suse.de>
17231 PR tree-optimization/54767
17232 PR tree-optimization/53465
17233 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17234 (vrp_visit_phi_node): For PHI arguments coming via backedges
17235 drop all symbolical range information.
17236 (execute_vrp): Compute backedges.
17238 2013-01-16 Richard Biener <rguenther@suse.de>
17240 * doc/install.texi: Update CLooG and ISL requirements to
17243 2013-01-16 Christian Bruel <christian.bruel@st.com>
17246 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17247 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17248 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17250 2013-01-16 DJ Delorie <dj@redhat.com>
17252 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17253 (UNSPECV_SP_SWITCH_E): New.
17254 (sp_switch_1): Change to an unspec.
17255 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17258 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17260 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17261 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17262 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17263 (expand_mem_thread_fence): Ditto.
17264 (expand_mem_signal_fence): Ditto.
17265 (expand_atomic_load): Ditto.
17266 (expand_atomic_store): Ditto.
17268 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17270 PR rtl-optimization/55547
17271 PR rtl-optimization/53827
17274 * alias.c (memrefs_conflict_p): Set sizes to negative after
17277 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17280 * function.c (thread_prologue_and_epilogue_insns): Always
17281 add crtl->drap_reg to set_up_by_prologue.set, even if
17282 stack_realign_drap is false.
17284 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17286 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17287 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17288 *call): Fix indention.
17290 2013-01-15 Tom de Vries <tom@codesourcery.com>
17293 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17296 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17298 PR rtl-optimization/55153
17299 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17301 2013-01-15 Martin Jambor <mjambor@suse.cz>
17303 PR tree-optimization/55920
17304 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17305 accesses as grp_to_be_debug_replaced.
17307 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17309 PR tree-optimization/55920
17310 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17311 there is non-useless type conversion needed from debug rhs to lhs,
17312 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17314 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17315 Mikael Pettersson <mikpe@it.uu.se>
17318 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17320 (ASM_OUTPUT_CASE_LABEL): Remove.
17321 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17322 * final.c (shorten_branches): Update alignment of labels before
17323 jump tables if CASE_VECTOR_SHORTEN_MODE.
17325 2013-01-15 Richard Biener <rguenther@suse.de>
17328 * system.h: Do not include gmp.h for building host tools.
17330 2013-01-15 Richard Biener <rguenther@suse.de>
17332 PR middle-end/55882
17333 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17334 account for bitpos when computing alignment.
17336 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17338 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17339 (ix86_target_macros_internal): Likewise.
17341 * config/i386/i386.c (m_CORE2I7): Removed.
17342 (m_CORE_HASWELL): New macro.
17343 (m_CORE_ALL): Likewise.
17344 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17345 (initial_ix86_arch_features): Likewise.
17346 (processor_target_table): Initializations for Core avx2.
17347 (cpu_names): New names "core-avx2".
17348 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17349 PROCESSOR_CORE_HASWELL.
17350 (ix86_issue_rate): New case.
17351 (ia32_multipass_dfa_lookahead): Likewise.
17352 (ix86_sched_init_global): Likewise.
17354 * config/i386/i386.h (TARGET_HASWELL): New macro.
17355 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17356 (processor_type): New PROCESSOR_HASWELL.
17358 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17360 PR tree-optimization/55955
17361 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17362 *SHIFT_EXPR and *ROTATE_EXPR codes.
17364 PR tree-optimization/48766
17365 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17366 -ftrapv disable -fwrapv.
17368 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17371 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17372 etc. to 1 and not to __flash.
17373 Use LL suffix for __INT24_MAX__ with -mint8.
17374 Use ULL suffix for __UINT24_MAX__ with -mint8.
17376 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17378 * config/avr/avr-arch.h
17379 (struct base_arch_s): Use typedef avr_arch_t instead.
17380 (struct arch_info_s): Use typedef avr_arch_info_t instead.
17381 (struct mcu_type_s): Use typedef avr_mcu_t instead.
17382 * config/avr/avr.c: Same.
17383 * config/avr/avr-devices.c: Same.
17384 * config/avr/driver-avr.c: Same.
17385 * config/avr/gen-avr-mmcu-texi.c: Same.
17386 * config/avr/avr-mcus.def: Adjust comment.
17388 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
17390 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
17391 * config/aarch64/iterators.md (VALLDI): New.
17393 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
17394 Andi Kleen <ak@linux.intel.com>
17397 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
17398 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
17401 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17403 * config/avr/avr-stdint.h: Remove trailing blanks.
17404 * config/avr/avr-log.h: Same.
17405 * config/avr/avr-arch.h: Same.
17406 * config/avr/avr-devices.c: Same.
17407 * config/avr/avr-dimode.md: Same.
17408 * config/avr/predicates.md: Same.
17409 * config/avr/avr-c.c: Same. And fix typo.
17411 * config/avr/avr-protos.h: Same. And:
17412 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
17413 (init_cumulative_args): Rename to avr_init_cumulative_args.
17414 (expand_prologue): Rename to avr_expand_prologue.
17415 (expand_epilogue): Rename to avr_expand_epilogue.
17416 (adjust_insn_length): Rename to avr_adjust_insn_length.
17417 (notice_update_cc): Rename to avr_notice_update_cc.
17418 (final_prescan_insn): Rename to avr_final_prescan_insn.
17419 * config/avr/avr.c: Same.
17420 * config/avr/avr.h: Same.
17421 * config/avr/avr.md: Remove trailing blanks.
17422 (prologue): Use avr_expand_prologue.
17423 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
17425 2013-01-14 Richard Biener <rguenther@suse.de>
17427 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
17428 verify_location, collect_subblocks): New functions.
17429 (verify_gimple_in_cfg): Verify that locations only reference
17430 BLOCKs in the functions BLOCK tree.
17432 2013-01-14 Richard Biener <rguenther@suse.de>
17434 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
17436 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
17438 (insert_out_of_ssa_copy_on_edge): Likewise.
17439 (rewrite_close_phi_out_of_ssa): Likewise.
17440 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
17442 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
17443 propagated constants.
17444 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
17447 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17449 * config/avr/avr-modes.def: Add GPL copyright notice.
17451 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
17453 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
17454 MEMMODEL_MASK to determine memory model.
17455 (atomic_store<mode>): Ditto from operands[2].
17456 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
17458 2013-01-13 Jakub Jelinek <jakub@redhat.com>
17461 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
17462 (fold_gimple_assign): Don't call unshare_expr here.
17463 (fold_ctor_reference): Call unshare_expr.
17465 2013-01-13 Terry Guo <terry.guo@arm.com>
17467 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
17468 * doc/fragments.texi: Document MULTILIB_REUSE.
17469 * gcc.c (multilib_reuse): New internal spec.
17470 (set_multilib_dir): Also search multilib from multilib_reuse.
17471 * genmultilib (tmpmultilib3): Refactor code.
17472 (tmpmultilib4): Ditto.
17473 (multilib_reuse): New multilib argument.
17475 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
17477 * Makefile.in: Update copyright.
17479 2013-01-12 Tom de Vries <tom@codesourcery.com>
17481 PR middle-end/55890
17482 * calls.c (expand_call): Check if arg_nr is valid.
17484 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17486 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
17487 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
17488 documentation. Add missing '__' in front of
17489 __builtin_ia32_packssdw256.
17491 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17494 * config/s390/s390.c (s390_preferred_reload_class): Do not return
17495 NO_REGS for larl operands.
17496 (s390_reload_larl_operand): Use s390_load_address instead of
17499 2013-01-11 Richard Biener <rguenther@suse.de>
17501 * tree-cfg.c (verify_node_sharing_1): Split out from ...
17502 (verify_node_sharing): ... here.
17503 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
17505 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
17507 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
17508 Substitute TREECHECKING.
17509 * configure: Regenerate.
17510 * Makefile.in (TREECHECKING): New.
17512 2013-01-11 Richard Guenther <rguenther@suse.de>
17514 PR tree-optimization/44061
17515 * tree-vrp.c (extract_range_basic): Compute zero as
17516 value-range for __builtin_constant_p of function parameters.
17518 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
17520 Update copyright years.
17522 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
17524 PR rtl-optimization/55672
17525 * lra-eliminations.c (mark_not_eliminable): Permit addition with
17526 const to be eliminable.
17528 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
17530 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
17531 * configure: Regenerate.
17533 2013-01-10 Richard Biener <rguenther@suse.de>
17535 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
17537 2013-01-10 Richard Biener <rguenther@suse.de>
17540 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
17541 locations for virtual PHI arguments.
17542 (rewrite_update_phi_arguments): Likewise.
17544 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
17546 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
17549 2013-01-10 Jakub Jelinek <jakub@redhat.com>
17551 PR tree-optimization/55921
17552 * tree-complex.c (expand_complex_asm): New function.
17553 (expand_complex_operations_1): Call it for GIMPLE_ASM.
17555 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17558 * config/s390/s390.c (s390_symref_operand_p)
17559 (s390_loadrelative_operand_p): Merge the two functions.
17560 (s390_check_qrst_address, print_operand_address): Add parameters
17561 to s390_loadrelative_operand_p invokation.
17562 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
17563 (s390_reload_larl_operand, s390_secondary_reload): Use
17564 s390_loadrelative_operand_p instead of s390_symref_operand_p.
17565 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
17567 2013-01-09 Mike Stump <mikestump@comcast.net>
17569 * dse.c (record_store): Remove unnecessary assert.
17571 2013-01-09 Jan Hubicka <jh@suse.cz>
17573 PR tree-optimization/55569
17574 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
17575 * cfgloop.h (scale_loop_profile): Likewise.
17577 2013-01-09 Jan Hubicka <jh@suse.cz>
17580 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
17582 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
17584 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
17586 PR middle-end/55114
17587 * expr.h (maybe_emit_group_store): Declare.
17588 * expr.c (maybe_emit_group_store): New function.
17589 * builtins.c (expand_builtin_int_roundingfn): Call it.
17590 (expand_builtin_int_roundingfn_2): Likewise.
17592 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
17594 PR rtl-optimization/55829
17595 * lra-constraints.c (match_reload): Add code for absent output.
17596 (curr_insn_transform): Add code for reloads of matched inputs
17599 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
17601 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
17602 attribute of movddup insn to DF.
17603 (*vec_interleave_lowv2df): Ditto.
17604 (vec_dupv2df): Ditto.
17606 2013-01-09 Jan Hubicka <jh@suse.cz>
17608 PR tree-optimiation/55875
17609 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
17610 EVERY_ITERATION parameter.
17611 (number_of_iterations_exit): Check if exit is executed every iteration.
17612 (idx_infer_loop_bounds): Similarly here.
17613 (n_of_executions_at_most): Simplify
17614 to only test for cases where statement is dominated by the
17615 particular bound; handle correctly the "postdominance" test.
17616 (scev_probably_wraps_p): Use max loop iterations info
17617 as a global bound first.
17619 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
17620 Nick Clifton <nickc@redhat.com>
17622 * config/v850/v850.md (cbranchsf4): New pattern.
17623 (cstoresf4): New pattern.
17624 (cbranchdf4): New pattern.
17625 (cstoredf4): New pattern.
17626 (movsicc): Disallow floating point comparisons.
17627 (cmpsf_le_insn): Fix order of operators.
17628 (cmpsf_lt_insn): Likewise.
17629 (cmpsf_eq_insn): Likewise.
17630 (cmpdf_le_insn): Likewise.
17631 (cmpdf_lt_insn): Likewise.
17632 (cmpdf_eq_insn): Likewise.
17633 (cmpsf_ge_insn): Use LE comparison.
17634 (cmpdf_ge_insn): Likewise.
17635 (cmpsf_gt_insn): Use LT comparison.
17636 (cmpdf_gt_insn): Likewise.
17637 (cmpsf_ne_insn): Delete pattern.
17638 (cmpdf_ne_insn): Delete pattern.
17639 * config/v850/v850.c (v850_gen_float_compare): Use
17640 gen_cmpdf_eq_insn for NE comparison.
17641 (v850_float_z_comparison_operator)
17642 (v850_float_nz_comparison_operator): Move from here ...
17643 * config/v850/predicates.md: ... to here. Move GT and GE
17644 comparisons into v850_float_z_comparison_operator.
17645 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
17647 (v850_float_nz_comparison_operator): Likewise.
17649 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17651 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
17652 with calls to gen_insvsi/gen_insvdi.
17654 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17656 * config/i386/i386.c (initial_ix86_tune_features): Set up
17657 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
17659 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
17660 Jakub Jelinek <jakub@redhat.com>
17662 PR tree-optimization/48189
17663 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
17664 If nitercst is 0, don't predict the exit edge.
17666 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17668 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
17669 in asm_fprintf with reg_names.
17670 (aarch64_print_operand_address): Likewise.
17671 (aarch64_return_addr): Likewise.
17672 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
17674 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17676 * config/pa/pa.h (VAL_U6_BITS_P): Define.
17677 (INT_U6_BITS): Likewise.
17678 * config/pa/predicates.md (uint6_operand): New predicate.
17679 (shift5_operand, shift6_operand): Likewise.
17680 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
17682 (lshrdi3): Use shift6_operand.
17683 (shrpsi4, shrpdi4): New insn patterns.
17684 (extzv): Delete expander.
17685 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
17686 predicates in unamed zero extract patterns. Tighten common constraint.
17687 (extv): Delete expander.
17688 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
17689 predicates in unamed sign extract patterns. Tighten common constraint.
17690 (insv): Delete expander.
17691 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
17692 predicates in unamed insert patterns. Tighten common constraint.
17693 Change uint32_operand predicate to uint6_operand predicate in unamed
17694 DImode pattern to insert constant values of type 1...1xxxx.
17696 2013-01-04 Jan Hubicka <jh@suse.cz>
17698 PR tree-optimization/55823
17699 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
17702 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17703 Uros Bizjak <ubizjak@gmail.com>
17705 PR rtl-optimization/55845
17706 * df-problems.c (can_move_insns_across): Stop scanning at
17707 volatile_insn_p source instruction or give up if
17708 across_from .. across_to range contains any volatile_insn_p
17711 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17713 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
17714 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
17716 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
17717 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
17719 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17722 * asan.c (asan_clear_shadow): New function.
17723 (asan_emit_stack_protection): Use it.
17725 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
17727 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
17728 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
17729 with tab instead of space.
17731 2013-01-08 Nick Clifton <nickc@redhat.com>
17733 * config/rl78/rl78.c (rl78_expand_prologue): Always select
17734 register bank 0 at the start of an interrupt handler.
17735 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
17738 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17740 * config/aarch64/aarch64-simd.md
17741 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
17742 (aarch64_simd_bsl): Likewise.
17743 (aarch64_vcond_internal<mode>): Likewise.
17744 (vcond<mode><mode>): Likewise.
17745 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
17746 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
17748 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
17750 * config/aarch64/aarch64-builtins.c
17751 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
17753 2013-01-08 Martin Jambor <mjambor@suse.cz>
17756 * tree-sra.c (analyze_access_subtree): Return true also after
17757 potentially creating a debug-only replacement.
17759 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17761 PR middle-end/55890
17762 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
17764 PR tree-optimization/54120
17765 * tree-vrp.c (range_fits_type_p): Don't allow
17766 src_precision < precision from signed vr to unsigned_p
17767 if vr->min or vr->max is negative.
17768 (simplify_float_conversion_using_ranges): Test can_float_p
17769 against CODE_FOR_nothing.
17771 2013-01-08 Jakub Jelinek <jakub@redhat.com>
17772 Richard Biener <rguenther@suse.de>
17774 PR middle-end/55851
17775 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
17776 types instead of just INTEGER_TYPE types.
17778 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
17780 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
17783 2013-01-07 Steve Ellcey <sellcey@mips.com>
17786 * config/mips/mips.opt: Change mad to mmad to match documentation.
17788 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17791 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
17792 .progmemx.data now.
17794 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17797 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
17798 (avr_addrspace_t): Add .section_name field.
17799 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
17801 (avr_addrspace): Same. Initialize .section_name. Remove last
17802 NULL entry. Put __memx into .progmemx.data.
17803 (progmem_section_prefix): Remove.
17804 (avr_asm_init_sections): No need to initialize progmem_section.
17805 (avr_asm_named_section): Use avr_addrspace[].section_name to get
17806 section name prefix.
17807 (avr_asm_select_section): Ditto. And use get_unnamed_section to
17808 retrieve the progmem section.
17809 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
17810 boundary to run over avr_addrspace[].
17811 (avr_register_target_pragmas): Ditto.
17813 2013-01-06 Jakub Jelinek <jakub@redhat.com>
17815 * varasm.c (output_constant_def_contents): For asan_protect_global
17816 protected strings, adjust DECL_ALIGN if needed, before testing for
17818 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
17819 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
17821 (output_object_block): For asan protected decls, emit asan padding
17822 after their contents.
17823 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
17824 (asan_finish_file): Test it here instead.
17826 2013-01-07 Nick Clifton <nickc@redhat.com>
17827 Matthias Klose <doko@debian.org>
17828 Doug Kwan <dougkwan@google.com>
17829 H.J. Lu <hongjiu.lu@intel.com>
17832 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
17834 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
17836 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
17838 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
17840 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
17842 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17845 * doc/install.texi (Cross-Compiler-Specific Options): Document
17848 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
17850 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
17851 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
17852 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
17853 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
17854 vqmovun_high_s64): Fix source operand number and update copyright.
17856 2013-01-07 Richard Biener <rguenther@suse.de>
17858 PR middle-end/55890
17859 * gimple.h (gimple_call_builtin_p): New overload.
17860 * gimple.c (validate_call): New function.
17861 (gimple_call_builtin_p): Likewise.
17862 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17863 Use gimple_call_builtin_p.
17864 (find_func_clobbers): Likewise.
17865 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
17866 (strlen_optimize_stmt): Likewise.
17868 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17870 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
17871 (vld1q_dup_*): Likewise.
17872 (vld1_*): Likewise.
17873 (vld1q_*): Likewise.
17874 (vld1_lane_*): Likewise.
17875 (vld1q_lane_*): Likewise.
17877 2013-01-07 Richard Biener <rguenther@suse.de>
17879 * lto-streamer.h (LTO_minor_version): Bump to 2.
17881 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
17883 * config/aarch64/aarch64-protos.h
17884 (aarch64_const_double_zero_rtx_p): Rename to...
17885 (aarch64_float_const_zero_rtx_p): ...this.
17886 (aarch64_float_const_representable_p): New.
17887 (aarch64_output_simd_mov_immediate): Likewise.
17888 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
17889 move immediate case.
17890 * config/aarch64/aarch64.c
17891 (aarch64_const_double_zero_rtx_p): Rename to...
17892 (aarch64_float_const_zero_rtx_p): ...this.
17893 (aarch64_print_operand): Allow printing of new constants.
17894 (aarch64_valid_floating_const): New.
17895 (aarch64_legitimate_constant_p): Check for valid floating-point
17897 (aarch64_simd_valid_immediate): Likewise.
17898 (aarch64_vect_float_const_representable_p): New.
17899 (aarch64_float_const_representable_p): Likewise.
17900 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
17901 (aarch64_output_simd_mov_immediate): New.
17902 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
17903 (*movdf_aarch64): Likewise.
17904 * config/aarch64/constraints.md (Ufc): New.
17905 (Y): call aarch64_float_const_zero_rtx.
17906 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
17908 2013-01-07 Richard Biener <rguenther@suse.de>
17910 PR tree-optimization/55888
17911 PR tree-optimization/55862
17912 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
17913 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
17914 not if it is contained therein.
17916 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17918 * config/avr/t-avr: Typo.
17920 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
17923 * config/avr/t-avr: Don't automatically rebuild
17924 $(srcdir)/config/avr/t-multilib
17925 $(srcdir)/config/avr/avr-tables.opt
17926 $(srcdir)/doc/avr-mmcu.texi
17927 (avr-mcus): New phony target to build them on request.
17928 (s-avr-mlib, s-avr-mmcu-texi): Remove.
17929 * avr/avr-mcus.def: Adjust comments.
17931 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
17933 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
17935 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17937 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
17939 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
17941 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
17943 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
17945 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
17946 to generate profiling.
17947 * config/rs6000/aix64.h (LIB_SPEC): Same.
17949 2013-01-04 Andrew Pinski <apinski@cavium.com>
17951 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
17953 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
17955 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
17957 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
17959 (ix86_expand_move): Ditto.
17960 (ix86_zero_extend_to_Pmode): Ditto.
17961 (ix86_expand_call): Ditto.
17962 (ix86_expand_special_args_builtin): Ditto.
17963 (ix86_expand_builtin): Ditto.
17965 2013-01-04 Richard Biener <rguenther@suse.de>
17967 PR tree-optimization/55862
17968 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
17969 translating them through PHI nodes.
17971 2013-01-04 Martin Jambor <mjambor@suse.cz>
17973 PR tree-optimization/55755
17974 * tree-sra.c (sra_modify_assign): Do not check that an access has no
17975 children when trying to avoid producing a VIEW_CONVERT_EXPR.
17977 2013-01-04 Marek Polacek <polacek@redhat.com>
17979 PR middle-end/55859
17980 * opts.c (default_options_optimization): Clarify error message.
17982 2013-01-04 Richard Biener <rguenther@suse.de>
17984 PR middle-end/55863
17985 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
17988 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17991 * config/pa/pa.md (movsi): Revert previous change.
17992 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
17995 2013-01-03 Richard Henderson <rth@redhat.com>
17997 * config/i386/i386.c (ix86_expand_move): Always assign to op1
17998 after eliminating TLS symbols.
18000 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18003 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18004 * graphite-poly.c (debug_gmp_value): Likewise.
18006 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18009 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18010 selected code model, define __code_mode_small__, __code_model_medium__,
18011 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18012 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18013 xchg temporary register with %k. Declare temporary register as
18015 [__x86_64__]: For medium and large code models, preserve %rbx register.
18017 2013-01-03 Richard Biener <rguenther@suse.de>
18019 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18020 (dump_subscript): Adjust.
18021 (finalize_ddr_dependent): Do not dump redundant info.
18022 (analyze_siv_subscript): Adjust.
18023 (subscript_dependence_tester): Likewise.
18024 (compute_affine_dependence): Likewise.
18026 2013-01-03 Richard Biener <rguenther@suse.de>
18029 2013-01-03 Richard Biener <rguenther@suse.de>
18031 PR tree-optimization/55857
18032 * tree-vect-stmts.c (vectorizable_load): Do not setup
18033 re-alignment for invariant loads.
18035 2013-01-02 Richard Biener <rguenther@suse.de>
18037 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18038 invariant load do not generate a vector load from the scalar location.
18040 2013-01-03 Richard Biener <rguenther@suse.de>
18042 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18043 for not vectorizing.
18044 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18045 not build INDIRECT_REFs, call get_name once only.
18046 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18047 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18049 2013-01-03 Richard Biener <rguenther@suse.de>
18051 PR tree-optimization/55857
18052 * tree-vect-stmts.c (vectorizable_load): Do not setup
18053 re-alignment for invariant loads.
18055 2013-01-03 Richard Biener <rguenther@suse.de>
18058 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18059 prefer a built-in decl.
18061 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18063 * gcc.c (process_command): Update copyright notice dates.
18064 * gcov.c (print_version): Likewise.
18065 * gcov-dump.c (print_version): Likewise.
18067 PR rtl-optimization/55838
18068 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18069 iv0.step, iv1.step and step.
18071 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18072 Marc Glisse <marc.glisse@inria.fr>
18074 PR tree-optimization/55832
18075 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18076 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18077 integer_{one,zero}_node.
18079 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18082 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18083 * var-tracking.c (reverse_op): Don't add reverse ops to
18084 VALUEs that have already
18085 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18087 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18089 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18091 2013-01-02 Teresa Johnson <tejohnson@google.com>
18093 * dumpfile.c (dump_loc): Print filename with location.
18094 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18095 new location_t parameter to emit complete unroll message with
18096 new dump framework.
18097 (canonicalize_loop_induction_variables): Compute loops location
18098 and pass to try_unroll_loop_completely.
18099 * loop-unroll.c (report_unroll_peel): New function.
18100 (peel_loops_completely): Use new dump format with location
18101 for main dumpfile message, and invoke report_unroll_peel on success.
18102 (decide_unrolling_and_peeling): Ditto.
18103 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18104 by report_unroll_peel.
18105 (decide_peel_completely): Ditto.
18106 (decide_unroll_constant_iterations): Ditto.
18107 (decide_unroll_runtime_iterations): Ditto.
18108 (decide_peel_simple): Ditto.
18109 (decide_unroll_stupid): Ditto.
18110 * cfgloop.c (get_loop_location): New function.
18111 * cfgloop.h (get_loop_location): Declare.
18113 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18115 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18118 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18120 PR middle-end/55198
18121 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18122 BLKmode objects when EXPAND_MEMORY is specified.
18124 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18126 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18128 (fold_builtin_cpu): Do not share cpu model decls across statements.
18130 2013-01-02 Jason Merrill <jason@redhat.com>
18133 * tree.c (build_array_type_1): Revert earlier change.
18135 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18137 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18139 * config/aarch64/aarch64-tune.md: Re-generate.
18141 2013-01-02 Richard Biener <rguenther@suse.de>
18143 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18144 invariant load do not generate a vector load from the scalar location.
18146 2013-01-02 Richard Biener <rguenther@suse.de>
18149 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18150 * configure: Regenerate.
18152 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18154 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18155 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18156 (expand_builtin_int_roundingfn_2): Keep the original target around
18157 for the fallback case.
18159 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18161 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18162 to be clear for sign changes.
18164 2013-01-01 Jan Hubicka <jh@suse.cz>
18166 * ipa-inline-analysis.c: Fix formatting.
18168 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18170 PR tree-optimization/55831
18171 * tree-vect-loop.c (get_initial_def_for_induction): Use
18172 gsi_after_labels instead of gsi_start_bb.
18174 Copyright (C) 2013 Free Software Foundation, Inc.
18176 Copying and distribution of this file, with or without modification,
18177 are permitted in any medium without royalty provided the copyright
18178 notice and this notice are preserved.