1 2013-09-23 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58464
4 * tree-ssa-pre.c (phi_trans_lookup): Remove.
5 (phi_trans_add): Change to add conditionally on being not
7 (phi_translate_1): Remove recursion detection here.
8 (phi_translate): Pre-seed the cache with NULL to catch
9 recursion here in a more generic way.
10 (bitmap_find_leader): Adjust comment.
11 (get_representative_for): Dump value-numbers.
12 (create_expression_by_pieces): Likewise.
13 (insert_into_preds_of_block): Likewise.
15 2013-09-23 Christian Bruel <christian.bruel@st.com>
18 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
19 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
21 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
24 2013-09-20 Renlin Li <renlin.li@arm.com>
26 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
27 (aarch64_expand_epilogue): Likewise.
28 (aarch64_legitimize_reload_address): Likewise.
31 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
33 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
34 NULL_TREE before pushing them onto the vector. Likewise for labels.
36 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
38 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
39 comparison operators when -fno-trapping-math is in effect.
40 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
41 comparison operators in TFmode and assert that unsupported operators
43 (ia64_print_operand): Likewise.
45 2013-09-21 Jan Hubicka <jh@suse.cz>
47 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
48 (sse_typeless_stores): Enable for core
49 (sse_load0_by_pxor): Likewise.
50 (four_jump_limit): Disable for core.
51 (pad_returns): Likewise.
52 (avoid_vector_decode): Likewise.
53 (fuse_cmp_and_branch): Enable for cores.
54 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
56 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
59 * config/pa/pa.c (pa_option_override): Disable auto increment and
60 decrement instructions until reload is completed.
62 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
63 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
64 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
66 2013-09-20 DJ Delorie <dj@redhat.com>
67 Nick Clifton <nickc@redhat.com>
69 * config/rl78/rl78.c: Various whitespace and comment tweaks.
70 (need_to_save): Save bank 0 on interrupts.
71 (characterize_address): Strip far address wrappers.
72 (rl78_as_legitimate_address): Likewise.
73 (transcode_memory_rtx): Likewise.
74 (rl78_peep_movhi_p): Disable this peephole after devirt.
75 (rl78_propogate_register_origins): Forget all origins when a
77 * config/rl78/rl78-virt.md: Various whitespace tweaks.
78 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
80 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
81 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
82 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
84 * config/rl78/constraints.md: Various whitespace and paren tweaks.
86 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
88 * config/pa/pa.md: In "scc" insn patterns, change output template to
89 handle const0_rtx in reg_or_0_operand operands.
91 2013-09-20 Martin Husemann <martin@NetBSD.org>
94 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
95 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
97 2013-09-20 Richard Biener <rguenther@suse.de>
100 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
101 remembering SSA name version and block index.
102 (new_scev_info_str): Adjust.
103 (hash_scev_info): Likewise. Also hash the block index.
104 (eq_scev_info): Adjust.
105 (find_var_scev_info): Likewise.
106 (struct instantiate_cache_entry): Remove.
107 (struct instantiate_cache_type): Use a htab to map name, block
109 (instantiate_cache_type::~instantiate_cache_type): Adjust.
110 (get_instantiated_value_entry): Likewise.
111 (hash_idx_scev_info, eq_idx_scev_info): New functions.
112 (instantiate_scev_name): Adjust.
114 2013-09-20 Jeff Law <law@redhat.com>
116 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
118 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
120 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
121 Call aarch64_simd_expand_args to update op[argc].
123 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
125 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
128 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
130 * gengtype.c (file_rules): Added rule for *.cc files.
131 (get_output_file_with_visibility): Give fatal message when no
134 2013-09-20 Renlin Li <renlin.li@arm.com>
136 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
137 (aarch64_expand_epilogue): Likewise.
138 (aarch64_legitimize_reload_address): Likewise.
140 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
143 * expr.c (expand_assignment): Remove misalignp code path.
145 2013-09-20 Marek Polacek <polacek@redhat.com>
148 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
149 TYPE_PRECISION. Add asserts.
151 2013-09-20 Richard Biener <rguenther@suse.de>
153 PR tree-optimization/58453
154 * tree-loop-distribution.c (distribute_loop): Apply the cost
155 model for -ftree-loop-distribute-patterns, too.
157 2013-09-20 Richard Biener <rguenther@suse.de>
160 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
161 make type comparison less strict.
163 2013-09-20 Alan Modra <amodra@gmail.com>
165 * configure: Regenerate.
166 * aclocal.m4: Regenerate.
168 2013-09-20 Marek Polacek <polacek@redhat.com>
171 * doc/extend.texi: Document that attribute used is meant to be used
172 on variables with static storage duration.
174 2013-09-19 Jakub Jelinek <jakub@redhat.com>
176 PR tree-optimization/58472
177 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
178 simd_lane_access set inv_p = false.
179 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
180 the simduid magic VAR_DECL.
182 2013-09-19 Jan Hubicka <jh@suse.cz>
184 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
186 2013-09-17 Jeff Law <law@redhat.com>
188 * tree-ssa-dom.c (record_temporary_equivalences): New function
189 split out of dom_opt_dom_walker::after_dom_children.
190 (dom_opt_dom_walker::thread_across_edge): Move common code
191 in here from dom_opt_dom_walker::after_dom_children.
192 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
194 2013-09-19 Jan Hubicka <jh@suse.cz>
196 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
197 (TARGET_GENERIC): Use PROCESOR_GENERIC
198 (enum processor_type): Unify generic32 and 64.
199 * i386.md (cpu): Likewise.
200 * x86-tune.def (use_leave): Enable for generic32.
201 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
202 * athlon.md: Change generic64 to generic in all occurences.
203 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
204 (ix86_target_macros_internal): Likewise.
205 * driver-i386.c (host_detect_local_cpu): Likewise.
206 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
207 (generic_memcpy, generic_memset, generic_cost): This one.
208 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
209 (m_GENERIC32, m_GENERIC64): Remove.
210 (m_GENERIC): Turn into one flag.
211 (processor_target): Unify generic tunnings.
212 (ix86_option_override_internal): Replace generic32/64 by generic.
213 (ix86_issue_rate): Likewise.
214 (ix86_adjust_cost): Likewise.
216 2013-09-19 Jan Hubicka <jh@suse.cz>
218 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
221 2013-09-19 Jakub Jelinek <jakub@redhat.com>
223 * omp-low.c (expand_omp_sections): Always pass len - 1 to
224 GOMP_sections_start, even if !exit_reachable.
226 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
228 * lra-constraints.c (need_for_all_save_p): Use macro
229 HARD_REGNO_CALL_PART_CLOBBERED.
230 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
231 set up pseudo conflict hard regs.
233 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
236 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
239 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
241 PR rtl-optimization/58438
242 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
243 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
244 from previous subpasses.
246 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
248 * arm.c (arm_get_frame_offsets): Validate architecture supports
249 LDRD/STRD before accepting the tuning preference.
250 (arm_expand_prologue): Likewise.
251 (arm_expand_epilogue): Likewise.
253 2013-09-18 Richard Biener <rguenther@suse.de>
255 PR tree-optimization/58417
256 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
257 have chrecs with symbols defined in the loop as operands.
258 (chrec_fold_multiply): Likewise.
259 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
260 parameters before folding binary operations.
261 (struct instantiate_cache_entry_hasher): Remove.
262 (struct instantiate_cache_type): Use a pointer-map.
263 (instantiate_cache_type::instantiate_cache_type): New function.
264 (instantiate_cache_type::get): Likewise.
265 (instantiate_cache_type::set): Likewise.
266 (instantiate_cache_type::~instantiate_cache_type): Adjust.
267 (get_instantiated_value_entry): Likewise.
268 (global_cache): New global.
269 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
270 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
271 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
272 (instantiate_scev_name): Adjust.
273 (instantiate_scev): Construct global instead of local cache.
274 (resolve_mixers): Likewise.
276 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
277 Paolo Carlini <paolo.carlini@oracle.com>
280 * doc/implement-cxx.texi: Fix references to the C++ standards.
282 2013-09-18 Jakub Jelinek <jakub@redhat.com>
284 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
285 * tree-vect-data-refs.c (vect_analyze_data_refs): For
286 simd_lane_access drs, update also DR_ALIGNED_TO.
288 2013-09-18 Marek Polacek <polacek@redhat.com>
291 * doc/extend.texi: Document no_sanitize_undefined attribute.
292 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
293 no_sanitize_undefined attribute.
295 2013-09-18 Nick Clifton <nickc@redhat.com>
297 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
298 (ASM_DECLARE_FUNCTION_NAME): Define.
300 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
302 * compare-elim.c (find_comparison_dom_walker): New class
303 (find_comparisons_in_bb): Rename to
304 find_comparison_dom_walker::before_dom_children
305 (find_comparisons): Adjust
306 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
308 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
309 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
310 (init_walk_dominator_tree): Remove declaration.
311 (fini_walk_dominator_tree): Remove declaration.
312 * fwprop.c (single_def_use_dom_walker): New class
313 (single_def_use_enter_block): Convert to
314 single_def_use_dom_walker::before_dom_children.
315 (single_def_use_leave_block): Convert to
316 single_def_use_dom_walker::after_dom_children.
317 (build_single_def_use_links): Adjust.
318 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
320 (find_candidates_in_block): Convert to
321 find_candidates_dom_walker::before_dom_children.
322 (execute_strength_reduction): Adjust.
323 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
324 (sese_dom_walker): New class.
325 (sese_dom_walker::sese_dom_walker): New constructor.
326 (sese_dom_walker::~sese_dom_walker): New destructor.
327 (build_sese_conditions_before): Convert to
328 sese_dom_walker::before_dom_children.
329 (build_sese_conditions_after): Convert to
330 sese_dom_walker::after_dom_children.
331 (build_poly_scop): Adjust
332 * tree-into-ssa.c (rewrite_dom_walker): New class
333 (rewrite_enter_block): Convert to
334 rewrite_dom_walker::before_dom_children.
335 (rewrite_leave_block): Convert to
336 rewrite_dom_walker::after_dom_children.
337 (rewrite_update_dom_walker): New class.
338 (rewrite_update_enter_block): Convert to
339 rewrite_update_dom_walker::before_dom_children.
340 (rewrite_update_leave_block): Convert to
341 rewrite_update_dom_walker::after_dom_children.
342 (rewrite_blocks, rewrite_into_ssa): Adjust.
343 (mark_def_dom_walker): New class.
344 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
345 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
346 (mark_def_sites_blocks): Convert to
347 mark_def_dom_walker::before_dom_children.
348 (mark_def_site_blocks): Remove.
349 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
350 (tree_ssa_dominator_optimize): Adjust.
351 (dom_thread_across_edge): Convert to method
352 dom_opt_dom_walker::thread_across_edge.
353 (dom_opt_enter_block): Convert to member function
354 dom_opt_dom_walker::before_dom_children.
355 (dom_opt_leave_block): Convert to member function
356 dom_opt_dom_walker::after_dom_children.
357 * tree-ssa-dse.c (dse_dom_walker): New class.
358 (dse_enter_block): Convert to member function
359 dse_dom_walker::before_dom_children.
360 (tree_ssa_dse): Adjust.
361 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
362 (determine_invariantness_stmt): Convert to method
363 invariantness_dom_walker::before_dom_children.
364 (determine_invariantness): Remove
365 (move_computations_dom_walker): New class.
366 (move_computations_stmt): Convert to method
367 move_computations_dom_walker::before_dom_children.
368 (move_computations, tree_ssa_lim): Adjust.
369 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
370 (nt_init_block): Make method
371 notrappping_dom_walker::before_dom_children.
372 (nt_fini_block): Make
373 method nontrapping_dom_walker::after_dom_children.
374 (get_non_trapping): Adjust.
375 * tree-ssa-pre.c (eliminate_dom_walker): New class.
376 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
377 (eliminate_leave_block): Make method.
378 eliminate_dom_walker::after_dom_children.
380 * tree-ssa-strlen.c (strlen_dom_walker): New class.
381 (strlen_enter_block): Make method
382 strlen_dom_walker::before_dom_children.
383 (strlen_leave_block): Make
384 method strlen_dom_walker::after_dom_children.
385 (tree_ssa_strlen): Adjust.
386 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
387 (tree_ssa_uncprop): Adjust.
388 (uncprop_leave_block): Make method
389 uncprop_dom_walker::after_dom_children.
390 (uncprop_leave_block): Make method
391 uncprop_dom_walker::before_dom_children.
393 2013-09-18 Bin Cheng <bin.cheng@arm.com>
395 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
396 before branch in same basic block. Check both src and dest of
399 2013-09-17 Nick Clifton <nickc@redhat.com>
401 * config/rl78/rl78-real.md (bf): New pattern.
403 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
404 (rl78_print_operand): Do not put a # before a %B.
405 * config/rl78/rl78.opt: Tweak doc strings.
407 2013-09-17 DJ Delorie <dj@redhat.com>
409 * config/rl78/constraints.md (Wcv): Allow up to $r31.
410 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
411 (rl78_option_override): Likewise, if -mallregs.
412 (is_virtual_register): Likewise.
413 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
414 (REGNO_OK_FOR_BASE_P): Likewise.
415 * config/rl78/rl78.opt (-mallregs): New.
417 2013-09-17 Nick Clifton <nickc@redhat.com>
419 * config/rl78/rl78.c (need_to_save): Change return type to bool.
420 For interrupt functions: save all call clobbered registers if the
421 interrupt handler is not a leaf function.
422 (rl78_expand_prologue): Always recompute the frame information.
423 For interrupt functions: only select bank 0 if one of the bank 0
424 registers is going to be psuhed.
426 2013-09-17 DJ Delorie <dj@redhat.com>
428 * config/rl78/constraints.md: For each W* constraint, rename to C*
429 and create a W* constraint that checks for an optional ES: prefix
431 * config/rl78/rl78.md (UNS_ES_ADDR): New.
432 (es_addr): New. Used to wrap far addresses.
433 * config/rl78/rl78-protos.h (rl78_es_addr): New.
435 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
436 wrapped far addresses.
437 (rl78_print_operand_1): Unwrap far addresses before processing.
438 (rl78_lo16): Wrap far addresses in unspecs.
441 (insn_ok_now): Check for not-yet-wrapped far addresses.
442 (transcode_memory_rtx): Properly re-wrap far addresses.
444 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
446 * config/sparc/t-rtems: Add leon3 multilibs.
448 2013-09-17 Cong Hou <congh@google.com>
450 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
451 when checking the dot production pattern. The type of rhs operand
452 of multiply is now checked correctly.
454 2013-09-17 Jeff Law <law@redhat.com>
456 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
457 edge implied equivalences into successor phis.
458 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
459 here from tree-ssa-threadedge.c.
460 (mark_threaded_blocks): When threading through a joiner, if both
461 successors of the joiner's clone reach the same block, verify the
462 PHI arguments are equal. If not, cancel the jump threading request.
463 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
464 tree-ssa-threadupdate.c
465 (thread_across_edge): Don't check PHI argument equality when
466 threading through joiner block here.
468 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
470 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
471 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
472 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
473 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
474 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
475 (has_undefined_value_p): New. Pass dependant parts of
476 ssa_undefined_value_p.
477 (uninit_undefined_value_p): Use has_undefined_value_p.
478 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
479 make_pass_early_warn_uninitialized): Move from tree-ssa.c
480 * tree-ssa.h: Adjust prototypes
482 2013-09-17 Jan Hubicka <jh@suse.cz>
485 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
486 * ipa-inline.c (can_inline_edge_p): Do not downgrade
487 FUNCTION_NOT_OPTIMIZED.
488 * ipa-inline-analysis.c (compute_inline_parameters): Function
489 not optimized is not inlinable unless it is alwaysinline.
490 (inline_analyze_function): Force calls in not optimized
491 function not inlinable.
493 2013-09-17 Jan Hubicka <jh@suse.cz>
496 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
498 * ipa.c (function_and_variable_visibility): Likewise.
499 * ipa-profile.c (ipa_profile): Likewise.
501 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
504 * cgraph.c (cgraph_function_body_availability): Check for ifunc
505 attribute, and don't inline the resolver in this case.
507 2013-09-17 Teresa Johnson <tejohnson@google.com>
509 * coverage.c (get_coverage_counts): Add missing newline.
511 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
513 PR tree-optimization/58088
514 * fold-const.c (mask_with_trailing_zeros): New function.
515 (fold_binary_loc): Make sure we don't recurse infinitely
516 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
517 Use mask_with_trailing_zeros where appropriate.
519 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
521 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
522 of 'prev' var to get better distance estimation.
524 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
526 * tree-inline.h (struct copy_body_data): Add transform_parameter.
527 * tree-inline.c (is_parameter_of): New predicate.
528 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
529 a parameter has been remapped.
530 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
531 (optimize_inline_calls): Initialize transform_parameter.
532 (copy_gimple_seq_and_replace_locals): Likewise.
533 (tree_function_versioning): Likewise.
534 (maybe_inline_call_in_expr): Likewise.
536 2013-09-17 Nick Clifton <nickc@redhat.com>
538 * config/msp430/msp430-protos.h: Add prototypes for new functions.
539 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
541 (is_attr_func): New function.
542 (msp430_is_interrupt_func): New function.
543 (is_naked_func): New function.
544 (is_reentrant_func): New function.
545 (is_critical_func): New function.
546 (msp430_start_function): Add annotations for function attributes.
547 (msp430_attr): New function.
548 (msp430_attribute_table): New.
549 (msp430_function_section): New function.
550 (TARGET_ASM_FUNCTION_SECTION): Define.
551 (msp430_builtin): New enum.
552 (msp430_init_builtins): New function.
553 (msp430_builtin_devl): New function.
554 (msp430_expand_builtin): New function.
555 (TARGET_INIT_BUILTINS): Define.
556 (TARGET_EXPAND_BUILTINS): Define.
557 (TARGET_BUILTIN_DECL): Define.
558 (msp430_expand_prologue): Add support for naked, interrupt,
559 critical and reentrant functions.
560 (msp430_expand_epilogue): Likewise.
561 (msp430_print_operand): Handle 'O' character.
562 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
564 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
565 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
566 (pushm): Use a 'n' rather than an 'i' constraint.
567 (msp_return): Add generation of the interrupt return instruction.
568 (disable_interrupts): New pattern.
569 (enable_interrupts): New pattern.
570 (push_intr_state): New pattern.
571 (pop_intr_state): New pattern.
572 (bic_SR): New pattern.
573 (bis_SR): New pattern.
574 * doc/extend.texi: Document MSP430 function attributes and builtin
577 2013-09-17 Richard Biener <rguenther@suse.de>
579 PR tree-optimization/58432
580 * tree-loop-distribution.c (tree_loop_distribution): Also
581 scan PHIs for outside loop uses and seed a partition from them.
583 2013-09-17 Bin Cheng <bin.cheng@arm.com>
585 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
586 (restructure_reference): Call backtrace_base_for_ref.
588 2013-09-17 Alan Modra <amodra@gmail.com>
591 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
593 2013-09-16 DJ Delorie <dj@redhat.com>
595 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
596 vregs location for RL78/G10.
597 (rl78_expand_prologue): Avoid SEL on G10.
598 (rl78_expand_epilogue): Likewise.
599 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
600 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
601 __RL78_G10__ when appropriate.
602 (ASM_SPEC): Pass -mg10 along to the assembler.
603 * config/rl78/rl78.md (sel_rb): Disable for G10.
604 * config/rl78/rl78.opt: Add -mg10 option.
605 * config/rl78/t-rl78: Add -mg10 multilib.
607 2013-09-16 Xinliang David Li <davidxl@google.com>
609 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
610 * omp-low.c (omp_max_vf): Ditto.
611 (expand_omp_simd): Ditto.
612 * tree-vectorizer.c (vectorize_loops): Ditto.
613 (gate_vect_slp): Ditto.
614 (gate_increase_alignment): Ditto.
615 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
616 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
617 (gate_tree_vectorize): Name change.
618 (tree_vectorize): Ditto.
619 (pass_vectorize::gate): Call new function.
620 (pass_vectorize::execute): Ditto.
621 opts.c: O3 default setting change.
622 (finish_options): Check new flag.
623 * doc/invoke.texi: Document new flags.
624 * common.opt: New flags.
626 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
628 * doc/tm.texi.in (Cond Exec Macros): Remove node.
629 (Condition Code): Don't reference it.
630 * doc/tm.texi: Regenerate.
632 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
635 * lra-constraints.c (undo_optional_reloads): Consider all optional
636 reload even if it did not get a hard reg.
638 2013-09-16 Teresa Johnson <tejohnson@google.com>
640 * dumpfile.c (dump_loc): Remove newline emission.
641 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
642 emission to dump_printf_loc calls where missing.
643 (vect_mark_for_runtime_alias_test): Ditto.
644 (vect_analyze_data_ref_dependence): Ditto.
645 (vect_analyze_data_ref_dependences): Ditto.
646 (vect_slp_analyze_data_ref_dependence): Ditto.
647 (vect_slp_analyze_data_ref_dependences): Ditto.
648 (vect_compute_data_ref_alignment): Ditto.
649 (vect_update_misalignment_for_peel): Ditto.
650 (vect_verify_datarefs_alignment): Ditto.
651 (vector_alignment_reachable_p): Ditto.
652 (vect_get_data_access_cost): Ditto.
653 (vect_enhance_data_refs_alignment): Ditto.
654 (vect_find_same_alignment_drs): Ditto.
655 (vect_analyze_data_refs_alignment): Ditto.
656 (vect_analyze_group_access): Ditto.
657 (vect_analyze_data_ref_access): Ditto.
658 (vect_analyze_data_ref_accesses): Ditto.
659 (vect_prune_runtime_alias_test_list): Ditto.
660 (vect_analyze_data_refs): Ditto.
661 (vect_create_addr_base_for_vector_ref): Ditto.
662 (vect_create_data_ref_ptr): Ditto.
663 (vect_grouped_store_supported): Ditto.
664 (vect_grouped_load_supported): Ditto.
665 * value-prof.c (check_counter): Ditto.
666 (check_ic_target): Ditto.
667 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
668 (vect_recog_widen_mult_pattern): Ditto.
669 (vect_recog_widen_sum_pattern): Ditto.
670 (vect_recog_over_widening_pattern): Ditto.
671 (vect_recog_widen_shift_pattern): Ditto.
672 (vect_recog_rotate_pattern): Ditto.
673 (vect_recog_vector_vector_shift_pattern): Ditto.
674 (vect_recog_divmod_pattern): Ditto.
675 (vect_recog_mixed_size_cond_pattern): Ditto.
676 (vect_recog_bool_pattern): Ditto.
677 (vect_pattern_recog_1): Ditto.
678 (vect_pattern_recog): Ditto.
679 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
680 (vect_is_simple_iv_evolution): Ditto.
681 (vect_analyze_scalar_cycles_1): Ditto.
682 (vect_get_loop_niters): Ditto.
683 (vect_analyze_loop_1): Ditto.
684 (vect_analyze_loop_form): Ditto.
685 (vect_analyze_loop_operations): Ditto.
686 (vect_analyze_loop_2): Ditto.
687 (vect_analyze_loop): Ditto.
688 (report_vect_op): Ditto.
689 (vect_is_slp_reduction): Ditto.
690 (vect_is_simple_reduction_1): Ditto.
691 (vect_get_known_peeling_cost): Ditto.
692 (vect_estimate_min_profitable_iters): Ditto.
693 (vect_model_reduction_cost): Ditto.
694 (vect_model_induction_cost): Ditto.
695 (get_initial_def_for_induction): Ditto.
696 (vect_create_epilog_for_reduction): Ditto.
697 (vectorizable_reduction): Ditto.
698 (vectorizable_induction): Ditto.
699 (vectorizable_live_operation): Ditto.
700 (vect_loop_kill_debug_uses): Ditto.
701 (vect_transform_loop): Ditto.
702 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
703 (vect_stmt_relevant_p): Ditto.
704 (process_use): Ditto.
705 (vect_mark_stmts_to_be_vectorized): Ditto.
706 (vect_model_simple_cost): Ditto.
707 (vect_model_promotion_demotion_cost): Ditto.
708 (vect_model_store_cost): Ditto.
709 (vect_get_store_cost): Ditto.
710 (vect_model_load_cost): Ditto.
711 (vect_get_load_cost): Ditto.
712 (vect_init_vector_1): Ditto.
713 (vect_get_vec_def_for_operand): Ditto.
714 (vect_finish_stmt_generation): Ditto.
715 (vectorizable_call): Ditto.
716 (vectorizable_conversion): Ditto.
717 (vectorizable_assignment): Ditto.
718 (vectorizable_shift): Ditto.
719 (vectorizable_operation): Ditto.
720 (vectorizable_store): Ditto.
721 (vectorizable_load): Ditto.
722 (vectorizable_condition): Ditto.
723 (vect_analyze_stmt): Ditto.
724 (vect_transform_stmt): Ditto.
725 (vect_is_simple_use): Ditto.
726 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
727 (vect_can_advance_ivs_p): Ditto.
728 (vect_update_ivs_after_vectorizer): Ditto.
729 (vect_do_peeling_for_loop_bound): Ditto.
730 (vect_gen_niters_for_prolog_loop): Ditto.
731 (vect_update_inits_of_drs): Ditto.
732 (vect_create_cond_for_alias_checks): Ditto.
733 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
734 (vect_build_slp_tree_1): Ditto.
735 (vect_supported_load_permutation_p): Ditto.
736 (vect_analyze_slp_instance): Ditto.
737 (vect_analyze_slp): Ditto.
738 (vect_make_slp_decision): Ditto.
739 (vect_detect_hybrid_slp): Ditto.
740 (vect_bb_vectorization_profitable_p): Ditto.
741 (vect_slp_analyze_bb_1): Ditto.
742 (vect_update_slp_costs_according_to_vf): Ditto.
743 (vect_get_mask_element): Ditto.
744 (vect_transform_slp_perm_load): Ditto.
745 (vect_schedule_slp_instance): Ditto.
746 (vect_schedule_slp): Ditto.
747 (vect_slp_transform_bb): Ditto.
748 * profile.c (read_profile_edge_counts): Ditto.
749 (compute_branch_probabilities): Ditto.
750 * coverage.c (get_coverage_counts): Ditto.
752 2013-09-16 Diego Novillo <dnovillo@google.com>
754 * tree-core.h: Add missing comment lines from refactoring
757 2013-09-16 Jan Hubicka <jh@suse.cz>
759 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
760 abstract functions; for static functions check the presence
763 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
765 * config/aarch64/aarch64-simd-builtins.def (fma): New.
766 * config/aarch64/aarch64-simd.md
767 (aarch64_mla_elt<mode>): New.
768 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
769 (aarch64_mls_elt<mode>): Likewise.
770 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
771 (aarch64_fma4_elt<mode>): Likewise.
772 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
773 (aarch64_fma4_elt_to_128v2df): Likewise.
774 (aarch64_fma4_elt_to_64df): Likewise.
775 (fnma<mode>4): Likewise.
776 (aarch64_fnma4_elt<mode>): Likewise.
777 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
778 (aarch64_fnma4_elt_to_128v2df): Likewise.
779 (aarch64_fnma4_elt_to_64df): Likewise.
780 * config/aarch64/iterators.md (VDQSF): New.
781 * config/aarch64/arm_neon.h
782 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
783 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
785 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
787 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
788 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
789 (aarch64_mul3_elt_to_128df): Likewise.
790 (aarch64_mul3_elt_to_64v2df): Likewise.
791 * config/aarch64/iterators.md (VEL): Also handle DFmode.
793 (VMUL_CHANGE_NLANES) Likewise.
796 * config/aarch64/arm_neon.h
797 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
799 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
801 * config/aarch64/arm_neon.h
802 (vcvtx_high_f32_f64): Fix parameters.
804 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
805 Uros Bizjak <ubizjak@gmail.com>
807 * config/alpha.c: Include tree-ssanames.h.
809 2013-09-16 Richard Biener <rguenther@suse.de>
811 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
812 (dot_rdg_1): Handle control_dd.
813 (create_edge_for_control_dependence): New function.
814 (create_rdg_edges): Add control dependences if asked for.
815 (build_rdg): Likewise.
816 (generate_loops_for_partition): If there are not necessary
817 control stmts remove all their dependencies.
818 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
819 (distribute_loop): Pass on control dependences.
820 (tree_loop_distribution): Compute control dependences and remove
821 restriction on number of loop nodes.
823 2013-09-16 Jakub Jelinek <jakub@redhat.com>
825 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
828 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
830 * cse.c (try_const_anchors): Punt on CC modes.
832 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
834 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
836 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
839 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
842 2013-09-14 Iain Sandoe <iain@codesourcery.com>
845 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
847 (darwin_objc1_section): Likewise.
848 (darwin_file_end): Emit Image Info section when required.
850 2013-09-14 Jan Hubicka <jh@suse.cz>
852 * tree-into-ssa.c (gate_into_ssa): New.
853 (pass_data_build_ssa): Use it.
854 * cgraph.h (expand_thunk): Update prototype.
855 * cgraphunit.c (analyze_function): Expand thunks early.
856 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
857 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
858 set lowered flag; do not add new function.
859 (assemble_thunks_and_aliases): Update.
860 * tree-ssa.c (gate_init_datastructures): New gate.
861 (pass_data_init_datastructures): Use it.
863 2013-09-14 Iain Sandoe <iain@codesourcery.com>
866 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
867 xmm register set described in the psABI.
869 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
871 * dwarf2out.c (should_emit_struct_debug): Add check
872 for type_decl variable is not NULL.
874 2013-09-13 Jacek Caban <jacek@codeweavers.com>
876 * config.gcc: Use new winnt-c.c target hooks
877 * config/t-winnt: New file
878 * config/winnt-c.c: New file
879 * doc/tm.texi.in: Document new hook
880 * doc/tm.texi: Regenerated
882 2013-09-13 Jan Hubicka <jh@suse.cz>
885 * ipa-inline.c (check_callers): New function.
886 (check_caller_edge): Remove.
887 (want_inline_function_to_all_callers_p): Also permit alises that are
889 (inline_to_all_callers): Terminate the walk when devirtualization
892 2013-09-13 Jan Hubicka <jh@suse.cz>
894 * ipa-inline-analysis.c (struct growth_data): Add node.
895 (do_estimate_growth_1): Fix detection of recursion.
897 2013-09-13 Jakub Jelinek <jakub@redhat.com>
899 PR tree-optimization/58392
900 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
901 to avoid shadowing of outer loop variable. If
902 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
903 replace_by_duplicate_decl simduid of loops that have it set and
904 set dest_cfun->has_simduid_loops and/or
905 dest_cfun->has_force_vect_loops.
906 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
907 instead of maybe_lookup_decl.
908 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
909 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
910 call. Copy over force_vect and copy and remap simduid. Set
911 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
912 (copy_cfg_body): Remove blocks_to_copy argument. Use
913 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
914 caller. Don't set cfun->has_simduid_loops and/or
915 cfun->has_force_vect_loops here.
916 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
918 (expand_call_inline, tree_function_versioning): Adjust copy_body
921 2013-09-13 Martin Jambor <mjambor@suse.cz>
924 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
925 the assert if the edge was a speculative one.
927 2013-09-13 Richard Biener <rguenther@suse.de>
929 * tree-data-ref.h (known_dependences_p): Move here ...
930 * tree-loop-distribution.c (known_dependences_p): ... from here.
931 (dump_rdg_component, debug_rdg_component): Remove.
933 (generate_loops_for_partition): Use gimple_uid instead of
934 relying on matching stmt visit order.
935 (rdg_build_partitions): Take starting stmt vector.
936 (ldist_gen): Merge into ...
937 (distribute_loop): ... this function. Do not compute starting
939 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
941 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
943 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
944 Set type attribute correctly. Set predicable_short_it attribute.
945 (cmpsi_shiftsi): Remove %? from output template.
947 2013-09-13 Richard Biener <rguenther@suse.de>
949 * tree-loop-distribution.c (struct rdg_component,
950 rdg_defs_used_in_other_loops_p, free_rdg_components,
951 rdg_build_components): Remove.
952 (stmts_from_loop): Do not record virtual PHIs.
953 (generate_loops_for_partition): Skip virtual PHIs.
954 (build_rdg_partition_for_component): Rename to ...
955 (build_rdg_partition_for_vertex): ... this and adjust.
956 (rdg_build_partitions): Take a vector of starting vertices
957 instead of components. Remove unnecessary leftover handling.
958 (ldist_gen): Do not build components or record other stores.
959 (distribute_loop): Do not distribute loops containing stmts
962 2013-09-13 Christian Bruel <christian.bruel@st.com>
965 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
967 2013-09-13 Kai Tietz <ktietz@redhat.com>
969 * config.gcc: Separate cases for mingw and cygwin targets,
970 and add 64-bit cygwin target case.
972 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
973 dll-export inline-functions.
974 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
977 2013-09-13 Jeff Law <law@redhat.com>
981 2013-09-06 Jeff Law <law@redhat.com>
983 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
984 edge implied equivalences into successor phis.
986 2013-09-12 DJ Delorie <dj@redhat.com>
988 * config/rl78/rl78-virt.md: Change from | to \; for asm line
991 2013-09-12 Brooks Moses <bmoses@google.com>
994 * Makefile.in: Do not install driver binaries in $(target)/bin.
996 2013-09-12 DJ Delorie <dj@redhat.com>
998 * config/rl78/rl78.opt (mrelax): New.
999 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1000 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1002 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1004 (rl78_expand_epilogue): Likewise.
1006 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1009 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1010 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1011 pointer elimination with using remove_reg_equal_offset_note.
1013 2013-09-12 DJ Delorie <dj@redhat.com>
1015 * config/msp430/: New port.
1016 * config.gcc (msp430): Added.
1017 * doc/invoke.texi: Document MSP430 options.
1018 * doc/install.texi: Document msp430-elf
1019 * doc/md.texi: Document msp430-elf
1020 * doc/contrib.texi: Document msp430-elf
1022 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1025 2013-09-12 Martin Jambor <mjambor@suse.cz>
1028 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1029 reference descriptor is NULL.
1030 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1033 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1035 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1036 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1037 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1038 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1039 make_temp_ssa_name): move to tree-ssanames.h
1040 * tree-ssa-alias.h: Move prototype.
1041 * tree-ssa.h: Include tree-ssanames.h.
1042 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1043 * tree-ssanames.h: New. Move items from tree-flow*.h
1044 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1046 2013-09-12 Richard Biener <rguenther@suse.de>
1048 PR tree-optimization/58404
1049 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1050 propagate non-invariant addresses into dereferences wrapped
1051 in component references.
1053 2013-09-12 Richard Biener <rguenther@suse.de>
1055 PR tree-optimization/58402
1056 * passes.def: Move pass_late_warn_uninitialized later.
1058 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1060 * tree-ssa.h: New. Move content from tree-flow.h and
1062 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1063 Move prototypes belonging to tree-ssa.c.
1064 * tree-flow-inline.h (redirect_edge_var_map_def,
1065 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1067 * gimple.h: Adjust prototypes.
1068 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1070 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1071 * tree.h: Move prototype to tree-ssa.h.
1072 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1073 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1074 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1075 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1076 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1077 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1078 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1079 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1080 graphite-blocking.c, graphite-clast-to-gimple.c,
1081 graphite-dependences.c, graphite-interchange.c,
1082 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1083 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1084 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1085 ipa-reference.c, ipa-split.c, ipa-utils.c,
1086 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1087 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1088 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1089 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1090 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1091 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1092 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1093 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1094 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1095 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1096 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1097 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1098 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1099 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1100 value-prof.c, var-tracking.c,
1101 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1103 2013-09-12 Richard Biener <rguenther@suse.de>
1105 PR tree-optimization/58396
1106 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1107 (build_rdg): Take a loop-nest parameter, fix memleaks.
1108 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1110 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1112 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1115 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1117 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1118 The FMA instruction names should have a 'v' prefix.
1120 2013-09-12 Richard Biener <rguenther@suse.de>
1122 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1123 (dot_rdg): Use popen instead of system in optional code.
1124 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1125 (already_processed_vertex_p): Adjust.
1126 (has_anti_or_output_dependence, predecessor_has_mem_write,
1127 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1128 rdg_flag_uses): Remove.
1129 (rdg_flag_vertex): Simplify.
1130 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1132 (build_rdg_partition_for_component): Process the first vertex
1133 of a component only.
1134 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1136 2013-09-12 Alan Modra <amodra@gmail.com>
1138 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1140 2013-09-11 DJ Delorie <dj@redhat.com>
1141 Nick Clifton <nickc@redhat.com>
1143 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1144 (rl78_stack_based_mem): New.
1145 * config/rl78/constraints.md (Iv08): New.
1154 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1156 (movsi): Change from expand to insn-and-split.
1157 (ashrsi3): Clobber AX.
1161 * config/rl78/rl78.md (CC_REG): Fix.
1162 (addsi3): Allow memory and immediate operands.
1163 (addsi3_internal): Split into...
1164 (addsi3_internal_virt): ...new, and ...
1165 (addsi3_internal_real): ...new.
1167 (subsi3_internal_virt): New.
1168 (subsi3_internal_real): New.
1169 (mulsi3): Add memory operand.
1170 (mulsi3_rl78): Likewise.
1171 (mulsi3_g13): Likewise.
1172 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1173 (cbranchqi4_real): Add more constraint options.
1174 (cbranchhi4_real): Expand pattern.
1175 (cbranchhi4_real_signed): New.
1176 (cbranchhi4_real_inverted): New.
1177 (cbranchsi4_real_lt): New.
1178 (cbranchsi4_real_ge): New.
1179 (cbranchsi4_real_signed): New.
1180 (cbranchsi4_real): New.
1182 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1184 (lshrsi3_virt): Likewise.
1185 (ashlsi3_virt): Likewise.
1186 (cbranchqi4_virt_signed): New.
1187 (cbranchhi4_virt_signed): New.
1188 (cbranchsi4_virt): New.
1189 * config/rl78/rl78.c: Whitespace fixes throughout.
1190 (move_elim_pass): New.
1191 (pass_data_rl78_move_elim): New.
1192 (pass_rl78_move_elim): New.
1193 (make_pass_rl78_move_elim): New.
1194 (rl78_devirt_info): Run devirt earlier.
1195 (rl78_move_elim_info): New.
1196 (rl78_asm_file_start): Register it.
1197 (rl78_split_movsi): New.
1198 (rl78_as_legitimate_address): Allow virtual base registers when
1200 (rl78_addr_space_convert): Remove spurious debug stuff.
1201 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1202 (rl78_print_operand): More cases for not printing '#'.
1203 (rl78_expand_compare): Remove most of the logic.
1204 (content_memory): New.
1205 (clear_content_memory): New.
1206 (get_content_index): New.
1207 (get_content_name): New.
1208 (display_content_memory): New.
1209 (update_content): New.
1210 (record_content): New.
1211 (already_contains): New.
1212 (insn_ok_now): Re-recog insns with virtual registers.
1213 (add_postponed_content_update): New.
1214 (process_postponed_content_update): New.
1215 (gen_and_emit_move): New.
1216 (transcode_memory_rtx): Record new location content.
1217 Use gen_and_emit_move.
1218 (force_into_acc): New.
1219 (move_to_acc): Use gen_and_emit_move.
1220 (move_from_acc): Likewise.
1221 (move_acc_to_reg): Likewise.
1222 (move_to_x): Likewise.
1223 (move_to_hl): Likewise.
1224 (move_to_de): Likewise.
1225 (rl78_alloc_physical_registers_op1): Record location content.
1226 (has_constraint): New.
1227 (rl78_alloc_physical_registers_op2): Record location content.
1229 (rl78_alloc_physical_registers_ro1): Likewise.
1230 (rl78_alloc_physical_registers_cmp): Likewise.
1231 (rl78_alloc_physical_registers_umul): Likewise.
1232 (rl78_alloc_address_registers_macax): New.
1233 (rl78_alloc_physical_registers): Initialize and set location
1234 content memory as needed.
1235 (rl78_reorg): Make sure split2 is called.
1236 (rl78_rtx_costs): New.
1238 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1240 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1241 for (not (neg ...)) and (neg (not ...)) cases.
1243 2013-09-11 Richard Biener <rguenther@suse.de>
1246 * passes.def: Split critical edges before late uninit warning passes.
1247 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1249 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1251 PR tree-optimization/58385
1252 * fold-const.c (build_range_check): If both low and high are NULL,
1253 use omit_one_operand_loc to preserve exp side-effects.
1255 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1257 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1259 2013-09-11 Richard Biener <rguenther@suse.de>
1261 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1262 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1263 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1264 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1265 stmts_from_loop, known_dependences_p, build_empty_rdg,
1266 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1267 * tree-loop-distribution.c: ... here.
1268 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1269 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1270 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1271 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1272 * tree-loop-distribution.c: ... here.
1273 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1274 (struct partition_s): Add loops member.
1275 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1276 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1277 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1279 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1280 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1281 Sergey Lega <sergey.s.lega@intel.com>
1282 Anna Tikhonova <anna.tikhonova@intel.com>
1283 Ilya Tocar <ilya.tocar@intel.com>
1284 Andrey Turetskiy <andrey.turetskiy@intel.com>
1285 Ilya Verbin <ilya.verbin@intel.com>
1286 Kirill Yukhin <kirill.yukhin@intel.com>
1287 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1289 * config/i386/constraints.md (k): New.
1291 * config/i386/i386.c (const regclass_map): Add new mask registers.
1292 (dbx_register_map): Ditto.
1293 (dbx64_register_map): Ditto.
1294 (svr4_dbx_register_map): Ditto.
1295 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1297 (ix86_preferred_reload_class): Disable constants for mask registers.
1298 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1299 (ix86_hard_regno_mode_ok): Support new mask registers.
1300 (x86_order_regs_for_local_alloc): Ditto.
1301 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1302 (FIXED_REGISTERS): Add new mask registers.
1303 (CALL_USED_REGISTERS): Ditto.
1304 (REG_ALLOC_ORDER): Ditto.
1305 (VALID_MASK_REG_MODE): New.
1306 (FIRST_MASK_REG): Ditto.
1307 (LAST_MASK_REG): Ditto.
1308 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1309 (MAYBE_MASK_CLASS_P): New.
1310 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1311 (REG_CLASS_CONTENTS): Ditto.
1312 (MASK_REGNO_P): New.
1313 (ANY_MASK_REG_P): Ditto.
1314 (HI_REGISTER_NAMES): Add new mask registers.
1315 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1316 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1318 (attribute "type"): Add mskmov, msklog.
1319 (attribute "length_immediate"): Support them.
1320 (attribute "memory"): Ditto.
1321 (attribute "prefix_0f"): Ditto.
1322 (*movhi_internal): Support new mask registers.
1323 (*movqi_internal): Ditto.
1324 (define_split): Split out clobber pattern is a logic
1325 insn on mask registers.
1326 (*k<logic><mode>): New.
1327 (*andhi_1): Extend to support mask regs.
1328 (*andqi_1): Extend to support mask regs.
1330 (define_split): Split and-not to and and not if operands
1332 (*<code><mode>_1): Separate HI mode to new pattern...
1333 (*<code>hi_1): This.
1334 (*<code>qi_1): Extend to support mask regs.
1336 (kortestzhi): Ditto.
1337 (kortestchi): Ditto.
1339 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1340 (*one_cmplhi2_1): ...Here, now with mask registers support.
1341 (*one_cmplqi2_1): Support new mask registers.
1342 (HI/QImode arithmetics splitter): Don't split if mask registers
1344 (HI/QImode not splitter): Ditto.
1345 * config/i386/predicated.md (mask_reg_operand): New.
1346 (general_reg_operand): Ditto.
1348 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1350 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1351 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1353 2013-09-10 Jeff Law <law@redhat.com>
1355 PR tree-optimization/58380
1356 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1357 of threading through a buried loop header.
1359 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1360 return value for single successor case.
1362 2013-09-10 Jan Hubicka <jh@suse.cz>
1364 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1366 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1369 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1370 support conditional execution.
1371 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
1373 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
1375 * lra.c (lra): Clear lra_optional_reload_pseudos before every
1377 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
1378 Check destination too to check move insn.
1379 (undo_optional_reloads): Add check that the original peudo did not
1380 changed its allocation and the optional reload was inherited on last
1381 inheritance pass. Break loop after deciding to keep optional reload.
1382 (lra_undo_inheritance): Add check that inherited pseudo still in
1385 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
1387 * config/aarch64/aarch64.md (generic_sched): New.
1388 * config/aarch64/aarch64-generic.md (load): Make conditional
1389 on generic_sched attribute.
1390 (nonload): Likewise.
1392 2013-09-10 Jan Hubicka <jh@suse.cz>
1394 * lto-cgraph.c: Include ipa-utils.h.
1395 (compute_ltrans_boundary): Also add possible targets into the boundary.
1397 2013-09-10 Jan Hubicka <jh@suse.cz>
1399 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
1400 VAR_DECL of vtable rather than full expression.
1402 2013-09-10 Jan Hubicka <jh@suse.cz>
1403 Paolo Carlini <paolo.carlini@oracle.com>
1405 * cgraphunit.c (analyze_functions): Save input_location, set it
1406 to UNKNOWN_LOCATION and restore it at the end.
1408 2013-09-10 Martin Jambor <mjambor@suse.cz>
1410 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
1411 represented by a thunk.
1413 2013-09-10 Jeff Law <law@redhat.com>
1415 PR tree-optimization/58343
1416 * tree-ssa-threadupdate.c (thread_block): Identify and disable
1417 jump threading requests through loop headers buried in the middle
1418 of a jump threading path.
1420 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
1421 in return value/type.
1423 2013-09-10 Jakub Jelinek <jakub@redhat.com>
1425 PR rtl-optimization/58365
1426 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
1427 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
1430 2013-09-10 Richard Biener <rguenther@suse.de>
1432 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
1433 * tree-data-ref.c (create_rdg_vertices): Collect all data
1434 references, signal failure to the caller, use data-ref API.
1435 (build_rdg): Compute data references only once. Maintain lifetime
1436 of data references and data dependences from within RDG.
1437 (free_rdg): Free dependence relations.
1438 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
1439 inventing extra dependences.
1440 (distribute_loop): Update for RDG API changes.
1442 2013-09-10 Kai Tietz <ktietz@redhat.com>
1444 * doc/invoke.texi (fms-extensions): Document changed
1445 behavior for ms-abi targets.
1446 * config/i386/i386.c (ix86_option_override_internal):
1447 Set default value of option -fms-extension for ms-abi targets.
1449 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1451 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
1453 2013-09-10 Alan Modra <amodra@gmail.com>
1456 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
1458 2013-09-10 Alan Modra <amodra@gmail.com>
1460 * config/rs6000/predicates.md (add_cint_operand): New.
1461 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
1462 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
1463 using add_cint_operand.
1464 (largetoc_high_plus_aix): Likewise.
1466 2013-09-09 Jakub Jelinek <jakub@redhat.com>
1468 PR tree-optimization/58364
1469 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
1470 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
1471 the current range can't be an unconditional true or false.
1473 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
1475 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
1477 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
1479 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
1481 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
1484 * doc/invoke.texi (-Wdelete-incomplete): Document it.
1486 2013-09-09 Ian Bolton <ian.bolton@arm.com>
1488 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
1489 NO_REGS for immediate that can't be moved directly into FP_REGS.
1491 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1493 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
1494 comparison with negated operand.
1495 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
1498 2013-09-09 Richard Biener <rguenther@suse.de>
1501 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
1502 of a subloop record all its block as affecting loop-closed SSA form.
1504 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1506 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
1507 of an rtx/bitpos pair.
1508 (store_fixed_bit_field): Update accordingly.
1510 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1512 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
1514 * builtins.c (expand_errno_check): Likewise.
1515 * dwarf2cfi.c (init_return_column_size): Likewise.
1516 * except.c (sjlj_mark_call_sites): Likewise.
1517 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
1518 * lra-constraints.c (emit_inc): Likewise.
1519 * ree.c (combine_set_extension): Likewise.
1520 * regmove.c (fixup_match_2): Likewise.
1521 * reload1.c (inc_for_reload): Likewise.
1523 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1525 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
1526 (force_to_mode, simplify_shift_const_1, simplify_comparison):
1527 Use gen_int_mode with the mode of the associated simplify_* call.
1528 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
1529 * expmed.c (expand_shift_1): Likewise.
1530 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1531 * loop-iv.c (iv_number_of_iterations): Likewise.
1532 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1533 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
1535 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1537 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
1538 of the associated expand_* call.
1539 (asan_emit_stack_protection): Likewise.
1540 * builtins.c (round_trampoline_addr): Likewise.
1541 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
1542 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
1543 (emit_store_flag): Likewise.
1544 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
1545 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
1547 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1548 * ifcvt.c (noce_try_store_flag_constants): Likewise.
1549 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1550 * modulo-sched.c (generate_prolog_epilog): Likewise.
1551 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
1552 (expand_ctz, expand_ffs, expand_unop): Likewise.
1554 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1556 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
1557 of the associated gen_rtx_* call.
1558 * caller-save.c (init_caller_save): Likewise.
1559 * combine.c (find_split_point, make_extraction): Likewise.
1560 (make_compound_operation): Likewise.
1561 * dwarf2out.c (mem_loc_descriptor): Likewise.
1562 * explow.c (plus_constant, probe_stack_range): Likewise.
1563 * expmed.c (expand_mult_const): Likewise.
1564 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
1565 * reload1.c (init_reload): Likewise.
1566 * valtrack.c (cleanup_auto_inc_dec): Likewise.
1567 * var-tracking.c (adjust_mems): Likewise.
1568 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
1569 rather than gen_rtx_fmt_ee.
1571 2013-09-09 Jan Hubicka <jh@suse.cz>
1574 * value-prof.c (gimple_ic): Copy also abnormal edges.
1576 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1578 * asan.c (asan_shadow_cst): Use gen_int_mode.
1580 2013-09-08 Jan Hubicka <jh@suse.cz>
1582 * ipa-profile.c: Add toplevel comment.
1583 (ipa_propagate_frequency_1): Be more conservative when profile is read.
1584 (contains_hot_call_p): New function.
1585 (ipa_propagate_frequency): Set frequencies based on counts when
1587 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
1588 profile; do not tamper with profile after inlining if it is read.
1590 2013-09-08 Jan Hubicka <jh@suse.cz>
1592 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
1595 2013-09-08 Jan Hubicka <jh@suse.cz>
1597 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
1600 2013-09-08 Jeff Law <law@redhat.com>
1603 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
1606 2013-09-08 Andi Kleen <ak@linux.intel.com>
1608 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
1610 2013-09-08 Jan Hubicka <jh@suse.cz>
1612 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
1614 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
1616 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
1617 for non-debug insns.
1618 * lra.c (new_insn_reg): Take the containing insn as a parameter.
1619 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
1620 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
1623 2013-09-08 Jan Hubicka <jh@suse.cz>
1625 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
1626 targets and devirtualize to BUILT_IN_UNREACHABLE.
1627 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
1628 * ipa.c (walk_polymorphic_call_targets): New function.
1629 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
1630 functions; use the new timevar.
1631 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
1632 was removed from the program.
1633 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
1634 not consider it in the walk when its vtable is dead.
1635 (possible_polymorphic_call_targets_1): Pass anonymous flag to
1637 (devirt_variable_node_removal_hook): New function.
1638 (possible_polymorphic_call_targets): Also register
1639 devirt_variable_node_removal_hook.
1640 (ipa_devirt): Do not do non-speculative devirtualization.
1641 (gate_ipa_devirt): One execute if devirtualizing speculatively.
1643 2013-09-08 Jan Hubicka <jh@suse.cz>
1645 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
1646 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
1647 varpool_remove_variable_insertion_hook): Declare.
1648 * varpool.c (varpool_node_hook_list): New structure.
1649 (first_varpool_node_removal_hook,
1650 first_varpool_variable_insertion_hook): New variables.
1651 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
1652 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
1653 varpool_remove_variable_insertion_hook,
1654 varpool_call_variable_insertion_hooks): New functions.
1655 (varpool_remove_node): Use it.
1657 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
1660 * diagnostic.c (diagnostic_build_prefix): When s.file is
1661 "<built-in>" don't output line and column numbers.
1663 2013-09-06 Jan Hubicka <jh@suse.cz>
1665 * cgraphunit.c (expand_thunk): Get body before touching arguments.
1666 * lto-streamer-out.c: Stream thunks, too.
1667 * lto-streamer-in.c (input_function): Pop cfun here
1668 (lto_read_body): Instead of here.
1670 2013-09-06 Caroline Tice <cmtice@google.com>
1672 * doc/install.texi: Add documentation for the --enable-vtable-verify
1673 and the --disable-libvtv configure options.
1675 2013-09-06 Jeff Law <law@redhat.com>
1677 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1678 edge implied equivalences into successor phis.
1680 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
1682 * resource.c (mark_referenced_resources): Handle COND_EXEC.
1684 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
1686 * resource.c (mark_target_live_regs): Compute resources taking
1687 into account if a call is predicated or not.
1689 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1691 * toplev.c (output_stack_usage): Be prepared for suffixes created by
1692 the compiler in the function names.
1694 2013-09-06 Jan Hubicka <jh@suse.cz>
1697 * ipa-inline.c (has_caller_p): New function.
1698 (want_inline_function_to_all_callers_p): Use it.
1699 (sum_callers, inline_to_all_callers): Break out from ...
1700 (ipa_inline): ... here.
1702 2013-09-06 Jan Hubicka <jh@suse.cz>
1704 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
1705 only when AVX is enabled.
1707 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1709 * config/aarch64/aarch64.md
1710 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
1711 is fpsimd_<load/store>2.
1712 (load_pair<mode>): Likewise.
1713 (store_pair<mode>): Likewise.
1715 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1717 * config/arm/types.md (type): Add "mrs" type.
1718 * config/aarch64/aarch64.md
1719 (aarch64_load_tp_hard): Make type "mrs".
1721 (load_tp_hard): Make type "mrs".
1722 * config/arm/cortex-a15.md: Update with new attributes.
1723 * config/arm/cortex-a5.md: Update with new attributes.
1724 * config/arm/cortex-a53.md: Update with new attributes.
1725 * config/arm/cortex-a7.md: Update with new attributes.
1726 * config/arm/cortex-a8.md: Update with new attributes.
1727 * config/arm/cortex-a9.md: Update with new attributes.
1728 * config/arm/cortex-m4.md: Update with new attributes.
1729 * config/arm/cortex-r4.md: Update with new attributes.
1730 * config/arm/fa526.md: Update with new attributes.
1731 * config/arm/fa606te.md: Update with new attributes.
1732 * config/arm/fa626te.md: Update with new attributes.
1733 * config/arm/fa726te.md: Update with new attributes.
1735 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1737 * config/aarch64/aarch64.md
1738 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
1739 (*movtf_aarch64): Likewise.
1741 (thumb1_movdi_insn): Use "multiple" for type where more than one
1742 instruction is used for a move.
1743 (*arm32_movhf): Likewise.
1744 (*thumb_movdf_insn): Likewise.
1746 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1748 * config/arm/types.md (type): Rename fcpys to fmov.
1750 (*arm_movsi_vfp): Rename type fcpys as fmov.
1751 (*thumb2_movsi_vfp): Likewise
1752 (*movhf_vfp_neon): Likewise
1753 (*movhf_vfp): Likewise
1754 (*movsf_vfp): Likewise
1755 (*thumb2_movsf_vfp): Likewise
1756 (*movsfcc_vfp): Likewise
1757 (*thumb2_movsfcc_vfp): Likewise
1758 * config/aarch64/aarch64-simd.md
1759 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
1760 * config/aarch64/aarch64.md
1761 (*movsi_aarch64): Replace type mov_reg with fmovs.
1762 (*movdi_aarch64): Likewise
1763 (*movsf_aarch64): Likewise
1764 (*movdf_aarch64): Likewise
1766 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
1767 * config/arm/iwmmxt.md
1768 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
1769 * config/arm/arm1020e.md: Update with new attributes.
1770 * config/arm/cortex-a15-neon.md: Update with new attributes.
1771 * config/arm/cortex-a5.md: Update with new attributes.
1772 * config/arm/cortex-a53.md: Update with new attributes.
1773 * config/arm/cortex-a7.md: Update with new attributes.
1774 * config/arm/cortex-a8-neon.md: Update with new attributes.
1775 * config/arm/cortex-a9.md: Update with new attributes.
1776 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1777 * config/arm/cortex-r4f.md: Update with new attributes.
1778 * config/arm/marvell-pj4.md: Update with new attributes.
1779 * config/arm/vfp11.md: Update with new attributes.
1781 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1783 * config/aarch64/aarch64.md
1784 (*madd<mode>): Fix type attribute.
1785 (*maddsi_uxtw): Likewise.
1786 (*msub<mode>): Likewise.
1787 (*msubsi_uxtw): Likewise.
1788 (<su_optab>maddsidi4): Likewise.
1789 (<su_optab>msubsidi4): Likewise.
1791 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1793 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
1794 * config/arm/arm.md (core_cycles): Remove fdiv.
1795 * config/arm/vfp.md:
1796 (*sqrtsf2_vfp): Update for attribute changes.
1797 (*sqrtdf2_vfp): Likewise.
1798 * config/aarch64/aarch64.md:
1799 (sqrt<mode>2): Update for attribute changes.
1800 * config/arm/arm1020e.md: Update with new attributes.
1801 * config/arm/cortex-a15-neon.md: Update with new attributes.
1802 * config/arm/cortex-a5.md: Update with new attributes.
1803 * config/arm/cortex-a53.md: Update with new attributes.
1804 * config/arm/cortex-a7.md: Update with new attributes.
1805 * config/arm/cortex-a8-neon.md: Update with new attributes.
1806 * config/arm/cortex-a9.md: Update with new attributes.
1807 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1808 * config/arm/cortex-r4f.md: Update with new attributes.
1809 * config/arm/marvell-pj4.md: Update with new attributes.
1810 * config/arm/vfp11.md: Update with new attributes.
1812 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1814 * config/arm/types.md
1815 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
1816 * config/aarch64/aarch64.md
1817 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
1819 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
1820 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
1821 (float<GPI:mode><GPF:mode>2): Likewise.
1823 (*truncsisf2_vfp): Update with new attributes.
1824 (*truncsidf2_vfp): Likewise.
1825 (fixuns_truncsfsi2): Likewise.
1826 (fixuns_truncdfsi2): Likewise.
1827 (*floatsisf2_vfp): Likewise.
1828 (*floatsidf2_vfp): Likewise.
1829 (floatunssisf2): Likewise.
1830 (floatunssidf2): Likewise.
1831 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
1832 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
1833 * config/arm/arm1020e.md: Update with new attributes.
1834 * config/arm/cortex-a15-neon.md: Update with new attributes.
1835 * config/arm/cortex-a5.md: Update with new attributes.
1836 * config/arm/cortex-a53.md: Update with new attributes.
1837 * config/arm/cortex-a7.md: Update with new attributes.
1838 * config/arm/cortex-a8-neon.md: Update with new attributes.
1839 * config/arm/cortex-a9.md: Update with new attributes.
1840 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1841 * config/arm/cortex-r4f.md: Update with new attributes.
1842 * config/arm/marvell-pj4.md: Update with new attributes.
1843 * config/arm/vfp11.md: Update with new attributes.
1845 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1847 * config/aarch64/arm_neon.h
1848 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
1849 (vqtbx<1,2,3,4><q>_s8): Likewise.
1851 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1853 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
1854 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
1855 (add<mode>3): Add type attribute.
1856 (add<mode>3): Likewise.
1857 (usadd<mode>3): Likewise.
1858 (ssadd<mode>3): Likewise.
1859 (sub<mode>3): Likewise.
1860 (sub<mode>3): Likewise.
1861 (ussub<mode>3): Likewise.
1862 (sssub<mode>3): Likewise.
1863 (ssmulsa3): Likewise.
1864 (usmulusa3): Likewise.
1865 (arm_usatsihi): Likewise.
1867 (*movdi_vfp): Add types for all instructions.
1868 (*movdi_vfp_cortexa8): Likewise.
1869 (*movhf_vfp_neon): Likewise.
1870 (*movhf_vfp): Likewise.
1871 (*movdf_vfp): Likewise.
1872 (*thumb2_movdf_vfp): Likewise.
1873 (*thumb2_movdfcc_vfp): Likewise.
1874 * config/arm/arm.md: Add type attribute to all insn patterns.
1875 (*thumb1_adddi3): Add type attribute.
1876 (*arm_adddi3): Likewise.
1877 (*adddi_sesidi_di): Likewise.
1878 (*adddi_zesidi_di): Likewise.
1879 (*thumb1_addsi3): Likewise.
1880 (addsi3_compare0): Likewise.
1881 (*addsi3_compare0_scratch): Likewise.
1882 (*compare_negsi_si): Likewise.
1883 (cmpsi2_addneg): Likewise.
1884 (*addsi3_carryin_<optab>): Likewise.
1885 (*addsi3_carryin_alt2_<optab>): Likewise.
1886 (*addsi3_carryin_clobercc_<optab>): Likewise.
1887 (*subsi3_carryin): Likewise.
1888 (*subsi3_carryin_const): Likewise.
1889 (*subsi3_carryin_compare): Likewise.
1890 (*subsi3_carryin_compare_const): Likewise.
1891 (*arm_subdi3): Likewise.
1892 (*thumb_subdi3): Likewise.
1893 (*subdi_di_zesidi): Likewise.
1894 (*subdi_di_sesidi): Likewise.
1895 (*subdi_zesidi_di): Likewise.
1896 (*subdi_sesidi_di): Likewise.
1897 (*subdi_zesidi_ze): Likewise.
1898 (thumb1_subsi3_insn): Likewise.
1899 (*arm_subsi3_insn): Likewise.
1900 (*anddi3_insn): Likewise.
1901 (*anddi_zesidi_di): Likewise.
1902 (*anddi_sesdi_di): Likewise.
1903 (*ne_zeroextracts): Likewise.
1904 (*ne_zeroextracts): Likewise.
1905 (*ite_ne_zeroextr): Likewise.
1906 (*ite_ne_zeroextr): Likewise.
1907 (*anddi_notdi_di): Likewise.
1908 (*anddi_notzesidi): Likewise.
1909 (*anddi_notsesidi): Likewise.
1910 (andsi_notsi_si): Likewise.
1911 (thumb1_bicsi3): Likewise.
1912 (*iordi3_insn): Likewise.
1913 (*iordi_zesidi_di): Likewise.
1914 (*iordi_sesidi_di): Likewise.
1915 (*thumb1_iorsi3_insn): Likewise.
1916 (*xordi3_insn): Likewise.
1917 (*xordi_zesidi_di): Likewise.
1918 (*xordi_sesidi_di): Likewise.
1919 (*arm_xorsi3): Likewise.
1920 (*andsi_iorsi3_no): Likewise.
1921 (*smax_0): Likewise.
1922 (*smax_m1): Likewise.
1923 (*arm_smax_insn): Likewise.
1924 (*smin_0): Likewise.
1925 (*arm_smin_insn): Likewise.
1926 (*arm_umaxsi3): Likewise.
1927 (*arm_uminsi3): Likewise.
1928 (*minmax_arithsi): Likewise.
1929 (*minmax_arithsi_): Likewise.
1930 (*satsi_<SAT:code>): Likewise.
1931 (arm_ashldi3_1bit): Likewise.
1932 (arm_ashrdi3_1bit): Likewise.
1933 (arm_lshrdi3_1bit): Likewise.
1934 (*arm_negdi2): Likewise.
1935 (*thumb1_negdi2): Likewise.
1936 (*arm_negsi2): Likewise.
1937 (*thumb1_negsi2): Likewise.
1938 (*negdi_extendsid): Likewise.
1939 (*negdi_zero_extend): Likewise.
1940 (*arm_abssi2): Likewise.
1941 (*thumb1_abssi2): Likewise.
1942 (*arm_neg_abssi2): Likewise.
1943 (*thumb1_neg_abss): Likewise.
1944 (one_cmpldi2): Likewise.
1945 (extend<mode>di2): Likewise.
1946 (*compareqi_eq0): Likewise.
1947 (*arm_extendhisi2addsi): Likewise.
1948 (*arm_movdi): Likewise.
1949 (*thumb1_movdi_insn): Likewise.
1950 (*arm_movt): Likewise.
1951 (*thumb1_movsi_insn): Likewise.
1952 (pic_add_dot_plus_four): Likewise.
1953 (pic_add_dot_plus_eight): Likewise.
1954 (tls_load_dot_plus_eight): Likewise.
1955 (*thumb1_movhi_insn): Likewise.
1956 (*thumb1_movsf_insn): Likewise.
1957 (*movdf_soft_insn): Likewise.
1958 (*thumb_movdf_insn): Likewise.
1959 (cbranchsi4_insn): Likewise.
1960 (cbranchsi4_scratch): Likewise.
1961 (*negated_cbranchsi4): Likewise.
1962 (*tbit_cbranch): Likewise.
1963 (*tlobits_cbranch): Likewise.
1964 (*tstsi3_cbranch): Likewise.
1965 (*cbranchne_decr1): Likewise.
1966 (*addsi3_cbranch): Likewise.
1967 (*addsi3_cbranch_scratch): Likewise.
1968 (*arm_cmpdi_insn): Likewise.
1969 (*arm_cmpdi_unsig): Likewise.
1970 (*arm_cmpdi_zero): Likewise.
1971 (*thumb_cmpdi_zero): Likewise.
1972 (*deleted_compare): Likewise.
1973 (*mov_scc): Likewise.
1974 (*mov_negscc): Likewise.
1975 (*mov_notscc): Likewise.
1976 (*cstoresi_eq0_thumb1_insn): Likewise.
1977 (cstoresi_nltu_thumb1): Likewise.
1978 (cstoresi_ltu_thu): Likewise.
1979 (thumb1_addsi3_addgeu): Likewise.
1980 (*arm_jump): Likewise.
1981 (*thumb_jump): Likewise.
1982 (*check_arch2): Likewise.
1983 (arm_casesi_internal): Likewise.
1984 (thumb1_casesi_dispatch): Likewise.
1985 (*arm_indirect_jump): Likewise.
1986 (*thumb1_indirect_jump): Likewise.
1988 (*and_scc): Likewise.
1989 (*ior_scc): Likewise.
1990 (*compare_scc): Likewise.
1991 (*cond_move): Likewise.
1992 (*cond_arith): Likewise.
1993 (*cond_sub): Likewise.
1994 (*cmp_ite0): Likewise.
1995 (*cmp_ite1): Likewise.
1996 (*cmp_and): Likewise.
1997 (*cmp_ior): Likewise.
1998 (*ior_scc_scc): Likewise.
1999 (*ior_scc_scc_cmp): Likewise.
2000 (*and_scc_scc): Likewise.
2001 (*and_scc_scc_cmp): Likewise.
2002 (*and_scc_scc_nod): Likewise.
2003 (*negscc): Likewise.
2004 (movcond_addsi): Likewise.
2005 (movcond): Likewise.
2006 (*ifcompare_plus_move): Likewise.
2007 (*if_plus_move): Likewise.
2008 (*ifcompare_move_plus): Likewise.
2009 (*if_move_plus): Likewise.
2010 (*ifcompare_arith_arith): Likewise.
2011 (*if_arith_arith): Likewise.
2012 (*ifcompare_arith_move): Likewise.
2013 (*if_arith_move): Likewise.
2014 (*ifcompare_move_arith): Likewise.
2015 (*if_move_arith): Likewise.
2016 (*ifcompare_move_not): Likewise.
2017 (*if_move_not): Likewise.
2018 (*ifcompare_not_move): Likewise.
2019 (*if_not_move): Likewise.
2020 (*ifcompare_shift_move): Likewise.
2021 (*if_shift_move): Likewise.
2022 (*ifcompare_move_shift): Likewise.
2023 (*if_move_shift): Likewise.
2024 (*ifcompare_shift_shift): Likewise.
2025 (*ifcompare_not_arith): Likewise.
2026 (*ifcompare_arith_not): Likewise.
2027 (*if_arith_not): Likewise.
2028 (*ifcompare_neg_move): Likewise.
2029 (*if_neg_move): Likewise.
2030 (*ifcompare_move_neg): Likewise.
2031 (*if_move_neg): Likewise.
2032 (prologue_thumb1_interwork): Likewise.
2033 (*cond_move_not): Likewise.
2034 (*sign_extract_onebit): Likewise.
2035 (*not_signextract_onebit): Likewise.
2036 (stack_tie): Likewise.
2037 (align_4): Likewise.
2038 (align_8): Likewise.
2039 (consttable_end): Likewise.
2040 (consttable_1): Likewise.
2041 (consttable_2): Likewise.
2042 (consttable_4): Likewise.
2043 (consttable_8): Likewise.
2044 (consttable_16): Likewise.
2045 (*thumb1_tablejump): Likewise.
2046 (prefetch): Likewise.
2047 (force_register_use): Likewise.
2048 (thumb_eh_return): Likewise.
2049 (load_tp_hard): Likewise.
2050 (load_tp_soft): Likewise.
2051 (tlscall): Likewise.
2052 (*arm_movtas_ze): Likewise.
2053 (*arm_rev): Likewise.
2054 (*arm_revsh): Likewise.
2055 (*arm_rev16): Likewise.
2056 * config/arm/thumb2.md
2057 (*thumb2_smaxsi3): Likewise.
2058 (*thumb2_sminsi3): Likewise.
2059 (*thumb32_umaxsi3): Likewise.
2060 (*thumb2_uminsi3): Likewise.
2061 (*thumb2_negdi2): Likewise.
2062 (*thumb2_abssi2): Likewise.
2063 (*thumb2_neg_abss): Likewise.
2064 (*thumb2_movsi_insn): Likewise.
2065 (tls_load_dot_plus_four): Likewise.
2066 (*thumb2_movhi_insn): Likewise.
2067 (*thumb2_mov_scc): Likewise.
2068 (*thumb2_mov_negs): Likewise.
2069 (*thumb2_mov_negs): Likewise.
2070 (*thumb2_mov_nots): Likewise.
2071 (*thumb2_mov_nots): Likewise.
2072 (*thumb2_movsicc_): Likewise.
2073 (*thumb2_movsfcc_soft_insn): Likewise.
2074 (*thumb2_indirect_jump): Likewise.
2075 (*thumb2_and_scc): Likewise.
2076 (*thumb2_ior_scc): Likewise.
2077 (*thumb2_ior_scc_strict_it): Likewise.
2078 (*thumb2_cond_move): Likewise.
2079 (*thumb2_cond_arith): Likewise.
2080 (*thumb2_cond_ari): Likewise.
2081 (*thumb2_cond_sub): Likewise.
2082 (*thumb2_negscc): Likewise.
2083 (*thumb2_movcond): Likewise.
2084 (thumb2_casesi_internal): Likewise.
2085 (thumb2_casesi_internal_pic): Likewise.
2086 (*thumb2_alusi3_short): Likewise.
2087 (*thumb2_mov<mode>_shortim): Likewise.
2088 (*thumb2_addsi_short): Likewise.
2089 (*thumb2_subsi_short): Likewise.
2090 (thumb2_addsi3_compare0): Likewise.
2091 (*thumb2_cbz): Likewise.
2092 (*thumb2_cbnz): Likewise.
2093 (*thumb2_one_cmplsi2_short): Likewise.
2094 (*thumb2_negsi2_short): Likewise.
2095 (*orsi_notsi_si): Likewise.
2096 * config/arm/arm1020e.md: Update with new attributes.
2097 * config/arm/arm1026ejs.md: Update with new attributes.
2098 * config/arm/arm1136jfs.md: Update with new attributes.
2099 * config/arm/arm926ejs.md: Update with new attributes.
2100 * config/arm/cortex-a15.md: Update with new attributes.
2101 * config/arm/cortex-a5.md: Update with new attributes.
2102 * config/arm/cortex-a53.md: Update with new attributes.
2103 * config/arm/cortex-a7.md: Update with new attributes.
2104 * config/arm/cortex-a8.md: Update with new attributes.
2105 * config/arm/cortex-a9.md: Update with new attributes.
2106 * config/arm/cortex-m4.md: Update with new attributes.
2107 * config/arm/cortex-r4.md: Update with new attributes.
2108 * config/arm/fa526.md: Update with new attributes.
2109 * config/arm/fa606te.md: Update with new attributes.
2110 * config/arm/fa626te.md: Update with new attributes.
2111 * config/arm/fa726te.md: Update with new attributes.
2113 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2115 * config/aarch64/aarch64-simd.md
2116 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2117 <vwx> iterator to ensure correct register choice.
2118 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2119 (aarch64_sqdmull_n<mode>): Likewise.
2120 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2121 * config/aarch64/arm_neon.h
2122 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2123 (vml<as><q>_n_<su>16): Likewise.
2124 (vml<as>l_high_lane<q>_<su>16): Likewise.
2125 (vml<as>l_high_n_<su>16): Likewise.
2126 (vml<as>l_lane<q>_<su>16): Likewise.
2127 (vml<as>l_n_<su>16): Likewise.
2128 (vmul<q>_lane<q>_<su>16): Likewise.
2129 (vmul<q>_n_<su>16): Likewise.
2130 (vmull_lane<q>_<su>16): Likewise.
2131 (vmull_n_<su>16): Likewise.
2132 (vmull_high_lane<q>_<su>16): Likewise.
2133 (vmull_high_n_<su>16): Likewise.
2134 (vqrdmulh<q>_n_s16): Likewise.
2136 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2138 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2139 have the correct lane parameter.
2141 2013-09-06 Richard Biener <rguenther@suse.de>
2143 * cfganal.c (control_dependences::~control_dependences):
2144 Properly free all of the vector.
2146 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2149 * config/i386/i386.c (ix86_conditional_register_usage):
2150 Proper initialize extended SSE registers.
2152 2013-09-06 Jan Hubicka <jh@suse.cz>
2154 PR tree-optimization/58311
2155 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2157 2013-09-06 Jan Hubicka <jh@suse.cz>
2159 * Makefile.in (tree-sra.o): Update dependencies.
2160 * tree-sra.c: Include ipa-utils.h
2161 (scan_function): Use recursive_call_p.
2162 (has_caller_p): New function.
2163 (cgraph_for_node_and_aliases): Count also callers of aliases.
2165 2013-09-06 Jan Hubicka <jh@suse.cz>
2168 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2169 * tree-tailcall.c: Include ipa-utils.h.
2170 (find_tail_calls): Use it.
2171 * ipa-pure-const.c (check_call): Likewise.
2172 * ipa-utils.c (recursive_call_p): New function.
2173 * ipa-utils.h (recursive_call_p): Dclare.
2174 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2175 (symtab_semantically_equivalent_p): New function.
2176 * Makefile.in (tree-tailcall.o): Update dependencies.
2178 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2180 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2183 2013-09-06 Richard Biener <rguenther@suse.de>
2185 * lto-streamer.h (lto_global_var_decls): Remove.
2186 * Makefile.in (OBJS): Remove lto-symtab.o.
2187 (lto-symtab.o): Remove.
2188 (GTFILES): Remove lto-symtab.c
2189 * lto-symtab.c: Move to lto/
2191 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2193 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2194 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2195 (UNSPEC_FPINT_RINT): New constant definitions.
2196 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2197 definition with 2 attributes.
2198 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2199 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2202 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2204 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2205 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2206 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2207 attribute to "z196_alone".
2208 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2211 2013-09-06 Richard Biener <rguenther@suse.de>
2213 * basic-block.h (class control_dependences): New.
2214 * tree-ssa-dce.c (control_dependence_map): Remove.
2216 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2217 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2218 find_pdom, find_control_dependence, find_all_control_dependences):
2220 (mark_control_dependent_edges_necessary,
2221 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2222 tree_dce_done, perform_tree_ssa_dce): Adjust.
2223 * cfganal.c (set_control_dependence_map_bit,
2224 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2225 find_all_control_dependences): Move from tree-ssa-dce.c and
2226 implement as methods of control_dependences class.
2227 (control_dependences::control_dependences): New.
2228 (control_dependences::~control_dependences): Likewise.
2229 (control_dependences::get_edges_dependent_on): Likewise.
2230 (control_dependences::get_edge): Likewise.
2232 2013-09-04 Jan Hubicka <jh@suse.cz>
2234 * tree.c (types_same_for_odr): Drop overactive check.
2235 * ipa-devirt.c (hash_type_name): Likewise.
2237 2013-09-04 Jan Hubicka <jh@suse.cz>
2239 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2240 (analyze_functions): ... here.
2242 2013-09-04 Jan Hubicka <jh@suse.cz>
2245 * cgraphunit.c (analyze_functions): Clear AUX fields
2246 after processing; initialize assembler name has.
2248 2013-09-05 Jeff Law <law@redhat.com>
2250 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2251 from thread_around_empty_block. Record threading path into PATH.
2252 Recurse if threading through the initial block is successful.
2253 (thread_across_edge): Corresponding changes to slightly simplify.
2255 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2257 * config/aarch64/aarch64.md
2258 (type): Remove frecpe, frecps, frecpx.
2259 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2260 fix to be a TARGET_SIMD instruction.
2261 (aarch64_frecps): Remove.
2262 * config/aarch64/aarch64-simd.md
2263 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2264 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2266 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2267 Sofiane Naci <sofiane.naci@arm.com>
2269 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2270 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2271 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2272 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2273 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2274 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2275 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2276 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2277 into "clz, "rbit". Rename "shift" to "shift_imm".
2278 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2279 changes. Update for attribute changes all occurrences of arlo_* and
2281 * config/arm/arm-fixed.md: Update for attribute changes
2282 all occurrences of arlo_* types.
2283 * config/arm/thumb2.md: Update for attribute changes all occurrences
2285 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2286 (cortexa7_older_only): Likewise.
2287 (cortexa7_younger): Likewise.
2288 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2289 (1020alu_shift_op): Likewise.
2290 (1020alu_shift_reg_op): Likewise.
2291 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2292 (alu_shift_op): Likewise.
2293 (alu_shift_reg_op): Likewise.
2294 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2295 (11_alu_shift_op): Likewise.
2296 (11_alu_shift_reg_op): Likewise.
2297 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2298 (9_alu_shift_reg_op): Likewise.
2299 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2301 (cortex_a15_alu_shift): Likewise.
2302 (cortex_a15_alu_shift_reg): Likewise.
2303 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2305 (cortex_a5_alu_shift): Likewise.
2306 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2308 (cortex_a53_alu_shift): Likewise.
2309 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2311 (cortex_a7_alu_reg): Likewise.
2312 (cortex_a7_alu_shift): Likewise.
2313 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2315 (cortex_a8_alu_shift): Likewise.
2316 (cortex_a8_alu_shift_reg): Likewise.
2317 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2318 (cortex_a9_dp_shift): Likewise.
2319 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2321 * config/arm/cortex-r4.md
2322 (cortex_r4_alu): Update for attribute changes.
2323 (cortex_r4_mov): Likewise.
2324 (cortex_r4_alu_shift_reg): Likewise.
2325 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2326 (526_alu_shift_op): Likewise.
2327 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2328 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2329 (626te_alu_shift_op): Likewise.
2330 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2331 (726te_alu_shift_op): Likewise.
2332 (726te_alu_shift_reg_op): Likewise.
2333 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2334 (mp626_alu_shift_op): Likewise.
2335 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2336 (pj4_alu_conds): Likewise.
2337 (pj4_shift): Likewise.
2338 (pj4_shift_conds): Likewise.
2339 (pj4_alu_shift): Likewise.
2340 (pj4_alu_shift_conds): Likewise.
2341 * config/aarch64/aarch64.md: Update for attribute change
2342 all occurrences of arlo_* and shift* types.
2344 2013-09-05 Mike Stump <mikestump@comcast.net>
2346 * tree.h: Move documentation for tree_function_decl to tree-core.h
2347 with the declaration.
2349 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2352 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2353 looking for widest mode.
2355 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2357 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2359 2013-09-05 Richard Biener <rguenther@suse.de>
2361 PR tree-optimization/58137
2362 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2363 Do not create vectors of pointers.
2364 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2365 types for the components of the vector initializer.
2366 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2367 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2369 2013-09-05 Martin Jambor <mjambor@suse.cz>
2371 * ipa-prop.c (remove_described_reference): Accept missing references,
2372 return false if that hppens, otherwise return true.
2373 (cgraph_node_for_jfunc): New function.
2374 (try_decrement_rdesc_refcount): Likewise.
2375 (try_make_edge_direct_simple_call): Use them.
2376 (ipa_edge_removal_hook): Remove references from rdescs.
2377 (ipa_edge_duplication_hook): Clone rdescs and their references
2378 when the new edge has the same caller as the old one.
2379 * cgraph.c (cgraph_resolve_speculation): Remove speculative
2380 reference before removing any edges.
2382 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
2384 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
2386 * thumb2.md (thumb2_storewb_parisi): New pattern.
2388 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
2390 * config/aarch64/aarch64-option-extensions.def: Add
2391 AARCH64_OPT_EXTENSION of 'crc'.
2392 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
2393 (AARCH64_ISA_CRC): Ditto.
2394 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
2395 description of the CRC extension.
2397 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
2399 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
2400 * config/rs6000/linux.h: Ditto.
2401 * alpha/linux.h: Ditto.
2402 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
2403 no_c99_libc_has_function.
2404 * config/c6x/uclinux-elf.h: Ditto.
2405 * config/lm32/uclinux-elf.h: Ditto.
2406 * config/m68k/uclinux.h: Ditto.
2407 * config/moxie/uclinux.h: Ditto.
2408 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
2409 (crisv32-*-linux*, cris-*-linux*): Ditto.
2410 * config/bfin/bfin.c: Include "tm_p.h".
2412 2013-09-05 Richard Biener <rguenther@suse.de>
2414 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
2415 check for a definition without a basic-block.
2417 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2418 Sofiane Naci <sofiane.naci@arm.com>
2420 * config/aarch64/aarch64.md
2421 (*movti_aarch64): Rename r_2_f and f_2_r.
2422 (*movsf_aarch64): Likewise.
2423 (*movdf_aarch64): Likewise.
2424 (*movtf_aarch64): Likewise.
2425 (aarch64_movdi_<mode>low): Likewise.
2426 (aarch64_movdi_<mode>high): Likewise.
2427 (aarch64_mov<mode>high_di): Likewise.
2428 (aarch64_mov<mode>low_di): Likewise.
2429 (aarch64_movtilow_tilow): Likewise.
2430 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
2431 values to config/arm/types.md
2432 (attribute "conds"): Update for attribute change.
2433 (anddi3_insn): Likewise.
2434 (iordi3_insn): Likewise.
2435 (xordi3_insn): Likewise.
2436 (one_cmpldi2): Likewise.
2437 * config/arm/types.md (type): Add Neon types.
2438 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
2439 use "type" attribute.
2440 (movmisalign<mode>_neon_store): Likewise.
2441 (movmisalign<mode>_neon_load): Likewise.
2442 (vec_set<mode>_internal): Likewise.
2443 (vec_setv2di_internal): Likewise.
2444 (vec_extract<mode>): Likewise.
2445 (vec_extractv2di): Likewise.
2446 (add<mode>3_neon): Likewise.
2447 (adddi3_neon): Likewise.
2448 (sub<mode>3_neon): Likewise.
2449 (subdi3_neon): Likewise.
2450 (mul<mode>3_neon): Likewise.
2451 (mul<mode>3add<mode>_neon): Likewise.
2452 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
2453 (fma<VCVTF:mode>4)): Likewise.
2454 (fma<VCVTF:mode>4_intrinsic): Likewise.
2455 (fmsub<VCVTF:mode>4)): Likewise.
2456 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2457 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2458 (ior<mode>3): Likewise.
2459 (and<mode>3): Likewise.
2460 (anddi3_neon): Likewise.
2461 (orn<mode>3_neon): Likewise.
2462 (orndi3_neon): Likewise.
2463 (bic<mode>3_neon): Likewise.
2464 (bicdi3_neon): Likewise.
2465 (xor<mode>3): Likewise.
2466 (one_cmpl<mode>2): Likewise.
2467 (abs<mode>2): Likewise.
2468 (neg<mode>2): Likewise.
2469 (umin<mode>3_neon): Likewise.
2470 (umax<mode>3_neon): Likewise.
2471 (smin<mode>3_neon): Likewise.
2472 (smax<mode>3_neon): Likewise.
2473 (vashl<mode>3): Likewise.
2474 (vashr<mode>3_imm): Likewise.
2475 (vlshr<mode>3_imm): Likewise.
2476 (ashl<mode>3_signed): Likewise.
2477 (ashl<mode>3_unsigned): Likewise.
2478 (neon_load_count): Likewise.
2479 (ashldi3_neon_noclobber): Likewise.
2480 (signed_shift_di3_neon): Likewise.
2481 (unsigned_shift_di3_neon): Likewise.
2482 (ashrdi3_neon_imm_noclobber): Likewise.
2483 (lshrdi3_neon_imm_noclobber): Likewise.
2484 (widen_ssum<mode>3): Likewise.
2485 (widen_usum<mode>3): Likewise.
2486 (quad_halves_<code>v4si): Likewise.
2487 (quad_halves_<code>v4sf): Likewise.
2488 (quad_halves_<code>v8hi): Likewise.
2489 (quad_halves_<code>v16qi): Likewise.
2490 (reduc_splus_v2di): Likewise.
2491 (neon_vpadd_internal<mode>): Likewise.
2492 (neon_vpsmin<mode>): Likewise.
2493 (neon_vpsmax<mode>): Likewise.
2494 (neon_vpumin<mode>): Likewise.
2495 (neon_vpumax<mode>): Likewise.
2496 (ss_add<mode>_neon): Likewise.
2497 (us_add<mode>_neon): Likewise.
2498 (ss_sub<mode>_neon): Likewise.
2499 (us_sub<mode>_neon): Likewise.
2500 (neon_vadd<mode>_unspec): Likewise.
2501 (neon_vaddl<mode>): Likewise.
2502 (neon_vaddw<mode>): Likewise.
2503 (neon_vhadd<mode>): Likewise.
2504 (neon_vqadd<mode>): Likewise.
2505 (neon_vaddhn<mode>): Likewise.
2506 (neon_vmul<mode>): Likewise.
2507 (neon_vmla<mode>): Likewise.
2508 (neon_vmlal<mode>): Likewise.
2509 (neon_vmls<mode>): Likewise.
2510 (neon_vmlsl<mode>): Likewise.
2511 (neon_vqdmulh<mode>): Likewise.
2512 (neon_vqdmlal<mode>): Likewise.
2513 (neon_vqdmlsl<mode>): Likewise.
2514 (neon_vmull<mode>): Likewise.
2515 (neon_vqdmull<mode>): Likewise.
2516 (neon_vsub<mode>_unspec): Likewise.
2517 (neon_vsubl<mode>): Likewise.
2518 (neon_vsubw<mode>): Likewise.
2519 (neon_vqsub<mode>): Likewise.
2520 (neon_vhsub<mode>): Likewise.
2521 (neon_vsubhn<mode>): Likewise.
2522 (neon_vceq<mode>): Likewise.
2523 (neon_vcge<mode>): Likewise.
2524 (neon_vcgeu<mode>): Likewise.
2525 (neon_vcgt<mode>): Likewise.
2526 (neon_vcgtu<mode>): Likewise.
2527 (neon_vcle<mode>): Likewise.
2528 (neon_vclt<mode>): Likewise.
2529 (neon_vcage<mode>): Likewise.
2530 (neon_vcagt<mode>): Likewise.
2531 (neon_vtst<mode>): Likewise.
2532 (neon_vabd<mode>): Likewise.
2533 (neon_vabdl<mode>): Likewise.
2534 (neon_vaba<mode>): Likewise.
2535 (neon_vabal<mode>): Likewise.
2536 (neon_vmax<mode>): Likewise.
2537 (neon_vmin<mode>): Likewise.
2538 (neon_vpaddl<mode>): Likewise.
2539 (neon_vpadal<mode>): Likewise.
2540 (neon_vpmax<mode>): Likewise.
2541 (neon_vpmin<mode>): Likewise.
2542 (neon_vrecps<mode>): Likewise.
2543 (neon_vrsqrts<mode>): Likewise.
2544 (neon_vqabs<mode>): Likewise.
2545 (neon_vqneg<mode>): Likewise.
2546 (neon_vcls<mode>): Likewise.
2547 (clz<mode>2): Likewise.
2548 (popcount<mode>2): Likewise.
2549 (neon_vrecpe): Likewise.
2550 (neon_vrsqrte): Likewise.
2551 (neon_vget_lane<mode>_sext_internal): Likewise.
2552 (neon_vget_lane<mode>_zext_internal): Likewise.
2553 (neon_vdup_n<mode>): Likewise.
2554 (neon_vdup_nv2di): Likewise.
2555 (neon_vdpu_lane<mode>_internal): Likewise.
2556 (neon_vswp<mode>): Likewise.
2557 (float<mode><V_cvtto>2): Likewise.
2558 (floatuns<mode><V_cvtto>2): Likewise.
2559 (fix_trunc<mode><V_cvtto>)2): Likewise
2560 (fixuns_trunc<mode><V_cvtto)2): Likewise.
2561 (neon_vcvt<mode>): Likewise.
2562 (neon_vcvtv4sfv4hf): Likewise.
2563 (neon_vcvtv4hfv4sf): Likewise.
2564 (neon_vcvt_n<mode>): Likewise.
2565 (neon_vmovn<mode>): Likewise.
2566 (neon_vqmovn<mode>): Likewise.
2567 (neon_vqmovun<mode>): Likewise.
2568 (neon_vmovl<mode>): Likewise.
2569 (neon_vmul_lane<mode>): Likewise.
2570 (neon_vmull_lane<mode>): Likewise.
2571 (neon_vqdmull_lane<mode>): Likewise.
2572 (neon_vqdmulh_lane<mode>): Likewise.
2573 (neon_vmla_lane<mode>): Likewise.
2574 (neon_vmlal_lane<mode>): Likewise.
2575 (neon_vqdmlal_lane<mode>): Likewise.
2576 (neon_vmls_lane<mode>): Likewise.
2577 (neon_vmlsl_lane<mode>): Likewise.
2578 (neon_vqdmlsl_lane<mode>): Likewise.
2579 (neon_vext<mode>): Likewise.
2580 (neon_vrev64<mode>): Likewise.
2581 (neon_vrev32<mode>): Likewise.
2582 (neon_vrev16<mode>): Likewise.
2583 (neon_vbsl<mode>_internal): Likewise.
2584 (neon_vshl<mode>): Likewise.
2585 (neon_vqshl<mode>): Likewise.
2586 (neon_vshr_n<mode>): Likewise.
2587 (neon_vshrn_n<mode>): Likewise.
2588 (neon_vqshrn_n<mode>): Likewise.
2589 (neon_vqshrun_n<mode>): Likewise.
2590 (neon_vshl_n<mode>): Likewise.
2591 (neon_vqshl_n<mode>): Likewise.
2592 (neon_vqshlu_n<mode>): Likewise.
2593 (neon_vshll_n<mode>): Likewise.
2594 (neon_vsra_n<mode>): Likewise.
2595 (neon_vsri_n<mode>): Likewise.
2596 (neon_vsli_n<mode>): Likewise.
2597 (neon_vtbl1v8qi): Likewise.
2598 (neon_vtbl2v8qi): Likewise.
2599 (neon_vtbl3v8qi): Likewise.
2600 (neon_vtbl4v8qi): Likewise.
2601 (neon_vtbx1v8qi): Likewise.
2602 (neon_vtbx2v8qi): Likewise.
2603 (neon_vtbx3v8qi): Likewise.
2604 (neon_vtbx4v8qi): Likewise.
2605 (neon_vtrn<mode>_internal): Likewise.
2606 (neon_vzip<mode>_internal): Likewise.
2607 (neon_vuzp<mode>_internal): Likewise.
2608 (neon_vld1<mode>): Likewise.
2609 (neon_vld1_lane<mode>): Likewise.
2610 (neon_vld1_dup<mode>): Likewise.
2611 (neon_vld1_dupv2di): Likewise.
2612 (neon_vst1<mode>): Likewise.
2613 (neon_vst1_lane<mode>): Likewise.
2614 (neon_vld2<mode>): Likewise.
2615 (neon_vld2_lane<mode>): Likewise.
2616 (neon_vld2_dup<mode>): Likewise.
2617 (neon_vst2<mode>): Likewise.
2618 (neon_vst2_lane<mode>): Likewise.
2619 (neon_vld3<mode>): Likewise.
2620 (neon_vld3qa<mode>): Likewise.
2621 (neon_vld3qb<mode>): Likewise.
2622 (neon_vld3_lane<mode>): Likewise.
2623 (neon_vld3_dup<mode>): Likewise.
2624 (neon_vst3<mode>): Likewise.
2625 (neon_vst3qa<mode>): Likewise.
2626 (neon_vst3qb<mode>): Likewise.
2627 (neon_vst3_lane<mode>): Likewise.
2628 (neon_vld4<mode>): Likewise.
2629 (neon_vld4qa<mode>): Likewise.
2630 (neon_vld4qb<mode>): Likewise.
2631 (neon_vld4_lane<mode>): Likewise.
2632 (neon_vld4_dup<mode>): Likewise.
2633 (neon_vst4<mode>): Likewise.
2634 (neon_vst4qa<mode>): Likewise.
2635 (neon_vst4qb<mode>): Likewise.
2636 (neon_vst4_lane<mode>): Likewise.
2637 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2638 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2639 (neon_vec_<US>mult_lo_<mode>): Likewise.
2640 (neon_vec_<US>mult_hi_<mode>): Likewise.
2641 (neon_vec_<US>shiftl_<mode>): Likewise.
2642 (neon_unpack<US>_<mode>): Likewise.
2643 (neon_vec_<US>mult_<mode>): Likewise.
2644 (vec_pack_trunc_<mode>): Likewise.
2645 (neon_vec_pack_trunk_<mode>): Likewise.
2646 (neon_vabd<mode>_2): Likewise.
2647 (neon_vabd<mode>_3): Likewise.
2648 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2649 (thumb2_movsi_vfp): Likewise.
2650 (movdi_vfp): Likewise.
2651 (movdi_vfp_cortexa8): Likewise.
2652 (movhf_vfp_neon): Likewise.
2653 (movhf_vfp): Likewiwse.
2654 (movsf_vfp): Likewiwse.
2655 (thumb2_movsf_vfp): Likewiwse.
2656 (movdf_vfp): Likewise.
2657 (thumb2_movdf_vfp): Likewise.
2658 (movsfcc_vfp): Likewise.
2659 (thumb2_movsfcc_vfp): Likewise.
2660 (movdfcc_vfp): Likewise.
2661 (thumb2_movdfcc_vfp): Likewise.
2662 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
2663 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
2664 (v10_v2c): Likewise.
2665 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
2667 (cortex_a15_neon_int_2): Likewise.
2668 (cortex_a15_neon_int_3): Likewise.
2669 (cortex_a15_neon_int_4): Likewise.
2670 (cortex_a15_neon_int_5): Likewise.
2671 (cortex_a15_neon_vqneg_vqabs): Likewise.
2672 (cortex_a15_neon_vmov): Likewise.
2673 (cortex_a15_neon_vaba): Likewise.
2674 (cortex_a15_neon_vaba_qqq): Likewise.
2675 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2676 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2677 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
2678 scalar_64_32_long_scalar): Likewise.
2679 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2680 (cortex_a15_neon_mla_qqq_8_16): Likewise.
2681 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2682 lotype_qdd_64_32_long): Likewise.
2683 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2684 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2685 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
2686 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2687 (cortex_a15_neon_shift_1): Likewise.
2688 (cortex_a15_neon_shift_2): Likewise.
2689 (cortex_a15_neon_shift_3): Likewise.
2690 (cortex_a15_neon_vshl_ddd): Likewise.
2691 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2692 (cortex_a15_neon_vsra_vrsra): Likewise.
2693 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
2694 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
2695 (cortex_a15_neon_fp_vmul_ddd): Likewise.
2696 (cortex_a15_neon_fp_vmul_qqd): Likewise.
2697 (cortex_a15_neon_fp_vmla_ddd): Likewise.
2698 (cortex_a15_neon_fp_vmla_qqq): Likewise.
2699 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2700 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2701 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2702 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2703 (cortex_a15_neon_bp_simple): Likewise.
2704 (cortex_a15_neon_bp_2cycle): Likewise.
2705 (cortex_a15_neon_bp_3cycle): Likewise.
2706 (cortex_a15_neon_vld1_1_2_regs): Likewise.
2707 (cortex_a15_neon_vld1_3_4_regs): Likewise.
2708 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2709 (cortex_a15_neon_vld2_4_regs): Likewise.
2710 (cortex_a15_neon_vld3_vld4): Likewise.
2711 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2712 (cortex_a15_neon_vst1_3_4_regs): Likewise.
2713 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
2714 (cortex_a15_neon_vst3_vst4): Likewise.
2715 (cortex_a15_neon_vld1_vld2_lane): Likewise.
2716 (cortex_a15_neon_vld3_vld4_lane" 10
2717 (cortex_a15_neon_vst1_vst2_lane): Likewise.
2718 (cortex_a15_neon_vst3_vst4_lane): Likewise.
2719 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
2720 (cortex_a15_neon_ldm_2): Likewise.0
2721 (cortex_a15_neon_stm_2): Likewise.
2722 (cortex_a15_neon_mcr): Likewise.
2723 (cortex_a15_neon_mcr_2_mcrr): Likewise.
2724 (cortex_a15_neon_mrc): Likewise.
2725 (cortex_a15_neon_mrrc): Likewise.
2726 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2728 (cortex_a15_alu_shift): Likewise.
2729 (cortex_a15_alu_shift_reg): Likewise.
2730 (cortex_a15_mult32): Likewise.
2731 (cortex_a15_mult64): Likewise.
2732 (cortex_a15_block): Likewise.
2733 (cortex_a15_branch): Likewise.
2734 (cortex_a15_load1): Likewise.
2735 (cortex_a15_load3): Likewise.
2736 (cortex_a15_store1): Likewise.
2737 (cortex_a15_store3): Likewise.
2738 (cortex_a15_call): Likewise.
2739 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
2740 (cortex_a5_f2r): Likewise.
2741 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
2743 (cortex_a53_f2r): Likewise.
2744 * config/arm/cortex-a7.md
2745 (cortex_a7_branch): Update for attribute change.
2746 (cortex_a7_call): Likewise.
2747 (cortex_a7_alu_imm): Likewise.
2748 (cortex_a7_alu_reg): Likewise.
2749 (cortex_a7_alu_shift): Likewise.
2750 (cortex_a7_mul): Likewise.
2751 (cortex_a7_load1): Likewise.
2752 (cortex_a7_store1): Likewise.
2753 (cortex_a7_load2): Likewise.
2754 (cortex_a7_store2): Likewise.
2755 (cortex_a7_load3): Likewise.
2756 (cortex_a7_store3): Likewise.
2757 (cortex_a7_load4): Likewise.
2758 (cortex_a7_store4): Likewise.
2759 (cortex_a7_fpalu): Likewise.
2760 (cortex_a7_fconst): Likewise.
2761 (cortex_a7_fpmuls): Likewise.
2762 (cortex_a7_neon_mul): Likewise.
2763 (cortex_a7_fpmacs): Likewise.
2764 (cortex_a7_neon_mla: Likewise.
2765 (cortex_a7_fpmuld: Likewise.
2766 (cortex_a7_fpmacd: Likewise.
2767 (cortex_a7_fpfmad: Likewise.
2768 (cortex_a7_fdivs: Likewise.
2769 (cortex_a7_fdivd: Likewise.
2770 (cortex_a7_r2f: Likewise.
2771 (cortex_a7_f2r: Likewise.
2772 (cortex_a7_f_flags: Likewise.
2773 (cortex_a7_f_loads: Likewise.
2774 (cortex_a7_f_loadd: Likewise.
2775 (cortex_a7_f_stores: Likewise.
2776 (cortex_a7_f_stored: Likewise.
2777 (cortex_a7_neon): Likewise.
2778 * config/arm/cortex-a8-neon.md
2779 (cortex_a8_neon_mrc): Update for attribute change.
2780 (cortex_a8_neon_mrrc): Likewise.
2781 (cortex_a8_neon_int_1): Likewise.
2782 (cortex_a8_neon_int_2): Likewise.
2783 (cortex_a8_neon_int_3): Likewise.
2784 (cortex_a8_neon_int_4): Likewise.
2785 (cortex_a8_neon_int_5): Likewise.
2786 (cortex_a8_neon_vqneg_vqabs): Likewise.
2787 (cortex_a8_neon_vmov): Likewise.
2788 (cortex_a8_neon_vaba): Likewise.
2789 (cortex_a8_neon_vaba_qqq): Likewise.
2790 (cortex_a8_neon_vsma): Likewise.
2791 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2792 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2793 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2794 long_scalar): Likewise.
2795 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2796 (cortex_a8_neon_mla_qqq_8_16): Likewise.
2797 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2798 long_scalar_qdd_64_32_long): Likewise.
2799 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2800 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2801 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
2802 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2803 (cortex_a8_neon_shift_1): Likewise.
2804 (cortex_a8_neon_shift_2): Likewise.
2805 (cortex_a8_neon_shift_3): Likewise.
2806 (cortex_a8_neon_vshl_ddd): Likewise.
2807 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2808 (cortex_a8_neon_vsra_vrsra): Likewise.
2809 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
2810 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
2811 (cortex_a8_neon_fp_vsum): Likewise.
2812 (cortex_a8_neon_fp_vmul_ddd): Likewise.
2813 (cortex_a8_neon_fp_vmul_qqd): Likewise.
2814 (cortex_a8_neon_fp_vmla_ddd): Likewise.
2815 (cortex_a8_neon_fp_vmla_qqq): Likewise.
2816 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
2817 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
2818 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2819 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2820 (cortex_a8_neon_bp_simple): Likewise.
2821 (cortex_a8_neon_bp_2cycle): Likewise.
2822 (cortex_a8_neon_bp_3cycle): Likewise.
2823 (cortex_a8_neon_ldr): Likewise.
2824 (cortex_a8_neon_str): Likewise.
2825 (cortex_a8_neon_vld1_1_2_regs): Likewise.
2826 (cortex_a8_neon_vld1_3_4_regs): Likewise.
2827 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2828 (cortex_a8_neon_vld2_4_regs): Likewise.
2829 (cortex_a8_neon_vld3_vld4): Likewise.
2830 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2831 (cortex_a8_neon_vst1_3_4_regs): Likewise.
2832 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
2833 (cortex_a8_neon_vst3_vst4): Likewise.
2834 (cortex_a8_neon_vld1_vld2_lane): Likewise.
2835 (cortex_a8_neon_vld3_vld4_lane): Likewise.
2836 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2837 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2838 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
2839 (cortex_a8_neon_mcr): Likewise.
2840 (cortex_a8_neon_mcr_2_mcrr): Likewise.
2841 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2842 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
2844 (ca9_neon_mrrc): Likewise.
2845 (cortex_a9_neon_int_1): Likewise.
2846 (cortex_a9_neon_int_2): Likewise.
2847 (cortex_a9_neon_int_3): Likewise.
2848 (cortex_a9_neon_int_4): Likewise.
2849 (cortex_a9_neon_int_5): Likewise.
2850 (cortex_a9_neon_vqneg_vqabs): Likewise.
2851 (cortex_a9_neon_vmov): Likewise.
2852 (cortex_a9_neon_vaba): Likewise.
2853 (cortex_a9_neon_vaba_qqq): Likewise.
2854 (cortex_a9_neon_vsma): Likewise.
2855 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2856 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2857 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2858 long_scalar): Likewise.
2859 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2860 (cortex_a9_neon_mla_qqq_8_16): Likewise.
2861 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2862 long_scalar_qdd_64_32_long): Likewise.
2863 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2864 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2865 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
2866 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2867 (cortex_a9_neon_shift_1): Likewise.
2868 (cortex_a9_neon_shift_2): Likewise.
2869 (cortex_a9_neon_shift_3): Likewise.
2870 (cortex_a9_neon_vshl_ddd): Likewise.
2871 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2872 (cortex_a9_neon_vsra_vrsra): Likewise.
2873 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
2874 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
2875 (cortex_a9_neon_fp_vsum): Likewise.
2876 (cortex_a9_neon_fp_vmul_ddd): Likewise.
2877 (cortex_a9_neon_fp_vmul_qqd): Likewise.
2878 (cortex_a9_neon_fp_vmla_ddd): Likewise.
2879 (cortex_a9_neon_fp_vmla_qqq): Likewise.
2880 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
2881 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
2882 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2883 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2884 (cortex_a9_neon_bp_simple): Likewise.
2885 (cortex_a9_neon_bp_2cycle): Likewise.
2886 (cortex_a9_neon_bp_3cycle): Likewise.
2887 (cortex_a9_neon_ldr): Likewise.
2888 (cortex_a9_neon_str): Likewise.
2889 (cortex_a9_neon_vld1_1_2_regs): Likewise.
2890 (cortex_a9_neon_vld1_3_4_regs): Likewise.
2891 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2892 (cortex_a9_neon_vld2_4_regs): Likewise.
2893 (cortex_a9_neon_vld3_vld4): Likewise.
2894 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2895 (cortex_a9_neon_vst1_3_4_regs): Likewise.
2896 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
2897 (cortex_a9_neon_vst3_vst4): Likewise.
2898 (cortex_a9_neon_vld1_vld2_lane): Likewise.
2899 (cortex_a9_neon_vld3_vld4_lane): Likewise.
2900 (cortex_a9_neon_vst1_vst2_lane): Likewise.
2901 (cortex_a9_neon_vst3_vst4_lane): Likewise.
2902 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2903 (cortex_a9_neon_mcr): Likewise.
2904 (cortex_a9_neon_mcr_2_mcrr): Likewise.
2905 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2906 (cortex_a9_fps): Likewise.
2907 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
2909 (cortex_m4_fmuls): Likewise.
2910 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
2912 (cortex_r4_mrc): Likewise.
2913 * config/arm/iterators.md: Update comment referring to neon_type.
2914 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
2915 (iwmmxt_movsi_insn): Likewise.
2916 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
2918 (pj4_core_to_vfp): Likewise.
2919 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
2921 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
2922 (vfp_fstore): Likewise.
2923 * doc/md.texi: Change references to neon_type to refer to type.
2925 2013-09-04 Dodji Seketeli <dodji@redhat.com>
2927 * tree.h (DECL_BUILT_IN): Fix typo in comment.
2929 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
2931 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
2934 2013-09-04 Easwaran Raman <eraman@google.com>
2937 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
2938 (build_and_add_sum): Use it.
2939 (appears_later_in_bb): Simplify code.
2941 2013-09-04 Teresa Johnson <tejohnson@google.com>
2943 * dumpfile.c (dump_finish): Don't close stderr/stdout.
2945 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
2947 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
2949 2013-09-04 Jan Hubicka <jh@suse.cz>
2951 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
2952 * ipa-devirt.c: Include diganostic.h
2953 (odr_type_d): Add types and types_set.
2954 (hash_type_name): Work for types with vtables during LTO.
2955 (odr_hasher::remove): Fix comment; destroy types_set.
2956 (add_type_duplicate): New function,
2957 (get_odr_type): Use it.
2958 (dump_type_inheritance_graph): Dump type duplicates.
2959 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
2961 * tree.c (types_same_for_odr): Give exact answers on types with
2964 2013-09-04 Dodji Seketeli <dodji@redhat.com>
2966 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
2967 explaining their differences.
2969 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
2971 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
2973 2013-09-03 Jeff Law <law@redhat.com>
2975 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
2976 when not threading through a joiner block. Pass joiner/no joiner
2977 state to register_jump_thread.
2978 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
2979 state from argument rather than implying on path length.
2980 Dump the entire jump thread path into debugging dump.
2981 * tree-flow.h (register_jump_thread): Update prototype.
2983 2013-08-29 Xinliang David Li <davidxl@google.com>
2985 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2986 Remove a trivial gcc_assert.
2988 2013-08-29 Xinliang David Li <davidxl@google.com>
2990 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
2991 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
2992 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2993 Delay base decl alignment adjustment.
2994 * tree-vectorizer.c (vect_destroy_datarefs): New function.
2995 * tree-vectorizer.h: New data structure.
2996 (set_dr_misalignment): New function.
2997 (dr_misalignment): Ditto.
2998 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
2999 (vectorizable_load): Ditto.
3000 (ensure_base_align): New function.
3001 (vectorize_loops): Add dbg_cnt support.
3002 (execute_vect_slp): Ditto.
3003 * dbgcnt.def: New debug counter.
3004 * Makefile: New dependency.
3006 2013-09-03 Meador Inge <meadori@codesourcery.com>
3010 2013-08-30 Meador Inge <meadori@codesourcery.com>
3012 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3014 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3016 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3017 function descriptor.
3019 2013-09-03 Richard Biener <rguenther@suse.de>
3021 * tree-affine.c (add_elt_to_tree): Fix association issue,
3022 avoid useless converts and make sure to always return a
3023 properly typed result.
3025 2013-09-03 Richard Biener <rguenther@suse.de>
3028 * fold-const.c (negate_expr_p): Fix division case.
3029 (negate_expr): Likewise.
3031 2013-09-03 Richard Biener <rguenther@suse.de>
3034 * tree-streamer-out.c: Include tm.h.
3035 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3037 2013-09-03 Jan Hubicka <jh@suse.cz>
3039 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3041 2013-09-03 Alan Modra <amodra@gmail.com>
3043 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3044 multilibs to big-endian target and vice versa.
3045 * config/rs6000/t-linux64: Use := assignment on all vars.
3046 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3047 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3048 * config/rs6000/t-linux64le: New file.
3049 * config/rs6000/t-linux64bele: New file.
3050 * config/rs6000/t-linux64lebe: New file.
3052 2013-09-02 Jan Hubicka <jh@suse.cz>
3054 * ipa-inline-transform.c (inline_transform): Do not
3055 optimize_inline_calls when not optimizing.
3057 2013-09-02 Jan Hubicka <jh@suse.cz>
3059 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3060 duplicated nodes for assembler names.
3061 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3062 hard registers from assembler name hash.
3064 2013-09-02 Jan Hubicka <jh@suse.cz>
3066 * ipa-split.c (execute_split_functions): Split externally visible
3067 functions called once.
3069 2013-09-02 Martin Jambor <mjambor@suse.cz>
3072 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3073 linked list. When finding the correct duplicate, also consider also
3074 the caller in additon to its inlined_to node.
3076 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3078 * config/aarch64/aarch64-simd-builtins.def
3079 (dup_lane_scalar): Remove.
3080 * config/aarch64/aarch64-simd.md
3081 (aarch64_simd_dup): Add 'w->w' alternative.
3082 (aarch64_dup_lane<mode>): Allow for VALL.
3083 (aarch64_dup_lane_scalar<mode>): Remove.
3084 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3085 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3086 (aarch64_get_lane_unsigned<mode>): Likewise.
3087 (aarch64_get_lane<mode>): Likewise.
3088 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3089 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3090 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3091 (VCON): Change container of V2SF.
3092 (vswap_width_name): Likewise.
3093 * config/aarch64/arm_neon.h
3094 (__aarch64_vdup_lane_any): New.
3095 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3096 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3097 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3099 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3102 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3103 the source or the destination is a single hard register.
3105 2013-09-02 Richard Biener <rguenther@suse.de>
3108 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3111 2013-09-02 Richard Biener <rguenther@suse.de>
3113 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3114 arithmetic to sizetype.
3116 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3118 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3119 Find auto-increment use both before and after candidate.
3121 2013-09-02 Marek Polacek <polacek@redhat.com>
3123 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3125 2013-09-01 Jan Hubicka <jh@suse.cz>
3127 * Makefile.in: Add ipa-profile.o
3128 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3129 * cgraph.c (struct cgraph_propagate_frequency_data,
3130 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3131 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3132 * cgraph.h (cgraph_propagate_frequency): Remove.
3133 * ipa-inline-analysis.c: Include ipa-utils.h;
3134 drop duplicated cfgloop.h.
3135 (inline_update_callee_summaries): Update.
3136 * ipa-profile.c: New file.
3137 * ipa-utils.h (ipa_propagate_frequency): Declare.
3138 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3139 data-streamer.h, value-prof.h.
3140 (symtab_remove_unreachable_nodes): Update profile.
3141 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3142 account_time_size, cmp_counts, dump_histogram,
3143 ipa_profile_generate_summary, ipa_profile_write_summary,
3144 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3145 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3146 Move to ipa-profile.c.
3148 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3150 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3152 2013-09-01 Jan Hubicka <jh@suse.cz>
3154 * common.opt (fdevirtualize-speculatively): New function.
3155 * invoke.texi (fdevirtualize-speculatively): Document.
3156 * ipa-devirt.c: Include ipa-inline.h
3157 (likely_target_p): New function.
3158 (ipa_devirt): New function.
3159 (gate_ipa_devirt): New function.
3160 (pass_data_ipa_devirt): New static var.
3161 (pass_ipa_devirt): Likewise.
3162 (make_pass_ipa_devirt): New function.
3163 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3164 (common_handle_option): Disable devirtualization when
3165 value range profiling is available.
3166 * passes.def (pass_ipa_devirt): Add.
3167 * timever.def (TV_IPA_DEVIRT): New timevar.
3168 * tree-pass.h (make_pass_ipa_devirt):
3170 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3172 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3173 include sanitize(undefined).
3175 2013-08-31 Diego Novillo <dnovillo@google.com>
3177 * Makefile.in (TREE_CORE_H): Define.
3179 (GTFILES): Add tree-core.h.
3180 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3182 * tree-core.h: New file.
3183 Move all data structures, enum, typedefs, global
3184 declarations and constants from ...
3187 2013-08-31 Jan Hubicka <jh@suse.cz>
3189 * bulitins.c (expand_builtin): Do not early exit for gcov
3190 instrumented functions.
3192 2013-08-31 Marek Polacek <polacek@redhat.com>
3194 * ubsan.c: Include tm_p.h.
3196 2013-08-31 Jan Hubicka <jh@suse.cz>
3198 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3201 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3202 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3203 * tree-cfg.c (verify_gimple_label): ... here.
3204 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3205 (ipa_merge_profiles): New function.
3206 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3207 (lto_input_function_body): Likewise.
3208 * ipa-utils.h (ipa_merge_profiles): Declare.
3209 * lto-streamer.h (lto_input_function_body): Update prototype.
3210 (emit_label_in_global_context_p): Remove.
3211 * lto-symtab.c: Include ipa-utils.h
3212 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3214 2013-08-31 Jan Hubicka <jh@suse.cz>
3216 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3218 2013-08-31 Jan Hubicka <jh@suse.cz>
3220 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3222 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3224 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3225 "cmp" RTX before signed_comparison_operator check to account
3228 2013-08-30 Jan Hubicka <jh@suse.cz>
3230 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3231 (detect_type_change_1): Rename to ...
3232 (detect_type_change): ... this one; early return on non-polymorphic
3234 (detect_type_change_ssa): Add comp_type parameter; update
3235 use of detect_type_change.
3236 (compute_complex_assign_jump_func): Add param_type parameter;
3237 update use of detect_type_change_ssa.
3238 (compute_complex_ancestor_jump_func): Likewise.
3239 (ipa_get_callee_param_type): New function.
3240 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3241 update calls to the jump function computation functions.
3243 2013-08-30 Teresa Johnson <tejohnson@google.com>
3244 Steven Bosscher <steven@gcc.gnu.org>
3246 * cfgrtl.c (fixup_new_cold_bb): New routine.
3247 (commit_edge_insertions): Invoke fixup_partitions.
3248 (find_partition_fixes): New routine.
3249 (fixup_partitions): Ditto.
3250 (verify_hot_cold_block_grouping): Update comments.
3251 (rtl_verify_edges): Invoke find_partition_fixes.
3252 (rtl_verify_bb_pointers): Update comments.
3253 (rtl_verify_bb_layout): Ditto.
3254 * basic-block.h (probably_never_executed_edge_p): Declare.
3255 (fixup_partitions): Ditto.
3256 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3257 * bb-reorder.c (sanitize_hot_paths): New function.
3258 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3260 * predict.c (probably_never_executed_edge_p): New routine.
3261 * cfg.c (check_bb_profile): Add partition insanity warnings.
3263 2013-08-30 Meador Inge <meadori@codesourcery.com>
3265 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3267 2013-08-30 Marek Polacek <polacek@redhat.com>
3269 * Makefile.in (ubsan.o): Add.
3270 (c-family/c-ubsan.o): Add.
3271 (builtins.o): Add ubsan.h dependency.
3272 * ubsan.h: New file.
3273 * ubsan.c: New file.
3274 * common.opt: Add -fsanitize=undefined option.
3275 (flag_sanitize): Add variable.
3276 (fsanitize=): Add option. Add Driver.
3277 (fsanitize=thread): Remove option.
3278 (fsanitize=address): Likewise.
3279 (static-libubsan): New option.
3280 * doc/invoke.texi: Document the new flag and -static-libubsan.
3281 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3282 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3283 * builtin-attrs.def (ATTR_COLD): Define.
3284 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3285 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3286 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3287 * flag-types.h (sanitize_code): New enum.
3288 * opts.c (common_handle_option): Parse command line arguments
3289 of -fsanitize=. Add -fsanitize=unreachable option.
3290 * varasm.c (get_variable_section): Adjust.
3291 (assemble_noswitch_variable): Likewise.
3292 (assemble_variable): Likewise.
3293 (output_constant_def_contents): Likewise.
3294 (categorize_decl_for_section): Likewise.
3295 (place_block_symbol): Likewise.
3296 (output_object_block): Likewise.
3297 * builtins.def: Likewise.
3298 * toplev.c (compile_file): Likewise.
3299 (process_options): Likewise.
3300 * cppbuiltin.c: Likewise.
3301 * tsan.c (tsan_pass): Likewise.
3302 (tsan_gate): Likewise.
3303 (tsan_gate_O0): Likewise.
3304 * cfgexpand.c (partition_stack_vars): Likewise.
3305 (expand_stack_vars): Likewise.
3306 (defer_stack_allocation): Likewise.
3307 (expand_used_vars): Likewise.
3308 * cfgcleanup.c (old_insns_match_p): Likewise.
3309 * asan.c (asan_finish_file): Likewise.
3310 (asan_instrument): Likewise.
3311 (gate_asan): Likewise.
3312 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3313 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3314 (asan_global_struct): Use pointer_sized_int_node instead
3315 calling build_nonstandard_integer_type.
3316 (initialize_sanitizer_builtins): Likewise.
3317 (asan_finish_file): Likewise.
3318 * gcc.c: Document %{%:function(args):X}.
3319 (static_spec_functions): Add sanitize.
3320 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3321 store funcval != NULL there.
3322 (do_spec_1): Adjust handle_spec_function caller.
3323 (handle_braces): Allow %:function(args) as condition.
3324 (sanitize_spec_function): New function.
3325 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3326 (LIBUBSAN_SPEC): Likewise.
3327 (LIBUBSAN_EARLY_SPEC): Likewise.
3328 (SANITIZER_SPEC): Handle libubsan.
3329 (SANITIZER_EARLY_SPEC): Likewise.
3330 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3331 instead of fsanitize=address.
3332 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3333 instead of fsanitize=address*.
3334 * builtins.c: Include ubsan.h.
3335 (fold_builtin_0): Instrument __builtin_unreachable.
3336 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3337 instead of flag_asan.
3338 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3339 (pointer_sized_int_node): Define.
3340 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3342 2013-08-30 Mike Stump <mikestump@comcast.net>
3344 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3347 2013-08-29 Jan Hubicka <jh@suse.cz>
3349 * cgraph.c (cgraph_function_body_availability): Handle weakref
3351 * passes.def: Remove pass_fixup_cfg.
3352 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3353 track when we need to remove functions.
3354 (gate_ipa_inline): Execute inlining always; add comment why.
3355 (pass_data_ipa_inline): Remove TODO_remove_functions.
3356 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3357 do not produce summaries.
3358 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3359 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3360 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3361 constant pool and vtable.
3363 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3365 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3366 New arm_neon.h's internal macros to specify 64-bit constants.
3367 Avoid using stdint.h's macros.
3369 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3371 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3372 before applying REG_P.
3374 2013-08-30 Jakub Jelinek <jakub@redhat.com>
3376 PR tree-optimization/58277
3377 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
3378 after seeing too many stmts with vdef in between dombb and current
3379 bb, invalidate everything.
3381 2013-08-30 Richard Biener <rguenther@suse.de>
3383 * fold-const.c (fold_single_bit_test): Fix overflow test.
3385 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
3387 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
3388 and which can live in a register, always retrieve the value on entry.
3389 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
3390 passed by invisible reference specially.
3391 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
3392 (vt_add_function_parameter): Correctly deal with a parameter passed by
3393 invisible reference.
3395 2013-08-30 Jan Hubicka <jh@suse.cz>
3397 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
3399 2013-08-30 Richard Biener <rguenther@suse.de>
3401 PR tree-optimization/58228
3402 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
3403 allow invariant loads in nested loop vectorization.
3405 2013-08-30 Richard Biener <rguenther@suse.de>
3407 PR tree-optimization/58223
3408 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
3409 (has_anti_or_output_dependence): ... this and adjust to also
3410 look for output dependences.
3411 (mark_nodes_having_upstream_mem_writes): Adjust.
3412 (rdg_flag_uses): Likewise.
3414 2013-08-30 Richard Biener <rguenther@suse.de>
3416 PR tree-optimization/58010
3417 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
3418 assert that we have a loop-closed PHI.
3420 2013-08-29 Jan Hubicka <jh@suse.cz>
3422 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
3423 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
3424 * lto-section-in.c (lto_free_function_in_decl_state): New function.
3425 (lto_free_function_in_decl_state_for_node): New function.
3427 2013-08-29 Xinliang David Li <davidxl@google.com>
3429 * loop-unroll.c (report_unroll_peel): Minor message change.
3430 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
3431 Emit alignment peeling message with default -fopt-info.
3432 (vect_loop_versioning): Emit loop version info message.
3433 * tree-vectorizer.c (vectorize_loops): Minor message change.
3434 (execute_vect_slp): Ditto.
3436 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
3438 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
3439 of the clone from the DECL_NAME of the original function.
3441 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
3443 * passes.c (register_pass): Add overload.
3444 * tree-pass.h (register_pass): Forward declare it. Add comment.
3446 2013-08-29 Jan Hubicka <jh@suse.cz>
3448 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
3449 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
3450 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
3451 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3452 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
3453 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
3454 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3455 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
3457 2013-08-29 Teresa Johnson <tejohnson@google.com>
3459 * dumpfile.c (dump_loc): Output column number.
3460 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
3461 * doc/invoke.texi: Document optall -fopt-info flag.
3462 * profile.c (read_profile_edge_counts): Use new dump framework.
3463 (compute_branch_probabilities): Ditto.
3464 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
3465 when pass not in any opt group.
3466 * pass_manager.h (pass_manager::get_pass_profile): New method.
3467 * value-prof.c (check_counter): Use new dump framework.
3468 (check_ic_target): Ditto.
3469 * coverage.c (get_coverage_counts): Ditto.
3470 (coverage_init): Setup new dump framework.
3472 2013-08-29 Richard Biener <rguenther@suse.de>
3474 PR tree-optimization/58246
3475 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
3476 handle the dominance check inside a basic-block.
3478 2013-08-29 Richard Biener <rguenther@suse.de>
3481 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
3482 of default defs that appear in abnormal PHI nodes.
3484 2013-08-29 Richard Biener <rguenther@suse.de>
3486 PR tree-optimization/57685
3487 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
3488 single-use operands to avoid exponential complexity.
3490 2013-08-28 Dehao Chen <dehao@google.com>
3492 * ipa-inline.c (edge_badness): Fix integer underflow.
3494 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
3496 * gtm-builtins.def (_ITM_free): Declare leaf.
3498 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3501 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
3502 (*tls_local_dynamic_base_64_largepic): Likewise.
3503 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
3504 Remove predicate from call operand.
3505 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
3506 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
3508 2013-08-28 Jeff Law <law@redhat.com>
3510 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
3511 checks for the number of predecessors and successors allowed.
3512 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
3513 which require copying a joiner block if there is a request which
3514 is a subpath that requires no joiner block copying.
3516 2013-08-28 Jan Hubicka <jh@suse.cz>
3518 * lto-streamer-out.c (DFS_write_tree_body): Drop
3519 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
3520 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
3521 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
3522 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
3523 Do not read DECL_ERROR_ISSUED.
3524 (unpack_ts_decl_with_vis_value_fields): Do not read
3526 (lto_input_ts_binfo_tree_pointers): Do not read
3527 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
3528 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
3529 write DECL_ERROR_ISSUED..
3530 (pack_ts_decl_with_vis_value_fields): Do not write
3532 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
3533 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
3534 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
3535 * tree.h (tree_decl_common): Update comment.
3536 (DECL_ERROR_ISSUED): Remove.
3538 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3541 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
3543 2013-08-28 Jan Hubicka <jh@suse.cz>
3545 * builtins.def (free): Declare leaf.
3547 2013-08-27 David Malcolm <dmalcolm@redhat.com>
3550 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
3551 * configure: Regenerate.
3553 2013-08-27 Martin Jambor <mjambor@suse.cz>
3555 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
3556 (ipa_ancestor_jf_data): Likewise.
3557 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
3558 (ipa_get_jf_pass_through_type_preserved): New function.
3559 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
3560 (ipa_get_jf_ancestor_type_preserved): New function.
3561 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
3562 (ipa_get_jf_ancestor_result): Likewise.
3563 (propagate_vals_accross_pass_through): Use
3564 ipa_get_jf_pass_through_result to do all the value mappings.
3565 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
3566 type_preserved flag.
3567 (ipa_set_jf_cst_copy): New function.
3568 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
3569 (ipa_set_jf_arith_pass_through): Likewise.
3570 (ipa_set_ancestor_jf): Likewise.
3571 (compute_complex_assign_jump_func): Set type_preserved instead of
3573 (ipa_compute_jump_functions_for_edge): Likewise.
3574 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
3575 (update_jump_functions_after_inlining): Update type_preserved.
3576 Explicitely create jump functions when combining one with pass_through.
3577 (ipa_write_jump_function): Stream the type_preserved flags.
3578 (ipa_read_jump_function): Likewise.
3580 2013-08-27 Jakub Jelinek <jakub@redhat.com>
3581 Aldy Hernandez <aldyh@redhat.com>
3583 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
3584 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
3585 * function.h (struct function): Add has_force_vect_loops and
3587 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
3588 * gimple.c (gimple_build_omp_critical): Add KIND argument and
3590 * gimple.def: Update CLAUSES comments.
3591 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
3592 (gimple_build_omp_for): Add argument to prototype.
3593 (gimple_omp_for_kind): New.
3594 (gimple_omp_for_set_kind): New.
3595 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
3596 GOVD_DATA_SHARE_CLASS.
3597 (enum omp_region_type): Add ORT_SIMD.
3598 (gimple_add_tmp_var): Handle ORT_SIMD.
3599 (gimplify_var_or_parm_decl): Same.
3600 (is_gimple_stmt): Same.
3601 (omp_firstprivatize_variable): Same.
3602 (omp_add_variable): Only use splay_tree_insert if lookup failed.
3603 (omp_notice_variable): Handle ORT_SIMD.
3604 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
3605 (omp_check_private): Handle ORT_SIMD.
3606 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3608 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
3609 Handle OMP_CLAUSE_LASTPRIVATE.
3610 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3612 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
3613 (gimplify_expr): Handle OMP_SIMD.
3614 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
3615 (expand_GOMP_SIMD_VF): New.
3616 (expand_GOMP_SIMD_LAST_LANE): New.
3617 * internal-fn.def (GOMP_SIMD_LANE): New.
3618 (GOMP_SIMD_VF): New.
3619 (GOMP_SIMD_LAST_LANE): New.
3620 * omp-low.c: Include target.h.
3621 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
3623 (check_omp_nesting_restrictions): Same.
3625 (lower_rec_simd_input_clauses): New.
3626 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
3628 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
3629 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
3630 (expand_omp_build_assign): New.
3631 (expand_omp_for_init_counts): New.
3632 (expand_omp_for_init_vars): New.
3633 (extract_omp_for_update_vars): New.
3634 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
3635 and rewrite accordingly.
3636 (expand_omp_simd): New.
3637 (expand_omp_for): Use expand_omp_simd.
3638 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
3639 (lower_omp_for): Do not lower the body.
3640 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
3642 * tree-flow.h (find_omp_clause): Remove prototype.
3643 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
3644 forcing vectorization of the loop, or if flag_tree_vectorize.
3645 (gate_tree_if_conversion): Similarly.
3646 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
3647 gimple_build_omp_for.
3648 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
3649 * tree-parloops (create_parallel_loop): Pass kind argument to
3650 gimple_build_omp_for.
3651 * tree-pretty-print.c (dump_omp_clause): Add cases for
3652 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
3653 OMP_CLAUSE__SIMDUID_.
3654 (dump_generic_node): Handle OMP_SIMD.
3655 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
3656 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
3657 unroll OMP_SIMD loops here.
3658 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
3659 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
3661 (vect_analyze_data_refs): Handle simd loops.
3662 * tree-vect-loop.c (vectorizable_live_operation): Handle
3664 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
3665 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
3666 (vectorizable_load): Same.
3667 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
3668 (struct simduid_to_vf): New.
3669 (simduid_to_vf::hash): New.
3670 (simduid_to-vf::equal): New.
3671 (struct simd_array_to_simduid): New.
3672 (simd_array_to_simduid::hash): New.
3673 (simd_array_to_simduid::equal): New.
3674 (adjust_simduid_builtins): New.
3675 (struct note_simd_array_uses_struct): New.
3676 (note_simd_array_uses_cb): New.
3677 (note_simd_array_uses): New.
3678 (vectorize_loops): Handle simd hints and adjust simd builtins
3680 * tree-vectorizer.h (struct _stmt_vec_info): Add
3681 simd_lane_access_p field.
3682 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
3683 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
3684 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
3685 (omp_clause_code_name): Same.
3686 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
3687 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
3688 * tree.def (OMP_SIMD): New entry.
3689 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
3690 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
3691 (OMP_CLAUSE_DECL): Adjust range for new clauses.
3692 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
3693 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
3694 (OMP_CLAUSE_LINEAR_STEP): New.
3695 (OMP_CLAUSE_SAFELEN_EXPR): New.
3696 (OMP_CLAUSE__SIMDUID__DECL): New.
3697 (find_omp_clause): New prototype.
3699 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
3701 * config/i386/driver-i386.c (host_detect_local_cpu): Update
3702 Haswell processor detection.
3704 2013-08-27 Christian Widmer <shadow@umbrox.de>
3707 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
3708 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
3709 AVX2 capable processors.
3711 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
3713 * config/aarch64/arm_neon.h: Replace all inline asm implementations
3714 of vget_low_* with implementations in terms of other intrinsics.
3716 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
3719 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
3720 values to -1, 0 and 1.
3722 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
3724 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
3725 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
3726 (<optab><mode>3_insn): Remove.
3727 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
3728 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
3729 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
3730 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
3731 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
3732 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
3733 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
3734 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
3735 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
3736 (ror<mode>3_insn): Likewise.
3737 * config/aarch64/predicates.md (aarch64_simd_register): New.
3739 2013-08-27 Richard Biener <rguenther@suse.de>
3741 PR tree-optimization/57521
3742 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
3743 one edge is non-critical.
3744 (find_phi_replacement_condition): Make sure to use a non-critical
3745 edge. Cleanup and remove old bug workarounds.
3746 (bb_postdominates_preds): Remove.
3747 (if_convertible_loop_p_1): Do not compute post-dominators.
3748 (combine_blocks): Do not free post-dominators.
3749 (main_tree_if_conversion): Likewise.
3750 (pass_data_if_conversion): Add TODO_verify_ssa.
3752 2013-08-27 DJ Delorie <dj@redhat.com>
3754 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
3756 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
3758 * function.c (assign_parm_find_data_types): Set passed_mode and
3759 nominal_mode to the TYPE_MODE of nominal_type for the built
3760 pointer type in case of the struct-pass-by-reference.
3762 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3764 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
3765 (UINT16_TYPE): Change default to "unsigned int".
3767 * config/avr/avr.opt (mfract-convert-truncate): New option.
3768 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
3769 is set, round negative fractional integers according to n1169
3770 when converting to integer types.
3772 2013-08-26 Jan Hubicka <jh@suse.cz>
3774 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
3775 methods can not be called indirectly when their address is not taken.
3777 2013-08-26 Jan Hubicka <jh@suse.cz>
3779 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
3782 2013-08-26 Jan Hubicka <jh@suse.cz>
3784 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
3787 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3789 * reload.c (find_valid_class): Allow classes that do not include
3790 FIRST_PSEUDO_REGISTER - 1.
3792 2013-08-26 Jan Hubicka <jh@suse.cz>
3794 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
3795 fix edge count/frequency when speculation failed; fix type check
3796 for the direct call.
3798 2013-08-26 Jan Hubicka <jh@suse.cz>
3800 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
3802 2013-08-26 Jan Hubicka <jh@suse.cz>
3804 * ipa-inline-transform.c (inline_transform): Be ready for basic block
3805 to be changed by edge redirection.
3807 2013-08-26 Jan Hubicka <jh@suse.cz>
3809 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
3810 formating; add sanity check.
3811 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
3812 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
3813 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
3814 (resolve_noninline_speculation): Update callee keys, too.
3816 2013-08-26 Jan Hubicka <jh@suse.cz>
3818 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
3819 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
3821 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3823 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
3826 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
3828 * config/i386/i386.c (ix86_debug_options): Remove prototype.
3829 (x86_64_elf_select_section): Ditto.
3830 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
3832 (ix86_pass_by_reference): Ditto.
3833 (output_set_got): Ditto.
3834 (ix86_unary_operator_ok): Ditto.
3835 (ix86_expand_builtin): Ditto.
3837 2013-08-23 Jan Hubicka <jh@suse.cz>
3839 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
3841 2013-08-23 Jan Hubicka <jh@suse.cz>
3843 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
3844 (tree_decl_with_vis): Add FINAL field.
3846 2013-08-23 Jeff Law <law@redhat.com>
3848 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
3849 the debugging dump when the expression is fully redundant.
3851 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3853 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
3854 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
3855 * pretty-print.c (pp_formatted_text_data): Likewise.
3856 (pp_write_text_to_stream): Likewise.
3857 (pp_write_text_as_dot_label_to_stream): Likewise.
3858 (pp_append_r): Likewise.
3859 (pp_format): Likewise.
3860 (pp_flush): Likewise.
3861 (pp_clear_output_area): Likewise.
3862 (pp_append_text): Likewise.
3863 (pp_formatted_text): Likewise.
3864 (pp_remaining_character_count_for_line): Likewise.
3865 (pp_newline): Likewise.
3866 (pp_character): Likewise.
3867 (output_buffer::~output_buffer): Define.
3868 (pretty_printer::~pretty_printer): Destruct output buffer.
3869 * pretty-print.h (output_buffer::~output_buffer): Declare.
3870 (pretty_printer::~pretty_printer): Declare virtual.
3872 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
3875 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
3876 HOST_WIDE_INT_M1U): New macros.
3877 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
3878 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
3879 unsigned -1 for lshift.
3880 * cse.c (cse_insn): Likewise.
3881 * double-int.c (rshift_double, lshift_double): Likewise.
3882 * builtins.c (fold_builtin_bitop): Likewise.
3883 * combine.c (force_to_mode): Likewise.
3884 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
3885 * simplify-rtx.c (simplify_const_unary_operation,
3886 simplify_const_binary_operation): Likewise.
3887 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
3888 check_va_list_escapes): Likewise.
3889 * rtlanal.c (nonzero_bits1): Likewise.
3890 * expmed.c (expand_smod_pow2): Likewise.
3891 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
3893 2013-08-23 Jan Hubicka <jh@suse.cz>
3895 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
3896 as having address taken.
3898 2013-08-23 Jan Hubicka <jh@suse.cz>
3900 * ipa-utils.h (method_class_type): Declare.
3901 * ipa-devirt.c (method_class_type): Export.
3903 * cgraphunit.c (analyze_functions): Do basic devirtualization;
3904 do not walk base classes of anonymous types.
3906 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
3908 PR rtl-optimization/58220
3910 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
3911 handle SEQUENCE insns properly.
3913 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3915 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
3917 (pp_newline_and_indent): Likewise.
3918 (pp_separate_with): Likewise.
3919 * pretty-print.c (pp_newline_and_flush): Define.
3920 (pp_newline_and_indent): Likewise.
3921 (pp_separate_with): Likewise.
3923 2013-08-23 Jakub Jelinek <jakub@redhat.com>
3926 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
3927 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
3929 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
3931 * config/i386/predicates.md (ext_sse_reg_operand): New.
3932 * config/i386/i386.md (*movti_internal): Use
3933 predicate to determine if EVEX is needed.
3934 (*movsi_internal): Ditto.
3935 (*movdf_internal): Ditto.
3936 (*movsf_internal): Ditto.
3937 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
3939 2013-08-23 Jakub Jelinek <jakub@redhat.com>
3941 PR tree-optimization/58209
3942 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
3943 (find_tail_calls): Give up for pointer result types if m is non-NULL.
3944 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
3945 emit POINTER_PLUS_EXPR.
3946 (create_tailcall_accumulator): For pointer result type accumulate in
3949 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
3951 * configure.ac: Add backslashes missing from the last change.
3952 * configure: Regenerate.
3954 2013-08-22 Jan Hubicka <jh@susue.cz>
3956 * ipa.c (function_and_variable_visibility): First remember function
3957 was global and then make it local.
3959 2013-08-22 Julian Brown <julian@codesourcery.com>
3961 * configure.ac: Add aarch64 to list of arches which use "nop" in
3963 * configure: Regenerate.
3965 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3967 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
3968 gnu_libc_has_function.
3969 * config/s390/tpf.h: Likewise.
3971 2013-08-22 Jan Hubicka <jh@susue.cz>
3973 * timevar.c (validate_phases): Add cast.
3975 2013-08-22 Jan Hubicka <jh@susue.cz>
3977 * timevar.c (validate_phases): Use size_t for memory.
3978 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
3980 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
3982 * pretty-print.h (output_buffer::output_buffer): Declare.
3983 (pretty_printer::pretty_printer): Likewise.
3984 (pp_construct): Remove.
3985 * pretty-print.c (output_buffer::output_buffer): Define.
3986 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
3987 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
3988 (print_gimple_expr): Likewise.
3989 (print_gimple_seq): Likewise.
3990 (gimple_dump_bb): Likewise.
3991 * sched-vis.c (dump_value_slim): Likewise.
3992 (dump_insn_slim): Likewise.
3993 (dump_rtl_slim): Likewise.
3994 (str_pattern_slim): Likewise.
3995 * tree-mudflap.c (mf_varname_tree): Likewise.
3996 * graph.c (print_graph_cfg): Likewise.
3997 (start_graph_dump): Likewise.
3998 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4000 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4001 pretty printer initialization.
4002 * coretypes.h (diagnostic_context): Remove superflous type alias
4004 (pretty_printer): Likewise. Declare directly as a class.
4005 (pretty_print_info): Remove declaration as class.
4006 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4007 and pp_clear_output_area.
4008 (asan_add_global): Likewise.
4010 2013-08-22 Jan Hubicka <jh@suse.cz>
4012 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4013 * ipa-utils.h (update_type_inheritance_graph): Declare.
4014 (possible_polymorphic_call_target_p): Declare.
4015 (possible_polymorphic_call_target_p): New.
4016 * ipa-devirt.c: Update toplevel comments.
4017 (cached_polymorphic_call_targets): Move up.
4018 (odr_type_d): Move ID down.
4019 (polymorphic_type_binfo_p): Update comment.
4020 (odr_hasher::remove): Likewise;
4021 (get_odr_type): Set anonymous_namespace.
4022 (dump_odr_type): Dump it.
4023 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4024 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4025 (record_binfo): Add comment.
4026 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4028 (devirt_node_removal_hook): Do not iCE when cache is freed.
4029 (possible_polymorphic_call_target_p): New predicate.
4030 (update_type_inheritance_graph): New function.
4032 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4033 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4034 Sergey Lega <sergey.s.lega@intel.com>
4035 Anna Tikhonova <anna.tikhonova@intel.com>
4036 Ilya Tocar <ilya.tocar@intel.com>
4037 Andrey Turetskiy <andrey.turetskiy@intel.com>
4038 Ilya Verbin <ilya.verbin@intel.com>
4039 Kirill Yukhin <kirill.yukhin@intel.com>
4040 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4042 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4043 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4044 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4045 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4046 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4047 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4048 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4049 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4050 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4051 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4052 OPT_mavx512pf, OPT_mavx512er cases.
4053 * config/i386/constraints.md (v): New constraint.
4054 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4055 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4056 (bit_AVX512CD): New.
4057 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4058 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4059 * config/i386/i386-c.c (ix86_target_macros_internal):
4060 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4062 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4063 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4064 * config/i386/i386.c (regclass_map, dbx_register_map)
4065 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4066 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4067 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4069 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4070 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4071 -mavx512pf options. Fix formatting.
4072 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4073 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4074 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4075 -mavx512cd, -mavx512pf options.
4076 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4077 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4078 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4080 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4081 (ix86_set_reg_reg_cost): Ditto.
4082 (x86_order_regs_for_local_alloc): Ditto.
4083 (MAX_VECT_LEN): Extend to 64-byte.
4084 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4085 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4086 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4087 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4088 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4089 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4090 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4091 (SSE_REG_MODE_P): Support new modes.
4092 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4093 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4094 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4095 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4096 (REG_CLASS_CONTENTS): Add new registers.
4097 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4098 (EXT_REX_SSE_REGNO_P): New.
4099 (HI_REGISTER_NAMES): Add new registers.
4100 * config/i386/i386.md: Define constants for new registers.
4101 (mode): Add new 512-bit modes.
4102 (prefix): Support evex prefix.
4103 (isa): Support avx512f, noavx512f, fma_avx512f.
4104 (ssemodesuffix): Add new 512-bit modes.
4106 (*movxi_internal_avx512f): Ditto.
4107 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4109 (*movsi_internal): Likewise.
4110 (*movdf_internal): Likewise.
4111 (*movsf_internal): Likewise.
4112 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4113 (<code><mode>3): Likewise.
4114 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4116 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4117 with the new constraint "v".
4118 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4120 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4121 with the new constraint "v".
4122 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4123 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4124 (*<plusminus_insn><mode>3): Likewise.
4125 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4126 (*mul<mode>3): Likewise.
4127 (<sse>_vmmul<mode>3): Likewise.
4128 (<sse>_div<mode>3): Likewise.
4129 (<sse>_vmdiv<mode>3): Likewise.
4130 (<sse>_sqrt<mode>2): Likewise.
4131 (<sse>_vmsqrt<mode>2): Likewise.
4132 (*<code><mode>3_finite): Likewise.
4133 (*<code><mode>3) <smaxmin>: Likewise.
4134 (<sse>_vm<code><mode>3): Likewise.
4135 (*<code><mode>3) <any_logic>: Likewise.
4136 (*fma_fmadd_<mode>): Likewise.
4137 (*fma_fmsub_<mode>): Likewise.
4138 (*fma_fnmadd_<mode>): Likewise.
4139 (*fma_fnmsub_<mode>): Likewise.
4140 (*fma_fmaddsub_<mode>): Likewise.
4141 (*fma_fmsubadd_<mode>): Likewise.
4142 (*fmai_fmadd_<mode>): Likewise.
4143 (*fmai_fmsub_<mode>): Likewise.
4144 (*fmai_fnmadd_<mode>): Likewise.
4145 (*fmai_fnmsub_<mode>): Likewise.
4146 (sse_cvtsi2ss): Likewise.
4147 (sse_cvtsi2ssq): Likewise.
4148 (sse_cvtss2si): Likewise.
4149 (sse_cvtss2si_2): Likewise.
4150 (sse_cvtss2siq): Likewise.
4151 (sse_cvtss2siq_2): Likewise.
4152 (sse_cvttss2si): Likewise.
4153 (sse_cvtss2siq_2): Likewise.
4154 (float<sseintvecmodelower><mode>2): Likewise.
4155 (sse2_cvtsd2si_2): Likewise.
4156 (sse2_cvtsd2siq_2): Likewise.
4157 (*<plusminus_insn><mode>3): Likewise.
4158 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4159 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4160 (ashr<mode>3): Likewise.
4161 (<shift_insn><mode>3): Likewise.
4162 (avx2_<code><mode>3): Likewise.
4163 (*avx2_<code><mode>3): Likewise.
4164 (*andnot<mode>3): Likewise.
4165 (*<code><mode>3) <any_logic>: Likewise.
4166 (abs<mode>2): Likewise.
4167 (avx2_permvar<mode>): Likewise.
4168 (avx2_perm<mode>_1): Likewise.
4169 (*avx_vpermilp<mode>): Likewise.
4170 (avx_vpermilvar<mode>3): Likewise.
4171 (avx2_ashrv<mode>): Likewise.
4172 (avx2_<shift_insn>v<mode>): Likewise.
4173 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4175 * doc/rtl.texi: Document XImode.
4177 2013-08-21 Jeff Law <law@redhat.com>
4179 * tree-flow.h (register_jump_thread): Pass vector of edges
4180 instead of each important edge.
4181 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4182 thread path into a vector and pass that to register_jump_thread.
4183 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4184 passed in edge vector to the current 3-edge form.
4187 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4189 * dce.c (fini_dce): Call df_analyze again just in case
4190 delete_unmarked_insns removed anything.
4192 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4194 * reload.h (struct reg_equivs): Rename to ..
4195 (struct reg_equivs_s): .. this.
4197 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4199 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4201 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4203 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4205 2013-08-21 Jeff Law <law@redhat.com>
4207 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4208 simplify assignments too. If the RHS collapses to a singleton
4209 range, then return the value for the range.
4211 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4213 * config/i386/sse.md (V16): Rename to...
4215 (mov<mode>): Update iterator name.
4216 (*mov<mode>_internal): Ditto.
4217 (push<mode>1): Ditto.
4218 (movmisalign<mode>): Ditto.
4220 2013-08-20 Jan Hubicka <jh@suse.cz>
4223 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4224 entry for direct edges.
4225 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4227 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4229 Revert my last two changes, r201865 and r201864:
4232 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4234 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4235 instances can own GC refs.
4237 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4238 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4239 (gcc::context::gt_pch_nx): Likewise.
4240 (gcc::context::gt_pch_nx): Likewise.
4241 * ggc.h (gt_ggc_mx <T>): New.
4242 (gt_pch_nx_with_op <T>): New.
4243 (gt_pch_nx <T>): New.
4244 * passes.c (opt_pass::gt_ggc_mx): New.
4245 (opt_pass::gt_pch_nx): New.
4246 (opt_pass::gt_pch_nx_with_op): New.
4247 (pass_manager::gt_ggc_mx): New.
4248 (pass_manager::gt_pch_nx): New.
4249 (pass_manager::gt_pch_nx_with_op): New.
4250 (pass_manager::operator new): Use
4251 ggc_internal_cleared_alloc_stat rather than xcalloc.
4252 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4253 (pass_manager::gt_ggc_mx): New.
4254 (pass_manager::gt_pch_nx): New.
4255 (pass_manager::gt_pch_nx_with_op): New.
4256 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4257 (opt_pass::operator new): New.
4258 (opt_pass::gt_ggc_mx): New.
4259 (opt_pass::gt_pch_nx): New.
4260 (opt_pass::gt_pch_nx_with_op): New.
4263 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4265 * Makefile.in (GTFILES): Add context.h.
4266 * context.c (gcc::context::operator new): New.
4267 (gcc::context::gt_ggc_mx): New.
4268 (gcc::context::gt_pch_nx): New.
4269 (gcc::context::gt_pch_nx): New.
4270 * context.h (gcc::context): Add GTY((user)) marking.
4271 (gcc::context::operator new): New.
4272 (gcc::context::gt_ggc_mx): New.
4273 (gcc::context::gt_pch_nx): New.
4274 (gcc::context::gt_pch_nx): New.
4275 (g): Add GTY marking.
4276 (gt_ggc_mx (gcc::context *)): New.
4277 (gt_pch_nx (gcc::context *)): New.
4278 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4279 void *cookie)): New.
4280 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4282 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4284 * dce.c (fini_dce): Call df_analyze again just in case
4285 delete_unmarked_insns removed anything.
4287 2013-08-20 Teresa Johnson <tejohnson@google.com>
4289 PR rtl-optimizations/57451
4290 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4291 from crossing split section boundaries.
4293 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4295 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4296 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4297 with MULTLIB_DEFAULTS.
4299 2013-08-20 Nick Clifton <nickc@redhat.com>
4301 * target.def (narrow_volatile_bitfield): Note that the default
4302 value is false, not !TARGET_STRICT_ALIGN.
4303 * doc/tm.texi: Regenerate.
4305 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4307 Fix LIB_SPEC for systems without libpthread.
4309 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4310 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4312 * config/i386/linux-common.h: Likewise.
4313 * config/mips/linux-common.h: Likewise.
4315 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4317 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4320 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4322 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4323 instances can own GC refs.
4325 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4326 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4327 (gcc::context::gt_pch_nx): Likewise.
4328 (gcc::context::gt_pch_nx): Likewise.
4329 * ggc.h (gt_ggc_mx <T>): New.
4330 (gt_pch_nx_with_op <T>): New.
4331 (gt_pch_nx <T>): New.
4332 * passes.c (opt_pass::gt_ggc_mx): New.
4333 (opt_pass::gt_pch_nx): New.
4334 (opt_pass::gt_pch_nx_with_op): New.
4335 (pass_manager::gt_ggc_mx): New.
4336 (pass_manager::gt_pch_nx): New.
4337 (pass_manager::gt_pch_nx_with_op): New.
4338 (pass_manager::operator new): Use
4339 ggc_internal_cleared_alloc_stat rather than xcalloc.
4340 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4341 (pass_manager::gt_ggc_mx): New.
4342 (pass_manager::gt_pch_nx): New.
4343 (pass_manager::gt_pch_nx_with_op): New.
4344 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4345 (opt_pass::operator new): New.
4346 (opt_pass::gt_ggc_mx): New.
4347 (opt_pass::gt_pch_nx): New.
4348 (opt_pass::gt_pch_nx_with_op): New.
4350 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4352 * Makefile.in (GTFILES): Add context.h.
4353 * context.c (gcc::context::operator new): New.
4354 (gcc::context::gt_ggc_mx): New.
4355 (gcc::context::gt_pch_nx): New.
4356 (gcc::context::gt_pch_nx): New.
4357 * context.h (gcc::context): Add GTY((user)) marking.
4358 (gcc::context::operator new): New.
4359 (gcc::context::gt_ggc_mx): New.
4360 (gcc::context::gt_pch_nx): New.
4361 (gcc::context::gt_pch_nx): New.
4362 (g): Add GTY marking.
4363 (gt_ggc_mx (gcc::context *)): New.
4364 (gt_pch_nx (gcc::context *)): New.
4365 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4366 void *cookie)): New.
4367 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4369 2013-08-20 Alan Modra <amodra@gmail.com>
4372 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
4373 (rs6000_emit_epilogue): Likewise.
4375 2013-08-19 Dehao Chen <dehao@google.com>
4377 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
4379 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
4380 Jakub Jelinek <jakub@redhat.com>
4382 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
4383 (BUILT_IN_FABSD64): Likewise.
4384 (BUILT_IN_FABSD128): Likewise.
4385 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
4386 (fold_builtin_1): Likewise.
4387 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
4388 destination and source operands.
4389 (*abstd2_fpr): Likewise.
4390 (*nabstd2_fpr): Likewise.
4392 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
4394 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
4397 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
4399 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
4402 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
4404 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
4405 * builtins.c (default_libc_has_function): New.
4406 (gnu_libc_has_function): Ditto.
4407 (no_c99_libc_has_function): Ditto.
4408 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
4409 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
4410 (fold_builtin_sincos): Likewise.
4411 (fold_builtin_cexp): Likewise.
4412 * builtins.def (DEF_C94_BUILTIN): Likewise.
4413 (DEF_C99_BUILTIN): Likewise.
4414 (DEF_C99_C90RES_BUILTIN): Likewise.
4415 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
4416 definitions to using this define.
4417 * config/darwin-protos.h (darwin_libc_has_function): New.
4418 * config/darwin.c (darwin_libc_has_function): Ditto.
4419 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
4420 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
4421 * config/darwin.h: Ditto.
4422 * config/elfos.h: Ditto.
4423 * config/freebsd.h: Ditto.
4424 * config/i386/cygming.h: Ditto.
4425 * config/i386/djgpp.h: Ditto.
4426 * config/i386/i386-interix.h: Ditto.
4427 * config/microblaze/microblaze.h: Ditto.
4428 * config/mmix/mmix.h: Ditto.
4429 * config/gnu-user.h: Ditto.
4430 * config/ia64/hpux.h: Ditto.
4431 * config/pa/pa-hpux.h: Ditto.
4432 * config/pdp11/pdp11.h: Ditto.
4433 * config/picochip/picochip.h: Ditto.
4434 * config/linux.h: Ditto.
4435 * config/netbsd.h: Ditto.
4436 * config/openbsd.h: Ditto.
4437 * config/rs6000/aix43.h: Ditto.
4438 * config/rs6000/aix51.h: Ditto.
4439 * config/rs6000/aix52.h: Ditto.
4440 * config/rs6000/aix53.h: Ditto.
4441 * config/rs6000/aix61.h: Ditto.
4442 * config/rs6000/darwin.h: Ditto.
4443 * config/rs6000/linux.h: Ditto.
4444 * config/rs6000/linux64.h: Ditto.
4445 * config/s390/tpf.h: Ditto.
4446 * config/sol2-10.h: Ditto.
4447 * config/sol2.h: Ditto.
4448 * config/vms/vms.h: Ditto.
4449 * config/vxworks.h: Ditto.
4450 * config/linux-android.c (linux_android_libc_has_function):
4451 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
4452 * config/linux-protos.h (linux_android_libc_has_function):
4454 * config/i386/i386.c (ix86_libc_has_function): New.
4455 * config/i386/i386-protos.h
4456 (ix86_libc_has_function): New declaration.
4457 * config/i386/i386.md
4458 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
4459 ("isinf<mode>2): Likewise.
4460 * convert.c (convert_to_integer): Using new target hook
4461 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
4462 TARGET_C99_FUNCTIONS.
4463 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
4464 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
4465 * coretypes.h (function_class): New enum for different
4466 classes of functions.
4467 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
4468 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
4469 (TARGET_HAS_SINCOS): Likewise.
4470 (TARGET_LIBC_HAS_FUNCTION): New.
4471 * doc/tm.texi: Regenerated.
4472 * targhooks.h (default_libc_has_function): New declaration.
4473 (no_c99_libc_has_function): Ditto.
4474 (gnu_libc_has_function): Ditto.
4475 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
4476 and TARGET_HAS_SINCOS.
4478 2013-08-18 Jan Hubicka <jh@suse.cz>
4480 * Makeifle-in (ipa-devirt.o): New.
4481 (GTFILES): Add ipa-utils.h and ipa-devirt.c
4482 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
4483 (analyze_functions): Look into possible targets of polymorphic call.
4484 * dumpfile.c (dump_files): Add type-inheritance dump.
4485 * dumpfile.h (TDI_inheritance): New.
4486 * ipa-devirt.c: New file.
4487 * ipa-utils.h (odr_type_d): Forward declare.
4488 (odr_type): New type.
4489 (build_type_inheritance_graph): Declare.
4490 (possible_polymorphic_call_targets): Declare and introduce inline
4491 variant when only edge is pased.
4492 (dump_possible_polymorphic_call_targets): Likewise.
4493 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
4494 * tree.c (type_in_anonymous_namespace_p): Break out from ...
4495 (types_same_for_odr): ... here.
4496 * tree.h (type_in_anonymous_namespace_p): Declare.
4498 2013-08-18 Jakub Jelinek <jakub@redhat.com>
4500 PR tree-optimization/58006
4501 * tree-parloops.c (take_address_of): Don't ICE if get_name
4503 (eliminate_local_variables_stmt): Remove clobber stmts.
4505 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
4507 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
4508 error message is issued for an alias to undefined symbol.
4510 2013-08-18 Jan Hubicka <jh@suse.cz>
4512 * cgraph.c (cgraph_create_indirect_edge): Discover
4513 polymorphic calls and record basic info into indirect_info.
4514 * gimple-fold.c (gimple_fold_call): When doing BINFO based
4515 devirtualization, ignore objc function calls.
4516 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
4517 call with no parm index info.
4518 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
4519 * tree.c (virtual_method_call_p): New function.
4520 * tree.h (virtual_method_call_p): Declare.
4522 2013-08-16 Jan Hubicka <jh@suse.cz>
4525 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
4527 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
4529 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
4532 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4534 * gengtype.c (type_for_name): Add special-case support for
4535 locating types within the "gcc::" namespace.
4536 (open_base_files): Emit a "using namespace gcc" directive.
4538 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
4541 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
4542 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
4544 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
4545 array instead of each individual operand as a separate argument.
4546 (emit_fusion_gpr_load): Likewise.
4547 (expand_fusion_gpr_load): Add new function declaration.
4549 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
4550 signature to have the operands passed as an array, instead of as
4551 separate arguments. Allow ZERO_EXTEND to be in the memory
4552 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
4553 depend on the register live/dead flags when peepholes are run.
4554 (expand_fusion_gpr_load): New function to be called from the
4555 peephole2 pass, to change the register that addis sets to be the
4557 (emit_fusion_gpr_load): Change the calling signature to have the
4558 operands passed as an array, instead of as separate arguments.
4559 Allow ZERO_EXTEND to be in the memory address, and also
4560 SIGN_EXTEND if -mpower8-fusion-sign.
4562 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
4564 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
4565 adjust the register addis loads up in the peephole2 pass. Do not
4566 depend on the register live/dead state when the peephole pass is done.
4568 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4570 * gengtype.c (create_user_defined_type): Ensure that the kind
4571 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
4572 declaration is seen before the GTY((user)) marking.
4574 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
4577 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
4579 2013-08-16 Jan Hubicka <jh@suse.cz>
4581 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
4582 arugment expected_type.
4583 (gimple_fold_call): Use it.
4584 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
4585 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
4586 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
4587 (try_make_edge_direct_virtual_call): Likewise.
4588 * tree.c (obj_type_ref_class): New.
4589 * tree.h (obj_type_ref_class): Use it.
4591 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
4593 * sched-vis.c (rtl_slim_pp_initialized): Remove.
4594 (rtl_slim_pp): Likewise.
4595 (init_rtl_slim_pretty_print): Likewise.
4596 (dump_value_slim): Don't call it. Use local pretty printer.
4597 (dump_insn_slim): Likewise.
4598 (dump_rtl_slim): Likewise.
4599 (str_pattern_slim): Likewise.
4600 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
4603 2013-08-16 Jakub Jelinek <jakub@redhat.com>
4605 PR tree-optimization/58164
4606 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
4607 walk gimple_goto_dest of GIMPLE_GOTO.
4609 PR tree-optimization/58165
4610 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
4611 bi_call must be the last stmt in a bb, don't split_block, instead
4612 use fallthru edge from it and give up if there is none.
4613 Release conds vector when returning early.
4615 2013-08-14 Xinliang David Li <davidxl@google.com>
4617 * config/i386/i386.c (ix86_option_override_internal):
4618 Remove unused variable and field.
4620 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4623 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
4624 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
4625 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
4626 and Linux, correct BLKmode alignment when 128-bit alignment is
4627 required and compatibility flag is not set.
4628 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
4629 for zero-size arguments when compatibility flag is not set.
4631 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4633 PR tree-optimization/58145
4634 * tree-sra.c (build_ref_for_offset): If prev_base has
4635 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
4637 2013-08-14 Xinliang David Li <davidxl@google.com>
4639 * config/i386/i386.c (ix86_option_override_internal):
4640 Fix uninitialized variable error.
4642 2013-08-14 Xinliang David Li <davidxl@google.com>
4644 * config/i386/i386.opt: Define two new options.
4645 * config/i386/x86-tune.def: Add arch selector field in macros.
4646 * config/i386/i386.h: Adjust macro definition.
4647 * config/i386/i386.c (ix86_option_override_internal):
4649 (parse_mtune_ctrl_str): New function.
4650 (set_ix86_tune_features): New function.
4651 (ix86_function_specific_restore): Call the new helper function.
4653 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
4655 PR rtl-optimization/57662
4656 * sel-sched.c (code_motion_process_successors): When the current insn
4657 is removed after the recursive traversal, break from the loop.
4658 Add comments and debug printouts.
4660 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4661 Alexandre Oliva <aoliva@redhat.com>
4664 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
4665 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
4666 there also UNSPEC_PLTOFF.
4668 2013-08-14 Marek Polacek <polacek@redhat.com>
4670 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
4671 NUM_CONDITIONS bit positions.
4673 2013-08-13 Cary Coutant <ccoutant@google.com>
4675 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
4676 (attr_checksum): Hash vector contents instead of pointer.
4677 (attr_checksum_ordered): Likewise.
4679 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
4681 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
4682 when Pmode != word_mode. Add length_address attribute.
4683 (sse3_monitor_<mode>): Merge from sse3_monitor and
4684 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
4685 Pmode != word_mode. Update insn length attribute.
4686 * config/i386/i386.c (ix86_option_override_internal): Update
4687 ix86_gen_monitor selection for merged sse3_monitor insn.
4689 2013-08-13 Julian Brown <julian@codesourcery.com>
4691 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
4692 perform invalid legitimization on greater-than-word-size modes for
4695 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
4697 * ira.c (setup_class_translate_array): Use aclass instead of cl
4698 for classes not fully covered by allocno classes.
4700 2013-08-13 Jakub Jelinek <jakub@redhat.com>
4702 PR tree-optimization/57661
4703 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
4704 * tree-inline.c (tree_function_versioning): Initialize it.
4705 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
4706 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
4707 that is not being copied.
4710 * asan.c (instrument_strlen_call): Fix typo in comment.
4711 Use char * type even for the lhs of POINTER_PLUS_EXPR.
4713 2013-08-13 Steve Ellcey <sellcey@mips.com>
4715 * config/mips/mips.md (prefetch): Use lw instead of ld on
4716 loongson in 32bit mode.
4718 2013-08-13 Nick Clifton <nickc@redhat.com>
4720 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
4722 2013-08-13 Jan Hubicka <jh@suse.cz>
4724 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
4725 introduced edge; fix typo in sanity check.
4726 (cgraph_resolve_speculation): Export; improve diagnostic.
4727 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
4728 speculation at type mismatch.
4729 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
4730 (cgraph_resolve_speculation): Declare.
4731 (symtab_can_be_discarded): New function.
4732 * value-prof.c (gimple_ic_transform): Remove actual transform code.
4733 * ipa-inline-transform.c (speculation_removed): New global var.
4734 (clone_inlined_nodes): See if speculation can be removed.
4735 (inline_call): If speculations was removed, we growths may not match.
4736 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
4737 (speculation_useful_p): New function.
4738 (resolve_noninline_speculation): New function.
4739 (inline_small_functions): Resolve useless speculations.
4740 * ipa-inline.h (speculation_useful_p): Declare
4741 * ipa.c (can_replace_by_local_alias): Simplify.
4742 (ipa_profile): Produce speculative calls in non-lto, too;
4743 add simple cost model; produce local aliases.
4745 2013-08-13 David Malcolm <dmalcolm@redhat.com>
4747 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
4750 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
4752 * config/i386/i386.c (ix86_function_versions): Use error + inform.
4754 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
4756 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
4757 iterator instead of X87MODEF.
4759 2013-08-12 Perez Read <netfirewall@gmail.com>
4762 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
4763 operand 0 for intel asm alternative.
4764 (*movabs<mode>_2): Ditto for operand 1.
4766 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
4768 * config/aarch64/arm_none.h
4769 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
4771 2013-08-12 Nick Clifton <nickc@redhat.com>
4773 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
4775 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4777 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
4778 expand for QI/HImode operand to produce more effictive code for
4779 unsigned char(short) --> float(double) conversion.
4781 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
4783 * doc/invoke.texi: Mention that -ftls-model does not force the final
4786 2013-08-12 Marek Polacek <polacek@redhat.com>
4787 Marc Glisse <marc.glisse@inria.fr>
4789 PR tree-optimization/57980
4790 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
4791 when creating -1 constant.
4793 2013-08-10 Jan Hubicka <jh@suse.cz>
4795 Workaround binutils PR14342.
4796 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
4797 (gimple_init_edge_profiler): Likewise.
4798 (gimple_gen_ic_func_profiler): Likewise.
4800 2013-08-09 Jan Hubicka <jh@suse.cz>
4802 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
4804 2013-08-09 Xinliang David Li <davidxl@google.com>
4806 * config/i386/stringop.def: New file.
4807 * config/i386/stringop.opt: New file.
4808 * config/i386/i386-opts.h: Include stringopt.def.
4809 * config/i386/i386.opt: Include stringopt.opt.
4810 * config/i386/i386.c (ix86_option_override_internal):
4811 Override default size based stringop inline strategies with options.
4812 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
4815 2013-08-09 Jan Hubicka <jh@suse.cz>
4817 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
4819 2013-08-09 Jan Hubicka <jh@suse.cz>
4821 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
4823 (dump_cgraph_node): Dump profile-id.
4824 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
4825 and common_target_probability.
4826 * lto-cgraph.c (lto_output_edge): Stream common targets.
4827 (lto_output_node): Stream profile ids.
4828 (input_node): Stream profile ids.
4829 (input_edge): Stream common targets.
4830 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
4831 * ipa.c: Include value-prof.h
4832 (ipa_profile_generate_summary): Turn indirect call statement histograms
4833 into common targets.
4834 (ipa_profile): Turn common targets into speculative edges.
4836 2013-08-09 Jan Hubicka <jh@suse.cz>
4838 * cgraph.h (cgraph_node): Add profile_id.
4839 * value-prof.c (cgraph_node_map): Turn into pointer_map.
4840 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
4841 (del_node_map): Update.
4842 (find_func_by_funcdef_no): Replace by ...
4843 (find_func_by_profile_id): ... this one.
4844 (gimple_ic_transform): Do not remove useful histograms when
4845 speculation is not done; dump info when indirect call removal
4847 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
4848 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
4849 (__gcov_indirect_call_profiler_v2): .. this one.
4850 * profile.h (init_node_map): Update.
4851 * coverage.c (coverage_compute_profile_id): New function.
4852 * coverage.h (coverage_compute_profile_id): Declare.
4853 * tree-profile.c (init_ic_make_global_vars): Make
4854 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
4855 (gimple_init_edge_profiler): Update prototype of
4856 __gcov_indirect_call_profiler.
4857 (gimple_gen_ic_func_profiler): Simplify.
4858 (tree_profiling): Use init_node_map
4860 2013-08-09 Jan Hubicka <jh@suse.cz>
4862 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
4863 non-speculative refs.
4864 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
4865 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
4866 (cgraph_set_call_stmt): Likewise.
4867 (cgraph_create_edge_1): Fix release checking compilatoin;
4869 (cgraph_free_edge): Free indirect info.
4870 (cgraph_turn_edge_to_speculative): New function.
4871 (cgraph_speculative_call_info): New function.
4872 (cgraph_make_edge_direct): Return direct edge; handle speculation.
4873 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
4874 (dump_cgraph_node): Dump speculation.
4875 (verify_edge_count_and_frequency): Accept speculative edges.
4876 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
4877 (verify_cgraph_node): Handle speculation.
4878 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
4879 (cgraph_set_call_stmt): Update prototype.
4880 (cgraph_make_edge_direct): Update prototype.
4881 (cgraph_speculative_call_info): Declare.
4882 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
4883 to change; update call of ipa_find_references.
4884 * ipa-ref.c (ipa_record_reference): Fix return value; clear
4885 lto_stmt_uid and speculative flags.
4886 (ipa_dump_references): Dump speculation.
4887 (ipa_clone_references): Clone speculative flag.
4888 (ipa_clone_referring): Likewise.
4889 (ipa_clone_ref): New function.
4890 (ipa_find_reference): Look into lto_stmt_uids
4891 (ipa_clear_stmts_in_references): Do not clear speculative calls.
4892 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
4893 (ipa_find_reference): Update declaration.
4894 (ipa_clone_ref): Declare.
4895 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
4896 stream speculative flag.
4897 (lto_output_ref): Stream statements uids and speculation.
4898 (input_ref): Likewise.
4899 (input_edge): Stream speuclation.
4900 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
4901 (cgraph_set_call_stmt_including_clones): Handle speculation.
4902 * ipa-inline.c (heap_edge_removal_hook): New function.
4903 (inline_small_functions): Register it.
4904 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
4905 also initialize refs.
4906 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
4908 (try_make_edge_direct_simple_call): Likewise.
4909 (try_make_edge_direct_simple_call): Likewise.
4910 (update_indirect_edges_after_inlining): Likewise.
4911 (remove_described_reference): Look proper lto_stmt_uid.
4912 (propagate_controlled_uses): Likewise.
4913 (propagate_controlled_uses): Liekwise.
4914 * tree-inline.c (copy_bb): Copy speculative edges.
4915 (redirect_all_calls): New function.
4916 (copy_cfg_body): Do redirection after loop info is updated.
4917 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
4919 2013-08-09 Jan Hubicka <jh@suse.cz>
4921 * lto-streamer-out.c (output_function): Renumber PHIs.
4922 * lto-streamer-in.c (input_function): Likewise.
4924 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
4926 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
4927 (get_lane_unsigned): Likewise.
4928 (dup_lane_scalar): Likewise.
4929 (get_lane): enable for VALL.
4930 * config/aarch64/aarch64-simd.md
4931 (aarch64_dup_lane_scalar<mode>): Remove.
4932 (aarch64_get_lane_signed<mode>): Likewise.
4933 (aarch64_get_lane_unsigned<mode>): Likewise.
4934 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
4935 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
4936 (aarch64_get_lane<mode>): Enable for all vector modes.
4937 (aarch64_get_lanedi): Remove misleading constraints.
4938 * config/aarch64/arm_neon.h
4939 (__aarch64_vget_lane_any): Define.
4940 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
4941 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
4942 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
4943 * config/aarch64/iterators.md (VDQQH): New.
4947 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
4949 * configure.ac: Add GAS check for LEON instructions on SPARC.
4950 * configure: Regenerate.
4951 * config.in: Likewise.
4952 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
4954 * config/sparc/sparc.opt (LEON, LEON3): New masks.
4955 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
4957 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
4958 (AS_LEON_FLAG): New macro.
4959 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
4960 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
4961 Deal with LEON and LEON3 for the memory model.
4962 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
4963 (atomic_compare_and_swap<mode>_1): Likewise.
4964 (*atomic_compare_and_swap<mode>_1): Likewise.
4966 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4968 * config/arm/neon.md (vcond): Fix floating-point vector
4969 comparisons against 0.
4971 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
4973 * lra-constraints.c (emit_spill_move): Remove assert.
4974 (process_alt_operands): Add more debugging
4975 output. Increase reject for spilling into memory. Decrease
4976 reject for reloading scratch.
4977 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
4979 2013-08-08 Steve Ellcey <sellcey@mips.com>
4981 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
4982 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
4983 micromips incompatible. Add nan2008.
4984 (MULTILIB_DIRNAMES): Add nan2008.
4985 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4986 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
4987 and micromips incompatible. Add nan2008.
4988 (MULTILIB_DIRNAMES): Add nan2008.
4989 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4991 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
4993 PR rtl-optimization/58079
4994 * combine.c (combine_simplify_rtx): Avoid using SUBST if
4995 simplify_comparison has widened a comparison with an integer.
4997 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4999 * config/arm/neon.md (movmisalign<mode>): Disable when we
5000 don't allow unaligned accesses.
5001 (*movmisalign<mode>_neon_store): Likewise.
5002 (*movmisalign<mode>_neon_load): Likewise.
5003 (*movmisalign<mode>_neon_store): Likewise.
5004 (*movmisalign<mode>_neon_load): Likewise.
5006 2013-08-08 Jan Hubicka <jh@suse.cz>
5008 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5009 (make_pass_rebuild_cgraph_edges): Also clear references.
5010 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5011 * ipa-inline-transform.c (inline_transform): Remove all references
5013 * cgraphunit.c (expand_function): Remove all references after
5015 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5016 (ipa_find_reference): Rewrite to iterator.
5017 (remove_stmt_references): Likewise.
5018 (ipa_clear_stmts_in_references): New function.
5019 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5020 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5022 * ipa-split.c (split_function): Remove references in split function.
5024 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5027 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5028 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5030 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5033 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5034 suggested mode for the assignment isn't compatible with the
5037 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5040 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5042 2013-08-07 Xinliang David Li <davidxl@google.com>
5044 * config/i386/i386.opt: New option -mtune-ctrl=.
5045 * config/i386/x86-tune.def: New file.
5046 * config/i386/i386.h: include x86-tune.def.
5047 * config/i386/i386.c (ix86_option_override_internal):
5048 Parsing -mtune-ctrl= option and set tune features.
5050 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5053 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5054 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5057 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5059 * rtl.h (update_alignments): Declare.
5060 * final.c (grow_label_align): New function extracted from...
5061 (shorten_branches): ...here. Call it.
5062 (update_alignments): New function.
5063 * reorg.c (sibling_labels): New variable.
5064 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5065 the new label along with it onto the sibling_labels vector.
5066 (fill_simple_delay_slots): Adjust call to get_label_before.
5067 (fill_slots_from_thread): Likewise.
5068 (relax_delay_slots): Likewise.
5069 (make_return_insns): Likewise.
5070 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5072 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5074 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5075 document its semantics.
5076 (diagnostic_report_diagnostic): Adjust accordingly.
5078 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5080 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5081 (sparc_option_override): ...and port to new C++ pass API.
5082 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5084 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5086 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5088 2013-08-06 Caroline Tice <cmtice@google.com>
5090 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5091 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5092 * tree-pass.h: Add pass_vtable_verify.
5093 * varasm.c (assemble_variable): Add code to properly set the comdat
5094 section and name for the .vtable_map_vars section.
5095 (assemble_vtyv_preinit_initializer): New function.
5096 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5098 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5099 * vtable-verify.c: New file.
5100 * vtable-verify.h: New file.
5101 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5102 initialiation levels.
5103 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5104 * passes.def: Insert pass_vtable_verify.
5105 * aclocal.m4: Reorder includes.
5106 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5107 -fvtv-counts options.
5108 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5109 as appropriate, if -fvtable-verify=... is used.
5110 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5111 -fvtable-verify=... is used.
5112 * Makefile.in (OBJS): Add vtable-verify.o to list.
5113 (vtable-verify.o): Add new build rule.
5114 (GTFILES): Add vtable-verify.c to list.
5115 * common.opt (fvtable-verify=): New flag.
5116 (vtv_priority): Values for fvtable-verify= flag.
5117 (fvtv-counts): New flag.
5118 (fvtv-debug): New flag.
5119 * tree.h (save_vtable_map_decl): New extern function decl.
5121 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5123 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5124 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5125 (pass_data_rl78_devirt): ...new pass_data instance and...
5126 (make_pass_rl78_devirt): ...new function.
5127 (rl78_asm_file_start): Port pass registration to new C++ API.
5129 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5131 * coretypes.h (rtl_opt_pass): Add.
5132 (gcc::context): Add.
5133 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5134 (epiphany_init): Port to new C++ pass API.
5135 (epiphany_optimize_mode_switching): Likewise.
5136 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5137 (pass_manager::get_pass_mode_switching): New.
5138 (pass_manager::get_pass_peephole2): New.
5139 * mode-switching.c (pass_mode_switching): Add clone method.
5140 * recog.c (pass_peephole2): Add clone method.
5141 (pass_split_all_insns): Add clone method.
5143 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5145 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5146 (mips_option_override): ...here, porting to new C++ API for passes.
5148 2013-08-06 Jan Hubicka <jh@suse.cz>
5150 * cgraph.c (cgraph_get_body): New function based on lto.c
5152 * cgraph.h (cgraph_get_body): Declare.
5153 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5155 * cgraphunit.c (expand_function): Get body prior expanding.
5156 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5157 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5159 * passes.c (do_per_function_toporder): Get body.
5160 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5161 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5163 2013-08-06 Martin Jambor <mjambor@suse.cz>
5166 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5167 is not re-finalized. Rename second parameter to no_collect.
5169 2013-08-06 Martin Jambor <mjambor@suse.cz>
5172 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5173 MEM_REF has proper alignment information.
5175 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5178 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5180 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5181 rtx (*) (rtx, ...) with insn_gen_fn.
5182 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5183 insn_gen_fn::stored_funcptr. Add initializer braces.
5185 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5187 Rewrite how instances of passes are cloned to remove assumptions
5188 about their sizes (thus allowing pass subclasses to have
5189 additional data fields, albeit non-GC-managed ones at this point).
5191 * passes.c (make_pass_instance): Now that passes have clone
5192 methods, rewrite this function to eliminate XNEW and memcpy
5193 calls that used hardcoded sizes. Since this function no longer
5194 creates pass instances, rename it to...
5195 (add_pass_instance): ...this. Document the old way that passes were
5196 numbered and flagged, and rework this function to continue using it.
5197 (next_pass_1): Add an initial_pass argument for use by
5199 (position_pass): When adding multiple instances of a pass, use
5200 the pass's clone method, rather than relying on the XNEW/memcpy
5201 within the former make_pass_instance (now add_pass_instance).
5202 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5203 the initial instance of the current pass within the pass manager.
5205 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5207 This is the automated part of the conversion of passes from C
5208 structs to C++ classes.
5210 Patch autogenerated by refactor_passes.py from
5211 https://github.com/davidmalcolm/gcc-refactoring-scripts
5212 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5214 * asan.c (pass_asan): Convert from a global struct to a subclass of
5215 gimple_opt_pass along with...
5216 (pass_data_asan): ...new pass_data instance and...
5217 (make_pass_asan): ...new function.
5218 (pass_asan_O0): Convert from a global struct to a subclass of
5219 gimple_opt_pass along with...
5220 (pass_data_asan_O0): ...new pass_data instance and...
5221 (make_pass_asan_O0): ...new function.
5222 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5223 subclass of rtl_opt_pass along with...
5224 (pass_data_inc_dec): ...new pass_data instance and...
5225 (make_pass_inc_dec): ...new function.
5226 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5227 a subclass of rtl_opt_pass along with...
5228 (pass_data_reorder_blocks): ...new pass_data instance and...
5229 (make_pass_reorder_blocks): ...new function.
5230 (pass_duplicate_computed_gotos): Convert from a global struct to a
5231 subclass of rtl_opt_pass along with...
5232 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5233 (make_pass_duplicate_computed_gotos): ...new function.
5234 (pass_partition_blocks): Convert from a global struct to a subclass of
5235 rtl_opt_pass along with...
5236 (pass_data_partition_blocks): ...new pass_data instance and...
5237 (make_pass_partition_blocks): ...new function.
5238 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5239 struct to a subclass of rtl_opt_pass along with...
5240 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5242 (make_pass_branch_target_load_optimize1): ...new function.
5243 (pass_branch_target_load_optimize2): Convert from a global struct to a
5244 subclass of rtl_opt_pass along with...
5245 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5247 (make_pass_branch_target_load_optimize2): ...new function.
5248 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5249 of rtl_opt_pass along with...
5250 (pass_data_jump): ...new pass_data instance and...
5251 (make_pass_jump): ...new function.
5252 (pass_jump2): Convert from a global struct to a subclass of
5253 rtl_opt_pass along with...
5254 (pass_data_jump2): ...new pass_data instance and...
5255 (make_pass_jump2): ...new function.
5256 * cfgexpand.c (pass_expand): Convert from a global struct to a
5257 subclass of rtl_opt_pass along with...
5258 (pass_data_expand): ...new pass_data instance and...
5259 (make_pass_expand): ...new function.
5260 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5261 of rtl_opt_pass along with...
5262 (pass_data_free_cfg): ...new pass_data instance and...
5263 (make_pass_free_cfg): ...new function.
5264 (pass_into_cfg_layout_mode): Convert from a global struct to a
5265 subclass of rtl_opt_pass along with...
5266 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5267 (make_pass_into_cfg_layout_mode): ...new function.
5268 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5269 subclass of rtl_opt_pass along with...
5270 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5271 (make_pass_outof_cfg_layout_mode): ...new function.
5272 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5273 struct to a subclass of gimple_opt_pass along with...
5274 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5275 (make_pass_build_cgraph_edges): ...new function.
5276 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5277 subclass of gimple_opt_pass along with...
5278 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5279 (make_pass_rebuild_cgraph_edges): ...new function.
5280 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5281 subclass of gimple_opt_pass along with...
5282 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5284 (make_pass_remove_cgraph_callee_edges): ...new function.
5285 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5286 struct to a subclass of rtl_opt_pass along with...
5287 (pass_data_stack_adjustments): ...new pass_data instance and...
5288 (make_pass_stack_adjustments): ...new function.
5289 * combine.c (pass_combine): Convert from a global struct to a subclass
5290 of rtl_opt_pass along with...
5291 (pass_data_combine): ...new pass_data instance and...
5292 (make_pass_combine): ...new function.
5293 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5294 global struct to a subclass of rtl_opt_pass along with...
5295 (pass_data_compare_elim_after_reload): ...new pass_data instance
5297 (make_pass_compare_elim_after_reload): ...new function.
5298 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5299 of rtl_opt_pass along with...
5300 (pass_data_rtl_cprop): ...new pass_data instance and...
5301 (make_pass_rtl_cprop): ...new function.
5302 * cse.c (pass_cse): Convert from a global struct to a subclass of
5303 rtl_opt_pass along with...
5304 (pass_data_cse): ...new pass_data instance and...
5305 (make_pass_cse): ...new function.
5306 (pass_cse2): Convert from a global struct to a subclass of
5307 rtl_opt_pass along with...
5308 (pass_data_cse2): ...new pass_data instance and...
5309 (make_pass_cse2): ...new function.
5310 (pass_cse_after_global_opts): Convert from a global struct to a
5311 subclass of rtl_opt_pass along with...
5312 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5313 (make_pass_cse_after_global_opts): ...new function.
5314 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5315 of rtl_opt_pass along with...
5316 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5317 (make_pass_ud_rtl_dce): ...new function.
5318 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5319 rtl_opt_pass along with...
5320 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5321 (make_pass_fast_rtl_dce): ...new function.
5322 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5323 a subclass of rtl_opt_pass along with...
5324 (pass_data_df_initialize_opt): ...new pass_data instance and...
5325 (make_pass_df_initialize_opt): ...new function.
5326 (pass_df_initialize_no_opt): Convert from a global struct to a
5327 subclass of rtl_opt_pass along with...
5328 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5329 (make_pass_df_initialize_no_opt): ...new function.
5330 (pass_df_finish): Convert from a global struct to a subclass of
5331 rtl_opt_pass along with...
5332 (pass_data_df_finish): ...new pass_data instance and...
5333 (make_pass_df_finish): ...new function.
5334 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5335 rtl_opt_pass along with...
5336 (pass_data_rtl_dse1): ...new pass_data instance and...
5337 (make_pass_rtl_dse1): ...new function.
5338 (pass_rtl_dse2): Convert from a global struct to a subclass of
5339 rtl_opt_pass along with...
5340 (pass_data_rtl_dse2): ...new pass_data instance and...
5341 (make_pass_rtl_dse2): ...new function.
5342 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5343 subclass of rtl_opt_pass along with...
5344 (pass_data_dwarf2_frame): ...new pass_data instance and...
5345 (make_pass_dwarf2_frame): ...new function.
5346 * except.c (pass_set_nothrow_function_flags): Convert from a global
5347 struct to a subclass of rtl_opt_pass along with...
5348 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5350 (make_pass_set_nothrow_function_flags): ...new function.
5351 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5352 subclass of rtl_opt_pass along with...
5353 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5355 (make_pass_convert_to_eh_region_ranges): ...new function.
5356 * final.c (pass_compute_alignments): Convert from a global struct to a
5357 subclass of rtl_opt_pass along with...
5358 (pass_data_compute_alignments): ...new pass_data instance and...
5359 (make_pass_compute_alignments): ...new function.
5360 (pass_final): Convert from a global struct to a subclass of
5361 rtl_opt_pass along with...
5362 (pass_data_final): ...new pass_data instance and...
5363 (make_pass_final): ...new function.
5364 (pass_shorten_branches): Convert from a global struct to a subclass of
5365 rtl_opt_pass along with...
5366 (pass_data_shorten_branches): ...new pass_data instance and...
5367 (make_pass_shorten_branches): ...new function.
5368 (pass_clean_state): Convert from a global struct to a subclass of
5369 rtl_opt_pass along with...
5370 (pass_data_clean_state): ...new pass_data instance and...
5371 (make_pass_clean_state): ...new function.
5372 * function.c (pass_instantiate_virtual_regs): Convert from a global
5373 struct to a subclass of rtl_opt_pass along with...
5374 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
5375 (make_pass_instantiate_virtual_regs): ...new function.
5376 (pass_leaf_regs): Convert from a global struct to a subclass of
5377 rtl_opt_pass along with...
5378 (pass_data_leaf_regs): ...new pass_data instance and...
5379 (make_pass_leaf_regs): ...new function.
5380 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
5381 subclass of rtl_opt_pass along with...
5382 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
5384 (make_pass_thread_prologue_and_epilogue): ...new function.
5385 (pass_match_asm_constraints): Convert from a global struct to a
5386 subclass of rtl_opt_pass along with...
5387 (pass_data_match_asm_constraints): ...new pass_data instance and...
5388 (make_pass_match_asm_constraints): ...new function.
5389 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
5390 subclass of rtl_opt_pass along with...
5391 (pass_data_rtl_fwprop): ...new pass_data instance and...
5392 (make_pass_rtl_fwprop): ...new function.
5393 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
5394 rtl_opt_pass along with...
5395 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
5396 (make_pass_rtl_fwprop_addr): ...new function.
5397 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
5398 rtl_opt_pass along with...
5399 (pass_data_rtl_pre): ...new pass_data instance and...
5400 (make_pass_rtl_pre): ...new function.
5401 (pass_rtl_hoist): Convert from a global struct to a subclass of
5402 rtl_opt_pass along with...
5403 (pass_data_rtl_hoist): ...new pass_data instance and...
5404 (make_pass_rtl_hoist): ...new function.
5405 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
5406 subclass of gimple_opt_pass along with...
5407 (pass_data_lower_cf): ...new pass_data instance and...
5408 (make_pass_lower_cf): ...new function.
5409 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
5410 from a global struct to a subclass of gimple_opt_pass along with...
5411 (pass_data_strength_reduction): ...new pass_data instance and...
5412 (make_pass_strength_reduction): ...new function.
5413 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
5414 of rtl_opt_pass along with...
5415 (pass_data_rtl_ifcvt): ...new pass_data instance and...
5416 (make_pass_rtl_ifcvt): ...new function.
5417 (pass_if_after_combine): Convert from a global struct to a subclass of
5418 rtl_opt_pass along with...
5419 (pass_data_if_after_combine): ...new pass_data instance and...
5420 (make_pass_if_after_combine): ...new function.
5421 (pass_if_after_reload): Convert from a global struct to a subclass of
5422 rtl_opt_pass along with...
5423 (pass_data_if_after_reload): ...new pass_data instance and...
5424 (make_pass_if_after_reload): ...new function.
5425 * init-regs.c (pass_initialize_regs): Convert from a global struct to
5426 a subclass of rtl_opt_pass along with...
5427 (pass_data_initialize_regs): ...new pass_data instance and...
5428 (make_pass_initialize_regs): ...new function.
5429 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
5430 of ipa_opt_pass_d along with...
5431 (pass_data_ipa_cp): ...new pass_data instance and...
5432 (make_pass_ipa_cp): ...new function.
5433 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
5434 global struct to a subclass of gimple_opt_pass along with...
5435 (pass_data_inline_parameters): ...new pass_data instance and...
5436 (make_pass_inline_parameters): ...new function.
5437 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
5438 subclass of gimple_opt_pass along with...
5439 (pass_data_early_inline): ...new pass_data instance and...
5440 (make_pass_early_inline): ...new function.
5441 (pass_ipa_inline): Convert from a global struct to a subclass of
5442 ipa_opt_pass_d along with...
5443 (pass_data_ipa_inline): ...new pass_data instance and...
5444 (make_pass_ipa_inline): ...new function.
5445 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
5446 struct to a subclass of gimple_opt_pass along with...
5447 (pass_data_local_pure_const): ...new pass_data instance and...
5448 (make_pass_local_pure_const): ...new function.
5449 (pass_ipa_pure_const): Convert from a global struct to a subclass of
5450 ipa_opt_pass_d along with...
5451 (pass_data_ipa_pure_const): ...new pass_data instance and...
5452 (make_pass_ipa_pure_const): ...new function.
5453 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
5454 to a subclass of ipa_opt_pass_d along with...
5455 (pass_data_ipa_reference): ...new pass_data instance and...
5456 (make_pass_ipa_reference): ...new function.
5457 * ipa-split.c (pass_split_functions): Convert from a global struct to
5458 a subclass of gimple_opt_pass along with...
5459 (pass_data_split_functions): ...new pass_data instance and...
5460 (make_pass_split_functions): ...new function.
5461 (pass_feedback_split_functions): Convert from a global struct to a
5462 subclass of gimple_opt_pass along with...
5463 (pass_data_feedback_split_functions): ...new pass_data instance and...
5464 (make_pass_feedback_split_functions): ...new function.
5465 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
5466 global struct to a subclass of simple_ipa_opt_pass along with...
5467 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
5469 (make_pass_ipa_function_and_variable_visibility): ...new function.
5470 (pass_ipa_free_inline_summary): Convert from a global struct to a
5471 subclass of simple_ipa_opt_pass along with...
5472 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
5473 (make_pass_ipa_free_inline_summary): ...new function.
5474 (pass_ipa_whole_program_visibility): Convert from a global struct to a
5475 subclass of ipa_opt_pass_d along with...
5476 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
5478 (make_pass_ipa_whole_program_visibility): ...new function.
5479 (pass_ipa_profile): Convert from a global struct to a subclass of
5480 ipa_opt_pass_d along with...
5481 (pass_data_ipa_profile): ...new pass_data instance and...
5482 (make_pass_ipa_profile): ...new function.
5483 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
5484 ipa_opt_pass_d along with...
5485 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
5486 (make_pass_ipa_cdtor_merge): ...new function.
5487 * ira.c (pass_ira): Convert from a global struct to a subclass of
5488 rtl_opt_pass along with...
5489 (pass_data_ira): ...new pass_data instance and...
5490 (make_pass_ira): ...new function.
5491 (pass_reload): Convert from a global struct to a subclass of
5492 rtl_opt_pass along with...
5493 (pass_data_reload): ...new pass_data instance and...
5494 (make_pass_reload): ...new function.
5495 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
5496 subclass of rtl_opt_pass along with...
5497 (pass_data_cleanup_barriers): ...new pass_data instance and...
5498 (make_pass_cleanup_barriers): ...new function.
5499 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
5500 of rtl_opt_pass along with...
5501 (pass_data_loop2): ...new pass_data instance and...
5502 (make_pass_loop2): ...new function.
5503 (pass_rtl_loop_init): Convert from a global struct to a subclass of
5504 rtl_opt_pass along with...
5505 (pass_data_rtl_loop_init): ...new pass_data instance and...
5506 (make_pass_rtl_loop_init): ...new function.
5507 (pass_rtl_loop_done): Convert from a global struct to a subclass of
5508 rtl_opt_pass along with...
5509 (pass_data_rtl_loop_done): ...new pass_data instance and...
5510 (make_pass_rtl_loop_done): ...new function.
5511 (pass_rtl_move_loop_invariants): Convert from a global struct to a
5512 subclass of rtl_opt_pass along with...
5513 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
5514 (make_pass_rtl_move_loop_invariants): ...new function.
5515 (pass_rtl_unswitch): Convert from a global struct to a subclass of
5516 rtl_opt_pass along with...
5517 (pass_data_rtl_unswitch): ...new pass_data instance and...
5518 (make_pass_rtl_unswitch): ...new function.
5519 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
5520 subclass of rtl_opt_pass along with...
5521 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
5523 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5524 (pass_rtl_doloop): Convert from a global struct to a subclass of
5525 rtl_opt_pass along with...
5526 (pass_data_rtl_doloop): ...new pass_data instance and...
5527 (make_pass_rtl_doloop): ...new function.
5528 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
5529 a subclass of rtl_opt_pass along with...
5530 (pass_data_lower_subreg): ...new pass_data instance and...
5531 (make_pass_lower_subreg): ...new function.
5532 (pass_lower_subreg2): Convert from a global struct to a subclass of
5533 rtl_opt_pass along with...
5534 (pass_data_lower_subreg2): ...new pass_data instance and...
5535 (make_pass_lower_subreg2): ...new function.
5536 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
5537 struct to a subclass of ipa_opt_pass_d along with...
5538 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
5539 (make_pass_ipa_lto_gimple_out): ...new function.
5540 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
5541 of ipa_opt_pass_d along with...
5542 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
5543 (make_pass_ipa_lto_finish_out): ...new function.
5544 * mode-switching.c (pass_mode_switching): Convert from a global struct
5545 to a subclass of rtl_opt_pass along with...
5546 (pass_data_mode_switching): ...new pass_data instance and...
5547 (make_pass_mode_switching): ...new function.
5548 * modulo-sched.c (pass_sms): Convert from a global struct to a
5549 subclass of rtl_opt_pass along with...
5550 (pass_data_sms): ...new pass_data instance and...
5551 (make_pass_sms): ...new function.
5552 * omp-low.c (pass_expand_omp): Convert from a global struct to a
5553 subclass of gimple_opt_pass along with...
5554 (pass_data_expand_omp): ...new pass_data instance and...
5555 (make_pass_expand_omp): ...new function.
5556 (pass_lower_omp): Convert from a global struct to a subclass of
5557 gimple_opt_pass along with...
5558 (pass_data_lower_omp): ...new pass_data instance and...
5559 (make_pass_lower_omp): ...new function.
5560 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
5561 of gimple_opt_pass along with...
5562 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
5563 (make_pass_diagnose_omp_blocks): ...new function.
5564 * passes.c (pass_early_local_passes): Convert from a global struct to
5565 a subclass of simple_ipa_opt_pass along with...
5566 (pass_data_early_local_passes): ...new pass_data instance and...
5567 (make_pass_early_local_passes): ...new function.
5568 (pass_all_early_optimizations): Convert from a global struct to a
5569 subclass of gimple_opt_pass along with...
5570 (pass_data_all_early_optimizations): ...new pass_data instance and...
5571 (make_pass_all_early_optimizations): ...new function.
5572 (pass_all_optimizations): Convert from a global struct to a subclass
5573 of gimple_opt_pass along with...
5574 (pass_data_all_optimizations): ...new pass_data instance and...
5575 (make_pass_all_optimizations): ...new function.
5576 (pass_all_optimizations_g): Convert from a global struct to a subclass
5577 of gimple_opt_pass along with...
5578 (pass_data_all_optimizations_g): ...new pass_data instance and...
5579 (make_pass_all_optimizations_g): ...new function.
5580 (pass_rest_of_compilation): Convert from a global struct to a subclass
5581 of rtl_opt_pass along with...
5582 (pass_data_rest_of_compilation): ...new pass_data instance and...
5583 (make_pass_rest_of_compilation): ...new function.
5584 (pass_postreload): Convert from a global struct to a subclass of
5585 rtl_opt_pass along with...
5586 (pass_data_postreload): ...new pass_data instance and...
5587 (make_pass_postreload): ...new function.
5588 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
5589 subclass of rtl_opt_pass along with...
5590 (pass_data_gcse2): ...new pass_data instance and...
5591 (make_pass_gcse2): ...new function.
5592 * postreload.c (pass_postreload_cse): Convert from a global struct to
5593 a subclass of rtl_opt_pass along with...
5594 (pass_data_postreload_cse): ...new pass_data instance and...
5595 (make_pass_postreload_cse): ...new function.
5596 * predict.c (pass_profile): Convert from a global struct to a subclass
5597 of gimple_opt_pass along with...
5598 (pass_data_profile): ...new pass_data instance and...
5599 (make_pass_profile): ...new function.
5600 (pass_strip_predict_hints): Convert from a global struct to a subclass
5601 of gimple_opt_pass along with...
5602 (pass_data_strip_predict_hints): ...new pass_data instance and...
5603 (make_pass_strip_predict_hints): ...new function.
5604 * recog.c (pass_peephole2): Convert from a global struct to a subclass
5605 of rtl_opt_pass along with...
5606 (pass_data_peephole2): ...new pass_data instance and...
5607 (make_pass_peephole2): ...new function.
5608 (pass_split_all_insns): Convert from a global struct to a subclass of
5609 rtl_opt_pass along with...
5610 (pass_data_split_all_insns): ...new pass_data instance and...
5611 (make_pass_split_all_insns): ...new function.
5612 (pass_split_after_reload): Convert from a global struct to a subclass
5613 of rtl_opt_pass along with...
5614 (pass_data_split_after_reload): ...new pass_data instance and...
5615 (make_pass_split_after_reload): ...new function.
5616 (pass_split_before_regstack): Convert from a global struct to a
5617 subclass of rtl_opt_pass along with...
5618 (pass_data_split_before_regstack): ...new pass_data instance and...
5619 (make_pass_split_before_regstack): ...new function.
5620 (pass_split_before_sched2): Convert from a global struct to a subclass
5621 of rtl_opt_pass along with...
5622 (pass_data_split_before_sched2): ...new pass_data instance and...
5623 (make_pass_split_before_sched2): ...new function.
5624 (pass_split_for_shorten_branches): Convert from a global struct to a
5625 subclass of rtl_opt_pass along with...
5626 (pass_data_split_for_shorten_branches): ...new pass_data instance
5628 (make_pass_split_for_shorten_branches): ...new function.
5629 * ree.c (pass_ree): Convert from a global struct to a subclass of
5630 rtl_opt_pass along with...
5631 (pass_data_ree): ...new pass_data instance and...
5632 (make_pass_ree): ...new function.
5633 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
5634 subclass of rtl_opt_pass along with...
5635 (pass_data_stack_regs): ...new pass_data instance and...
5636 (make_pass_stack_regs): ...new function.
5637 (pass_stack_regs_run): Convert from a global struct to a subclass of
5638 rtl_opt_pass along with...
5639 (pass_data_stack_regs_run): ...new pass_data instance and...
5640 (make_pass_stack_regs_run): ...new function.
5641 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
5642 subclass of rtl_opt_pass along with...
5643 (pass_data_cprop_hardreg): ...new pass_data instance and...
5644 (make_pass_cprop_hardreg): ...new function.
5645 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
5646 subclass of rtl_opt_pass along with...
5647 (pass_data_reginfo_init): ...new pass_data instance and...
5648 (make_pass_reginfo_init): ...new function.
5649 * regmove.c (pass_regmove): Convert from a global struct to a subclass
5650 of rtl_opt_pass along with...
5651 (pass_data_regmove): ...new pass_data instance and...
5652 (make_pass_regmove): ...new function.
5653 * regrename.c (pass_regrename): Convert from a global struct to a
5654 subclass of rtl_opt_pass along with...
5655 (pass_data_regrename): ...new pass_data instance and...
5656 (make_pass_regrename): ...new function.
5657 * reorg.c (pass_delay_slots): Convert from a global struct to a
5658 subclass of rtl_opt_pass along with...
5659 (pass_data_delay_slots): ...new pass_data instance and...
5660 (make_pass_delay_slots): ...new function.
5661 (pass_machine_reorg): Convert from a global struct to a subclass of
5662 rtl_opt_pass along with...
5663 (pass_data_machine_reorg): ...new pass_data instance and...
5664 (make_pass_machine_reorg): ...new function.
5665 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
5666 of rtl_opt_pass along with...
5667 (pass_data_sched): ...new pass_data instance and...
5668 (make_pass_sched): ...new function.
5669 (pass_sched2): Convert from a global struct to a subclass of
5670 rtl_opt_pass along with...
5671 (pass_data_sched2): ...new pass_data instance and...
5672 (make_pass_sched2): ...new function.
5673 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
5674 to a subclass of rtl_opt_pass along with...
5675 (pass_data_stack_ptr_mod): ...new pass_data instance and...
5676 (make_pass_stack_ptr_mod): ...new function.
5677 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
5678 to a subclass of rtl_opt_pass along with...
5679 (pass_data_rtl_store_motion): ...new pass_data instance and...
5680 (make_pass_rtl_store_motion): ...new function.
5681 * tracer.c (pass_tracer): Convert from a global struct to a subclass
5682 of gimple_opt_pass along with...
5683 (pass_data_tracer): ...new pass_data instance and...
5684 (make_pass_tracer): ...new function.
5685 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
5686 to a subclass of gimple_opt_pass along with...
5687 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
5688 (make_pass_diagnose_tm_blocks): ...new function.
5689 (pass_lower_tm): Convert from a global struct to a subclass of
5690 gimple_opt_pass along with...
5691 (pass_data_lower_tm): ...new pass_data instance and...
5692 (make_pass_lower_tm): ...new function.
5693 (pass_tm_init): Convert from a global struct to a subclass of
5694 gimple_opt_pass along with...
5695 (pass_data_tm_init): ...new pass_data instance and...
5696 (make_pass_tm_init): ...new function.
5697 (pass_tm_mark): Convert from a global struct to a subclass of
5698 gimple_opt_pass along with...
5699 (pass_data_tm_mark): ...new pass_data instance and...
5700 (make_pass_tm_mark): ...new function.
5701 (pass_tm_edges): Convert from a global struct to a subclass of
5702 gimple_opt_pass along with...
5703 (pass_data_tm_edges): ...new pass_data instance and...
5704 (make_pass_tm_edges): ...new function.
5705 (pass_tm_memopt): Convert from a global struct to a subclass of
5706 gimple_opt_pass along with...
5707 (pass_data_tm_memopt): ...new pass_data instance and...
5708 (make_pass_tm_memopt): ...new function.
5709 (pass_ipa_tm): Convert from a global struct to a subclass of
5710 simple_ipa_opt_pass along with...
5711 (pass_data_ipa_tm): ...new pass_data instance and...
5712 (make_pass_ipa_tm): ...new function.
5713 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
5714 subclass of gimple_opt_pass along with...
5715 (pass_data_call_cdce): ...new pass_data instance and...
5716 (make_pass_call_cdce): ...new function.
5717 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
5718 subclass of gimple_opt_pass along with...
5719 (pass_data_build_cfg): ...new pass_data instance and...
5720 (make_pass_build_cfg): ...new function.
5721 (pass_split_crit_edges): Convert from a global struct to a subclass of
5722 gimple_opt_pass along with...
5723 (pass_data_split_crit_edges): ...new pass_data instance and...
5724 (make_pass_split_crit_edges): ...new function.
5725 (pass_warn_function_return): Convert from a global struct to a
5726 subclass of gimple_opt_pass along with...
5727 (pass_data_warn_function_return): ...new pass_data instance and...
5728 (make_pass_warn_function_return): ...new function.
5729 (pass_warn_function_noreturn): Convert from a global struct to a
5730 subclass of gimple_opt_pass along with...
5731 (pass_data_warn_function_noreturn): ...new pass_data instance and...
5732 (make_pass_warn_function_noreturn): ...new function.
5733 (pass_warn_unused_result): Convert from a global struct to a subclass
5734 of gimple_opt_pass along with...
5735 (pass_data_warn_unused_result): ...new pass_data instance and...
5736 (make_pass_warn_unused_result): ...new function.
5737 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
5738 a subclass of gimple_opt_pass along with...
5739 (pass_data_merge_phi): ...new pass_data instance and...
5740 (make_pass_merge_phi): ...new function.
5741 * tree-complex.c (pass_lower_complex): Convert from a global struct to
5742 a subclass of gimple_opt_pass along with...
5743 (pass_data_lower_complex): ...new pass_data instance and...
5744 (make_pass_lower_complex): ...new function.
5745 (pass_lower_complex_O0): Convert from a global struct to a subclass of
5746 gimple_opt_pass along with...
5747 (pass_data_lower_complex_O0): ...new pass_data instance and...
5748 (make_pass_lower_complex_O0): ...new function.
5749 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
5750 subclass of gimple_opt_pass along with...
5751 (pass_data_lower_eh): ...new pass_data instance and...
5752 (make_pass_lower_eh): ...new function.
5753 (pass_refactor_eh): Convert from a global struct to a subclass of
5754 gimple_opt_pass along with...
5755 (pass_data_refactor_eh): ...new pass_data instance and...
5756 (make_pass_refactor_eh): ...new function.
5757 (pass_lower_resx): Convert from a global struct to a subclass of
5758 gimple_opt_pass along with...
5759 (pass_data_lower_resx): ...new pass_data instance and...
5760 (make_pass_lower_resx): ...new function.
5761 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
5762 of gimple_opt_pass along with...
5763 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
5764 (make_pass_lower_eh_dispatch): ...new function.
5765 (pass_cleanup_eh): Convert from a global struct to a subclass of
5766 gimple_opt_pass along with...
5767 (pass_data_cleanup_eh): ...new pass_data instance and...
5768 (make_pass_cleanup_eh): ...new function.
5769 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
5770 to a subclass of simple_ipa_opt_pass along with...
5771 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
5772 (make_pass_ipa_lower_emutls): ...new function.
5773 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
5774 a subclass of gimple_opt_pass along with...
5775 (pass_data_if_conversion): ...new pass_data instance and...
5776 (make_pass_if_conversion): ...new function.
5777 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
5778 subclass of gimple_opt_pass along with...
5779 (pass_data_build_ssa): ...new pass_data instance and...
5780 (make_pass_build_ssa): ...new function.
5781 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
5782 global struct to a subclass of gimple_opt_pass along with...
5783 (pass_data_loop_distribution): ...new pass_data instance and...
5784 (make_pass_loop_distribution): ...new function.
5785 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
5786 subclass of gimple_opt_pass along with...
5787 (pass_data_mudflap_1): ...new pass_data instance and...
5788 (make_pass_mudflap_1): ...new function.
5789 (pass_mudflap_2): Convert from a global struct to a subclass of
5790 gimple_opt_pass along with...
5791 (pass_data_mudflap_2): ...new pass_data instance and...
5792 (make_pass_mudflap_2): ...new function.
5793 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
5794 subclass of gimple_opt_pass along with...
5795 (pass_data_mudflap_1): ...new pass_data instance and...
5796 (make_pass_mudflap_1): ...new function.
5797 (pass_mudflap_2): Convert from a global struct to a subclass of
5798 gimple_opt_pass along with...
5799 (pass_data_mudflap_2): ...new pass_data instance and...
5800 (make_pass_mudflap_2): ...new function.
5801 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
5802 gimple_opt_pass along with...
5803 (pass_data_nrv): ...new pass_data instance and...
5804 (make_pass_nrv): ...new function.
5805 (pass_return_slot): Convert from a global struct to a subclass of
5806 gimple_opt_pass along with...
5807 (pass_data_return_slot): ...new pass_data instance and...
5808 (make_pass_return_slot): ...new function.
5809 * tree-object-size.c (pass_object_sizes): Convert from a global struct
5810 to a subclass of gimple_opt_pass along with...
5811 (pass_data_object_sizes): ...new pass_data instance and...
5812 (make_pass_object_sizes): ...new function.
5813 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
5814 global struct to a subclass of gimple_opt_pass along with...
5815 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
5817 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5818 (pass_fixup_cfg): Convert from a global struct to a subclass of
5819 gimple_opt_pass along with...
5820 (pass_data_fixup_cfg): ...new pass_data instance and...
5821 (make_pass_fixup_cfg): ...new function.
5822 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
5823 (make_pass_mudflap_1): ...new function.
5824 (pass_mudflap_2): Replace declaration with that of...
5825 (make_pass_mudflap_2): ...new function.
5826 (pass_asan): Replace declaration with that of...
5827 (make_pass_asan): ...new function.
5828 (pass_asan_O0): Replace declaration with that of...
5829 (make_pass_asan_O0): ...new function.
5830 (pass_tsan): Replace declaration with that of...
5831 (make_pass_tsan): ...new function.
5832 (pass_tsan_O0): Replace declaration with that of...
5833 (make_pass_tsan_O0): ...new function.
5834 (pass_lower_cf): Replace declaration with that of...
5835 (make_pass_lower_cf): ...new function.
5836 (pass_refactor_eh): Replace declaration with that of...
5837 (make_pass_refactor_eh): ...new function.
5838 (pass_lower_eh): Replace declaration with that of...
5839 (make_pass_lower_eh): ...new function.
5840 (pass_lower_eh_dispatch): Replace declaration with that of...
5841 (make_pass_lower_eh_dispatch): ...new function.
5842 (pass_lower_resx): Replace declaration with that of...
5843 (make_pass_lower_resx): ...new function.
5844 (pass_build_cfg): Replace declaration with that of...
5845 (make_pass_build_cfg): ...new function.
5846 (pass_early_tree_profile): Replace declaration with that of...
5847 (make_pass_early_tree_profile): ...new function.
5848 (pass_cleanup_eh): Replace declaration with that of...
5849 (make_pass_cleanup_eh): ...new function.
5850 (pass_sra): Replace declaration with that of...
5851 (make_pass_sra): ...new function.
5852 (pass_sra_early): Replace declaration with that of...
5853 (make_pass_sra_early): ...new function.
5854 (pass_early_ipa_sra): Replace declaration with that of...
5855 (make_pass_early_ipa_sra): ...new function.
5856 (pass_tail_recursion): Replace declaration with that of...
5857 (make_pass_tail_recursion): ...new function.
5858 (pass_tail_calls): Replace declaration with that of...
5859 (make_pass_tail_calls): ...new function.
5860 (pass_tree_loop): Replace declaration with that of...
5861 (make_pass_tree_loop): ...new function.
5862 (pass_tree_loop_init): Replace declaration with that of...
5863 (make_pass_tree_loop_init): ...new function.
5864 (pass_lim): Replace declaration with that of...
5865 (make_pass_lim): ...new function.
5866 (pass_tree_unswitch): Replace declaration with that of...
5867 (make_pass_tree_unswitch): ...new function.
5868 (pass_predcom): Replace declaration with that of...
5869 (make_pass_predcom): ...new function.
5870 (pass_iv_canon): Replace declaration with that of...
5871 (make_pass_iv_canon): ...new function.
5872 (pass_scev_cprop): Replace declaration with that of...
5873 (make_pass_scev_cprop): ...new function.
5874 (pass_empty_loop): Replace declaration with that of...
5875 (make_pass_empty_loop): ...new function.
5876 (pass_record_bounds): Replace declaration with that of...
5877 (make_pass_record_bounds): ...new function.
5878 (pass_graphite): Replace declaration with that of...
5879 (make_pass_graphite): ...new function.
5880 (pass_graphite_transforms): Replace declaration with that of...
5881 (make_pass_graphite_transforms): ...new function.
5882 (pass_if_conversion): Replace declaration with that of...
5883 (make_pass_if_conversion): ...new function.
5884 (pass_loop_distribution): Replace declaration with that of...
5885 (make_pass_loop_distribution): ...new function.
5886 (pass_vectorize): Replace declaration with that of...
5887 (make_pass_vectorize): ...new function.
5888 (pass_slp_vectorize): Replace declaration with that of...
5889 (make_pass_slp_vectorize): ...new function.
5890 (pass_complete_unroll): Replace declaration with that of...
5891 (make_pass_complete_unroll): ...new function.
5892 (pass_complete_unrolli): Replace declaration with that of...
5893 (make_pass_complete_unrolli): ...new function.
5894 (pass_parallelize_loops): Replace declaration with that of...
5895 (make_pass_parallelize_loops): ...new function.
5896 (pass_loop_prefetch): Replace declaration with that of...
5897 (make_pass_loop_prefetch): ...new function.
5898 (pass_iv_optimize): Replace declaration with that of...
5899 (make_pass_iv_optimize): ...new function.
5900 (pass_tree_loop_done): Replace declaration with that of...
5901 (make_pass_tree_loop_done): ...new function.
5902 (pass_ch): Replace declaration with that of...
5903 (make_pass_ch): ...new function.
5904 (pass_ccp): Replace declaration with that of...
5905 (make_pass_ccp): ...new function.
5906 (pass_phi_only_cprop): Replace declaration with that of...
5907 (make_pass_phi_only_cprop): ...new function.
5908 (pass_build_ssa): Replace declaration with that of...
5909 (make_pass_build_ssa): ...new function.
5910 (pass_build_alias): Replace declaration with that of...
5911 (make_pass_build_alias): ...new function.
5912 (pass_build_ealias): Replace declaration with that of...
5913 (make_pass_build_ealias): ...new function.
5914 (pass_dominator): Replace declaration with that of...
5915 (make_pass_dominator): ...new function.
5916 (pass_dce): Replace declaration with that of...
5917 (make_pass_dce): ...new function.
5918 (pass_dce_loop): Replace declaration with that of...
5919 (make_pass_dce_loop): ...new function.
5920 (pass_cd_dce): Replace declaration with that of...
5921 (make_pass_cd_dce): ...new function.
5922 (pass_call_cdce): Replace declaration with that of...
5923 (make_pass_call_cdce): ...new function.
5924 (pass_merge_phi): Replace declaration with that of...
5925 (make_pass_merge_phi): ...new function.
5926 (pass_split_crit_edges): Replace declaration with that of...
5927 (make_pass_split_crit_edges): ...new function.
5928 (pass_pre): Replace declaration with that of...
5929 (make_pass_pre): ...new function.
5930 (pass_profile): Replace declaration with that of...
5931 (make_pass_profile): ...new function.
5932 (pass_strip_predict_hints): Replace declaration with that of...
5933 (make_pass_strip_predict_hints): ...new function.
5934 (pass_lower_complex_O0): Replace declaration with that of...
5935 (make_pass_lower_complex_O0): ...new function.
5936 (pass_lower_complex): Replace declaration with that of...
5937 (make_pass_lower_complex): ...new function.
5938 (pass_lower_vector): Replace declaration with that of...
5939 (make_pass_lower_vector): ...new function.
5940 (pass_lower_vector_ssa): Replace declaration with that of...
5941 (make_pass_lower_vector_ssa): ...new function.
5942 (pass_lower_omp): Replace declaration with that of...
5943 (make_pass_lower_omp): ...new function.
5944 (pass_diagnose_omp_blocks): Replace declaration with that of...
5945 (make_pass_diagnose_omp_blocks): ...new function.
5946 (pass_expand_omp): Replace declaration with that of...
5947 (make_pass_expand_omp): ...new function.
5948 (pass_expand_omp_ssa): Replace declaration with that of...
5949 (make_pass_expand_omp_ssa): ...new function.
5950 (pass_object_sizes): Replace declaration with that of...
5951 (make_pass_object_sizes): ...new function.
5952 (pass_strlen): Replace declaration with that of...
5953 (make_pass_strlen): ...new function.
5954 (pass_fold_builtins): Replace declaration with that of...
5955 (make_pass_fold_builtins): ...new function.
5956 (pass_stdarg): Replace declaration with that of...
5957 (make_pass_stdarg): ...new function.
5958 (pass_early_warn_uninitialized): Replace declaration with that of...
5959 (make_pass_early_warn_uninitialized): ...new function.
5960 (pass_late_warn_uninitialized): Replace declaration with that of...
5961 (make_pass_late_warn_uninitialized): ...new function.
5962 (pass_cse_reciprocals): Replace declaration with that of...
5963 (make_pass_cse_reciprocals): ...new function.
5964 (pass_cse_sincos): Replace declaration with that of...
5965 (make_pass_cse_sincos): ...new function.
5966 (pass_optimize_bswap): Replace declaration with that of...
5967 (make_pass_optimize_bswap): ...new function.
5968 (pass_optimize_widening_mul): Replace declaration with that of...
5969 (make_pass_optimize_widening_mul): ...new function.
5970 (pass_warn_function_return): Replace declaration with that of...
5971 (make_pass_warn_function_return): ...new function.
5972 (pass_warn_function_noreturn): Replace declaration with that of...
5973 (make_pass_warn_function_noreturn): ...new function.
5974 (pass_cselim): Replace declaration with that of...
5975 (make_pass_cselim): ...new function.
5976 (pass_phiopt): Replace declaration with that of...
5977 (make_pass_phiopt): ...new function.
5978 (pass_forwprop): Replace declaration with that of...
5979 (make_pass_forwprop): ...new function.
5980 (pass_phiprop): Replace declaration with that of...
5981 (make_pass_phiprop): ...new function.
5982 (pass_tree_ifcombine): Replace declaration with that of...
5983 (make_pass_tree_ifcombine): ...new function.
5984 (pass_dse): Replace declaration with that of...
5985 (make_pass_dse): ...new function.
5986 (pass_nrv): Replace declaration with that of...
5987 (make_pass_nrv): ...new function.
5988 (pass_rename_ssa_copies): Replace declaration with that of...
5989 (make_pass_rename_ssa_copies): ...new function.
5990 (pass_sink_code): Replace declaration with that of...
5991 (make_pass_sink_code): ...new function.
5992 (pass_fre): Replace declaration with that of...
5993 (make_pass_fre): ...new function.
5994 (pass_check_data_deps): Replace declaration with that of...
5995 (make_pass_check_data_deps): ...new function.
5996 (pass_copy_prop): Replace declaration with that of...
5997 (make_pass_copy_prop): ...new function.
5998 (pass_vrp): Replace declaration with that of...
5999 (make_pass_vrp): ...new function.
6000 (pass_uncprop): Replace declaration with that of...
6001 (make_pass_uncprop): ...new function.
6002 (pass_return_slot): Replace declaration with that of...
6003 (make_pass_return_slot): ...new function.
6004 (pass_reassoc): Replace declaration with that of...
6005 (make_pass_reassoc): ...new function.
6006 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6007 (make_pass_rebuild_cgraph_edges): ...new function.
6008 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6009 (make_pass_remove_cgraph_callee_edges): ...new function.
6010 (pass_build_cgraph_edges): Replace declaration with that of...
6011 (make_pass_build_cgraph_edges): ...new function.
6012 (pass_local_pure_const): Replace declaration with that of...
6013 (make_pass_local_pure_const): ...new function.
6014 (pass_tracer): Replace declaration with that of...
6015 (make_pass_tracer): ...new function.
6016 (pass_warn_unused_result): Replace declaration with that of...
6017 (make_pass_warn_unused_result): ...new function.
6018 (pass_diagnose_tm_blocks): Replace declaration with that of...
6019 (make_pass_diagnose_tm_blocks): ...new function.
6020 (pass_lower_tm): Replace declaration with that of...
6021 (make_pass_lower_tm): ...new function.
6022 (pass_tm_init): Replace declaration with that of...
6023 (make_pass_tm_init): ...new function.
6024 (pass_tm_mark): Replace declaration with that of...
6025 (make_pass_tm_mark): ...new function.
6026 (pass_tm_memopt): Replace declaration with that of...
6027 (make_pass_tm_memopt): ...new function.
6028 (pass_tm_edges): Replace declaration with that of...
6029 (make_pass_tm_edges): ...new function.
6030 (pass_split_functions): Replace declaration with that of...
6031 (make_pass_split_functions): ...new function.
6032 (pass_feedback_split_functions): Replace declaration with that of...
6033 (make_pass_feedback_split_functions): ...new function.
6034 (pass_strength_reduction): Replace declaration with that of...
6035 (make_pass_strength_reduction): ...new function.
6036 (pass_ipa_lower_emutls): Replace declaration with that of...
6037 (make_pass_ipa_lower_emutls): ...new function.
6038 (pass_ipa_function_and_variable_visibility): Replace declaration with
6040 (make_pass_ipa_function_and_variable_visibility): ...new function.
6041 (pass_ipa_tree_profile): Replace declaration with that of...
6042 (make_pass_ipa_tree_profile): ...new function.
6043 (pass_early_local_passes): Replace declaration with that of...
6044 (make_pass_early_local_passes): ...new function.
6045 (pass_ipa_whole_program_visibility): Replace declaration with that
6047 (make_pass_ipa_whole_program_visibility): ...new function.
6048 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6049 (make_pass_ipa_lto_gimple_out): ...new function.
6050 (pass_ipa_increase_alignment): Replace declaration with that of...
6051 (make_pass_ipa_increase_alignment): ...new function.
6052 (pass_ipa_inline): Replace declaration with that of...
6053 (make_pass_ipa_inline): ...new function.
6054 (pass_ipa_free_lang_data): Replace declaration with that of...
6055 (make_pass_ipa_free_lang_data): ...new function.
6056 (pass_ipa_free_inline_summary): Replace declaration with that of...
6057 (make_pass_ipa_free_inline_summary): ...new function.
6058 (pass_ipa_cp): Replace declaration with that of...
6059 (make_pass_ipa_cp): ...new function.
6060 (pass_ipa_reference): Replace declaration with that of...
6061 (make_pass_ipa_reference): ...new function.
6062 (pass_ipa_pure_const): Replace declaration with that of...
6063 (make_pass_ipa_pure_const): ...new function.
6064 (pass_ipa_pta): Replace declaration with that of...
6065 (make_pass_ipa_pta): ...new function.
6066 (pass_ipa_lto_finish_out): Replace declaration with that of...
6067 (make_pass_ipa_lto_finish_out): ...new function.
6068 (pass_ipa_tm): Replace declaration with that of...
6069 (make_pass_ipa_tm): ...new function.
6070 (pass_ipa_profile): Replace declaration with that of...
6071 (make_pass_ipa_profile): ...new function.
6072 (pass_ipa_cdtor_merge): Replace declaration with that of...
6073 (make_pass_ipa_cdtor_merge): ...new function.
6074 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6076 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6077 (pass_init_datastructures): Replace declaration with that of...
6078 (make_pass_init_datastructures): ...new function.
6079 (pass_fixup_cfg): Replace declaration with that of...
6080 (make_pass_fixup_cfg): ...new function.
6081 (pass_expand): Replace declaration with that of...
6082 (make_pass_expand): ...new function.
6083 (pass_instantiate_virtual_regs): Replace declaration with that of...
6084 (make_pass_instantiate_virtual_regs): ...new function.
6085 (pass_rtl_fwprop): Replace declaration with that of...
6086 (make_pass_rtl_fwprop): ...new function.
6087 (pass_rtl_fwprop_addr): Replace declaration with that of...
6088 (make_pass_rtl_fwprop_addr): ...new function.
6089 (pass_jump): Replace declaration with that of...
6090 (make_pass_jump): ...new function.
6091 (pass_jump2): Replace declaration with that of...
6092 (make_pass_jump2): ...new function.
6093 (pass_lower_subreg): Replace declaration with that of...
6094 (make_pass_lower_subreg): ...new function.
6095 (pass_cse): Replace declaration with that of...
6096 (make_pass_cse): ...new function.
6097 (pass_fast_rtl_dce): Replace declaration with that of...
6098 (make_pass_fast_rtl_dce): ...new function.
6099 (pass_ud_rtl_dce): Replace declaration with that of...
6100 (make_pass_ud_rtl_dce): ...new function.
6101 (pass_rtl_dce): Replace declaration with that of...
6102 (make_pass_rtl_dce): ...new function.
6103 (pass_rtl_dse1): Replace declaration with that of...
6104 (make_pass_rtl_dse1): ...new function.
6105 (pass_rtl_dse2): Replace declaration with that of...
6106 (make_pass_rtl_dse2): ...new function.
6107 (pass_rtl_dse3): Replace declaration with that of...
6108 (make_pass_rtl_dse3): ...new function.
6109 (pass_rtl_cprop): Replace declaration with that of...
6110 (make_pass_rtl_cprop): ...new function.
6111 (pass_rtl_pre): Replace declaration with that of...
6112 (make_pass_rtl_pre): ...new function.
6113 (pass_rtl_hoist): Replace declaration with that of...
6114 (make_pass_rtl_hoist): ...new function.
6115 (pass_rtl_store_motion): Replace declaration with that of...
6116 (make_pass_rtl_store_motion): ...new function.
6117 (pass_cse_after_global_opts): Replace declaration with that of...
6118 (make_pass_cse_after_global_opts): ...new function.
6119 (pass_rtl_ifcvt): Replace declaration with that of...
6120 (make_pass_rtl_ifcvt): ...new function.
6121 (pass_into_cfg_layout_mode): Replace declaration with that of...
6122 (make_pass_into_cfg_layout_mode): ...new function.
6123 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6124 (make_pass_outof_cfg_layout_mode): ...new function.
6125 (pass_loop2): Replace declaration with that of...
6126 (make_pass_loop2): ...new function.
6127 (pass_rtl_loop_init): Replace declaration with that of...
6128 (make_pass_rtl_loop_init): ...new function.
6129 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6130 (make_pass_rtl_move_loop_invariants): ...new function.
6131 (pass_rtl_unswitch): Replace declaration with that of...
6132 (make_pass_rtl_unswitch): ...new function.
6133 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6134 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6135 (pass_rtl_doloop): Replace declaration with that of...
6136 (make_pass_rtl_doloop): ...new function.
6137 (pass_rtl_loop_done): Replace declaration with that of...
6138 (make_pass_rtl_loop_done): ...new function.
6139 (pass_web): Replace declaration with that of...
6140 (make_pass_web): ...new function.
6141 (pass_cse2): Replace declaration with that of...
6142 (make_pass_cse2): ...new function.
6143 (pass_df_initialize_opt): Replace declaration with that of...
6144 (make_pass_df_initialize_opt): ...new function.
6145 (pass_df_initialize_no_opt): Replace declaration with that of...
6146 (make_pass_df_initialize_no_opt): ...new function.
6147 (pass_reginfo_init): Replace declaration with that of...
6148 (make_pass_reginfo_init): ...new function.
6149 (pass_inc_dec): Replace declaration with that of...
6150 (make_pass_inc_dec): ...new function.
6151 (pass_stack_ptr_mod): Replace declaration with that of...
6152 (make_pass_stack_ptr_mod): ...new function.
6153 (pass_initialize_regs): Replace declaration with that of...
6154 (make_pass_initialize_regs): ...new function.
6155 (pass_combine): Replace declaration with that of...
6156 (make_pass_combine): ...new function.
6157 (pass_if_after_combine): Replace declaration with that of...
6158 (make_pass_if_after_combine): ...new function.
6159 (pass_ree): Replace declaration with that of...
6160 (make_pass_ree): ...new function.
6161 (pass_partition_blocks): Replace declaration with that of...
6162 (make_pass_partition_blocks): ...new function.
6163 (pass_match_asm_constraints): Replace declaration with that of...
6164 (make_pass_match_asm_constraints): ...new function.
6165 (pass_regmove): Replace declaration with that of...
6166 (make_pass_regmove): ...new function.
6167 (pass_split_all_insns): Replace declaration with that of...
6168 (make_pass_split_all_insns): ...new function.
6169 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6170 (make_pass_fast_rtl_byte_dce): ...new function.
6171 (pass_lower_subreg2): Replace declaration with that of...
6172 (make_pass_lower_subreg2): ...new function.
6173 (pass_mode_switching): Replace declaration with that of...
6174 (make_pass_mode_switching): ...new function.
6175 (pass_sms): Replace declaration with that of...
6176 (make_pass_sms): ...new function.
6177 (pass_sched): Replace declaration with that of...
6178 (make_pass_sched): ...new function.
6179 (pass_ira): Replace declaration with that of...
6180 (make_pass_ira): ...new function.
6181 (pass_reload): Replace declaration with that of...
6182 (make_pass_reload): ...new function.
6183 (pass_clean_state): Replace declaration with that of...
6184 (make_pass_clean_state): ...new function.
6185 (pass_branch_prob): Replace declaration with that of...
6186 (make_pass_branch_prob): ...new function.
6187 (pass_value_profile_transformations): Replace declaration with that
6189 (make_pass_value_profile_transformations): ...new function.
6190 (pass_postreload_cse): Replace declaration with that of...
6191 (make_pass_postreload_cse): ...new function.
6192 (pass_gcse2): Replace declaration with that of...
6193 (make_pass_gcse2): ...new function.
6194 (pass_split_after_reload): Replace declaration with that of...
6195 (make_pass_split_after_reload): ...new function.
6196 (pass_branch_target_load_optimize1): Replace declaration with that
6198 (make_pass_branch_target_load_optimize1): ...new function.
6199 (pass_thread_prologue_and_epilogue): Replace declaration with that
6201 (make_pass_thread_prologue_and_epilogue): ...new function.
6202 (pass_stack_adjustments): Replace declaration with that of...
6203 (make_pass_stack_adjustments): ...new function.
6204 (pass_peephole2): Replace declaration with that of...
6205 (make_pass_peephole2): ...new function.
6206 (pass_if_after_reload): Replace declaration with that of...
6207 (make_pass_if_after_reload): ...new function.
6208 (pass_regrename): Replace declaration with that of...
6209 (make_pass_regrename): ...new function.
6210 (pass_cprop_hardreg): Replace declaration with that of...
6211 (make_pass_cprop_hardreg): ...new function.
6212 (pass_reorder_blocks): Replace declaration with that of...
6213 (make_pass_reorder_blocks): ...new function.
6214 (pass_branch_target_load_optimize2): Replace declaration with that
6216 (make_pass_branch_target_load_optimize2): ...new function.
6217 (pass_leaf_regs): Replace declaration with that of...
6218 (make_pass_leaf_regs): ...new function.
6219 (pass_split_before_sched2): Replace declaration with that of...
6220 (make_pass_split_before_sched2): ...new function.
6221 (pass_compare_elim_after_reload): Replace declaration with that of...
6222 (make_pass_compare_elim_after_reload): ...new function.
6223 (pass_sched2): Replace declaration with that of...
6224 (make_pass_sched2): ...new function.
6225 (pass_stack_regs): Replace declaration with that of...
6226 (make_pass_stack_regs): ...new function.
6227 (pass_stack_regs_run): Replace declaration with that of...
6228 (make_pass_stack_regs_run): ...new function.
6229 (pass_df_finish): Replace declaration with that of...
6230 (make_pass_df_finish): ...new function.
6231 (pass_compute_alignments): Replace declaration with that of...
6232 (make_pass_compute_alignments): ...new function.
6233 (pass_duplicate_computed_gotos): Replace declaration with that of...
6234 (make_pass_duplicate_computed_gotos): ...new function.
6235 (pass_variable_tracking): Replace declaration with that of...
6236 (make_pass_variable_tracking): ...new function.
6237 (pass_free_cfg): Replace declaration with that of...
6238 (make_pass_free_cfg): ...new function.
6239 (pass_machine_reorg): Replace declaration with that of...
6240 (make_pass_machine_reorg): ...new function.
6241 (pass_cleanup_barriers): Replace declaration with that of...
6242 (make_pass_cleanup_barriers): ...new function.
6243 (pass_delay_slots): Replace declaration with that of...
6244 (make_pass_delay_slots): ...new function.
6245 (pass_split_for_shorten_branches): Replace declaration with that of...
6246 (make_pass_split_for_shorten_branches): ...new function.
6247 (pass_split_before_regstack): Replace declaration with that of...
6248 (make_pass_split_before_regstack): ...new function.
6249 (pass_convert_to_eh_region_ranges): Replace declaration with that
6251 (make_pass_convert_to_eh_region_ranges): ...new function.
6252 (pass_shorten_branches): Replace declaration with that of...
6253 (make_pass_shorten_branches): ...new function.
6254 (pass_set_nothrow_function_flags): Replace declaration with that of...
6255 (make_pass_set_nothrow_function_flags): ...new function.
6256 (pass_dwarf2_frame): Replace declaration with that of...
6257 (make_pass_dwarf2_frame): ...new function.
6258 (pass_final): Replace declaration with that of...
6259 (make_pass_final): ...new function.
6260 (pass_rtl_seqabstr): Replace declaration with that of...
6261 (make_pass_rtl_seqabstr): ...new function.
6262 (pass_release_ssa_names): Replace declaration with that of...
6263 (make_pass_release_ssa_names): ...new function.
6264 (pass_early_inline): Replace declaration with that of...
6265 (make_pass_early_inline): ...new function.
6266 (pass_inline_parameters): Replace declaration with that of...
6267 (make_pass_inline_parameters): ...new function.
6268 (pass_update_address_taken): Replace declaration with that of...
6269 (make_pass_update_address_taken): ...new function.
6270 (pass_convert_switch): Replace declaration with that of...
6271 (make_pass_convert_switch): ...new function.
6272 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6273 to a subclass of simple_ipa_opt_pass along with...
6274 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6275 (make_pass_ipa_tree_profile): ...new function.
6276 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6277 subclass of gimple_opt_pass along with...
6278 (pass_data_sra_early): ...new pass_data instance and...
6279 (make_pass_sra_early): ...new function.
6280 (pass_sra): Convert from a global struct to a subclass of
6281 gimple_opt_pass along with...
6282 (pass_data_sra): ...new pass_data instance and...
6283 (make_pass_sra): ...new function.
6284 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6285 gimple_opt_pass along with...
6286 (pass_data_early_ipa_sra): ...new pass_data instance and...
6287 (make_pass_early_ipa_sra): ...new function.
6288 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6289 subclass of gimple_opt_pass along with...
6290 (pass_data_ccp): ...new pass_data instance and...
6291 (make_pass_ccp): ...new function.
6292 (pass_fold_builtins): Convert from a global struct to a subclass of
6293 gimple_opt_pass along with...
6294 (pass_data_fold_builtins): ...new pass_data instance and...
6295 (make_pass_fold_builtins): ...new function.
6296 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6297 subclass of gimple_opt_pass along with...
6298 (pass_data_copy_prop): ...new pass_data instance and...
6299 (make_pass_copy_prop): ...new function.
6300 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6301 global struct to a subclass of gimple_opt_pass along with...
6302 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6303 (make_pass_rename_ssa_copies): ...new function.
6304 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6305 subclass of gimple_opt_pass along with...
6306 (pass_data_dce): ...new pass_data instance and...
6307 (make_pass_dce): ...new function.
6308 (pass_dce_loop): Convert from a global struct to a subclass of
6309 gimple_opt_pass along with...
6310 (pass_data_dce_loop): ...new pass_data instance and...
6311 (make_pass_dce_loop): ...new function.
6312 (pass_cd_dce): Convert from a global struct to a subclass of
6313 gimple_opt_pass along with...
6314 (pass_data_cd_dce): ...new pass_data instance and...
6315 (make_pass_cd_dce): ...new function.
6316 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6317 subclass of gimple_opt_pass along with...
6318 (pass_data_dominator): ...new pass_data instance and...
6319 (make_pass_dominator): ...new function.
6320 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6321 gimple_opt_pass along with...
6322 (pass_data_phi_only_cprop): ...new pass_data instance and...
6323 (make_pass_phi_only_cprop): ...new function.
6324 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6325 subclass of gimple_opt_pass along with...
6326 (pass_data_dse): ...new pass_data instance and...
6327 (make_pass_dse): ...new function.
6328 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6329 a subclass of gimple_opt_pass along with...
6330 (pass_data_forwprop): ...new pass_data instance and...
6331 (make_pass_forwprop): ...new function.
6332 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6333 struct to a subclass of gimple_opt_pass along with...
6334 (pass_data_tree_ifcombine): ...new pass_data instance and...
6335 (make_pass_tree_ifcombine): ...new function.
6336 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6337 subclass of gimple_opt_pass along with...
6338 (pass_data_ch): ...new pass_data instance and...
6339 (make_pass_ch): ...new function.
6340 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6341 subclass of gimple_opt_pass along with...
6342 (pass_data_tree_loop): ...new pass_data instance and...
6343 (make_pass_tree_loop): ...new function.
6344 (pass_tree_loop_init): Convert from a global struct to a subclass of
6345 gimple_opt_pass along with...
6346 (pass_data_tree_loop_init): ...new pass_data instance and...
6347 (make_pass_tree_loop_init): ...new function.
6348 (pass_lim): Convert from a global struct to a subclass of
6349 gimple_opt_pass along with...
6350 (pass_data_lim): ...new pass_data instance and...
6351 (make_pass_lim): ...new function.
6352 (pass_tree_unswitch): Convert from a global struct to a subclass of
6353 gimple_opt_pass along with...
6354 (pass_data_tree_unswitch): ...new pass_data instance and...
6355 (make_pass_tree_unswitch): ...new function.
6356 (pass_predcom): Convert from a global struct to a subclass of
6357 gimple_opt_pass along with...
6358 (pass_data_predcom): ...new pass_data instance and...
6359 (make_pass_predcom): ...new function.
6360 (pass_vectorize): Convert from a global struct to a subclass of
6361 gimple_opt_pass along with...
6362 (pass_data_vectorize): ...new pass_data instance and...
6363 (make_pass_vectorize): ...new function.
6364 (pass_graphite): Convert from a global struct to a subclass of
6365 gimple_opt_pass along with...
6366 (pass_data_graphite): ...new pass_data instance and...
6367 (make_pass_graphite): ...new function.
6368 (pass_graphite_transforms): Convert from a global struct to a subclass
6369 of gimple_opt_pass along with...
6370 (pass_data_graphite_transforms): ...new pass_data instance and...
6371 (make_pass_graphite_transforms): ...new function.
6372 (pass_check_data_deps): Convert from a global struct to a subclass of
6373 gimple_opt_pass along with...
6374 (pass_data_check_data_deps): ...new pass_data instance and...
6375 (make_pass_check_data_deps): ...new function.
6376 (pass_iv_canon): Convert from a global struct to a subclass of
6377 gimple_opt_pass along with...
6378 (pass_data_iv_canon): ...new pass_data instance and...
6379 (make_pass_iv_canon): ...new function.
6380 (pass_scev_cprop): Convert from a global struct to a subclass of
6381 gimple_opt_pass along with...
6382 (pass_data_scev_cprop): ...new pass_data instance and...
6383 (make_pass_scev_cprop): ...new function.
6384 (pass_record_bounds): Convert from a global struct to a subclass of
6385 gimple_opt_pass along with...
6386 (pass_data_record_bounds): ...new pass_data instance and...
6387 (make_pass_record_bounds): ...new function.
6388 (pass_complete_unroll): Convert from a global struct to a subclass of
6389 gimple_opt_pass along with...
6390 (pass_data_complete_unroll): ...new pass_data instance and...
6391 (make_pass_complete_unroll): ...new function.
6392 (pass_complete_unrolli): Convert from a global struct to a subclass of
6393 gimple_opt_pass along with...
6394 (pass_data_complete_unrolli): ...new pass_data instance and...
6395 (make_pass_complete_unrolli): ...new function.
6396 (pass_parallelize_loops): Convert from a global struct to a subclass
6397 of gimple_opt_pass along with...
6398 (pass_data_parallelize_loops): ...new pass_data instance and...
6399 (make_pass_parallelize_loops): ...new function.
6400 (pass_loop_prefetch): Convert from a global struct to a subclass of
6401 gimple_opt_pass along with...
6402 (pass_data_loop_prefetch): ...new pass_data instance and...
6403 (make_pass_loop_prefetch): ...new function.
6404 (pass_iv_optimize): Convert from a global struct to a subclass of
6405 gimple_opt_pass along with...
6406 (pass_data_iv_optimize): ...new pass_data instance and...
6407 (make_pass_iv_optimize): ...new function.
6408 (pass_tree_loop_done): Convert from a global struct to a subclass of
6409 gimple_opt_pass along with...
6410 (pass_data_tree_loop_done): ...new pass_data instance and...
6411 (make_pass_tree_loop_done): ...new function.
6412 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
6413 struct to a subclass of gimple_opt_pass along with...
6414 (pass_data_cse_reciprocals): ...new pass_data instance and...
6415 (make_pass_cse_reciprocals): ...new function.
6416 (pass_cse_sincos): Convert from a global struct to a subclass of
6417 gimple_opt_pass along with...
6418 (pass_data_cse_sincos): ...new pass_data instance and...
6419 (make_pass_cse_sincos): ...new function.
6420 (pass_optimize_bswap): Convert from a global struct to a subclass of
6421 gimple_opt_pass along with...
6422 (pass_data_optimize_bswap): ...new pass_data instance and...
6423 (make_pass_optimize_bswap): ...new function.
6424 (pass_optimize_widening_mul): Convert from a global struct to a
6425 subclass of gimple_opt_pass along with...
6426 (pass_data_optimize_widening_mul): ...new pass_data instance and...
6427 (make_pass_optimize_widening_mul): ...new function.
6428 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
6429 subclass of gimple_opt_pass along with...
6430 (pass_data_phiopt): ...new pass_data instance and...
6431 (make_pass_phiopt): ...new function.
6432 (pass_cselim): Convert from a global struct to a subclass of
6433 gimple_opt_pass along with...
6434 (pass_data_cselim): ...new pass_data instance and...
6435 (make_pass_cselim): ...new function.
6436 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
6437 subclass of gimple_opt_pass along with...
6438 (pass_data_phiprop): ...new pass_data instance and...
6439 (make_pass_phiprop): ...new function.
6440 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
6441 subclass of gimple_opt_pass along with...
6442 (pass_data_pre): ...new pass_data instance and...
6443 (make_pass_pre): ...new function.
6444 (pass_fre): Convert from a global struct to a subclass of
6445 gimple_opt_pass along with...
6446 (pass_data_fre): ...new pass_data instance and...
6447 (make_pass_fre): ...new function.
6448 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
6449 subclass of gimple_opt_pass along with...
6450 (pass_data_reassoc): ...new pass_data instance and...
6451 (make_pass_reassoc): ...new function.
6452 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
6453 subclass of gimple_opt_pass along with...
6454 (pass_data_sink_code): ...new pass_data instance and...
6455 (make_pass_sink_code): ...new function.
6456 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
6457 subclass of gimple_opt_pass along with...
6458 (pass_data_strlen): ...new pass_data instance and...
6459 (make_pass_strlen): ...new function.
6460 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
6461 struct to a subclass of gimple_opt_pass along with...
6462 (pass_data_build_alias): ...new pass_data instance and...
6463 (make_pass_build_alias): ...new function.
6464 (pass_build_ealias): Convert from a global struct to a subclass of
6465 gimple_opt_pass along with...
6466 (pass_data_build_ealias): ...new pass_data instance and...
6467 (make_pass_build_ealias): ...new function.
6468 (pass_ipa_pta): Convert from a global struct to a subclass of
6469 simple_ipa_opt_pass along with...
6470 (pass_data_ipa_pta): ...new pass_data instance and...
6471 (make_pass_ipa_pta): ...new function.
6472 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
6473 subclass of gimple_opt_pass along with...
6474 (pass_data_uncprop): ...new pass_data instance and...
6475 (make_pass_uncprop): ...new function.
6476 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
6477 global struct to a subclass of gimple_opt_pass along with...
6478 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
6479 (make_pass_late_warn_uninitialized): ...new function.
6480 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
6481 to a subclass of gimple_opt_pass along with...
6482 (pass_data_init_datastructures): ...new pass_data instance and...
6483 (make_pass_init_datastructures): ...new function.
6484 (pass_early_warn_uninitialized): Convert from a global struct to a
6485 subclass of gimple_opt_pass along with...
6486 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
6487 (make_pass_early_warn_uninitialized): ...new function.
6488 (pass_update_address_taken): Convert from a global struct to a
6489 subclass of gimple_opt_pass along with...
6490 (pass_data_update_address_taken): ...new pass_data instance and...
6491 (make_pass_update_address_taken): ...new function.
6492 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
6493 struct to a subclass of gimple_opt_pass along with...
6494 (pass_data_release_ssa_names): ...new pass_data instance and...
6495 (make_pass_release_ssa_names): ...new function.
6496 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
6497 subclass of gimple_opt_pass along with...
6498 (pass_data_stdarg): ...new pass_data instance and...
6499 (make_pass_stdarg): ...new function.
6500 * tree-switch-conversion.c (pass_convert_switch): Convert from a
6501 global struct to a subclass of gimple_opt_pass along with...
6502 (pass_data_convert_switch): ...new pass_data instance and...
6503 (make_pass_convert_switch): ...new function.
6504 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
6505 to a subclass of gimple_opt_pass along with...
6506 (pass_data_tail_recursion): ...new pass_data instance and...
6507 (make_pass_tail_recursion): ...new function.
6508 (pass_tail_calls): Convert from a global struct to a subclass of
6509 gimple_opt_pass along with...
6510 (pass_data_tail_calls): ...new pass_data instance and...
6511 (make_pass_tail_calls): ...new function.
6512 * tree-vect-generic.c (pass_lower_vector): Convert from a global
6513 struct to a subclass of gimple_opt_pass along with...
6514 (pass_data_lower_vector): ...new pass_data instance and...
6515 (make_pass_lower_vector): ...new function.
6516 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
6517 gimple_opt_pass along with...
6518 (pass_data_lower_vector_ssa): ...new pass_data instance and...
6519 (make_pass_lower_vector_ssa): ...new function.
6520 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
6521 to a subclass of gimple_opt_pass along with...
6522 (pass_data_slp_vectorize): ...new pass_data instance and...
6523 (make_pass_slp_vectorize): ...new function.
6524 (pass_ipa_increase_alignment): Convert from a global struct to a
6525 subclass of simple_ipa_opt_pass along with...
6526 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
6527 (make_pass_ipa_increase_alignment): ...new function.
6528 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
6529 gimple_opt_pass along with...
6530 (pass_data_vrp): ...new pass_data instance and...
6531 (make_pass_vrp): ...new function.
6532 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
6533 subclass of simple_ipa_opt_pass along with...
6534 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
6535 (make_pass_ipa_free_lang_data): ...new function.
6536 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
6537 gimple_opt_pass along with...
6538 (pass_data_tsan): ...new pass_data instance and...
6539 (make_pass_tsan): ...new function.
6540 (pass_tsan_O0): Convert from a global struct to a subclass of
6541 gimple_opt_pass along with...
6542 (pass_data_tsan_O0): ...new pass_data instance and...
6543 (make_pass_tsan_O0): ...new function.
6544 * var-tracking.c (pass_variable_tracking): Convert from a global
6545 struct to a subclass of rtl_opt_pass along with...
6546 (pass_data_variable_tracking): ...new pass_data instance and...
6547 (make_pass_variable_tracking): ...new function.
6548 * web.c (pass_web): Convert from a global struct to a subclass of
6549 rtl_opt_pass along with...
6550 (pass_data_web): ...new pass_data instance and...
6551 (make_pass_web): ...new function.
6552 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
6553 declaration with that of...
6554 (make_pass_mode_switch_use): ...new function.
6555 (pass_resolve_sw_modes): Replace declaration with that of...
6556 (make_pass_resolve_sw_modes): ...new function.
6557 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
6558 from a global struct to a subclass of rtl_opt_pass along with...
6559 (pass_data_mode_switch_use): ...new pass_data instance and...
6560 (make_pass_mode_switch_use): ...new function.
6561 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
6562 from a global struct to a subclass of rtl_opt_pass along with...
6563 (pass_data_resolve_sw_modes): ...new pass_data instance and...
6564 (make_pass_resolve_sw_modes): ...new function.
6565 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
6566 struct to a subclass of rtl_opt_pass along with...
6567 (pass_data_insert_vzeroupper): ...new pass_data instance and...
6568 (make_pass_insert_vzeroupper): ...new function.
6569 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
6570 global struct to a subclass of rtl_opt_pass along with...
6571 (pass_data_work_around_errata): ...new pass_data instance and...
6572 (make_pass_work_around_errata): ...new function.
6573 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
6574 struct to a subclass of rtl_opt_pass along with...
6575 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
6576 (make_pass_mips_machine_reorg2): ...new function.
6578 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6580 * passes.c (pass_manager::operator new): New.
6582 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6584 Handwritten part of conversion of passes to C++ classes.
6586 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
6587 (toplev.o): Add dep on PASS_MANAGER_H.
6588 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
6589 of early local pases to reflect this moving from a global to a
6590 member of gcc::pass_manager.
6591 (cgraph_add_new_function): Likewise.
6592 * lto-cgraph.c (lto_output_node): Update for conversion of
6593 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
6594 * passes.c (opt_pass::clone): New.
6595 (opt_pass::gate): New.
6596 (opt_pass::execute): New.
6597 (opt_pass::opt_pass): New.
6598 (pass_manager::execute_early_local_passes): New.
6599 (pass_manager::execute_pass_mode_switching): new.
6600 (finish_optimization_passes): Convert to...
6601 (pass_manager::finish_optimization_passes): ...this.
6602 (finish_optimization_passes): Update for conversion of passes to
6604 (register_dump_files_1): Use has_gate since we cannot portably
6605 check a vtable entry against NULL.
6606 (dump_one_pass): Likewise.
6607 (ipa_write_summaries_2): Likewise.
6608 (ipa_write_optimization_summaries_1): Likewise.
6609 (ipa_read_summaries_1): Likewise.
6610 (ipa_read_optimization_summaries_1): Likewise.
6611 (execute_ipa_stmt_fixups): Likewise.
6612 (pass_manager::pass_manager): Rewrite pass-creation, invoking
6613 pass-creation functions rather than wiring up globals, and
6614 storing the results in fields of pass_manager generated using
6616 (pass_manager::dump_profile_report): Update for conversion of
6617 passes to C++ classes.
6618 (pass_manager::execute_ipa_summary_passes): Likewise.
6619 (execute_one_ipa_transform_pass): Likewise.
6620 (execute_one_pass): Use has_gate and has_execute since we cannot
6621 portably check a vtable entry against NULL.
6622 * pass_manager.h (pass_manager::finish_optimization_passes): New.
6623 (pass_manager): Use pass-instances.def to add fields for the
6624 various pass instances.
6625 * toplev.c (finalize): Update for move of
6626 finish_optimization_passes to a method of gcc::pass_manager.
6627 * toplev.h (finish_optimization_passes): Move to method of class
6629 * tree-pass.h (struct pass_data): New.
6630 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
6631 (opt_pass::gate): Convert to virtual function.
6632 (opt_pass::~opt_pass): New.
6633 (opt_pass::clone): New.
6634 (opt_pass::execute): Convert to virtual function.
6635 (opt_pass::opt_pass): New.
6636 (opt_pass::ctxt_): new.
6637 (gimple_opt_pass): Convert to subclass of opt_pass.
6638 (gimple_opt_pass::gimple_opt_pass): New.
6639 (rtl_opt_pass): Convert to subclass of opt_pass.
6640 (rtl_opt_pass::rtl_opt_pass): New.
6641 (ipa_opt_pass_d): Convert to subclass of opt_pass.
6642 (ipa_opt_pass_d::ipa_opt_pass_d): New.
6643 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
6644 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
6645 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
6646 invocation of pass_mode_switching to reflect this moving from a
6647 global to a member of gcc::pass_manager.
6648 (ix86_option_override): Rework how pass_insert_vzeroupper is
6649 added to the pass_manager to reflect autogenerated changes.
6650 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
6652 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
6654 PR rtl-optimization/57708
6655 * recog.c (peep2_find_free_register): Validate all regs in a
6658 2013-08-05 Jan Hubicka <jh@suse.cz>
6661 * cgraph.c (verify_cgraph_node): Accept local flags from other
6663 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
6664 (function_and_variable_visibility): Likewise.
6665 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
6667 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6669 * graph.c (init_graph_slim_pretty_print): Remove.
6670 (print_graph_cfg): Do not call it. Use local pretty printer.
6671 (start_graph_dump): Likewise.
6673 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6675 * gimple-pretty-print.c (buffer): Remove.
6676 (initialized): Likewise.
6677 (maybe_init_pretty_print): Likewise.
6678 (print_gimple_stmt): Do not call it. Use non-static local
6679 pretty_printer variable.
6680 (print_gimple_expr): Likewise.
6681 (print_gimple_seq): Likewise.
6682 (gimple_dump_bb): Likewise.
6684 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6686 * asan.c (asan_pp): Remove.
6687 (asan_pp_initialized): Likewise.
6688 (asan_pp_initialize): Likewise.
6689 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
6690 (asan_emit_stack_protection): Tidy. Use local pretty printer.
6691 (asan_add_global): Likewise.
6693 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
6695 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
6696 * diagnostic.h (diagnostic_flush_buffer): Adjust.
6697 * pretty-print.c (pp_formatted_text_data): Likewise.
6698 (pp_indent): Rename from pp_base_indent.
6699 (pp_format): Rename from pp_base_format.
6700 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
6701 (pp_format_verbatim): Rename from pp_base_format_verbatim.
6702 (pp_flush): Rename from pp_base_flush.
6703 (pp_set_line_maximum_length): Rename from
6704 pp_base_set_line_maximum_length.
6705 (pp_clear_output_area): Rename from pp_base_clear_output_area.
6706 (pp_set_prefix): Rename from pp_base_set_prefix.
6707 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
6708 (pp_emit_prefix): Rename from pp_base_emit_prefix.
6709 (pp_append_text): Rename from pp_base_append_text.
6710 (pp_formatted_text): Rename from pp_base_formatted_text.
6711 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
6712 (pp_remaining_character_count_for_line): Rename from
6713 pp_base_remaining_character_count_for_line.
6714 (pp_newline): Rename from pp_base_newline.
6715 (pp_character): Rename from pp_base_character.
6716 (pp_string): Rename from pp_base_string.
6717 (pp_maybe_space): Rename from pp_base_maybe_space.
6718 * asan.c (asan_pp_string): Adjust.
6719 (asan_emit_stack_protection): Likewise.
6720 (asan_add_global): Likewise.
6721 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
6722 * tree-mudflap.c (mf_varname_tree): Likewise.
6723 * tree-pretty-print.c (pp_tree_identifier): Rename from
6724 pp_base_tree_identifier.
6725 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
6726 Declare as function.
6728 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6730 * pretty-print.h (pp_bar_bar): New.
6731 (pp_ampersand_ampersand): Likewise.
6732 (pp_less_equal): Likewise.
6733 (pp_greater_equal): Likewise.
6734 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
6735 printer functions instead of pp_string or operators and punctuators.
6736 (dump_gimple_call): Likewise.
6737 (dump_gimple_omp_for): Likewise.
6738 (dump_gimple_transaction): Likewise.
6739 (dump_gimple_phi): Likewise.
6740 (pp_gimple_stmt_1): Likewise.
6741 * sched-vis.c (print_insn): Likewise.
6742 * tree-mudflap.c (mf_varname_tree): Likewise.
6743 * tree-pretty-print.c (dump_block_node): Likewise.
6744 (dump_generic_node): Likewise.
6746 2013-08-02 Jan Hubicka <jh@suse.cz>
6748 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
6750 * lto-streamer-out.c (tree_is_indexable): Results decls and
6751 parm decls are not indexable.
6752 (DFS_write_tree_body): Do not follow args and results.
6753 (hash_tree): Likewise.
6754 (output_functions): Rearrange so struct function is needed
6755 only when real body is output; be able to also ouptut abstract
6756 functions; output DECL_ARGUMENTS and DECL_RESULT.
6757 (lto_output): When not in WPA, ale store abstract functions.
6758 (write_symbol): Do not care about RESULT_DECL.
6759 (output_symbol_p): Handle correctly sbtract decls.
6760 * lto-streamer-in.c (input_function): Rearrange so struct
6761 function can be NULL at entry; allow streaming of
6762 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
6763 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
6764 sanity check during LTO.
6765 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
6766 RESULT_DECl and DECL_ARGUMENTS.
6767 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
6770 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6772 * pretty-print.h (pp_underscore): New.
6774 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
6775 printer functions instead of pp_character.
6776 (dump_binary_rhs): Likewise.
6777 (dump_ternary_rhs): Likewise.
6778 (dump_gimple_call_args): Likewise.
6779 (pp_points_to_solution): Likewise.
6780 (dump_gimple_call): Likewise.
6781 (dump_gimple_switch): Likewise.
6782 (dump_gimple_cond): Likewise.
6783 (dump_gimple_bind): Likewise.
6784 (dump_gimple_try): Likewise.
6785 (dump_gimple_omp_for): Likewise.
6786 (dump_gimple_omp_continue): Likewise.
6787 (dump_gimple_omp_single): Likewise.
6788 (dump_gimple_omp_sections): Likewise.
6789 (dump_gimple_omp_block): Likewise.
6790 (dump_gimple_omp_critical): Likewise.
6791 (dump_gimple_transaction): Likewise.
6792 (dump_gimple_asm): Likewise.
6793 (dump_gimple_phi): Likewise.
6794 (dump_gimple_omp_parallel): Likewise.
6795 (dump_gimple_omp_task): Likewise.
6796 (dump_gimple_omp_atomic_load): Likewise.
6797 (dump_gimple_omp_atomic_store): Likewise.
6798 (dump_gimple_mem_ops): Likewise.
6799 (pp_gimple_stmt_1): Likewise.
6800 (pp_cfg_jump): Likewise.
6801 (dump_implicit_edges): Likewise.
6802 (gimple_dump_bb_for_graph): Likewise.
6803 * graph.c (draw_cfg_node): Likewise.
6804 * langhooks.c (lhd_print_error_function): Likewise.
6805 * sched-vis.c (print_exp): Likewise.
6806 (print_value): Likewise.
6807 (print_pattern): Likewise.
6808 (print_insn): Likewise.
6809 (rtl_dump_bb_for_graph): Likewise.
6810 * tree-pretty-print.c (dump_function_declaration): Likewise.
6811 (dump_array_domain): Likewise.
6812 (dump_omp_clause): Likewise.
6813 (dump_location): Likewise.
6814 (dump_generic_node): Likewise.
6815 (print_struct_decl): Likewise.
6816 * diagnostic.c (diagnostic_show_locus): Use pp_space.
6818 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6820 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
6821 candidate table when replacing a candidate statement.
6822 (replace_rhs_if_not_dup): Likewise.
6823 (replace_one_candidate): Likewise.
6825 2013-08-02 Jan Hubicka <jh@suse.cz>
6826 Martin Liska <marxin.liska@gmail.com>
6828 * cgraphunit.c (add_new_function): Fix logic when adding from
6830 (assemble_thunk): Rename to ...
6831 (expand_thunk); .. this one; export; get it working with
6832 general functions; make produced gimple valid.
6833 * cgraph.h (expand_thunk): Declare.
6835 2013-08-02 Jan Hubicka <jh@suse.cz>
6837 * ipa-cp.c (gather_context_independent_values): Use
6838 ipa_get_param_move_cost.
6839 (get_replacement_map): Remove PARAM; move parameter folding
6841 (create_specialized_node): Update.
6842 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
6843 assert that we have gimple body; update move_cost.
6844 (count_formal_params): Assert that we have gimple body.
6845 (ipa_dump_param): New function.
6846 (ipa_alloc_node_params): Break out from ...
6847 (ipa_initialize_node_params): ... here.
6848 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
6849 (ipa_write_node_info): Stream move costs.
6850 (ipa_read_node_info): Read move costs.
6851 (ipa_update_after_lto_read): Do not recompute node params.
6852 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
6853 (ipa_get_param): Check we are not in WPA.
6854 (ipa_get_param_move_cost): New.
6855 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
6856 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
6857 parm numbers to be present.
6859 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6861 PR rtl-optimization/58048
6862 * lra-constraints.c (process_alt_operands): Don't check asm
6863 operand on register.
6865 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
6867 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
6868 the implied StoreLoad barrier for atomic operations if before.
6870 2013-08-02 Jan Hubicka <jh@suse.cz>
6871 Martin Liska <marxin.liska@gmail.com>
6873 * cgraph.c (cgraph_function_body_availability): Do not check
6875 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
6876 symtab_node_availability): Declare.
6877 * ipa.c (can_replace_by_local_alias): New.
6878 (function_and_variable_visibility): Use it.
6879 * symtab.c (symtab_for_node_and_aliases,
6880 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
6882 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6884 PR rtl-optimization/57963
6885 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
6886 (lra_constraints): Use them.
6888 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
6890 * config/arm/types.md (define_attr "type"): Add "load_acq"
6892 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
6894 (cortex_a53_store1): Likewise.
6896 2013-08-01 Jan Hubicka <jh@suse.cz>
6898 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
6899 partitions are not needed.
6901 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
6903 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
6904 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
6905 MAYBE_NON_Q_CLASS_P where appropriate.
6907 2013-08-01 Jan Hubicka <jh@suse.cz>
6909 * cgraph.h (release_function_body): Declare.
6910 * tree.c (free_lang_data_in_decl): Free, parameters and return values
6911 of unused delcarations.
6913 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6915 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
6916 RTL form when subtracting a constant.
6918 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6920 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
6921 Generate canonical plus rtx with negated immediate instead of minus
6923 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
6925 2013-08-01 Jan Hubicka <jh@suse.cz>
6927 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
6928 (cgraph_release_function_body): Likewise.
6929 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
6930 * cgraph.h (cgrpah_node): Rename abstract_and_needed
6931 to used_as_abstract_origin.
6932 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
6933 symbols used as abstract origins.
6934 * cgraphunit.c (analyze_functions): Update.
6935 * ipa.c (symtab_remove_unreachable_nodes): Recompute
6936 used_as_abstract_origin.
6937 * tree-inline.c (tree_function_versioning): Update
6938 used_as_abstract_origin; be ready for DECL_RESULT and
6939 DECL_ARGUMENTS to be NULL.
6941 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
6942 for abstract functions.
6943 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
6946 2013-08-01 Jan Hubicka <jh@suse.cz>
6948 * profile.c (compute_value_histograms): Fix thinko.
6950 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6952 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
6953 aarch-common-protos.h to extra_headers.
6954 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6955 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
6956 * config/aarch64/t-aarch64 (aarch-common.o): Define.
6958 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6960 * config/aarch64/aarch64.md (define_attr "type"): Delete.
6961 Include "../arm/types.md". Define "type" attribute for all patterns.
6962 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
6965 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
6967 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
6968 to support power8 load fusion.
6969 (fusion_gpr_mem_load): Likewise.
6971 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
6973 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
6974 declarations for power8 load fusion.
6975 (emit_fusion_gpr_load): Likewise.
6977 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
6978 tuning for power8, turn on fusion mode by default. Turn on sign
6979 extending fusion mode if normal fusion mode is on, and we are at
6981 (fusion_gpr_load_p): New function, return true if we can fuse an
6982 addis instruction with a dependent load to a GPR.
6983 (emit_fusion_gpr_load): Emit the instructions for power8 load
6986 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
6987 (VSX load fusion peepholes): New peepholes to fuse together an
6988 addi instruction with a VSX load instruction.
6990 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
6991 peepholes to fuse an addis instruction with a load to a GPR base
6992 register. If we are supporting sign extending fusions, convert
6993 sign extending loads to zero extending loads and add an explicit
6996 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6998 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
6999 aarch-common-protos.h to extra_headers.
7000 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7001 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7002 (arm_early_store_addr_dep): Likewise.
7003 (arm_no_early_alu_shift_dep): Likewise.
7004 (arm_no_early_alu_shift_value_dep): Likewise.
7005 (arm_no_early_mul_dep): Likewise.
7006 (arm_no_early_store_addr_dep): Likewise.
7007 (arm_mac_accumulator_is_mul_result): Likewise.
7008 (arm_mac_accumulator_is_result): Likewise.
7009 * config/arm/aarch-common.c: ... here. New file.
7010 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7012 (arm_early_store_addr_dep): Likewise.
7013 (arm_no_early_alu_shift_dep): Likewise.
7014 (arm_no_early_alu_shift_value_dep): Likewise.
7015 (arm_no_early_mul_dep): Likewise.
7016 (arm_no_early_store_addr_dep): Likewise.
7017 (arm_mac_accumulator_is_mul_result): Likewise.
7018 (arm_mac_accumulator_is_result): Likewise.
7019 * config/arm/aarch-common-protos.h: ... here. New file.
7020 * config/arm/t-arm (aarch-common.o): Define.
7022 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7024 * config/arm/arm.md: Include new file "types.md".
7025 (define_attr "type"): Move from here to ...
7026 (define_attr "mul32"): Likewise.
7027 (define_attr "mul64"): Likewise.
7028 * config/arm/types.md: ... here. New file.
7030 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7032 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7033 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7035 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7037 * gen-pass-instances.awk: Fix offset of substr().
7039 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7041 * Makefile.in (pass-instances.def): New.
7042 (passes.o): Replace dependency on passes.def with one on
7045 * gen-pass-instances.awk: New.
7047 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7048 rather than passes.def, updating local definition of NEXT_PASS
7049 macro to add an extra NUM parameter (currently unused).
7051 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7053 * Makefile.in (PASS_MANAGER_H): New.
7054 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7055 (passes.o): Likewise.
7056 (statistics.o): Likewise.
7057 (cgraphunit.o): Likewise.
7058 (context.o): Depend on PASS_MANAGER_H.
7060 * pass_manager.h: New.
7062 * cgraphunit.c (cgraph_add_new_function): Update for moves
7063 of globals to fields of pass_manager.
7064 (analyze_function): Likewise.
7065 (expand_function): Likewise.
7066 (ipa_passes): Likewise.
7067 (compile): Likewise.
7069 * context.c (context::context): New.
7070 * context.h (context::context): New.
7071 (context::get_passes): New.
7072 (context::passes_): New.
7074 * lto-cgraph.c (input_node): Update for moves of globals to
7075 fields of pass_manager.
7077 * passes.c (all_passes): Remove, in favor of a field of the
7078 same name within the new class pass_manager.
7079 (all_small_ipa_passes): Likewise.
7080 (all_lowering_passes): Likewise.
7081 (all_regular_ipa_passes): Likewise.
7082 (all_late_ipa_passes): Likewise.
7083 (all_lto_gen_passes): Likewise.
7084 (passes_by_id): Likewise.
7085 (passes_by_id_size): Likewise.
7086 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7087 the new class pass_manager.
7088 (set_pass_for_id): Convert to...
7089 (pass_manager::set_pass_for_id): ...method.
7090 (get_pass_for_id): Convert to...
7091 (pass_manager::get_pass_for_id): ...method.
7092 (register_one_dump_file): Move body of implementation into...
7093 (pass_manager::register_one_dump_file): ...here.
7094 (register_dump_files_1): Convert to...
7095 (pass_manager::register_dump_files_1): ...method.
7096 (register_dump_files): Convert to...
7097 (pass_manager::register_dump_files): ...method.
7098 (create_pass_tab): Update for moves of globals to fields of
7100 (dump_passes): Move body of implementation into...
7101 (pass_manager::dump_passes): ...here.
7102 (register_pass): Move body of implementation into...
7103 (pass_manager::register_pass): ...here.
7104 (init_optimization_passes): Convert into...
7105 (pass_manager::pass_manager): ...constructor for new
7106 pass_manager class, and initialize the pass_lists array.
7107 (check_profile_consistency): Update for moves of globals to
7108 fields of pass_manager.
7109 (dump_profile_report): Move body of implementation into...
7110 (pass_manager::dump_profile_report): ...here.
7111 (ipa_write_summaries_1): Update for moves of pass lists from
7112 being globals to fields of pass_manager.
7113 (ipa_write_optimization_summaries): Likewise.
7114 (ipa_read_summaries): Likewise.
7115 (ipa_read_optimization_summaries): Likewise.
7116 (execute_all_ipa_stmt_fixups): Likewise.
7118 * statistics.c (statistics_fini): Update for moves of globals to
7119 fields of pass_manager.
7121 * toplev.c (general_init): Replace call to
7122 init_optimization_passes with construction of the pass_manager
7125 * tree-pass.h (all_passes): Remove, in favor of a field of the
7126 same name within the new class pass_manager.
7127 (all_small_ipa_passes): Likewise.
7128 (all_lowering_passes): Likewise.
7129 (all_regular_ipa_passes): Likewise.
7130 (all_lto_gen_passes): Likewise.
7131 (all_late_ipa_passes): Likewise.
7132 (passes_by_id): Likewise.
7133 (passes_by_id_size): Likewise.
7134 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7135 the new class pass_manager.
7136 (get_pass_for_id): Remove.
7138 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7140 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7143 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7145 * arm.md (mulhi3): New expand pattern.
7147 2013-07-30 Jan Hubicka <jh@suse.cz>
7148 Martin Liska <marxin.liska@gmail.com>
7150 * profile.c (compute_value_histograms): Do not ICE when
7151 there is mismatch only on some counters.
7153 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7155 PR rtl-optimization/57637
7156 * function.c (move_insn_for_shrink_wrap): Also check the
7157 GEN set of the LIVE problem for the liveness analysis
7158 if it exists, otherwise give up.
7160 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7162 PR tree-optimization/57993
7163 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7164 replaced statement in the candidate table.
7165 (phi_add_costs): Return infinite cost when the hidden basis does
7166 not dominate all phis on which the candidate is dependent.
7167 (replace_one_candidate): Record replaced statement in the
7170 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7172 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7173 (ashlv2si3): New expander.
7174 (*ashlv2si3_i): New define_insn_and_split.
7175 * predicates.md (float_operation): Allow patterns with three
7178 PR rtl-optimization/58021
7179 * mode-switching.c (create_pre_exit): Always split off preceding
7180 insns if we are not at the basic block head.
7182 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7184 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7185 (UCLIBC_DYNAMIC_LINKER): New macro.
7186 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7188 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7189 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7191 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7192 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7193 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7194 for SF and DF modes. Use ieee_quad_format for TF mode.
7195 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7196 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7197 (mips_option_override): Handle `-mnan=legacy'.
7198 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7199 `-mabs=2008' and `-mnan=2008'.
7200 (OPTION_DEFAULT_SPECS): Add "nan" default.
7201 (ASM_SPEC): Handle `-mnan='.
7202 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7203 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7204 comment accordingly.
7205 (neg<mode>2): Likewise.
7206 * config/mips/mips.opt (mabs, mnan): New options.
7207 * doc/install.texi (Configuration): Document `--with-nan=' option.
7208 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7210 (MIPS Options): Document them.
7211 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7212 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7213 * configure: Regenerate.
7214 * config.in: Regenerate.
7216 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7218 * config/i386/i386.md (float post-reload splitters): Do not check
7219 for subregs of SSE registers.
7221 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7222 H.J. Lu <hongjiu.lu@intel.com>
7226 * config/i386/i386.md (post-reload splitter
7227 to avoid partial SSE reg dependency stalls): New pattern.
7229 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7231 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7233 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7235 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7236 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7238 2013-07-26 Cary Coutant <ccoutant@google.com>
7240 * dwarf2out.c (die_checksum_ordered): Don't include template
7241 instantiations in signature.
7242 (is_template_parameter): New function.
7243 (is_template_instantiation): New function.
7244 (generate_skeleton_bottom_up): Don't include template instantiations
7246 (generate_skeleton): Likewise.
7247 (break_out_comdat_types): Move recursive call to break out nested
7249 (prune_unused_types_mark_generic_parms_dies): Call
7250 is_template_parameter.
7252 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7254 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7255 uses vector registers.
7256 * config/aarch64/iterators.md: Add attributes rtn and vas.
7258 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7259 Richard Earnshaw <richard.earnshaw@arm.com>
7261 * combine.c (simplify_comparison): Re-canonicalize operands
7263 * config/arm/arm.md (movcond_addsi): New splitter.
7265 2013-07-25 Sterling Augustine <saugustine@google.com>
7267 * dwarf2out.c (size_of_pubnames): Move code to...
7268 (include_pubname_in_output): ...here. New.
7269 (want_pubnames): Rearrange.
7270 (output_pubnames): Call include_pubname_in_output. Move assertion.
7272 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7274 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7276 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7279 * doc/extend.texi: Remove obsolete builtins. Fix
7280 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7282 2013-07-25 Jan Hubicka <jh@suse.cz>
7284 * cgraph.c (release_function_body): Break out from ...
7285 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7287 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7288 old_tree in the map.
7289 (create_specialized_node): Update.
7290 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7292 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7293 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7294 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7295 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7298 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7300 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7301 addsi3_carryin_alt2_<optab>): Correct output template.
7303 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7305 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7306 Adjust for arm_restrict_it.
7307 Remove trailing whitespace.
7309 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7311 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7312 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7314 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7316 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7318 PR rtl-optimization/57960
7319 * lra-constraints.c (process_alt_operands): Use the right mode
7320 when checking strict_low.
7322 2013-07-25 Jan Hubicka <jh@suse.cz>
7324 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7325 * cgraph.c (cgraph_remove_node): Do not release function body
7326 when in cgraph streaming.
7327 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7328 in other partitions are not considered reachable; fix handling of
7331 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7333 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7335 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7340 * config/arm/arm.md ("*sibcall_insn): Replace use of
7341 Ss with US. Adjust output for v5 and v4t.
7342 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7344 * config/arm/constraints.md ("Ss"): Rename to US.
7346 2013-07-25 Terry Guo <terry.guo@arm.com>
7348 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7349 shift_add/shift_sub0/shift_sub1 RTXs.
7351 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7352 Anton Blanchard <anton@au1.ibm.com>
7354 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7355 (altivec_vpks<VI_char>ss): Likewise.
7356 (altivec_vpks<VI_char>us): Likewise.
7357 (altivec_vpku<VI_char>us): Likewise.
7358 (altivec_vpku<VI_char>um): Likewise.
7360 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7362 Introduce context class.
7364 * Makefile.in (CONTEXT_H): New.
7365 (OBJS): Add context.o.
7366 (toplev.o): Add CONTEXT_H to dependencies.
7369 * toplev.c (general_init): Create the singleton gcc::context instance.
7375 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
7377 PR rtl-optimization/57968
7378 * mode-switching.c (create_pre_exit): Allow instructions that
7379 don't set a return register to need a non-exit mode.
7381 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7382 Anton Blanchard <anton@au1.ibm.com>
7384 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
7385 operands to vperm for little endian.
7386 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
7387 of lvsl to create the control mask for a vperm for little endian.
7389 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7390 Anton Blanchard <anton@au1.ibm.com>
7392 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
7393 two operands for little-endian.
7395 2013-07-23 Steve Ellcey <sellcey@mips.com>
7397 * config/mips/mips.c (mips_case_values_threshold): New.
7398 (TARGET_CASE_VALUES_THRESHOLD): Define.
7400 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7401 Anton Blanchard <anton@au1.ibm.com>
7403 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
7404 selection of field for vector splat in little endian mode.
7406 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7408 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
7409 expanders to rs6000.md.
7410 (ior<mode>3): Likewise.
7411 (and<mode>3): Likewise.
7412 (one_cmpl<mode>2): Likewise.
7413 (nor<mode>3): Likewise.
7414 (andc<mode>3): Likewise.
7415 (eqv<mode>3): Likewise.
7416 (nand<mode>3): Likewise.
7417 (orc<mode>3): Likewise.
7419 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
7422 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
7423 to split multi-word logical operations.
7424 (rs6000_split_logical_di): Likewise.
7425 (rs6000_split_logical): Likewise.
7427 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
7428 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
7429 and allow TImode operations in 32-bit.
7430 (vsx_and<mode>3_64bit): Likewise.
7431 (vsx_ior<mode>3_32bit): Likewise.
7432 (vsx_ior<mode>3_64bit): Likewise.
7433 (vsx_xor<mode>3_32bit): Likewise.
7434 (vsx_xor<mode>3_64bit): Likewise.
7435 (vsx_one_cmpl<mode>2_32bit): Likewise.
7436 (vsx_one_cmpl<mode>2_64bit): Likewise.
7437 (vsx_nor<mode>3_32bit): Likewise.
7438 (vsx_nor<mode>3_64bit): Likewise.
7439 (vsx_andc<mode>3_32bit): Likewise.
7440 (vsx_andc<mode>3_64bit): Likewise.
7441 (vsx_eqv<mode>3_32bit): Likewise.
7442 (vsx_eqv<mode>3_64bit): Likewise.
7443 (vsx_nand<mode>3_32bit): Likewise.
7444 (vsx_nand<mode>3_64bit): Likewise.
7445 (vsx_orc<mode>3_32bit): Likewise.
7446 (vsx_orc<mode>3_64bit): Likewise.
7448 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
7449 logical types in GPRs.
7451 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
7452 logical insns to rs6000.md, and allow TImode operations in
7454 (altivec_ior<mode>3): Likewise.
7455 (altivec_xor<mode>3): Likewise.
7456 (altivec_one_cmpl<mode>2): Likewise.
7457 (altivec_nor<mode>3): Likewise.
7458 (altivec_andc<mode>3): Likewise.
7460 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
7461 attributes for moving the 128-bit logical operations into
7463 (BOOL_REGS_OUTPUT): Likewise.
7464 (BOOL_REGS_OP1): Likewise.
7465 (BOOL_REGS_OP2): Likewise.
7466 (BOOL_REGS_UNARY): Likewise.
7467 (BOOL_REGS_AND_CR0): Likewise.
7468 (one_cmpl<mode>2): Add support for DI logical operations on
7469 32-bit, splitting the operations to 32-bit.
7473 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
7474 changes to combine the 32/64-bit code, allow logical operations on
7475 TI mode in 32-bit, and to use similar match_operator patterns like
7476 scalar mode uses. Combine the Altivec and VSX code for logical
7477 operations, and move it here.
7478 (ior<mode>3, 128-bit types): Likewise.
7479 (xor<mode>3, 128-bit types): Likewise.
7480 (one_cmpl<mode>3, 128-bit types): Likewise.
7481 (nor<mode>3, 128-bit types): Likewise.
7482 (andc<mode>3, 128-bit types): Likewise.
7483 (eqv<mode>3, 128-bit types): Likewise.
7484 (nand<mode>3, 128-bit types): Likewise.
7485 (orc<mode>3, 128-bit types): Likewise.
7486 (and<mode>3_internal): Likewise.
7487 (bool<mode>3_internal): Likewise.
7488 (boolc<mode>3_internal1): Likewise.
7489 (boolc<mode>3_internal2): Likewise.
7490 (boolcc<mode>3_internal1): Likewise.
7491 (boolcc<mode>3_internal2): Likewise.
7492 (eqv<mode>3_internal1): Likewise.
7493 (eqv<mode>3_internal2): Likewise.
7494 (one_cmpl1<mode>3_internal): Likewise.
7496 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7498 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7499 Rename flag_stack_usage to flag_stack_usage_info.
7501 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7503 * config/microblaze/sync.md: New file.
7504 * config/microblaze/microblaze.md: Include sync.md
7505 * config/microblaze/microblaze.c: Add print_operand 'y'.
7506 * config/microblaze/constraints.md: Add memory_contraint
7507 'Q' which is a single register.
7509 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
7511 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
7513 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7515 * reload.c (find_reloads): Exit loop once we find this operand
7516 cannot be reloaded somehow for this alternative.
7518 * reload.c (find_reloads): Exit loop once we find a hard register.
7520 * rtlanal.c (computed_jump_p): Exit loop once we find label
7523 * i386.c (ix86_pad_returns): Exit loop after setting replace.
7525 * cfgloopmanip.c (remove_path): Exit loop after setting
7528 * gensupport.c (subst_dup): Avoid loop if code is not
7529 MATCH_DUP nor MATCH_OP_DUP.
7531 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
7533 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
7535 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7537 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
7538 true for SP_REGNUM if mode == ptr_mode.
7539 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
7540 with value R0_REGNUM + 31.
7542 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7544 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
7545 pad pointer-typed argument downward.
7547 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7549 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
7550 and __ILP32__ when the ILP32 model is in use.
7552 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7554 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
7555 (aarch64_load_symref_appropriately): In the case of
7556 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
7557 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
7558 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
7559 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
7560 if 'mode' doesn't equal to 'ptr_mode'.
7561 (aarch64_output_mi_thunk): Add an assertion on the alignment of
7562 'vcall_offset'; change to call aarch64_emit_move differently depending
7563 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
7564 to calculate the upper bound of 'vcall_offset'.
7565 (aarch64_cannot_force_const_mem): Change to also return true if
7567 (aarch64_legitimize_reload_address): In the case of large
7568 displacements, add new local variable 'xmode' and an assertion
7569 based on it; change to use 'xmode' to generate the new rtx and
7571 (aarch64_asm_trampoline_template): Change to generate the template
7572 differently depending on TARGET_ILP32 or not; change to use
7573 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
7574 (aarch64_trampoline_size): Removed.
7575 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
7576 and replace immediate literals with it. Change to use 'ptr_mode'
7577 instead of 'DImode' and call convert_memory_address if the mode
7578 of 'fnaddr' doesn't equal to 'ptr_mode'.
7579 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
7581 (aarch64_elf_asm_destructor): Likewise.
7582 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
7583 on TARGET_ILP32 instead of aarch64_trampoline_size.
7584 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
7585 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
7586 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
7587 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
7588 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
7589 (storewb_pair<GPI:mode>_<P:mode>): ... this.
7590 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
7591 depending on the value of 'mode'.
7592 (add_losym_<mode>): New.
7593 (ldr_got_small_<mode>): New, based on ldr_got_small.
7594 (ldr_got_small): Remove.
7595 (ldr_got_small_sidi): New.
7596 * config/aarch64/iterators.md (P): New.
7597 (PTR): Change to 'ptr_mode' in the condition.
7599 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7601 * config.gcc (aarch64*-*-*): Support --with-abi.
7602 (aarch64*-*-elf): Support --with-multilib-list.
7603 (aarch64*-*-linux*): Likewise.
7604 (supported_defaults): Add abi to aarch64*-*-*.
7605 * configure.ac: Mention AArch64 for --with-multilib-list.
7606 * configure: Re-generated.
7607 * config/aarch64/biarchilp32.h: New file.
7608 * config/aarch64/biarchlp64.h: New file.
7609 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
7611 (MULTILIB_DEFAULTS): Ditto.
7612 (DRIVER_SELF_SPECS): Ditto.
7613 (ASM_SPEC): Update to also substitute -mabi.
7614 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
7615 file whose name depends on -mabi= and -mbig-endian.
7616 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
7618 (POINTER_SIZE): New define.
7619 (POINTERS_EXTEND_UNSIGNED): Ditto.
7620 (enum aarch64_abi_type): New enumeration tag.
7621 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
7622 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
7623 (TARGET_ILP32): New define.
7624 * config/aarch64/aarch64.opt (mabi): New.
7626 (ilp32, lp64): New values for -mabi.
7627 * config/aarch64/t-aarch64 (comma): New define.
7628 (MULTILIB_OPTIONS): Ditto.
7629 (MULTILIB_DIRNAMES): Ditto.
7630 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
7631 * doc/invoke.texi: Document -mabi for AArch64.
7633 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
7635 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
7637 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7638 Anton Blanchard <anton@au1.ibm.com>
7640 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
7641 endianness when selecting field to splat.
7643 2013-07-22 Eric Christopher <echristo@gmail.com>
7645 * dwarf2out.c (die_odr_checksum): New function to use
7646 CHECKSUM_ macros and ULEB128 for DIE tag.
7647 (generate_type_signature): Use.
7649 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
7651 * config.gcc (sparc*-*-*): Accept leon3 processor.
7652 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
7653 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
7654 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
7655 * config/sparc/sparc.opt (enum processor_type): Add leon3.
7656 (mfix-ut699): Adjust comment.
7657 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
7658 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
7659 (CPP_CPU_SPEC): Likewise.
7660 (ASM_CPU_SPEC): Likewise.
7661 * config/sparc/sparc.c (leon3_cost): New constant.
7662 (sparc_option_override): Add leon3 support.
7663 (mem_ref): New function.
7664 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
7665 (sparc_do_work_around_errata): Look into the instruction in the delay
7666 slot and adjust accordingly. Add fix for the data cache nullify issues
7667 of the UT699. Change insertion position for the NOP.
7668 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
7669 (leon3_load): New reservation.
7670 (leon_store): Bump latency to 2.
7671 (grfpu): New automaton.
7672 (grfpu_alu): New unit.
7673 (grfpu_ds): Likewise.
7674 (leon_fp_alu): Adjust.
7675 (leon_fp_mult): Delete.
7676 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
7677 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
7678 * config/sparc/sparc.md (cpu): Add leon3.
7679 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
7681 (atomic_test_and_set): Likewise.
7684 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
7686 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
7687 default for R5900 targets.
7688 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
7689 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
7690 * config/mips/mips.c (mips_option_override): Report an error for
7691 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
7692 for -march=r5900 -mhard-float.
7694 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7696 * df-problems.c (can_move_insns_across): Exit loop once we
7697 find a non-fixed, non-global register.
7699 * ipa-pure-const.c (propagate_nothrow): Exit loop after
7702 * omega.c (omega_eliminate_red): Break after setting red_found.
7703 (omega_problem_has_red_equations): Similarly after setting found.
7704 (omega_query_variable): Similarly after setting coupled.
7706 2013-07-22 Marek Polacek <polacek@redhat.com>
7708 * gimplify.c: Don't include gimple.h twice.
7710 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7712 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
7713 instead of TARGET_THUMB1.
7714 (Pz): New constraint.
7715 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
7717 (compare_negsi_si): Likewise.
7718 (compare_addsi2_op0): Likewise.
7719 (compare_addsi2_op1): Likewise.
7720 (addsi3_carryin_<optab>): Likewise.
7721 (addsi3_carryin_alt2_<optab>): Likewise.
7722 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
7723 for arm_restrict_it.
7724 (subsi3_carryin): Likewise.
7725 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
7726 (minmax_arithsi): Disable for arm_restrict_it.
7727 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
7728 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
7729 (satsi_<SAT:code>_shift): Likewise.
7730 (arm_shiftsi3): Add alternative for 16-bit encoding.
7731 (arm32_movhf): Disable for arm_restrict_it.
7732 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
7733 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
7735 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
7737 * config/arm/arm.md (attribute "insn"): Delete.
7738 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
7739 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
7740 (not_shiftsi): Update for attribute change.
7741 (not_shiftsi_compare0): Likewise.
7742 (not_shiftsi_compare0_scratch): Likewise.
7743 (arm_one_cmplsi2): Likewise.
7744 (thumb1_one_cmplsi2): Likewise.
7745 (notsi_compare0): Likewise.
7746 (notsi_compare0_scratch): Likewise.
7747 (thumb1_movdi_insn): Likewise.
7748 (arm_movsi_insn): Likewise.
7749 (movhi_insn_arch4): Likewise.
7750 (movhi_bytes): Likewise.
7751 (arm_movqi_insn): Likewise.
7752 (thumb1_movqi_insn): Likewise.
7753 (arm32_movhf): Likewise.
7754 (thumb1_movhf): Likewise.
7755 (arm_movsf_soft_insn): Likewise.
7756 (thumb1_movsf_insn): Likewise.
7757 (thumb_movdf_insn): Likewise.
7758 (movsicc_insn): Likewise.
7759 (movsfcc_soft_insn): Likewise.
7760 (and_scc): Likewise.
7761 (cond_move): Likewise.
7762 (if_move_not): Likewise.
7763 (if_not_move): Likewise.
7764 (if_shift_move): Likewise.
7765 (if_move_shift): Likewise.
7766 (if_shift_shift): Likewise.
7767 (if_not_arith): Likewise.
7768 (if_arith_not): Likewise.
7769 (cond_move_not): Likewise.
7770 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
7771 (neon_mov<mode>): Likewise.
7772 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
7773 (thumb2_movsi_vfp): Likewise.
7774 (movsf_vfp): Likewise.
7775 (thumb2_movsf_vfp): Likewise.
7776 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7778 (cortexa7_older_only): Likewise.
7779 (cortexa7_younger): Likewise.
7780 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
7781 (1020alu_shift_op): Likewise.
7782 (1020alu_shift_reg_op): Likewise.
7783 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
7784 (alu_shift_op): Likewise.
7785 (alu_shift_reg_op): Likewise.
7786 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
7787 (11_alu_shift_op): Likewise.
7788 (11_alu_shift_reg_op): Likewise.
7789 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
7790 (9_alu_shift_reg_op): Likewise.
7791 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7793 (cortex_a15_alu_shift): Likewise.
7794 (cortex_a15_alu_shift_reg): Likewise.
7795 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
7796 (cortex_a5_alu_shift): Likewise.
7797 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
7799 (cortex_a53_alu_shift): Likewise.
7800 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7802 (cortex_a7_alu_reg): Likewise.
7803 (cortex_a7_alu_shift): Likewise.
7804 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7805 (cortex_a8_alu_shift): Likewise.
7806 (cortex_a8_alu_shift_reg): Likewise.
7807 (cortex_a8_mov): Likewise.
7808 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7809 (cortex_a9_dp_shift): Likewise.
7810 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
7811 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
7812 (cortex_r4_mov): Likewise.
7813 (cortex_r4_alu_shift): Likewise.
7814 (cortex_r4_alu_shift_reg): Likewise.
7815 * config/arm/fa526.md (526_alu_op): Update for attribute change.
7816 (526_alu_shift_op): Likewise.
7817 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
7818 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
7819 (626te_alu_shift_op): Likewise.
7820 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
7821 (726te_alu_op): Likewise.
7822 (726te_alu_shift_op): Likewise.
7823 (726te_alu_shift_reg_op): Likewise.
7824 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
7825 (mp626_alu_shift_op): Likewise.
7826 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
7827 (pj4_alu_e1_conds): Likewise.
7828 (pj4_alu): Likewise.
7829 (pj4_alu_conds): Likewise.
7830 (pj4_shift): Likewise.
7831 (pj4_shift_conds): Likewise.
7832 (pj4_alu_shift): Likewise.
7833 (pj4_alu_shift_conds): Likewise.
7835 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7837 * config/arm/predicates.md (shiftable_operator_strict_it):
7839 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
7840 Disable cond_exec version for arm_restrict_it.
7841 (thumb2_smaxsi3): Convert to generate cond_exec.
7842 (thumb2_sminsi3): Likewise.
7843 (thumb32_umaxsi3): Likewise.
7844 (thumb2_uminsi3): Likewise.
7845 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
7846 (thumb2_neg_abssi2): Likewise.
7847 (thumb2_mov_scc): Add alternative for 16-bit encoding.
7848 (thumb2_movsicc_insn): Adjust alternatives.
7849 (thumb2_mov_negscc): Disable for arm_restrict_it.
7850 (thumb2_mov_negscc_strict_it): New pattern.
7851 (thumb2_mov_notscc_strict_it): New pattern.
7852 (thumb2_mov_notscc): Disable for arm_restrict_it.
7853 (thumb2_ior_scc): Likewise.
7854 (thumb2_ior_scc_strict_it): New pattern.
7855 (thumb2_cond_move): Adjust for arm_restrict_it.
7856 (thumb2_cond_arith): Disable for arm_restrict_it.
7857 (thumb2_cond_arith_strict_it): New pattern.
7858 (thumb2_cond_sub): Adjust for arm_restrict_it.
7859 (thumb2_movcond): Likewise.
7860 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
7861 (thumb2_zero_extendhisi2_v6): Likewise.
7862 (thumb2_zero_extendqisi2_v6): Likewise.
7863 (orsi_notsi_si): Likewise.
7864 (orsi_not_shiftsi_si): Likewise.
7866 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
7868 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
7869 instruction sequence is 1 byte shorter.
7871 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
7873 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
7874 it is not needed after split.
7876 2013-07-20 Iain Sandoe <iain@codesourcery.com>
7879 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
7880 second label for nonlocal goto receivers. Don't output pic base labels
7881 unless we're producing PIC; mark that action unreachable().
7882 (ix86_save_reg): If the function contains a nonlocal label, save the
7884 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
7885 * config/darwin.c (emitted_pic_label_num): New GTY.
7886 (update_pic_label_number_if_needed): New.
7887 (machopic_output_function_base_name): Adjust for nonlocal receiver
7889 (machopic_should_output_picbase_label): New.
7890 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
7891 (nonlocal_goto_receiver): New insn and split.
7893 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
7895 * config/aarch64/aarch64-builtins.c
7896 (aarch64_fold_builtin): Fold abs in all modes.
7897 * config/aarch64/aarch64-simd-builtins.def
7898 (abs): Enable for all modes.
7899 * config/aarch64/arm_neon.h
7900 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
7901 (vabs_f64): Add missing intrinsic.
7903 2013-07-19 Ian Bolton <ian.bolton@arm.com>
7905 * config/aarch64/arm_neon.h (vabs_s64): New function
7907 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
7910 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
7911 * config/avr/avr.md (adjust_len): Add `round'.
7912 * config/avr/avr-protos.h (avr_out_round): New prototype.
7913 (avr_out_plus): Add `out_label' argument.
7914 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
7915 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
7916 Handle the case where `insn' is just a pattern.
7917 (avr_out_bitop): Handle the case where `insn' is just a pattern.
7918 (avr_out_round): New function.
7919 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
7921 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
7923 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7924 Add check for flag_stack_usage to handle -fstack-usage support
7926 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
7928 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
7929 interaction for new Power8 flags and VSX.
7931 2013-07-18 Sriraman Tallam <tmsriram@google.com>
7934 * tree-inline.c (expand_call_inline): Emit errors during
7935 early_inlining only if optimization is not turned on.
7937 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7941 * passes.c (init_optimization_passes): Move the construction of
7942 the pass hierarchy into a new passes.def file.
7944 * Makefile.in (passes.o): Add dependency on passes.def.
7946 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7948 * passes.c (init_optimization_passes): Introduce macros for
7949 constructing the tree of passes (INSERT_PASSES_AFTER,
7950 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
7951 TERMINATE_PASS_LIST).
7953 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
7954 Wei Mi <wmi@google.com>
7956 PR rtl-optimization/57878
7957 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
7959 (reload_pseudo_compare_func): Check nregs first for reload
7962 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7964 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
7966 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
7968 * read-rtl.c (validate_const_int): Once an invalid character is
7969 seen, quit the loop.
7971 * gengtype.c (write_roots): Similarly once we find the "deletable"
7972 or "if_marked" option.
7974 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7976 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
7977 "xtab" and "sat". Move value "clz" from here to ...
7978 (attriubte "type"): ... here.
7979 (satsi_<SAT:code>): Delete "insn" attribute.
7980 (satsi_<SAT:code>_shift): Likewise.
7981 (arm_zero_extendqisi2addsi): Likewise.
7982 (arm_extendqisi2addsi): Likewise.
7983 (clzsi2): Update for attribute changes.
7984 (rbitsi2): Likewise.
7985 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
7987 (arm_usatsihi): Likewise.
7988 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7990 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7992 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
7993 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
7994 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
7995 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
7996 in alphabetical order.
7997 (attribute "core_cycles"): Update for attribute changes.
7998 (arm_addsi3): Likewise.
7999 (addsi3_compare0): Likewise.
8000 (addsi3_compare0_scratch): Likewise.
8001 (addsi3_compare_op1): Likewise.
8002 (addsi3_compare_op2): Likewise.
8003 (compare_addsi2_op0): Likewise.
8004 (compare_addsi2_op1): Likewise.
8005 (addsi3_carryin_shift_<optab>): Likewise.
8006 (subsi3_carryin_shift): Likewise.
8007 (rsbsi3_carryin_shift): Likewise.
8008 (arm_subsi3_insn): Likewise.
8009 (subsi3_compare0): Likewise.
8010 (subsi3_compare): Likewise.
8011 (arm_andsi3_insn): Likewise.
8012 (thumb1_andsi3_insn): Likewise.
8013 (andsi3_compare0): Likewise.
8014 (andsi3_compare0_scratch): Likewise.
8015 (zeroextractsi_compare0_scratch
8016 (andsi_not_shiftsi_si): Likewise.
8017 (iorsi3_insn): Likewise.
8018 (iorsi3_compare0): Likewise.
8019 (iorsi3_compare0_scratch): Likewise.
8020 (arm_xorsi3): Likewise.
8021 (thumb1_xorsi3_insn): Likewise.
8022 (xorsi3_compare0): Likewise.
8023 (xorsi3_compare0_scratch): Likewise.
8024 (satsi_<SAT:code>_shift): Likewise.
8026 (arm_shiftsi3): Likewise.
8027 (shiftsi3_compare0): Likewise.
8028 (not_shiftsi): Likewise.
8029 (not_shiftsi_compare0): Likewise.
8030 (not_shiftsi_compare0_scratch): Likewise.
8031 (arm_one_cmplsi2): Likewise.
8032 (thumb_one_complsi2): Likewise.
8033 (notsi_compare0): Likewise.
8034 (notsi_compare0_scratch): Likewise.
8035 (thumb1_zero_extendhisi2): Likewise.
8036 (arm_zero_extendhisi2): Likewise.
8037 (arm_zero_extendhisi2_v6): Likewise.
8038 (arm_zero_extendhisi2addsi): Likewise.
8039 (thumb1_zero_extendqisi2): Likewise.
8040 (thumb1_zero_extendqisi2_v6): Likewise.
8041 (arm_zero_extendqisi2): Likewise.
8042 (arm_zero_extendqisi2_v6): Likewise.
8043 (arm_zero_extendqisi2addsi): Likewise.
8044 (thumb1_extendhisi2): Likewise.
8045 (arm_extendhisi2): Likewise.
8046 (arm_extendhisi2_v6): Likewise.
8047 (arm_extendqisi): Likewise.
8048 (arm_extendqisi_v6): Likewise.
8049 (arm_extendqisi2addsi): Likewise.
8050 (thumb1_extendqisi2): Likewise.
8051 (thumb1_movdi_insn): Likewise.
8052 (arm_movsi_insn): Likewise.
8053 (movsi_compare0): Likewise.
8054 (movhi_insn_arch4): Likewise.
8055 (movhi_bytes): Likewise.
8056 (arm_movqi_insn): Likewise.
8057 (thumb1_movqi_insn): Likewise.
8058 (arm32_movhf): Likewise.
8059 (thumb1_movhf): Likewise.
8060 (arm_movsf_soft_insn): Likewise.
8061 (thumb1_movsf_insn): Likewise.
8062 (movdf_soft_insn): Likewise.
8063 (thumb_movdf_insn): Likewise.
8064 (arm_cmpsi_insn): Likewise.
8065 (cmpsi_shiftsi): Likewise.
8066 (cmpsi_shiftsi_swp): Likewise.
8067 (arm_cmpsi_negshiftsi_si): Likewise.
8068 (movsicc_insn): Likewise.
8069 (movsfcc_soft_insn): Likewise.
8070 (arith_shiftsi): Likewise.
8071 (arith_shiftsi_compare0
8072 (arith_shiftsi_compare0_scratch
8073 (sub_shiftsi): Likewise.
8074 (sub_shiftsi_compare0
8075 (sub_shiftsi_compare0_scratch
8076 (and_scc): Likewise.
8077 (cond_move): Likewise.
8078 (if_plus_move): Likewise.
8079 (if_move_plus): Likewise.
8080 (if_move_not): Likewise.
8081 (if_not_move): Likewise.
8082 (if_shift_move): Likewise.
8083 (if_move_shift): Likewise.
8084 (if_shift_shift): Likewise.
8085 (if_not_arith): Likewise.
8086 (if_arith_not): Likewise.
8087 (cond_move_not): Likewise.
8088 (thumb1_ashlsi3): Set type attribute.
8089 (thumb1_ashrsi3): Likewise.
8090 (thumb1_lshrsi3): Likewise.
8091 (thumb1_rotrsi3): Likewise.
8092 (shiftsi3_compare0_scratch): Likewise.
8093 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8094 (neon_mov<mode>): Likewise.
8095 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8097 (thumb2_movsi_insn): Likewise.
8098 (thumb2_cmpsi_neg_shiftsi): Likewise.
8099 (thumb2_extendqisi_v6): Likewise.
8100 (thumb2_zero_extendhisi2_v6): Likewise.
8101 (thumb2_zero_extendqisi2_v6): Likewise.
8102 (thumb2_shiftsi3_short): Likewise.
8103 (thumb2_addsi3_compare0_scratch): Likewise.
8104 (orsi_not_shiftsi_si): Likewise.
8105 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8106 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8108 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8109 (1020alu_shift_op): Likewise.
8110 (1020alu_shift_reg_op): Likewise.
8111 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8112 (alu_shift_op): Likewise.
8113 (alu_shift_reg_op): Likewise.
8114 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8115 (11_alu_shift_op): Likewise.
8116 (11_alu_shift_reg_op): Likewise.
8117 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8118 (9_alu_shift_reg_op): Likewise.
8119 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8121 (cortex_a15_alu_shift): Likewise.
8122 (cortex_a15_alu_shift_reg): Likewise.
8123 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8125 (cortex_a5_alu_shift): Likewise.
8126 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8128 (cortex_a53_alu_shift): Likewise.
8129 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8131 (cortex_a7_alu_reg): Likewise.
8132 (cortex_a7_alu_shift): Likewise.
8133 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8135 (cortex_a8_alu_shift): Likewise.
8136 (cortex_a8_alu_shift_reg): Likewise.
8137 (cortex_a8_mov): Likewise.
8138 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8139 (cortex_a9_dp_shift): Likewise.
8140 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8142 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8144 (cortex_r4_mov): Likewise.
8145 (cortex_r4_alu_shift): Likewise.
8146 (cortex_r4_alu_shift_reg): Likewise.
8147 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8148 (526_alu_shift_op): Likewise.
8149 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8150 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8151 (626te_alu_shift_op): Likewise.
8152 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8153 (726te_alu_op): Likewise.
8154 (726te_alu_shift_op): Likewise.
8155 (726te_alu_shift_reg_op): Likewise.
8156 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8157 (mp626_alu_shift_op): Likewise.
8158 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8159 (pj4_alu_e1_conds): Likewise.
8160 (pj4_alu): Likewise.
8161 (pj4_alu_conds): Likewise.
8162 (pj4_shift): Likewise.
8163 (pj4_shift_conds): Likewise.
8164 (pj4_alu_shift): Likewise.
8165 (pj4_alu_shift_conds): Likewise.
8166 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8168 (cortexa7_older_only): Likewise.
8169 (cortexa7_younger): Likewise.
8171 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8173 * ipa-pure-const.c (generate_summary): Rename to...
8174 (pure_const_generate_summary): ... this.
8176 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8178 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8180 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8183 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8186 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8188 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8189 enabled without -march=zEC12.
8190 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8193 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8195 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8196 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8197 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8198 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8199 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8200 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8201 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8203 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8204 and ISA_HAS_NMADD3_NMSUB3.
8205 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8206 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8207 (nmsub4<mode>, nmsub3<mode>): Likewise.
8208 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8210 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8212 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8213 TARGET_MIPS5400 checking.
8215 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8216 Peter Bergner <bergner@vnet.ibm.com>
8218 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8219 registers in the comment.
8220 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8221 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8222 rather than FIRST_PSEUDO_REGISTERS.
8224 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8226 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8227 enable extra ISA flags with TARGET_HTM.
8229 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8231 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8234 2013-07-15 Cong Hou <congh@google.com>
8236 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8237 in compare function for sorting.
8239 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8241 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8242 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8243 * config/rs6000/rs6000.opt: Add -mhtm option.
8244 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8245 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8246 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8247 __HTM__ if the HTM instructions are available.
8248 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8249 htm_spr_reg_operand): New define_predicates.
8250 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8251 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8253 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8254 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8255 HTM builtin functions.
8256 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8257 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8258 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8259 (rs6000_builtin_mask_calculate): Likewise.
8260 (rs6000_option_override_internal): Likewise.
8261 (bdesc_htm): Add new HTM builtin support.
8262 (htm_spr_num): New function.
8263 (htm_spr_regno): Likewise.
8264 (rs6000_htm_spr_icode): Likewise.
8265 (htm_expand_builtin): Likewise.
8266 (htm_init_builtins): Likewise.
8267 (rs6000_expand_builtin): Add support for HTM builtin functions.
8268 (rs6000_init_builtins): Likewise.
8269 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8271 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8272 (TARGET_HTM, MASK_HTM): Define macros.
8273 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8274 (FIXED_REGISTERS): Likewise.
8275 (CALL_USED_REGISTERS): Likewise.
8276 (CALL_REALLY_USED_REGISTERS): Likewise.
8277 (REG_ALLOC_ORDER): Likewise.
8278 (enum reg_class): Likewise.
8279 (REG_CLASS_NAMES): Likewise.
8280 (REG_CLASS_CONTENTS): Likewise.
8281 (REGISTER_NAMES): Likewise.
8282 (ADDITIONAL_REGISTER_NAMES): Likewise.
8283 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8284 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8285 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8286 * config/rs6000/htm.md: New file.
8287 * config/rs6000/htmintrin.h: New file.
8288 * config/rs6000/htmxlintrin.h: New file.
8290 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8292 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8293 Define SYMBOL_TINY_GOT, update comment.
8294 * config/aarch64/aarch64.c
8295 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8296 (aarch64_expand_mov_immediate): Likewise.
8297 (aarch64_print_operand): Likewise.
8298 (aarch64_classify_symbol): Likewise.
8299 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8300 (ldr_got_tiny): Define.
8302 2013-07-13 Tobias Grosser <tobias@grosser.es>
8304 PR tree-optimization/54094
8305 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8306 scheduling dimension for the parallelism check from the polyhedral
8307 information in the AST.
8308 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8311 2013-07-13 Jason Merrill <jason@redhat.com>
8313 * print-tree.c (debug_vec_tree): Use debug_raw.
8314 (debug_raw (vec<tree, va_gc> &)): New.
8315 (debug_raw (vec<tree, va_gc> *)): New.
8316 * tree.h: Declare them.
8318 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8320 * ifcvt.c (ifcvt_after_combine): New static variable.
8321 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8323 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8324 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8325 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8327 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8329 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8331 2013-07-12 Michael Matz <matz@suse.de>
8334 * convert.c (convert_to_real): Reject non-float inner types.
8336 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8338 * config/aarch64/aarch64-protos.h
8339 (aarch64_simd_immediate_valid_for_move): Remove.
8340 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8341 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8343 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8345 2013-07-11 Steve Ellcey <sellcey@mips.com>
8347 * config/mips/mips.c (mips_conditional_register_usage): Do not
8348 use t[0-7] registers in MIPS16 mode when optimizing for size.
8350 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8352 * config/i386/i386.c (dispatch_function_versions): Fix array
8353 indexing of function_version_info to match actual_versions.
8355 2013-07-11 Teresa Johnson <tejohnson@google.com>
8357 * vec.h (struct va_gc): Move release out-of-line.
8358 (va_gc::release): Call ggc_free on released vec.
8360 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8362 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8363 Require GOT register as additional operand in UNSPEC.
8364 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8365 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8366 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8367 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8368 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8369 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8370 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8372 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
8375 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
8376 name seen by assembler/linker rather if available.
8378 2013-07-11 Andreas Schwab <schwab@suse.de>
8380 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
8382 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
8384 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
8386 2013-07-10 Joseph Myers <joseph@codesourcery.com>
8388 * doc/tm.texi.in: Move hook documentation to ....
8389 * target.def: ... here.
8391 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
8393 * doc/tm.texi: Regenerate.
8395 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
8398 * doc/invoke.texi: Document -Wconditionally-supported.
8400 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8403 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
8406 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8409 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
8410 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
8411 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
8412 Remove duplicate devices.
8413 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
8414 * config/avr/t-multilib: Regenerate.
8415 * config/avr/avr-tables.opt: Regenerate.
8416 * doc/avr-mmcu.texi: Regenerate.
8418 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8421 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
8423 2013-07-10 Graham Stott <graham.stott@btinternet.com>
8425 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
8426 the cost of MULT when optimizing for size.
8428 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8430 * config/cr16/cr16-protos.h: Don't include target.h.
8432 2013-07-09 Joseph Myers <joseph@codesourcery.com>
8434 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
8435 adjust register size for TDmode and TFmode for VSX registers.
8437 2013-07-08 Kai Tietz <ktietz@redhat.com>
8440 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
8441 hook_bool_const_tree_true.
8443 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8445 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
8446 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
8447 * config/s390/s390.md: Define FPR*_REGNUM constants.
8448 Fix FPR2_REGNUM constant (18 -> 17).
8449 ("*trunc<BFP:mode><DFP_ALL:mode>2")
8450 ("*trunc<DFP_ALL:mode><BFP:mode>2")
8451 ("trunc<BFP:mode><DFP_ALL:mode>2")
8452 ("trunc<DFP_ALL:mode><BFP:mode>2")
8453 ("*extend<BFP:mode><DFP_ALL:mode>2")
8454 ("*extend<DFP_ALL:mode><BFP:mode>2")
8455 ("extend<BFP:mode><DFP_ALL:mode>2")
8456 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
8459 2013-07-08 Graham Stott <graham.stott@btinternet.com>
8461 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
8463 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8465 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
8466 and cfun_fpr_bit_p to cfun_fpr_save_p.
8467 (s390_frame_area, s390_register_info, s390_frame_info)
8468 (s390_emit_prologue, s390_emit_epilogue)
8469 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
8471 * config/s390/s390.h: Define *_REGNUM macros for floating point
8474 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
8476 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
8478 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
8480 PR rtl-optimization/57786
8481 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
8482 and break out of the loop when it is set to false.
8484 2013-07-08 Jakub Jelinek <jakub@redhat.com>
8487 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
8488 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
8489 (const_int 63)) 0)).
8490 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
8491 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
8492 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
8494 PR rtl-optimization/57829
8495 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
8496 mask bits outside of mode are just sign-extension from mode to HWI.
8498 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8500 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
8501 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
8502 adjust_address instead of change_address to keep info about alignment.
8503 (emit_strmov): Remove.
8504 (emit_memmov): New function.
8505 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
8506 (expand_movmem_epilogue): Likewise and return updated rtx for
8508 (expand_constant_movmem_prologue): Likewise and return updated rtx for
8509 destination and source.
8510 (decide_alignment): Refactor, handle vector_loop.
8511 (ix86_expand_movmem): Likewise.
8512 (ix86_expand_setmem): Likewise.
8513 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
8515 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
8517 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
8518 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
8520 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8522 * config/i386/sse.md (sse_movlhps): Change alternative 3
8523 of operand 2 to "m".
8525 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8528 * config/i386/sse.md (iptr): New mode attribute.
8529 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
8530 (<sse>_vm<plusminus_insn><mode>3): Ditto.
8531 (<sse>_vmmul<mode>3): Ditto.
8532 (<sse>_vmdiv<mode>3): Ditto.
8533 (sse_vmrcpv4sf2): Ditto.
8534 (<sse>_vmsqrt<mode>2): Ditto.
8535 (sse_vmrsqrtv4sf2): Ditto.
8536 (<sse>_vm<code><mode>3): Ditto.
8537 (avx_vmcmp<mode>3): Ditto.
8538 (<sse>_vmmaskcmp<mode>3): Ditto.
8539 (<sse>_comi): Ditto.
8540 (<sse>_ucomi): Ditto.
8541 (*xop_vmfrcz_<mode>): Ditto.
8542 (*fmai_fmadd_<mode>): Ditto.
8543 (*fmai_fmsub_<mode>): Ditto.
8544 (*fmai_fnmadd_<mode>): Ditto.
8545 (*fmai_fnmsub_<mode>): Ditto.
8546 (*fma4i_vmfmadd_<mode>): Ditto.
8547 (*fma4i_vmfmsub_<mode>): Ditto.
8548 (*fma4i_vmfnmadd_<mode>): Ditto.
8549 (*fma4i_vmfnmsub_<mode>): Ditto.
8550 (*xop_vmfrcz_<mode>): Ditto.
8551 (sse_cvtps2pi): Ditto.
8552 (sse_cvttps2pi): Ditto.
8553 (sse_cvtss2si): Ditto.
8554 (sse_cvtss2si_2): Ditto.
8555 (sse_cvtss2siq_2): Ditto.
8556 (sse_cvttss2si): Ditto.
8557 (sse_cvttss2siq): Ditto.
8558 (sse_cvtsd2si): Ditto.
8559 (sse_cvtsd2si_2): Ditto.
8560 (sse_cvtsd2siq_2): Ditto.
8561 (sse_cvttsd2si): Ditto.
8562 (sse_cvttsd2siq): Ditto.
8563 (sse_cvtsd2ss): Ditto.
8564 (sse_cvtss2sd): Ditto.
8565 (avx2_pbroadcast<mode>): Ditto.
8566 (avx2_pbroadcast<mode>_1): Ditto.
8567 (*avx_vperm_broadcast_v4sf): Ditto.
8569 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
8570 (sse_movlhps): Ditto.
8571 (sse_storehps): Ditto.
8572 (sse_loadhps): Ditto.
8573 (sse_storelps): Ditto.
8574 (sse_loadlps): Ditto.
8575 (*vec_concatv4sf): Ditto.
8576 (*vec_interleave_highv2df): Ditto.
8577 (*vec_interleave_lowv2df): Ditto.
8578 (*vec_extractv2df_1_sse): Ditto.
8579 (*vec_extractv2df_0_sse): Ditto.
8580 (sse2_storelpd): Ditto.
8581 (sse2_loadlpd): Ditto.
8582 (sse2_movsd): Ditto.
8583 (*vec_concatv4si): Ditto.
8584 (vec_concatv2di): Ditto.
8586 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
8587 for Intel asm dialect.
8588 (mmx_punpcklwd): Ditto.
8589 (mmx_punpckldq): Ditto.
8591 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
8592 for intel assembler dialect.
8594 2013-07-06 Jakub Jelinek <jakub@redhat.com>
8597 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
8598 for BUILT_IN_C{LZ,LRSB}*.
8599 * tree.h (CASE_INT_FN): Add FN##IMAX case.
8600 * tree-vrp.c (extract_range_basic): Handle
8601 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
8602 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
8603 fall thru to code calling set_value*.
8604 * builtins.c (expand_builtin): Remove *IMAX cases.
8605 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
8606 if width is bigger than 2*HWI.
8608 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
8610 PR rtl-optimization/55342
8611 * lra-int.h (lra_subreg_reload_pseudos): New.
8612 * lra.c: Add undoing optional reloads to the block diagram.
8613 (lra_subreg_reload_pseudos): New.
8614 (lra_optional_reload_pseudos): Change comments.
8615 (lra): Init and clear lra_subreg_reload_pseudos. Clear
8616 lra_optional_reload_pseudos after undo transformations.
8617 * lra-assigns.c (pseudo_prefix_title): New.
8618 (lra_setup_reg_renumber): Use it.
8619 (spill_for): Ditto. Check subreg reload pseudos too.
8620 (assign_by_spills): Consider subreg reload pseudos too.
8621 * lra-constraints.c (simplify_operand_subreg): Use
8622 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
8623 (curr_insn_transform): Recognize and do optional reloads.
8624 (undo_optional_reloads): New.
8625 (lra_undo_inheritance): Call undo_optional_reloads.
8627 2013-07-05 Thomas Quinot <quinot@adacore.com>
8629 * tree-complex.c (expand_complex_operations_1): Fix typo.
8631 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
8633 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
8634 (tune_params): New member 'const vec_costs'.
8635 * config/aarch64/aarch64.c (generic_vector_cost): New.
8636 (generic_tunings): New member 'generic_vector_cost'.
8637 (aarch64_builtin_vectorization_cost): New.
8638 (aarch64_add_stmt_cost): New.
8639 (TARGET_VECTORIZE_ADD_STMT_COST): New.
8640 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
8642 2013-07-03 Jakub Jelinek <jakub@redhat.com>
8645 * config/i386/predicates.md (vsib_address_operand): Disallow
8646 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
8648 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
8651 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
8652 expand_builtin_setjmp_receiver.
8653 (expand_label): Adjust, call expand_builtin_setjmp_receiver
8654 with NULL for the label parameter.
8655 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
8656 the frame-pointer. Adjust comments.
8657 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
8658 only if LABEL is non-NULL.
8660 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
8662 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
8663 (ARM_ABI_AAPCS64): Ditto.
8665 (ARM_DEFAULT_ABI): Ditto.
8667 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
8669 * config/aarch64/aarch64-builtins.c
8670 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
8671 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
8673 * config/aarch64/aarch64-simd.md
8674 (aarch64_ld1<VALL:mode>): New.
8675 (aarch64_st1<VALL:mode>): Likewise.
8676 * config/aarch64/arm_neon.h
8677 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
8679 2013-07-02 Sriraman Tallam <tmsriram@google.com>
8681 * config/i386/i386.c (gate_insert_vzeroupper): Check if
8683 (ix86_option_override_internal):Turn on all -mavx target flags by
8684 default as they are dependent on AVX anyway.
8686 2013-07-02 Cary Coutant <ccoutant@google.com>
8688 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
8690 (loc_checksum_ordered): Likewise.
8691 (hash_loc_operands): Remove inline keyword.
8693 2013-07-02 Jakub Jelinek <jakub@redhat.com>
8695 PR tree-optimization/57741
8696 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
8697 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
8698 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
8699 Allow REAL_CST step_exprs if flag_associative_math.
8700 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
8702 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8704 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
8706 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8708 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
8710 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8712 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
8714 (iorsi3_insn): Likewise.
8715 (arm_xorsi3): Likewise.
8717 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
8719 * arm.md (attribute "wtype"): Delete. Move attribute values from here
8721 (attribute "type"): ... here, and prefix with "wmmx_".
8722 (attribute "core_cycles"): Update for attribute changes.
8723 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
8724 (tbcstv4hi): Likewise.
8725 (tbcstv2si): Likewise.
8726 (iwmmxt_iordi3): Likewise.
8727 (iwmmxt_xordi3): Likewise.
8728 (iwmmxt_anddi3): Likewise.
8729 (iwmmxt_nanddi3): Likewise.
8730 (iwmmxt_arm_movdi): Likewise.
8731 (iwmmxt_movsi_insn): Likewise.
8732 (mov<mode>_internal): Likewise.
8733 (and<mode>3_iwmmxt): Likewise.
8734 (ior<mode>3_iwmmxt): Likewise.
8735 (xor<mode>3_iwmmxt): Likewise.
8736 (add<mode>3_iwmmxt): Likewise.
8737 (ssaddv8qi3): Likewise.
8738 (ssaddv4hi3): Likewise.
8739 (ssaddv2si3): Likewise.
8740 (usaddv8qi3): Likewise.
8741 (usaddv4hi3): Likewise.
8742 (usaddv2si3): Likewise.
8743 (sub<mode>3_iwmmxt): Likewise.
8744 (sssubv8qi3): Likewise.
8745 (sssubv4hi3): Likewise.
8746 (sssubv2si3): Likewise.
8747 (ussubv8qi3): Likewise.
8748 (ussubv4hi3): Likewise.
8749 (ussubv2si3): Likewise.
8750 (mulv4hi3_iwmmxt): Likewise.
8751 (smulv4hi3_highpart): Likewise.
8752 (umulv4hi3_highpart): Likewise.
8753 (iwmmxt_wmacs): Likewise.
8754 (iwmmxt_wmacsz): Likewise.
8755 (iwmmxt_wmacu): Likewise.
8756 (iwmmxt_wmacuz): Likewise.
8757 (iwmmxt_clrdi): Likewise.
8758 (iwmmxt_clrv8qi): Likewise.
8759 (iwmmxt_clr4hi): Likewise.
8760 (iwmmxt_clr2si): Likewise.
8761 (iwmmxt_uavgrndv8qi3): Likewise.
8762 (iwmmxt_uavgrndv4hi3): Likewise.
8763 (iwmmxt_uavgv8qi3): Likewise.
8764 (iwmmxt_uavgv4hi3): Likewise.
8765 (iwmmxt_tinsrb): Likewise.
8766 (iwmmxt_tinsrh): Likewise.
8767 (iwmmxt_tinsrw): Likewise.
8768 (iwmmxt_textrmub): Likewise.
8769 (iwmmxt_textrmsb): Likewise.
8770 (iwmmxt_textrmuh): Likewise.
8771 (iwmmxt_textrmsh): Likewise.
8772 (iwmmxt_textrmw): Likewise.
8773 (iwmxxt_wshufh): Likewise.
8774 (eqv8qi3): Likewise.
8775 (eqv4hi3): Likewise.
8776 (eqv2si3): Likewise.
8777 (gtuv8qi3): Likewise.
8778 (gtuv4hi3): Likewise.
8779 (gtuv2si3): Likewise.
8780 (gtv8qi3): Likewise.
8781 (gtv4hi3): Likewise.
8782 (gtv2si3): Likewise.
8783 (smax<mode>3_iwmmxt): Likewise.
8784 (umax<mode>3_iwmmxt): Likewise.
8785 (smin<mode>3_iwmmxt): Likewise.
8786 (umin<mode>3_iwmmxt): Likewise.
8787 (iwmmxt_wpackhss): Likewise.
8788 (iwmmxt_wpackwss): Likewise.
8789 (iwmmxt_wpackdss): Likewise.
8790 (iwmmxt_wpackhus): Likewise.
8791 (iwmmxt_wpackwus): Likewise.
8792 (iwmmxt_wpackdus): Likewise.
8793 (iwmmxt_wunpckihb): Likewise.
8794 (iwmmxt_wunpckihh): Likewise.
8795 (iwmmxt_wunpckihw): Likewise.
8796 (iwmmxt_wunpckilb): Likewise.
8797 (iwmmxt_wunpckilh): Likewise.
8798 (iwmmxt_wunpckilw): Likewise.
8799 (iwmmxt_wunpckehub): Likewise.
8800 (iwmmxt_wunpckehuh): Likewise.
8801 (iwmmxt_wunpckehuw): Likewise.
8802 (iwmmxt_wunpckehsb): Likewise.
8803 (iwmmxt_wunpckehsh): Likewise.
8804 (iwmmxt_wunpckehsw): Likewise.
8805 (iwmmxt_wunpckelub): Likewise.
8806 (iwmmxt_wunpckeluh): Likewise.
8807 (iwmmxt_wunpckeluw): Likewise.
8808 (iwmmxt_wunpckelsb): Likewise.
8809 (iwmmxt_wunpckelsh): Likewise.
8810 (iwmmxt_wunpckelsw): Likewise.
8811 (ror<mode>3): Likewise.
8812 (ashr<mode>3_iwmmxt): Likewise.
8813 (lshr<mode>3_iwmmxt): Likewise.
8814 (ashl<mode>3_iwmmxt): Likewise.
8815 (ror<mode>3_di): Likewise.
8816 (ashr<mode>3_di): Likewise.
8817 (lshr<mode>3_di): Likewise.
8818 (ashl<mode>3_di): Likewise.
8819 (iwmmxt_wmadds): Likewise.
8820 (iwmmxt_wmaddu): Likewise.
8821 (iwmmxt_tmia): Likewise.
8822 (iwmmxt_tmiaph): Likewise.
8823 (iwmmxt_tmiabb): Likewise.
8824 (iwmmxt_tmiatb): Likewise.
8825 (iwmmxt_tmiabt): Likewise.
8826 (iwmmxt_tmiatt): Likewise.
8827 (iwmmxt_tmovmskb): Likewise.
8828 (iwmmxt_tmovmskh): Likewise.
8829 (iwmmxt_tmovmskw): Likewise.
8830 (iwmmxt_waccb): Likewise.
8831 (iwmmxt_wacch): Likewise.
8832 (iwmmxt_waccw): Likewise.
8833 (iwmmxt_waligni): Likewise.
8834 (iwmmxt_walignr): Likewise.
8835 (iwmmxt_walignr0): Likewise.
8836 (iwmmxt_walignr1): Likewise.
8837 (iwmmxt_walignr2): Likewise.
8838 (iwmmxt_walignr3): Likewise.
8839 (iwmmxt_wsadb): Likewise.
8840 (iwmmxt_wsadh): Likewise.
8841 (iwmmxt_wsadbz): Likewise.
8842 (iwmmxt_wsadhz): Likewise.
8843 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
8844 (iwmmxt_wabsdiffb): Likewise.
8845 (iwmmxt_wabsdiffh): Likewise.
8846 (iwmmxt_wabsdiffw): Likewise.
8847 (iwmmxt_waddsubhx): Likewise
8848 (iwmmxt_wsubaddhx): Likewise.
8849 (addc<mode>3): Likewise.
8850 (iwmmxt_avg4): Likewise.
8851 (iwmmxt_avg4r): Likewise.
8852 (iwmmxt_wmaddsx): Likewise.
8853 (iwmmxt_wmaddux): Likewise.
8854 (iwmmxt_wmaddsn): Likewise.
8855 (iwmmxt_wmaddun): Likewise.
8856 (iwmmxt_wmulwsm): Likewise.
8857 (iwmmxt_wmulwum): Likewise.
8858 (iwmmxt_wmulsmr): Likewise.
8859 (iwmmxt_wmulumr): Likewise.
8860 (iwmmxt_wmulwsmr): Likewise.
8861 (iwmmxt_wmulwumr): Likewise.
8862 (iwmmxt_wmulwl): Likewise.
8863 (iwmmxt_wqmulm): Likewise.
8864 (iwmmxt_wqmulwm): Likewise.
8865 (iwmmxt_wqmulmr): Likewise.
8866 (iwmmxt_wqmulwmr): Likewise.
8867 (iwmmxt_waddbhusm): Likewise.
8868 (iwmmxt_waddbhusl): Likewise.
8869 (iwmmxt_wqmiabb): Likewise.
8870 (iwmmxt_wqmiabt): Likewise.
8871 (iwmmxt_wqmiatb): Likewise.
8872 (iwmmxt_wqmiatt): Likewise.
8873 (iwmmxt_wqmiabbn): Likewise.
8874 (iwmmxt_wqmiabtn): Likewise.
8875 (iwmmxt_wqmiatbn): Likewise.
8876 (iwmmxt_wqmiattn): Likewise.
8877 (iwmmxt_wmiabb): Likewise.
8878 (iwmmxt_wmiabt): Likewise.
8879 (iwmmxt_wmiatb): Likewise.
8880 (iwmmxt_wmiatt): Likewise.
8881 (iwmmxt_wmiabbn): Likewise.
8882 (iwmmxt_wmiabtn): Likewise.
8883 (iwmmxt_wmiatbn): Likewise.
8884 (iwmmxt_wmiattn): Likewise.
8885 (iwmmxt_wmiawbb): Likewise.
8886 (iwmmxt_wmiawbt): Likewise.
8887 (iwmmxt_wmiawtb): Likewise.
8888 (iwmmxt_wmiawtt): Likewise.
8889 (iwmmxt_wmiawbbn): Likewise.
8890 (iwmmxt_wmiawbtn): Likewise.
8891 (iwmmxt_wmiawtbn): Likewise.
8892 (iwmmxt_wmiawttn): Likewise.
8893 (iwmmxt_wmerge): Likewise.
8894 (iwmmxt_tandc<mode>3): Likewise.
8895 (iwmmxt_torc<mode>3): Likewise.
8896 (iwmmxt_torvsc<mode>3): Likewise.
8897 (iwmmxt_textrc<mode>3): Likewise.
8898 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
8899 (wmmxt_pack): Likewise.
8900 (wmmxt_mult_c1): Likewise.
8901 (wmmxt_mult_c2): Likewise.
8902 (wmmxt_alu_c1): Likewise.
8903 (wmmxt_alu_c2): Likewise.
8904 (wmmxt_alu_c3): Likewise.
8905 (wmmxt_transfer_c1): Likewise.
8906 (wmmxt_transfer_c2): Likewise.
8907 (wmmxt_transfer_c3): Likewise.
8908 (marvell_f_iwmmxt_wstr): Likewise.
8909 (marvell_f_iwmmxt_wldr): Likewise.
8911 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
8913 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
8915 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8918 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8919 * lra-constraints.c (need_for_split_p): Check call used hard regs
8920 living through calls.
8922 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
8923 call used regs for call insn.
8925 2013-06-28 Jakub Jelinek <jakub@redhat.com>
8928 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
8929 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
8931 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8933 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
8935 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8937 * lra-constraints.c (need_for_split_p): Check call used hard regs
8938 living through calls.
8940 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
8943 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
8944 to tie with any other modes. Eliminate Altivec vector mode tests,
8945 since these are a subset of ALTIVEC or VSX vector modes. Simplify
8946 code, to return 0 if testing MODE2 for a condition, if we've
8947 already tested MODE1 for the same condition.
8949 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8951 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
8954 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8956 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8957 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
8959 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8961 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
8963 (aarch64_symbolic_constant_p): Remove.
8964 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
8965 static. Fix line length and white space.
8966 (aarch64_symbolic_constant_p): Remove.
8967 * config/aarch64/predicates.md (aarch64_valid_symref):
8968 Use aarch64_classify_symbol_expression.
8970 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8972 * config/arm/constraints.md (Ts): New constraint.
8973 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
8975 (compare_scc): Use "Ts" constraint for operand 0.
8976 (ior_scc_scc): Likewise.
8977 (and_scc_scc): Likewise.
8978 (and_scc_scc_nodom): Likewise.
8979 (ior_scc_scc_cmp): Likewise for operand 7.
8980 (and_scc_scc_cmp): Likewise.
8981 * config/arm/thumb2.md (thumb2_movsi_insn):
8982 Add alternatives for 16-bit encodings.
8983 (thumb2_movhi_insn): Likewise.
8984 (thumb2_movsicc_insn): Likewise.
8985 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
8986 (thumb2_negscc): Use "Ts" constraint.
8987 Move mvn instruction outside cond_exec block.
8988 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
8989 for 16-bit encodings.
8991 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8993 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
8995 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
8996 (mulsi3subsi): Likewise.
8997 (mulsidi3adddi): Likewise.
8998 (mulsidi3_v6): Likewise.
8999 (umulsidi3_v6): Likewise.
9000 (umulsidi3adddi_v6): Likewise.
9001 (smulsi3_highpart_v6): Likewise.
9002 (umulsi3_highpart_v6): Likewise.
9003 (mulhisi3tb): Likewise.
9004 (mulhisi3bt): Likewise.
9005 (mulhisi3tt): Likewise.
9006 (maddhisi4): Likewise.
9007 (maddhisi4tb): Likewise.
9008 (maddhisi4tt): Likewise.
9009 (maddhidi4): Likewise.
9010 (maddhidi4tb): Likewise.
9011 (maddhidi4tt): Likewise.
9012 (zeroextractsi_compare0_scratch): Likewise.
9013 (insv_zero): Likewise.
9014 (insv_t2): Likewise.
9015 (anddi_notzesidi_di): Likewise.
9016 (anddi_notsesidi_di): Likewise.
9017 (andsi_notsi_si): Likewise.
9018 (iordi_zesidi_di): Likewise.
9019 (xordi_zesidi_di): Likewise.
9020 (andsi_iorsi3_notsi): Likewise.
9022 (smax_m1): Likewise.
9024 (not_shiftsi): Likewise.
9025 (unaligned_loadsi): Likewise.
9026 (unaligned_loadhis): Likewise.
9027 (unaligned_loadhiu): Likewise.
9028 (unaligned_storesi): Likewise.
9029 (unaligned_storehi): Likewise.
9030 (extv_reg): Likewise.
9031 (extzv_t2): Likewise.
9033 (udivsi3): Likewise.
9034 (arm_zero_extendhisi2addsi): Likewise.
9035 (arm_zero_extendqisi2addsi): Likewise.
9036 (compareqi_eq0): Likewise.
9037 (arm_extendhisi2_v6): Likewise.
9038 (arm_extendqisi2addsi): Likewise.
9039 (arm_movt): Likewise.
9040 (thumb2_ldrd): Likewise.
9041 (thumb2_ldrd_base): Likewise.
9042 (thumb2_ldrd_base_neg): Likewise.
9043 (thumb2_strd): Likewise.
9044 (thumb2_strd_base): Likewise.
9045 (thumb2_strd_base_neg): Likewise.
9046 (arm_negsi2): Add alternative for 16-bit encoding.
9047 (arm_one_cmplsi2): Likewise.
9049 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9051 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9052 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9053 (movdfcc): Likewise.
9054 * config/arm/vfp.md (*thumb2_movsf_vfp):
9055 Disable predication for arm_restrict_it.
9056 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9057 (*thumb2_movdfcc_vfp): Likewise.
9058 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9059 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9060 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9061 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9062 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9063 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9064 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9065 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9066 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9067 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9068 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9069 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9070 Disable predication for arm_restrict_it.
9072 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9074 * config/i386/bmiintrin.h (_bextr_u32): New.
9075 (_bextr_u64): Ditto.
9077 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9079 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9080 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9081 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9082 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9083 * config/mips/n32-elf.h: ...this new file.
9085 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9088 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9089 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9091 2013-06-27 Catherine Moore <clm@codesourcery.com>
9093 * config/mips/mips-tables.opt: Regenerate.
9094 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9095 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9096 * doc/invoke.texi: Add -m14kc.
9098 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9101 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9102 constraints of operand 1 and 2.
9105 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9106 to match RTL canonicalization. Swap predicates and
9107 constraints of operand 1 and 2.
9109 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9111 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9112 Process OP_INOUT regs for splitting too.
9114 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9116 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9117 decl before the loop, initialize to NULL.
9118 (vectorizable_load): Initialize ptr_incr to NULL.
9120 2013-06-27 Martin Jambor <mjambor@suse.cz>
9123 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9124 * ipa-ref.c (ipa_maybe_record_reference): New function.
9125 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9126 * ipa-cp.c (create_specialized_node): Record potential references from
9128 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9130 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9132 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9133 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9134 'mode' to force_reg.
9135 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9136 (aarch64_expand_mov_immediate): Likewise.
9138 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9140 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9141 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9143 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9145 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9146 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9147 (struct machine_function): Add tbegin_p.
9148 (s390_canonicalize_comparison): Fold CC mode compares to
9149 conditional jump if possible.
9150 (s390_emit_jump): Return the emitted jump.
9151 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9152 Handle CCRAWmode compares.
9153 (s390_option_override): Default to -mhtm if available.
9154 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9155 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9156 FPRs instead of df_regs_ever_live_p.
9157 (s390_optimize_nonescaping_tx): New function.
9158 (s390_init_frame_layout): Extend clobbered_regs array to cover
9160 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9161 (s390_expand_tbegin): New function.
9162 (enum s390_builtin): New enum definition.
9163 (code_for_builtin): New array definition.
9164 (s390_init_builtins): New function.
9165 (s390_expand_builtin): New function.
9166 (TARGET_INIT_BUILTINS): Define.
9167 (TARGET_EXPAND_BUILTIN): Define.
9168 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9169 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9170 (s390_alc_comparison): Likewise.
9171 * config/s390/s390-modes.def: Add CCRAWmode.
9172 * config/s390/s390.h (processor_flags): Add PF_TX.
9173 (TARGET_CPU_HTM): Define macro.
9174 (TARGET_HTM): Define macro.
9175 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9176 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9177 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9178 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9179 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9180 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9181 ("*cc_to_int"): Move up.
9182 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9183 constants other than 0.
9184 ("*ccraw_to_int"): New insn and splitter definition.
9185 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9186 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9187 ("tx_assist"): New expander.
9188 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9189 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9190 * config/s390/s390.opt: Add -mhtm option.
9191 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9192 * config/s390/htmxlintrin.h: New file.
9193 * config/s390/htmintrin.h: New file.
9194 * config/s390/s390intrin.h: New file.
9195 * doc/extend.texi: Document htm builtins.
9196 * config.gcc: Add the new header files to extra_headers.
9198 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9200 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9201 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9203 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9204 Pat Haugen <pthaugen@us.ibm.com>
9205 Peter Bergner <bergner@vnet.ibm.com>
9207 * config/rs6000/power8.md: New.
9208 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9209 setting for power8 entry.
9210 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9211 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9212 test for Power4/Power5 only.
9213 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9215 (force_new_group): Adjust comment.
9216 * config/rs6000/rs6000.md: Include power8.md.
9218 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9220 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9221 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9223 (tune_params): Update comment.
9224 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9225 (arm_max_conditional_execute): New function.
9226 (thumb2_final_prescan_insn): Use max_insn_skipped and
9227 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9229 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9231 PR tree-optimization/57705
9232 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9233 SSA_NAME step, provided that it is not defined inside the loop.
9234 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9235 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9237 2013-06-25 Martin Jambor <mjambor@suse.cz>
9240 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9241 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9243 (ipa_note_param_call): Initialize member_ptr flag.
9244 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9245 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9246 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9247 (ipa_read_indirect_edge_info): Likewise.
9249 2013-06-25 Richard Biener <rguenther@suse.de>
9252 * passes.c (init_optimization_passes): Move pass_fold_builtins
9253 and pass_dce earlier with -Og.
9255 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9257 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9258 <BIT_FIELD_REF>: Remove trailing TAB.
9259 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9262 2013-06-24 Martin Jambor <mjambor@suse.cz>
9264 PR tree-optimization/57358
9265 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9266 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9267 (ipa_analyze_params_uses): Generate pessimistic info when true.
9269 2013-06-24 Martin Jambor <mjambor@suse.cz>
9271 PR tree-optimization/57539
9272 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9273 global.inlined_to of the new node to it. All callers changed.
9274 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9275 inlining_into, pass it to cgraph_clone_node.
9276 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9277 ipa_free_edge_args_substructures.
9278 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9279 rdesc linked list. Do not assert rdesc edges have inlined caller.
9280 Assert we have found an rdesc in the rdesc list.
9282 2013-06-24 Richard Biener <rguenther@suse.de>
9284 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9285 (pointer_set_lookup): Declare.
9286 (class pointer_map): New template class implementing a
9287 generic pointer to T map.
9288 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9289 pointer_map<T>::contains, pointer_map<T>::insert,
9290 pointer_map<T>::traverse): New functions.
9291 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9292 (pointer_set_lookup): New function.
9293 (pointer_set_contains): Use pointer_set_lookup.
9294 (pointer_set_insert): Likewise.
9295 (insert_aux): Remove.
9296 (struct pointer_map_t): Embed a pointer_set_t.
9297 (pointer_map_create): Adjust.
9298 (pointer_map_destroy): Likewise.
9299 (pointer_map_contains): Likewise.
9300 (pointer_map_insert): Likewise.
9301 (pointer_map_traverse): Likewise.
9302 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9303 pointer_map<unsigned> instead of a pointer_map_t.
9304 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9305 (streamer_tree_cache_lookup): Likewise.
9306 (streamer_tree_cache_create): Likewise.
9307 (streamer_tree_cache_delete): Likewise.
9308 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9309 pointer_map<unsigned> instead of a pointer_map_t.
9310 (lto_init_tree_ref_encoder): Adjust.
9311 (lto_destroy_tree_ref_encoder): Likewise.
9312 * lto-section-out.c (lto_output_decl_index): Likewise.
9313 (lto_record_function_out_decl_state): Likewise.
9314 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9316 2013-06-24 Richard Biener <rguenther@suse.de>
9318 PR tree-optimization/57488
9319 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9321 2013-06-24 Alan Modra <amodra@gmail.com>
9323 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9324 (gen_easy_altivec_constant): Likewise.
9325 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9326 easy_vector_constant_msb): Likewise.
9328 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9331 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9332 add missing return true.
9334 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9337 * config/sh/predicates.md (general_extend_operand): Invoke
9338 general_movsrc_operand for memory operands.
9339 (general_movsrc_operand): Allow reg+reg addressing, do not use
9340 general_operand for memory operands.
9342 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9344 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9345 when current target options does not apply.
9346 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9347 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9348 * config/i386/bmiintrin.h: Pass appropriate target
9349 attributes to header.
9350 * config/i386/mmintrin.h: Ditto.
9351 * config/i386/nmmintrin.h: Ditto.
9352 * config/i386/avx2intrin.h: Ditto.
9353 * config/i386/fxsrintrin.h: Ditto.
9354 * config/i386/tbmintrin.h: Ditto.
9355 * config/i386/xsaveintrin.h: Ditto.
9356 * config/i386/f16cintrin.h: Ditto.
9357 * config/i386/xtestintrin.h: Ditto.
9358 * config/i386/xsaveoptintrin.h: Ditto.
9359 * config/i386/bmi2intrin.h: Ditto.
9360 * config/i386/lzcntintrin.h: Ditto.
9361 * config/i386/smmintrin.h: Ditto.
9362 * config/i386/wmmintrin.h: Ditto.
9363 * config/i386/x86intrin.h: Remove all header include guards.
9364 * config/i386/prfchwintrin.h: Ditto.
9365 * config/i386/pmmintrin.h: Ditto.
9366 * config/i386/tmmintrin.h: Ditto.
9367 * config/i386/xmmintrin.h: Ditto.
9368 * config/i386/popcntintrin.h: Ditto.
9369 * config/i386/rdseedintrin.h: Ditto.
9370 * config/i386/ammintrin.h: Ditto.
9371 * config/i386/emmintrin.h: Ditto.
9372 * config/i386/immintrin.h: Remove all header include guards.
9373 * config/i386/fma4intrin.h: Ditto.
9374 * config/i386/lwpintrin.h: Ditto.
9375 * config/i386/xopintrin.h: Ditto.
9376 * config/i386/ia32intrin.h: Ditto.
9377 * config/i386/avxintrin.h: Ditto.
9378 * config/i386/rtmintrin.h: Ditto.
9379 * config/i386/fmaintrin.h: Ditto.
9380 * config/i386/mm3dnow.h: Ditto.
9382 2013-06-22 Sriraman Tallam <tmsriram@google.com>
9384 * common/config/i386/i386-common.c: Handle LZCNT.
9386 2013-06-22 Andi Kleen <ak@linux.intel.com>
9388 * doc/extend.texi: Use __atomic_store_n instead of
9389 __atomic_store in HLE example.
9391 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
9393 * config/sh/sh.c: Remove <cstdlib> workaround.
9395 2013-06-21 Andi Kleen <ak@linux.intel.com>
9397 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
9399 2013-06-21 Andi Kleen <ak@linux.intel.com>
9401 * doc/extend.texi: Document that __atomic_clear and
9402 __atomic_test_and_set should only be used with bool.
9404 2013-06-20 Jan Hubicka <jh@suse.cz>
9406 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
9408 * tree.c (decls_same_for_odr): New function.
9409 (same_for_edr): New function.
9410 (types_same_for_odr): New function.
9411 (get_binfo_at_offset): Use it.
9412 * tree.h (types_same_for_odr): Declare.
9414 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
9415 Jason Merrill <jason@redhat.com>
9417 * system.h: Include <cstdlib> as well as <stdlib.h>.
9419 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
9422 * config/i386/i386.c (construct_container): Report error if
9423 long double is used with disabled x87 float returns.
9425 2013-06-20 Jan Hubicka <jh@suse.cz>
9427 * lto-cgraph.c (input_symtab): Do not set cgraph state.
9429 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
9431 PR rtl-optimization/57425
9432 PR rtl-optimization/57569
9433 * alias.c (write_dependence_p): Remove parameters mem_mode and
9434 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
9435 Changed all callers.
9436 (canon_anti_dependence): Get comments and semantics in sync.
9437 Add parameter mem_canonicalized. Changed all callers.
9438 * rtl.h (canon_anti_dependence): Update prototype.
9440 2013-06-20 Richard Biener <rguenther@suse.de>
9442 * data-streamer-in.c (streamer_read_uhwi): Optimize single
9443 byte case, inline streamer_read_uchar and defer section
9446 2013-06-20 Richard Biener <rguenther@suse.de>
9448 PR tree-optimization/57584
9449 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
9450 SSA names into the expanded expression that take part in
9451 abnormal coalescing.
9453 2013-06-19 Sharad Singhai <singhai@google.com>
9455 * gcov.c (print_usage): Handle new option.
9456 (process_args): Ditto.
9457 (get_gcov_intermediate_filename): New function.
9458 (output_intermediate_file): New function.
9459 (output_gcov_file): New function
9460 (generate_results): Handle new option.
9461 (release_function): Relase demangled name.
9462 (read_graph_file): Handle demangled name.
9463 (output_lines): Ditto.
9464 * doc/gcov.texi: Document gcov intermediate format.
9466 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
9469 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
9470 (lra_emit_add): Use the functions. Add comment about Y as an
9473 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
9476 * collect2.c (collect_atexit): New.
9477 (collect_exit): Delete.
9478 (main): Register collect_atexit with atexit.
9479 (collect_wait): Change collect_exit to exit.
9481 * collect2.h (collect_exit): Delete.
9482 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
9484 2013-06-19 Wei Mi <wmi@google.com>
9486 PR rtl-optimization/57518
9487 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
9488 if regno is used in paradoxical subreg.
9489 (update_equiv_regs): Check pdx_subregs[regno] before
9490 set a reg to be equivalent with a mem.
9492 2013-06-19 Matthias Klose <doko@ubuntu.com>
9495 * file-find.h (find_a_file): Add a mode parameter.
9496 * file-find.c (find_a_file): Likewise.
9497 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
9498 with X_OK for the executables.
9499 * collect2.c (main): Call find_a_file with X_OK.
9501 2013-06-19 Steve Ellcey <sellcey@mips.com>
9504 * config/mips/mips.md (casesi_internal_mips16_<mode>):
9505 Use NEXT_INSN instead of next_real_insn.
9507 2013-06-19 Jan Hubicka <jh@suse.cz>
9509 * cgraph.h (const_value_known_p): Replace by ...
9510 (ctor_for_folding): .. this one.
9511 * cgraphunit.c (process_function_and_variable_attributes): Use it.
9512 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
9513 * expr.c (expand_expr_real_1): Likewise.
9514 (string_constant): Likewise.
9515 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
9516 * ipa.c (process_references): Likewise.
9517 (symtab_remove_unreachable_nodes): Likewise.
9518 * ipa-inline-analysis.c (param_change_prob): Likewise.
9519 * gimple-fold.c (canonicalize_constructor_val): Likewise.
9520 (get_base_constructor): Likwise.
9521 * varpool.c (varpool_remove_node): Likewise.
9522 (varpool_remove_initializer): LIkewise.
9523 (dump_varpool_node): LIkwise.
9524 (const_value_known_p): Rewrite to ...
9525 (ctor_for_folding): ... this one.
9527 2013-06-19 Jakub Jelinek <jakub@redhat.com>
9530 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
9531 PERSONALITY in $PATH derived prefixes.
9533 2013-06-19 Jeff Law <law@redhat.com>
9535 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
9538 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
9539 (simplify_bitwise_binary): Use it to simpify certain binary ops on
9542 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
9544 * config/arm/vfp.md: Move VFP instruction classification documentation
9546 * config/arm/arm.md: ... here. Update instruction classification
9549 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
9551 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
9552 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
9553 pattern. Use more efficient sequences on ARMv5 and Thumb2.
9555 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
9558 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9559 with NEXT_INSN. Use tablejump_p to check for jump table data
9562 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
9565 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
9566 that now in C++ the value is correct per the C++ standards.
9568 2013-06-19 Richard Biener <rguenther@suse.de>
9570 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
9573 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9576 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9579 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9580 with next_active_insn.
9582 2013-06-18 Sriraman Tallam <tmsriram@google.com>
9584 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
9585 functions are inlined during failures to flag an error.
9586 * tree-inline.c (expand_call_inline): Allow the error to be flagged
9587 in early inline pass.
9589 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
9591 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
9594 2013-06-18 Julian Brown <julian@codesourcery.com>
9596 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
9597 Permit virtual register pre-reload if !strict.
9598 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
9600 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
9602 * config/arm/neon.md (movmisalign<mode>): Use
9603 neon_perm_struct_or_reg_operand instead of
9604 neon_struct_or_register_operand.
9605 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
9606 neon_permissive_struct_operand instead of neon_struct_operand.
9607 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
9608 neon_vector_mem_operand.
9609 * config/arm/predicates.md (neon_struct_operand): Adjust call to
9610 neon_vector_mem_operand.
9611 (neon_permissive_struct_operand): New.
9612 (neon_struct_or_register_operand): Rename to...
9613 (neon_perm_struct_or_reg_operand): This. Adjust call to
9614 neon_vector_mem_operand.
9616 2013-06-18 Richard Biener <rguenther@suse.de>
9618 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
9619 * lto-streamer.h: Include pointer-set.h.
9620 (struct lto_decl_slot): Remove.
9621 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
9622 Remove next_index entry.
9623 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9624 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9625 (lto_init_tree_ref_encoder): Adjust.
9626 (lto_destroy_tree_ref_encoder): Likewise.
9627 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9628 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9629 (lto_output_decl_index): Adjust.
9630 (lto_new_out_decl_state): Likewise.
9631 (lto_record_function_out_decl_state): Likewise.
9632 * lto-streamer-out.c (copy_function): Likewise.
9634 2013-06-18 Richard Biener <rguenther@suse.de>
9636 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
9637 * cgraphunit.c: Include cfgloop.h.
9638 (init_lowered_empty_function): Initialize the loop tree.
9639 (assemble_thunk): Insert new BBs into loops.
9641 2013-06-18 Richard Biener <rguenther@suse.de>
9643 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
9644 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
9645 the map from cache entry to cache index optional.
9646 (streamer_tree_cache_replace_tree): Adjust accordingly.
9647 (streamer_tree_cache_append): Likewise.
9648 (streamer_tree_cache_delete): Likewise.
9649 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
9650 streamer cache map from cache entry to cache index.
9651 * lto-streamer-out.c (create_output_block): Adjust.
9653 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
9655 * config/arm/arm.md (attribute "insn"): Move multiplication and
9656 division attributes to...
9657 (attribute "type"): ... here. Remove mult.
9658 (attribute "mul32"): New attribute.
9659 (attribute "mul64"): Add umaal.
9660 (*arm_mulsi3): Update attributes.
9661 (*arm_mulsi3_v6): Likewise.
9662 (*thumb_mulsi3): Likewise.
9663 (*thumb_mulsi3_v6): Likewise.
9664 (*mulsi3_compare0): Likewise.
9665 (*mulsi3_compare0_v6): Likewise.
9666 (*mulsi_compare0_scratch): Likewise.
9667 (*mulsi_compare0_scratch_v6): Likewise.
9668 (*mulsi3addsi): Likewise.
9669 (*mulsi3addsi_v6): Likewise.
9670 (*mulsi3addsi_compare0): Likewise.
9671 (*mulsi3addsi_compare0_v6): Likewise.
9672 (*mulsi3addsi_compare0_scratch): Likewise.
9673 (*mulsi3addsi_compare0_scratch_v6): Likewise.
9674 (*mulsi3subsi): Likewise.
9675 (*mulsidi3adddi): Likewise.
9676 (*mulsi3addsi_v6): Likewise.
9677 (*mulsidi3adddi_v6): Likewise.
9678 (*mulsidi3_nov6): Likewise.
9679 (*mulsidi3_v6): Likewise.
9680 (*umulsidi3_nov6): Likewise.
9681 (*umulsidi3_v6): Likewise.
9682 (*umulsidi3adddi): Likewise.
9683 (*umulsidi3adddi_v6): Likewise.
9684 (*smulsi3_highpart_nov6): Likewise.
9685 (*smulsi3_highpart_v6): Likewise.
9686 (*umulsi3_highpart_nov6): Likewise.
9687 (*umulsi3_highpart_v6): Likewise.
9688 (mulhisi3): Likewise.
9689 (*mulhisi3tb): Likewise.
9690 (*mulhisi3bt): Likewise.
9691 (*mulhisi3tt): Likewise.
9692 (maddhisi4): Likewise.
9693 (*maddhisi4tb): Likewise.
9694 (*maddhisi4tt): Likewise.
9695 (maddhidi4): Likewise.
9696 (*maddhidi4tb): Likewise.
9697 (*maddhidi4tt): Likewise.
9699 (udivsi3): Likewise.
9700 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
9701 (thumb2_mulsi_short_compare0): Likewise.
9702 (thumb2_mulsi_short_compare0_scratch): Likewise.
9703 * config/arm/arm1020e.md (1020mult1): Update attribute change.
9704 (1020mult2): Likewise.
9705 (1020mult3): Likewise.
9706 (1020mult4): Likewise.
9707 (1020mult5): Likewise.
9708 (1020mult6): Likewise.
9709 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
9711 (cortex_a15_mult64): Likewise.
9712 (cortex_a15_sdiv): Likewise.
9713 (cortex_a15_udiv): Likewise.
9714 * config/arm/arm1026ejs.md (mult1): Update attribute change.
9720 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
9721 (pj4_ir_div): Likewise.
9722 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
9723 (11_mult2): Likewise.
9724 (11_mult3): Likewise.
9725 (11_mult4): Likewise.
9726 (11_mult5): Likewise.
9727 (11_mult6): Likewise.
9728 (11_mult7): Likewise.
9729 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
9730 (cortex_a8_mla): Likewise.
9731 (cortex_a8_mull): Likewise.
9732 (cortex_a8_smulwy): Likewise.
9733 (cortex_a8_smlald): Likewise.
9734 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
9735 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
9736 (cortex_r4_mul_3): Likewise.
9737 (cortex_r4_mla_4): Likewise.
9738 (cortex_r4_mla_3): Likewise.
9739 (cortex_r4_smlald): Likewise.
9740 (cortex_r4_mull): Likewise.
9741 (cortex_r4_sdiv): Likewise.
9742 (cortex_r4_udiv): Likewise.
9743 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
9744 (cortex_a7_idiv): Likewise.
9745 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
9746 (9_mult2): Likewise.
9747 (9_mult3): Likewise.
9748 (9_mult4): Likewise.
9749 (9_mult5): Likewise.
9750 (9_mult6): Likewise.
9751 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
9752 (cortex_a53_sdiv): Likewise.
9753 (cortex_a53_udiv): Likewise.
9754 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
9755 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
9756 (mp626_mult2): Likewise.
9757 (mp626_mult3): Likewise.
9758 (mp626_mult4): Likewise.
9759 * config/arm/fa526.md (526_mult1): Update attribute change.
9760 (526_mult2): Likewise.
9761 * config/arm/arm-generic.md (mult): Update attribute change.
9762 (mult_ldsched_strongarm): Likewise.
9763 (mult_ldsched): Likewise.
9764 (multi_cycle): Likewise.
9765 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
9766 * config/arm/fa606te.md (606te_mult1): Update attribute change.
9767 (606te_mult2): Likewise.
9768 (606te_mult3): Likewise.
9769 (606te_mult4): Likewise.
9770 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
9771 (cortex_a9_mac16): Likewise.
9772 (cortex_a9_multiply): Likewise.
9773 (cortex_a9_mac): Likewise.
9774 (cortex_a9_multiply_long): Likewise.
9775 * config/arm/fa626te.md (626te_mult1): Update attribute change.
9776 (626te_mult2): Likewise.
9777 (626te_mult3): Likewise.
9778 (626te_mult4): Likewise.
9780 2013-06-18 Richard Biener <rguenther@suse.de>
9783 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
9785 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9788 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9789 with next_active_insn.
9791 2013-06-18 Alan Modra <amodra@gmail.com>
9793 * config/rs6000/rs6000.h (enum data_align): New.
9794 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
9795 (DATA_ABI_ALIGNMENT): Define.
9796 (CONSTANT_ALIGNMENT): Correct comment.
9797 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
9798 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
9800 2013-06-17 David Malcolm <dmalcolm@redhat.com>
9802 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
9803 ATTRIBUTE_UNUSED marking.
9805 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
9807 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
9808 alternative and update.
9809 (aarch64_dup_lanedi): Delete.
9810 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
9811 * config/aarch64/aarch64-simd-builtins.def: Update.
9813 2013-06-17 Richard Biener <rguenther@suse.de>
9815 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
9816 (lto_input_scc): Declare.
9817 (lto_input_tree_1): Likewise.
9818 (struct lto_stats_d): Add num_tree_bodies_output and
9819 num_pickle_refs_output.
9820 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
9821 (lto_read_tree_1): Split out from ...
9822 (lto_read_tree): ... this.
9823 (lto_input_scc): New function.
9824 (lto_input_tree_1): Split out from ...
9825 (lto_input_tree): ... this. Handle LTO_tree_scc.
9826 (lto_data_in_create): Create the streamer cache without hashes.
9827 * lto-streamer-out.c (create_output_block): Create the streamer
9828 cache with hashes when not doing WPA.
9829 (lto_write_tree_1): Split out from ...
9830 (lto_write_tree): ... this.
9831 (get_symbol_initial_value): New function.
9832 (lto_output_tree_1): Split out from ...
9833 (lto_output_tree): ... this. Write trees as series of SCCs
9834 using a DFS walk via DFS_write_tree.
9835 (struct sccs, struct scc_entry): New types.
9836 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
9837 (DFS_write_tree_body): New function.
9838 (DFS_write_tree): Likewise.
9839 (hash_tree): Likewise.
9840 (scc_entry_compare): Likewise.
9841 (hash_scc): Likewise.
9842 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
9843 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
9844 TREE_CHAIN as regular reference.
9845 (streamer_read_integer_cst): Remove.
9846 (streamer_get_pickled_tree): Adjust.
9847 * tree-streamer-out.c (streamer_write_chain): Disable streaming
9848 of DECL_EXTERNALs in BLOCK_VARS for now.
9849 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
9851 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
9852 Add hash value argument and record that if hashes are recorded
9854 (streamer_tree_cache_insert_1): Adjust.
9855 (streamer_tree_cache_insert): Likewise.
9856 (streamer_tree_cache_insert_at): Rename to ...
9857 (streamer_tree_cache_replace_tree): ... this and adjust.
9858 (streamer_tree_cache_append): Adjust.
9859 (record_common_node): Likewise.
9860 (streamer_tree_cache_create): Add argument whether to
9861 record hash values together with trees.
9862 (streamer_tree_cache_delete): Adjust.
9863 * tree-streamer.h (struct streamer_tree_cache_d): Add
9865 (streamer_read_integer_cst): Remove.
9866 (streamer_tree_cache_insert): Adjust.
9867 (streamer_tree_cache_append): Likewise.
9868 (streamer_tree_cache_insert_at): Rename to ...
9869 (streamer_tree_cache_replace_tree): ... this and adjust.
9870 (streamer_tree_cache_create): Add argument whether to record hashes.
9871 (streamer_tree_cache_get): Rename to ...
9872 (streamer_tree_cache_get_tree): ... this.
9873 (streamer_tree_cache_get_hash): New function.
9874 * tree.c (cache_integer_cst): New function.
9875 * tree.h (cache_integer_cst): Declare.
9876 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
9877 * lto-symtab.c (lto_varpool_replace_node): Only release
9878 DECL_INITIAL of non-prevailing decls.
9879 * varpool.c (varpool_remove_initializer): Do not release
9880 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
9882 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9884 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
9885 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
9886 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
9887 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
9888 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
9889 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
9890 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
9891 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
9892 instead of TARGET_64BIT.
9893 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
9894 Require ISA_HAS_<D>DIV.
9896 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
9898 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
9899 (mips*-*-linux*): Move default with_llsc setting to where other
9901 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
9903 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
9904 Likewise. Remove default with_tune setting. Move default float
9905 setting to its own block. Handle with_llsc in the same block as above.
9907 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
9909 PR rtl-optimization/57425
9910 PR rtl-optimization/57569
9911 * alias.c (write_dependence_p): Add new parameters mem_mode,
9912 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
9913 Changed all callers.
9914 (canon_anti_dependence): New function.
9915 * cse.c (check_dependence): Use canon_anti_dependence.
9916 * cselib.c (cselib_invalidate_mem): Likewise.
9917 * rtl.h (canon_anti_dependence): Declare.
9919 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9921 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
9922 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
9923 ".set mips3" for 64-bit targets.
9925 2013-06-15 Dehao Chen <dehao@google.com>
9927 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
9928 * gimple-low.c (gimple_check_call_matching_types): Likewise.
9929 (gimple_check_call_args): Likewise.
9930 * value-prof.c (check_ic_target): Likewise.
9931 * ipa-inline.c (early_inliner): Likewise.
9932 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
9933 * cgraph.c (cgraph_create_edge_1): Likewise.
9934 (cgraph_make_edge_direct): Likewise.
9936 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
9939 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
9940 rs6000_output_move_128bit to handle emitting quad memory
9941 operations. Set attribute length to 8 bytes.
9943 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
9945 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
9947 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
9948 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
9949 (aarch64_<su>mlsl<mode>): Likewise.
9951 2013-06-14 Mike Stump <mikestump@comcast.net>
9953 * Makefile.in (TARGET_H): Add insn-codes.h.
9955 2013-06-14 Alan Modra <amodra@gmail.com>
9959 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
9960 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
9961 bitfield expansion when EXPAND_MEMORY.
9962 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
9964 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
9966 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
9967 test for clearing quad memory on 32-bit later.
9969 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
9971 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
9972 (fold_negate_expr): Likewise.
9973 (fold_real_zero_addition_p): Handle vectors.
9974 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
9976 2013-06-14 Alan Modra <amodra@gmail.com>
9978 * varasm.c (force_const_mem): Revert 2013-06-07 change.
9980 2013-06-13 Jan Hubicka <jh@suse.cz>
9982 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
9983 Local comdats are not externally visible.
9984 * symtab.c (dump_symtab_base): Dump externally visible.
9985 (verify_symtab_base): Verify back links in the symtab hash.
9987 2013-06-13 Bin Cheng <bin.cheng@arm.com>
9989 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
9990 CONVERT_EXPR as equal nodes.
9992 2013-06-13 Bin Cheng <bin.cheng@arm.com>
9994 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
9996 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
9998 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
9999 Generalize to complex and vector.
10000 * tree.c (build_all_ones_cst): New function.
10001 * tree.h (build_all_ones_cst): Declare it.
10003 2013-06-13 Alan Modra <amodra@gmail.com>
10005 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10006 * config/rs6000/rs6000.md (signbittf2): New insn.
10007 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10008 (abstf2_internal, cmptf_internal2): Likewise.
10009 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10011 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10012 Pat Haugen <pthaugen@us.ibm.com>
10013 Peter Bergner <bergner@vnet.ibm.com>
10015 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10016 power8 byte, half-word, and quad-word atomic instructions.
10017 (emit_store_conditional): Likewise.
10018 (rs6000_expand_atomic_compare_and_swap): Likewise.
10019 (rs6000_expand_atomic_op): Likewise.
10021 * config/rs6000/sync.md (larx): Add new modes for power8.
10023 (AINT): New mode iterator to include TImode as well as normal
10024 integer modes on power8.
10025 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10026 that VSX registers are not considered. Use AINT mode iterator
10027 instead of INT1 to allow inclusion of quad word atomic operations
10029 (load_locked<mode>): Likewise.
10030 (store_conditional<mode>): Likewise.
10031 (atomic_compare_and_swap<mode>): Likewise.
10032 (atomic_exchange<mode>): Likewise.
10033 (atomic_nand<mode>): Likewise.
10034 (atomic_fetch_<fetchop_name><mode>): Likewise.
10035 (atomic_nand_fetch<mode>): Likewise.
10036 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10038 (ATOMIC): On power8, add QImode, HImode modes.
10039 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10040 modes that promote to SImode.
10041 (load_lockedti): Convert TImode arguments to PTImode, so that we
10042 get a guaranteed even/odd register pair.
10043 (load_lockedpti): Likewise.
10044 (store_conditionalti): Likewise.
10045 (store_conditionalpti): Likewise.
10047 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10048 atomic load/store instructions.
10051 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10053 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10055 (insn_count): New attribute, with most cases extracted from...
10056 (length): ...here. Redefine most cases in terms of insn_count.
10057 (single_insn): Delete.
10058 (can_delay): Use insn_count to check for single instructions.
10059 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10060 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10061 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10062 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10063 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10064 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10065 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10066 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10067 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10068 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10069 rather than "length".
10070 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10071 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10072 Use "insn_count" rather than "length".
10073 * config/mips/mips-dsp.md
10074 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10075 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10078 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10080 PR tree-optimization/57361
10081 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10083 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10085 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10087 (aarch64_simd_combine<mode>): New instruction expansion.
10088 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10089 function prototype.
10090 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10091 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10093 2013-06-12 Jan Hubicka <jh@suse.cz>
10095 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10096 decl has when in streaming stage.
10097 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10098 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10100 2013-06-12 Roland Stigge <stigge@antcom.de>
10103 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10105 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10107 PR tree-optimization/57537
10108 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10109 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10111 2013-06-12 Richard Biener <rguenther@suse.de>
10113 * data-streamer.h (streamer_write_char_stream): CSE
10114 obs->current_pointer.
10115 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10116 streamer_write_char_stream manually and optimize the resulting loop.
10117 (streamer_write_hwi_stream): Likewise.
10119 2013-06-12 Jan Hubicka <jh@suse.cz>
10121 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10122 * cgraph.h (varpool_create_empty_node): Declare.
10123 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10125 * symtab.c (symtab_unregister_node): Be lax about missin entries
10127 (symtab_get_node): Update comment.
10128 * varpool.c (varpool_create_empty_node): Break out from ...
10129 (varpool_node_for_decl): ... here.
10130 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10132 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10134 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10135 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10136 part. Use straight-line flow at the end.
10137 <COMPONENT_REF>: Remove superfluous else.
10138 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10140 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10143 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10144 target hook even for !TREE_PUBLIC decls. If no resolution info
10145 is available, return false for common and external decls.
10147 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10149 * config/rl78/constraints.md (U): New constraint.
10150 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10153 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10156 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10157 to allow returning address to AT_PLATFORM name.
10159 2013-06-11 Jan Hubicka <jh@suse.cz>
10161 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10162 * cgraph.h (symtab_node_base): Add weakref flag.
10163 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10164 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10165 (output_weakrefs): Use weakref flag.
10166 * fold-const.c (simple_operand_p): Handle WEAK.
10167 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10168 * ipa.c (varpool_externally_visible_p): Drop weakref.
10169 (function_and_variable_visibility): Update comment; fix weakref
10170 sanity checks; do not clear DECL_WEAK on them.
10171 * lto-cgraph.c (lto_output_node): update.
10172 (lto_output_varpool_node): Update.
10173 (input_overwrite_node): Update.
10174 (input_node): Update.
10175 (input_varpool_node): Update.
10176 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10177 (lto_symtab_merge_symbols): Add sanity check.
10178 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10179 * passes.c (rest_of_decl_compilation): Set static flag, too.
10180 * symtab.c (dump_symtab_base): Dump weakref.
10181 (verify_symtab_base): Sanity check weakrefs.
10182 (symtab_make_decl_local): Remove duplicated code.
10183 (symtab_alias_ultimate_target): Simplify.
10184 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10186 2013-06-11 Tom de Vries <tom@codesourcery.com>
10188 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10189 sequence_vect == NULL.
10191 2013-06-11 DJ Delorie <dj@redhat.com>
10193 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10194 (rl78_unwind_word_mode): New.
10196 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10198 * final.c (debug_prefix_maps): Make static.
10200 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10202 * function.c (initial_trampoline): Remove stray copy.
10204 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10206 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10208 2013-06-11 Martin Jambor <mjambor@suse.cz>
10210 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10211 within bounds at the beginning of the function.
10213 2013-06-11 Alan Modra <amodra@gmail.com>
10215 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10217 (get_named_section): Don't NULL !DECL_P decl.
10219 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10221 * doc/invoke.texi (core-avx2): Document.
10223 (atom): Updated with MOVBE.
10225 2013-06-11 Richard Biener <rguenther@suse.de>
10227 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10229 2013-06-11 Anton Blanchard <anton@samba.org>
10231 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10232 correct shift value in little-endian mode.
10234 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10237 * varasm.c (get_variable_align): Move #endif to the right place.
10239 2013-06-10 Cary Coutant <ccoutant@google.com>
10241 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10242 for hash so that hash table traversal order is deterministic.
10244 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10245 Pat Haugen <pthaugen@us.ibm.com>
10246 Peter Bergner <bergner@vnet.ibm.com>
10248 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10249 of vectors in GPRS if they are direct moves or quad word load or
10252 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10254 (direct_move_p): Likewise.
10255 (quad_load_store_p): Likewise.
10257 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10258 classes into bins based on the physical register type.
10259 (reg_class_to_reg_type): Likewise.
10260 (IS_STD_REG_TYPE): Likewise.
10261 (IS_FP_VECT_REG_TYPE): Likewise.
10262 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10263 use direct move instructions.
10264 (reload_gpr_vsx): Likewise.
10265 (reload_vsx_gpr): Likewise.
10266 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10267 information that is a simplification of register classes. Also
10268 precalculate direct move reload helpers.
10269 (direct_move_p): New function to return true if the operation can
10270 be done as a direct move instruciton.
10271 (quad_load_store_p): New function to return true if the operation
10272 is a quad memory operation.
10273 (rs6000_legitimize_address): If quad memory, only allow register
10274 indirect for TImode addresses.
10275 (rs6000_legitimate_address_p): Likewise.
10276 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10277 (rs6000_reload_register_type): Likewise.
10278 (register_to_reg_type): Return register type.
10279 (rs6000_secondary_reload_simple_move): New helper function for
10280 secondary reload and secondary memory needed to identify anything
10281 that is a simple move, and does not need reloading.
10282 (rs6000_secondary_reload_direct_move): New helper function for
10283 secondary reload to identify cases that can be done with several
10284 instructions via the direct move instructions.
10285 (rs6000_secondary_reload_move): New helper function for secondary
10286 reload to identify moves between register types that can be done.
10287 (rs6000_secondary_reload): Add support for quad memory operations
10288 and for direct move.
10289 (rs6000_secondary_memory_needed): Likewise.
10290 (rs6000_debug_secondary_memory_needed): Change argument names.
10291 (rs6000_output_move_128bit): New function to return the move to
10292 use for 128-bit moves, including knowing about the various
10293 limitations of quad memory operations.
10295 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10296 memory operations. call rs6000_output_move_128bit for the actual
10297 instruciton(s) to generate.
10298 (vsx_movti_64bit): Likewise.
10300 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10301 (UNSPEC_P8V_MTVSRWZ): Likewise.
10302 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10303 (UNSPEC_P8V_MTVSRD): Likewise.
10304 (UNSPEC_P8V_XXPERMDI): Likewise.
10305 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10306 (UNSPEC_FUSION_GPR): Likewise.
10307 (FMOVE128_GPR): New iterator for direct move.
10308 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10309 (f32_sv): Likewise.
10310 (f32_dm): Likewise.
10311 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10312 loads and direct move instructions.
10313 (zero_extendsidi2_lfiwzx): Likewise.
10314 (extendsidi2_lfiwax): Likewise.
10315 (extendsidi2_nocell): Likewise.
10316 (floatsi<mode>2_lfiwax): Likewise.
10317 (lfiwax): Likewise.
10318 (floatunssi<mode>2_lfiwzx): Likewise.
10319 (lfiwzx): Likewise.
10320 (fix_trunc<mode>_stfiwx): Likewise.
10321 (fixuns_trunc<mode>_stfiwx): Likewise.
10322 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10323 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10324 (parity<mode>2_cmpb): Set length/type attr.
10325 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10326 for 'mr.' to fast_compare.
10327 (bpermd_<mode>): Change type attr to popcnt.
10328 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10329 (p8_mtvsrwz_1): Likewise.
10330 (p8_mtvsrwz_2): Likewise.
10331 (reload_fpr_from_gpr<mode>): Likewise.
10332 (p8_mtvsrd_1): Likewise.
10333 (p8_mtvsrd_2): Likewise.
10334 (p8_xxpermdi_<mode>): Likewise.
10335 (reload_vsx_from_gpr<mode>): Likewise.
10336 (reload_vsx_from_gprsf): Likewise.
10337 (p8_mfvsrd_3_<mode>): LIkewise.
10338 (reload_gpr_from_vsx<mode>): Likewise.
10339 (reload_gpr_from_vsxsf): Likewise.
10340 (p8_mfvsrd_4_disf): Likewise.
10341 (multi-word GPR splits): Do not split direct moves or quad memory
10344 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10346 * tree-into-ssa.c (interesting_blocks): Make static.
10348 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10351 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10352 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10353 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10354 (get_variable_align): New function.
10355 (get_variable_section, emit_bss, emit_common,
10356 assemble_variable_contents, place_block_symbol): Use
10357 get_variable_align instead of DECL_ALIGN.
10358 (assemble_noswitch_variable): Add align argument, use it
10359 instead of DECL_ALIGN.
10360 (assemble_variable): Adjust caller. Use get_variable_align
10361 instead of DECL_ALIGN.
10362 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10364 (DATA_ABI_ALIGNMENT): Define.
10365 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10366 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10367 opt is false, only return the psABI mandated alignment increase.
10368 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10369 (DATA_ABI_ALIGNMENT): ... this.
10370 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10371 (DATA_ABI_ALIGNMENT): ... this.
10372 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
10373 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
10374 (DATA_ABI_ALIGNMENT): ... this.
10375 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
10376 * doc/tm.texi: Regenerated.
10378 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
10380 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
10381 cmp_code to construct REG_EQUAL note.
10383 2013-06-09 Jakub Jelinek <jakub@redhat.com>
10386 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
10387 that operands[2] doesn't overlap with operands[0].
10389 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
10390 Jan Hubicka <jh@suse.cz>
10392 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
10393 hack to mark symbols as used.
10395 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
10397 PR rtl-optimization/57559
10398 * lra-constraints.c (process_alt_operands): Don't discourage
10399 memory with known offset for offsetable memory constraint.
10400 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
10402 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
10404 * varasm.c (struct oc_local_state): Reorder fields.
10405 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
10406 and adjust accordingly.
10407 (output_constructor): Reorder initialization code and adjust call to
10408 output_constructor_bitfield.
10410 2013-06-07 Jan Hubicka <jh@suse.cz>
10412 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
10414 2013-06-07 David Malcolm <dmalcolm@redhat.com>
10416 * tree-object-size.c (unknown): Make const.
10418 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10420 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
10421 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
10422 for last alternative in the cpu_facility attribute.
10424 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10427 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
10428 (xordi3): Change operand 2 constraint to arm_xordi_operand.
10429 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
10430 * config/arm/constraints.md (Dg): New constraint.
10431 * config/arm/neon.md (xordi3_neon): Remove.
10432 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
10433 * config/arm/predicates.md (arm_xordi_operand): New predicate.
10435 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10437 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
10438 Clean up alternatives.
10440 2013-06-07 Alan Modra <amodra@gmail.com>
10442 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
10445 2013-06-07 Alan Modra <amodra@gmail.com>
10447 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
10449 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10451 * config/arm/constraints.md (Df): New constraint.
10452 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
10453 Correct length attribute for last two alternatives.
10455 2013-06-07 Alan Modra <amodra@gmail.com>
10457 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10458 override user -mfp-in-toc.
10459 (offsettable_ok_by_alignment): Consider just the current access
10460 rather than the whole object, unless BLKmode. Handle
10461 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
10462 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
10463 for -mcmodel=medium.
10464 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
10465 override user -mfp-in-toc or -msum-in-toc. Default to
10466 -mno-fp-in-toc for -mcmodel=medium.
10468 2013-06-06 DJ Delorie <dj@redhat.com>
10470 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
10471 TARGET_VALID_POINTER_MODE.
10473 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
10474 Pat Haugen <pthaugen@us.ibm.com>
10475 Peter Bergner <bergner@vnet.ibm.com>
10477 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
10478 Document new power8 builtins.
10480 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
10481 condition code register, to allow 128-bit logical operations to be
10482 done in the VSX or GPR registers.
10483 (nor<mode>3): Use the canonical form for nor.
10484 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
10485 vclz*, and vpopcnt* vector instructions.
10486 (nand<mode>3): Likewise.
10487 (orc<mode>3): Likewise.
10488 (clz<mode>2): LIkewise.
10489 (popcount<mode>2): Likewise.
10491 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
10492 that only the GPRs are recognized.
10494 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10495 support for new power8 builtins.
10497 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
10499 (xscvdpspn): Likewise.
10505 (vpopcnt): Likewise.
10506 (vpopcntb): Likewise.
10507 (vpopcnth): Likewise.
10508 (vpopcntw): Likewise.
10509 (vpopcntd): Likewise.
10511 (vmrgew): Likewise.
10512 (vmrgow): Likewise.
10514 (eqv_v16qi3): Likewise.
10515 (eqv_v8hi3): Likewise.
10516 (eqv_v4si3): Likewise.
10517 (eqv_v2di3): Likewise.
10518 (eqv_v4sf3): Likewise.
10519 (eqv_v2df3): Likewise.
10521 (nand_v16qi3): Likewise.
10522 (nand_v8hi3): Likewise.
10523 (nand_v4si3): Likewise.
10524 (nand_v2di3): Likewise.
10525 (nand_v4sf3): Likewise.
10526 (nand_v2df3): Likewise.
10528 (orc_v16qi3): Likewise.
10529 (orc_v8hi3): Likewise.
10530 (orc_v4si3): Likewise.
10531 (orc_v2di3): Likewise.
10532 (orc_v4sf3): Likewise.
10533 (orc_v2df3): Likewise.
10535 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
10536 allow power8 quad mode in 64-bit.
10537 (rs6000_builtin_vectorized_function): Add support to vectorize
10538 ISA 2.07 count leading zeros, population count builtins.
10539 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
10540 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
10541 (builtin_function_type): Add vgbbd builtin function which takes an
10543 (altivec_expand_vec_perm_const): Add support for new power8 merge
10546 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
10547 that does not include TImdoe for use with 32-bit.
10548 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
10550 (UNSPEC_VSX_CVDPSPN): Likewise.
10551 (vsx_xscvdpspn): Likewise.
10552 (vsx_xscvspdpn): Likewise.
10553 (vsx_xscvdpspn_scalar): Likewise.
10554 (vsx_xscvspdpn_directmove): Likewise.
10555 (vsx_and<mode>3): Split logical operations into 32-bit and
10556 64-bit. Add support to do logical operations on TImode as well as
10557 VSX vector types. Allow logical operations to be done in either
10558 VSX registers or in general purpose registers in 64-bit mode. Add
10559 splitters if GPRs were used. For AND, add clobber of CCmode to
10560 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
10562 (vsx_and<mode>3_32bit): Likewise.
10563 (vsx_and<mode>3_64bit): Likewise.
10564 (vsx_ior<mode>3): Likewise.
10565 (vsx_ior<mode>3_32bit): Likewise.
10566 (vsx_ior<mode>3_64bit): Likewise.
10567 (vsx_xor<mode>3): Likewise.
10568 (vsx_xor<mode>3_32bit): Likewise.
10569 (vsx_xor<mode>3_64bit): Likewise.
10570 (vsx_one_cmpl<mode>2): Likewise.
10571 (vsx_one_cmpl<mode>2_32bit): Likewise.
10572 (vsx_one_cmpl<mode>2_64bit): Likewise.
10573 (vsx_nor<mode>3): Likewise.
10574 (vsx_nor<mode>3_32bit): Likewise.
10575 (vsx_nor<mode>3_64bit): Likewise.
10576 (vsx_andc<mode>3): Likewise.
10577 (vsx_andc<mode>3_32bit): Likewise.
10578 (vsx_andc<mode>3_64bit): Likewise.
10579 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
10580 and xxlorc instructions.
10581 (vsx_eqv<mode>3_64bit): Likewise.
10582 (vsx_nand<mode>3_32bit): Likewise.
10583 (vsx_nand<mode>3_64bit): Likewise.
10584 (vsx_orc<mode>3_32bit): Likewise.
10585 (vsx_orc<mode>3_64bit): Likewise.
10587 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
10589 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
10591 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
10592 (p8_vmrgow): Likewise.
10593 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
10594 GPRs to be split under VSX.
10595 (p8v_clz<mode>2): Add power8 count leading zero support.
10596 (p8v_popcount<mode>2): Add power8 population count support.
10597 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
10600 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
10603 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
10605 (vec_nand): Likewise.
10606 (vec_vclz): Likewise.
10607 (vec_vclzb): Likewise.
10608 (vec_vclzd): Likewise.
10609 (vec_vclzh): Likewise.
10610 (vec_vclzw): Likewise.
10611 (vec_vgbbd): Likewise.
10612 (vec_vmrgew): Likewise.
10613 (vec_vmrgow): Likewise.
10614 (vec_vpopcnt): Likewise.
10615 (vec_vpopcntb): Likewise.
10616 (vec_vpopcntd): Likewise.
10617 (vec_vpopcnth): Likewise.
10618 (vec_vpopcntw): Likewise.
10620 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10622 PR rtl-optimization/57468
10623 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
10626 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10628 PR rtl-optimization/57459
10629 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
10630 type when setting live regs.
10632 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10634 * config/s390/s390.opt (mlra): New option.
10635 * config/s390/s390.c (s390_decompose_address): Check displacement
10636 for all registers for LRA.
10637 (s390_secondary_reload): Don't used secondary reloads for LRA.
10638 (s390_lra_p): New function.
10639 (TARGET_LRA_P): Define.
10640 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
10641 of attribute cpu_facility to zarch for the last alternative.
10642 (*cmpmem_short): Ditto.
10644 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
10646 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
10647 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
10648 (arm_expand_prologue): Likewise.
10650 2013-06-06 Teresa Johnson <tejohnson@google.com>
10653 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
10654 as this is now done by redirect_edge_and_branch_force.
10655 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
10656 barriers, and fix interaction with splitting.
10657 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
10658 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
10659 reflect changes made in the routine.
10660 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
10661 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
10662 since this is called in cfglayout mode, and replace partition fixup
10663 with assert as that is now done by force_nonfallthru_and_redirect.
10664 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
10665 already be marked with region crossing note.
10666 (insert_section_boundary_note): Make non-static, gate on flag
10667 has_bb_partition, rewrite to also check for multiple partitions.
10668 (rest_of_handle_reorder_blocks): Remove call to
10669 insert_section_boundary_note, now done later during free_cfg.
10670 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
10671 * bb-reorder.h (insert_section_boundary_note): Declare.
10672 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
10673 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
10674 invoke insert_section_boundary_note.
10675 (try_redirect_by_replacing_jump): Remove unnecessary
10676 check for region crossing note.
10677 (fixup_partition_crossing): New function.
10678 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
10679 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
10680 in non-cfglayout mode.
10681 (force_nonfallthru_and_redirect): Fixup partition boundaries,
10682 remove old code that tried to do this. Emit barrier correctly
10683 when we are in cfglayout mode.
10684 (last_bb_in_partition): New function.
10685 (rtl_split_edge): Correctly fixup partition boundaries.
10686 (commit_one_edge_insertion): Remove old code that tried to
10687 fixup region crossing edge since this is now handled in
10688 split_block, and set up insertion point correctly since
10689 block may now end in a jump.
10690 (verify_hot_cold_block_grouping): Guard against checking when not in
10691 linearized RTL mode.
10692 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
10694 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
10695 rtl_verify_flow_info, so not called in cfglayout mode.
10696 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
10697 (fixup_reorder_chain): Remove old code that attempted to fixup region
10698 crossing note as this is now handled in force_nonfallthru_and_redirect.
10699 (duplicate_insn_chain): Don't duplicate switch section notes.
10700 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
10702 * basic-block.h (emit_barrier_after_bb): Declare.
10704 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10706 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
10707 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
10708 arm_usatsihi): Adjust alternatives for arm_restrict_it.
10710 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10712 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
10714 * config/arm/ldmstm.md: Regenerate.
10716 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10718 * config/arm/sync.md (atomic_loaddi_1):
10719 Disable predication for arm_restrict_it.
10720 (arm_load_exclusive<mode>): Likewise.
10721 (arm_load_exclusivesi): Likewise.
10722 (arm_load_exclusivedi): Likewise.
10723 (arm_load_acquire_exclusive<mode>): Likewise.
10724 (arm_load_acquire_exclusivesi): Likewise.
10725 (arm_load_acquire_exclusivedi): Likewise.
10726 (arm_store_exclusive<mode>): Likewise.
10727 (arm_store_exclusive<mode>): Likewise.
10728 (arm_store_release_exclusivedi): Likewise.
10729 (arm_store_release_exclusive<mode>): Likewise.
10731 2013-06-06 Richard Biener <rguenther@suse.de>
10733 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
10735 (lto_tag_is_tree_code_p): Adjust.
10736 (lto_tag_is_gimple_code_p): Likewise.
10737 (lto_gimple_code_to_tag): Likewise.
10738 (lto_tag_to_gimple_code): Likewise.
10739 (lto_tree_code_to_tag): Likewise.
10740 (lto_tag_to_tree_code): Likewise.
10741 * data-streamer.h (streamer_write_hwi_in_range): Use
10742 uhwi streaming to stream the normalized range.
10743 (streamer_read_hwi_in_range): Likewise.
10745 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10747 * config/arm/arm.md (enabled_for_depr_it): New attribute.
10748 (predicable_short_it): Likewise.
10749 (predicated): Likewise.
10750 (enabled): Handle above.
10751 (define_cond_exec): Set predicated attribute to yes.
10753 2013-06-05 Mike Stump <mikestump@comcast.net>
10755 * gdbinit.in (__FUNCTION__): Add.
10757 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
10759 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
10760 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
10762 2013-06-05 Jan Hubicka <jh@suse.cz>
10764 * varasm.c (mark_decl_referenced): Revert the removal until targets
10767 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
10769 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
10770 instead of mark_decl_referenced.
10772 2013-06-05 Jan Hubicka <jh@suse.cz>
10774 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
10775 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
10776 and symtab_used_from_object_file_p.
10777 (cgraph_make_node_local_1): Clear forced_by_abi.
10778 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
10779 * cgraph.h (symtab_node_base): Add forced_by_abi.
10780 (decide_is_variable_needed): Remove.
10781 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
10782 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
10783 (decide_is_symbol_needed): ... this one; handle symbols in general;
10784 always analyze virtuals; honnor forced_by_abi.
10785 (cgraph_finalize_function): Update.
10786 (varpool_finalize_decl): Update.
10787 (symbol_defined_and_needed): Remove.
10788 (analyze_functions): Update.
10789 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10790 output_refs, input_overwrite_node): Handle forced_by_abi.
10791 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
10792 (address_taken_from_non_vtable_p): ... this one.
10793 (comdat_can_be_unshared_p_1): New function.
10794 (cgraph_comdat_can_be_unshared_p): Rename to ...
10795 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
10796 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
10797 (function_and_variable_visibility): Clear forced_by_abi as needed.
10798 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
10799 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
10800 * symtab.c (dump_symtab_base): Dump forced_by_abi.
10801 * varpool.c (decide_is_variable_needed): Remove.
10803 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10805 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
10806 (arm_option_override): Override arm_restrict_it where appropriate.
10807 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
10808 * config/arm/arm.opt (mrestrict-it): New command-line option.
10809 * doc/invoke.texi: Document -mrestrict-it.
10811 2013-06-05 David Malcolm <dmalcolm@redhat.com>
10813 * tsan.c (tsan_atomic_table): Make const.
10815 2013-06-05 Richard Biener <rguenther@suse.de>
10817 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
10818 index associated with the tree we are supposed to replace.
10819 * tree-streamer-out.c (pack_ts_base_value_fields): Output
10820 TREE_ASM_WRITTEN as zero for everything but SSA names.
10822 2013-06-05 David Malcolm <dmalcolm@redhat.com>
10824 * tree-ssa-structalias.c (call_stmt_vars): Make static.
10826 2013-06-04 Jan Hubicka <jh@suse.cz>
10828 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
10829 (input_node, input_varpool_node): Handle correctly external same
10831 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
10832 nodes at ltrans stage.
10834 2013-06-04 Jan Hubicka <jh@suse.cz>
10836 * ipa-inline.c (update_caller_keys): Fix availability test.
10837 (update_callee_keys): Likewise.
10838 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
10839 to follow ELF standard.
10841 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
10843 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
10844 (mips64r5900el-*-elf*): New configurations.
10845 * config/mips/mips-cpus.def (r5900): New processor.
10846 * config/mips/mips-tables.opt: Regenerate.
10847 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
10848 (mips_issue_rate): Handle PROCESSOR_R5900.
10849 (mips_reorg_process_insns): Force reorder mode for the R5900.
10850 * config/mips/mips.h (TARGET_MIPS5900): Define.
10851 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
10853 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
10855 * config/mips/mips.md (processor): Add r5900.
10856 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
10858 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10860 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
10861 into function to generate MOVI instruction.
10862 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
10863 (aarch64_preferred_simd_mode): Turn into wrapper.
10864 (aarch64_output_scalar_simd_mov_immediate): New function.
10865 * config/aarch64/aarch64-protos.h: Add prototype for above.
10867 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10869 * config/aarch64/aarch64.c (simd_immediate_info): Remove
10870 element_char member.
10871 (sizetochar): Return signed char.
10872 (aarch64_simd_valid_immediate): Remove elchar and other
10873 unnecessary variables.
10874 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
10875 Calculate element_char as required.
10876 * config/aarch64/aarch64-protos.h: Update and move prototype
10877 for aarch64_output_simd_mov_immediate.
10878 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
10881 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10883 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
10884 information completed by aarch64_simd_valid_immediate.
10885 (aarch64_legitimate_constant_p): Update arguments.
10886 (aarch64_simd_valid_immediate): Work with struct rather than many
10888 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
10889 (aarch64_simd_make_constant): Update arguments.
10890 (aarch64_output_simd_mov_immediate): Work with struct rather than
10891 many pointers. Output immediate directly rather than as operand.
10892 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10894 * config/aarch64/constraints.md (Dn): Update arguments.
10896 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10898 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
10900 (aarch64_simd_immediate_valid_for_move): Remove.
10901 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
10902 (aarch64_simd_make_constant): Update call.
10903 (aarch64_output_simd_mov_immediate): Update call.
10904 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10906 * config/aarch64/constraints.md (Dn): Update call.
10908 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10910 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
10911 return type to bool for prototype.
10912 (aarch64_legitimate_constant_p): Check for true instead of not -1.
10913 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
10914 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
10916 2013-06-04 Catherine Moore <clm@codesourcery.com>
10918 * config/mips/mips.opt (meva): New.
10919 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
10920 (ASM_SPEC): Handle -meva.
10921 * doc/invoke.texi (meva): Document.
10923 2013-06-04 Alan Modra <amodra@gmail.com>
10925 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
10928 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10930 * rtl.def: Add extra fourth optional field to define_cond_exec.
10931 * gensupport.c (process_one_cond_exec): Process attributes from
10933 * doc/md.texi: Document fourth field in define_cond_exec.
10935 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
10937 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
10938 out the processing order as in store_bit_field_1.
10940 2013-06-04 Jan Hubicka <jh@suse.cz>
10942 PR middle-end/57500
10943 * cgraphunit.c (cgraph_process_same_body_aliases): Create
10944 non-VAR_DECL node if it does not exist yet.
10946 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
10948 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
10949 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
10950 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
10951 target_cpu_default setting.
10953 2013-06-03 Teresa Johnson <tejohnson@google.com>
10955 * dumpfile.c (opt_info_switch_p): Change -fopt-info
10956 default to -fopt-info=optimized instead of all.
10957 * doc/invoke.texi: Ditto.
10958 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
10959 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
10960 (execute_vect_slp): Emit BB vectorization success under
10961 MSG_OPTIMIZED_LOCATIONS.
10962 * tree-vect-slp.c (vect_slp_transform_bb): Change
10963 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
10964 * tree-vect-loop.c (vect_transform_loop): Ditto.
10966 2013-06-03 Jason Merrill <jason@redhat.com>
10969 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10970 Use TARGET_EXPR for C++.
10972 2013-06-03 Jakub Jelinek <jakub@redhat.com>
10974 PR rtl-optimization/57268
10975 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
10976 if DEBUG_INSN_P (insn).
10979 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10981 PR rtl-optimization/57268
10982 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10983 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10985 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10987 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
10988 (ix86_avoid_lea_for_addr): Likewise.
10989 (exact_dependency_1): Likewise.
10990 (ix86_adjust_cost): Likewise.
10991 (swap_top_of_ready_list): Fix formatting and !reload_completed check
10993 (do_reorder_for_imul): Fix typo, formatting and
10994 !reload_completed check removed.
10995 (ix86_sched_reorder): Fix typo and formatting.
10996 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
10999 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11001 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11003 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11005 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11006 <STRING_CST>: Likewise.
11007 <VECTOR_CST>: Likewise.
11009 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11010 Mikael Morin <mikael@gcc.gnu.org>
11012 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11013 * config.in: Regenerated.
11014 * configure: Regenerated.
11016 2013-06-01 Jan Hubicka <jh@suse.cz>
11018 PR middle-end/57366
11019 * cgraphunit.c (compile): When weakref is not supported,
11020 set up transparent aliases before final output pass.
11021 * varasm.c (assemble_alias): Do not try to do it here.
11023 2013-06-01 Jan Hubicka <jh@suse.cz>
11025 PR middle-end/57467
11026 * passes.c (for_per_function): Skip unanalyzed functions.
11028 2013-06-01 Jan Hubicka <jh@suse.cz>
11030 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11031 (lto_symtab_merge_symbols_1): ... this one.
11032 (lto_symtab_merge_cgraph_nodes): Rename to ...
11033 (lto_symtab_merge_symbols): ... this one; simplify.
11034 * cgraph.c (same_body_aliases_done): Rename to ...
11035 (cpp_implicit_aliases_done): ... this one.
11036 (cgraph_create_function_alias): Update.
11037 (cgraph_same_body_alias): Update.
11038 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11039 (verify_edge_corresponds_to_fndecl): Simplify.
11040 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11041 (cgraph_node): Remove same_body_alias.
11042 (varpool_node): Remove alias_of and extra_name_alias.
11043 (same_body_aliases_done): Rename to ..
11044 (cpp_implicit_aliases_done): ... this one.
11045 (symtab_alias_ultimate_target): Add default parameter.
11046 (symtab_resolve_alias): New function.
11047 (fixup_same_cpp_alias_visibility): Declare.
11048 (cgraph_function_node): Add default parameter.
11049 (cgraph_node_asm_name): Likewise.
11050 (cgraph_function_or_thunk_node): Add default parameter; do
11051 not ICE when it is NULL.
11052 (varpool_variable_node): Likewise.
11053 * tree-emutls.c (create_emultls_var): Update.
11054 (ipa_lower_emutls): Update.
11055 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11056 (cgraph_reset_node): Reset alias info.
11057 (cgraph_finalize_function): Update.
11058 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11059 (analyze_function): Simplify.
11060 (cgraph_process_same_body_aliases): Simplify.
11061 (analyze_functions): Fixup same body aliases.
11062 (handle_alias_pairs): Simplify.
11063 (assemble_thunk): Update.
11064 (assemble_thunks_and_aliases): Update.
11065 (output_weakrefs): Rewrite.
11066 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11067 (lto_output_varpool_node): Likewise.
11068 (compute_ltrans_boundary): Remve assert.
11069 (get_alias_symbol): New functoin.
11070 (input_node): Rewrite alias handling.
11071 (input_varpool_node): Likewise.
11072 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11073 * ipa.c (process_references): Handle weakrefs correctly.
11074 (symtab_remove_unreachable_nodes): Likewise.
11075 * trans-mem.c (get_cg_data): Update.
11076 (ipa_tm_create_version_alias): Update.
11077 (ipa_tm_execute): Update.
11078 * symtab.c (dump_symtab_base): Dump aliases.
11079 (verify_symtab_base): Verify aliases.
11080 (symtab_node_availability): New function.
11081 (symtab_alias_ultimate_target): Simplify.
11082 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11083 handle all the fixup cases.
11084 (symtab_resolve_alias): New function.
11085 * passes.c (ipa_write_summaries): Handle weakrefs.
11086 * varpool.c (varpool_analyze_node): Simplify.
11087 (assemble_aliases): Update.
11088 (varpool_create_variable_alias): Simplify.
11089 (varpool_extra_name_alias): Simplify.
11090 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11091 (lto_symtab_merge_symbols): ... this one.
11093 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11096 PR rtl-optimization/57268
11097 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11098 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11100 2013-06-01 Tobias Burnus <burnus@net-b.de>
11102 Partially reverted:
11103 2013-05-31 Tobias Burnus <burnus@net-b.de>
11105 PR middle-end/57073
11106 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11109 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11111 PR rtl-optimization/57268
11112 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11113 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11115 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11117 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11118 unordered comparison operators when -fno-trapping-math is in effect
11120 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11121 and implement unordered comparison operators properly on the e500.
11123 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11125 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11126 for constant scalar integers.
11127 (simplify_relational_operation_1): Likewise.
11129 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11131 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11132 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11135 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11136 Igor Zamyatin <igor.zamyatin@intel.com>
11138 Silvermont (SLM) architecture performance tuning.
11139 * config/i386/i386.h (enum ix86_tune_indices): Add
11140 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11141 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11143 * config/i386/i386.c (initial_ix86_tune_features)
11144 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11145 (ix86_lea_outperforms): Handle Silvermont tuning.
11146 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11148 (ix86_use_lea_for_mov): Likewise.
11149 (ix86_avoid_lea_for_addr): Likewise.
11150 (ix86_lea_for_add_ok): Likewise.
11151 (exact_dependency_1): New function.
11152 (exact_store_load_dependency): Likewise.
11153 (ix86_adjust_cost): Handle Silvermont tuning.
11154 (do_reoder_for_imul): Likewise.
11155 (swap_top_of_ready_list): New function.
11156 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11158 * config/i386/i386.md (peepholes that split memory operand in fp
11161 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11163 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11164 Remove un-necessary braces.
11166 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11168 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11169 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11171 2013-05-31 Tobias Burnus <burnus@net-b.de>
11173 PR middle-end/57073
11174 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11176 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11179 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11180 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11181 * config/arm/neon.md (iordi3_neon): Remove.
11182 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11183 * config/arm/predicates.md (imm_for_neon_logic_operand):
11184 Move to earlier in the file.
11185 (neon_logic_op2): Likewise.
11186 (arm_iordi_operand_neon): New predicate.
11188 2013-05-31 Richard Biener <rguenther@suse.de>
11190 PR tree-optimization/57478
11191 PR tree-optimization/57453
11192 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11195 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11197 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11198 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11200 2013-05-30 Tobias Burnus <burnus@net-b.de>
11201 Thomas Koenig <tkoenig@gcc.gnu.org>
11203 PR middle-end/57073
11204 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11205 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11207 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11209 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11211 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11213 * target.def (register_usage_leveling_p): New hook.
11214 * targhooks.c (default_register_usage_leveling_p): New.
11215 * targhooks.h (default_register_usage_leveling_p): New prototype.
11216 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11217 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11218 * doc/tm.texi: Update.
11219 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11221 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11223 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11224 (*insv_reg<mode>): New define_insn.
11226 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11228 PR rtl-optimization/57439
11229 * postreload.c (move2add_valid_value_p): Check that we have
11230 a zero subreg_regno_offset when accessing the register in
11231 the requested mode.
11233 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11234 Igor Zamyatin <igor.zamyatin@intel.com>
11236 Silvermont (SLM) architecture pipeline model, tuning and
11238 * config.gcc: Add slm config options and target.
11240 * config/i386/slm.md: New.
11242 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11244 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11246 (ix86_target_macros_internal): Likewise.
11248 * config/i386/i386.c (slm_cost): New cost.
11249 (m_SLM): New macro flag.
11250 (initial_ix86_tune_features): Set m_SLM.
11251 (x86_accumulate_outgoing_args): Likewise.
11252 (x86_arch_always_fancy_math_387): Likewise.
11253 (processor_target_table): Add slm cost.
11254 (cpu_names): Add slm cpu name.
11255 (x86_option_override_internal): Set SLM ISA.
11256 (ix86_issue_rate): New case PROCESSOR_SLM.
11257 (ia32_multipass_dfa_lookahead): Likewise.
11258 (fold_builtin_cpu): Add slm.
11260 * config/i386/i386.h (TARGET_SLM): New target macro.
11261 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11262 (processor_type): Add PROCESSOR_SLM.
11264 * config/i386/i386.md (cpu): Add new value "slm".
11265 (slm.md): Include slm.md.
11267 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11268 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11270 * config/arm/arm-protos.h: Add and update function protos.
11271 * config/arm/arm.c (use_simple_return_p): New added.
11272 (thumb2_expand_return): Check simple_return flag.
11273 * config/arm/arm.md: Add simple_return and conditional simple_return.
11274 * config/arm/iterators.md: Add iterator for return and simple_return.
11276 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11278 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11279 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11280 (arm_emit_vfp_multi_reg_pop): Likewise.
11281 (thumb2_emit_ldrd_pop): Likewise.
11282 (arm_expand_epilogue): Add misc REG_CFA notes.
11283 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11285 2013-05-29 Lawrence Crowl <crowl@google.com>
11287 * config/arm/t-arm: Update for below.
11289 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11290 Change type to hash_table. Update dependent calls and types.
11292 * config/i386/t-cygming: Update for below.
11294 * config/i386/t-interix: Update for below.
11296 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11297 Change type to hash_table. Update dependent calls and types.
11298 (i386_find_on_wrapper_list::wrappers): Likewise.
11300 * config/ia64/t-ia64: Update for below.
11302 * config/ia64/ia64.c (bundle_state_table):
11303 Change type to hash_table. Update dependent calls and types.
11305 * config/mips/mips.c (mips_reorg_process_insns::htab):
11306 Change type to hash_table. Update dependent calls and types.
11308 * config/sol2.c (solaris_comdat_htab):
11309 Change type to hash_table. Update dependent calls and types.
11311 * config/t-sol2: Update for above.
11313 2013-05-29 Teresa Johnson <tejohnson@google.com>
11315 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11316 functions are not yet marked as defined.
11318 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11319 Pat Haugen <pthaugen@us.ibm.com>
11320 Peter Bergner <bergner@vnet.ibm.com>
11322 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11326 (vrotl<mode>3): Likewise.
11327 (vashl<mode>3): Likewise.
11328 (vlshr<mode>3): Likewise.
11329 (vashr<mode>3): Likewise.
11331 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11332 support for power8 V2DI builtins.
11334 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11335 power8 V2DI builtins.
11336 (vupkhsw): Likewise.
11337 (vupklsw): Likewise.
11338 (vaddudm): Likewise.
11339 (vminsd): Likewise.
11340 (vmaxsd): Likewise.
11341 (vminud): Likewise.
11342 (vmaxud): Likewise.
11343 (vpkudum): Likewise.
11344 (vpksdss): Likewise.
11345 (vpkudus): Likewise.
11346 (vpksdus): Likewise.
11351 (vsubudm): Likewise.
11352 (vcmpequd): Likewise.
11353 (vcmpgtsd): Likewise.
11354 (vcmpgtud): Likewise.
11355 (vcmpequd_p): Likewise.
11356 (vcmpgtsd_p): Likewise.
11357 (vcmpgtud_p): Likewise.
11358 (vupkhsw): Likewise.
11359 (vupklsw): Likewise.
11360 (vaddudm): Likewise.
11361 (vmaxsd): Likewise.
11362 (vmaxud): Likewise.
11363 (vminsd): Likewise.
11364 (vminud): Likewise.
11365 (vpksdss): Likewise.
11366 (vpksdus): Likewise.
11367 (vpkudum): Likewise.
11368 (vpkudus): Likewise.
11373 (vsubudm): Likewise.
11375 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
11376 support for power8 V2DI instructions.
11378 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
11379 power8 V2DI instructions. Combine pack and unpack insns to use an
11380 iterator for each mode. Check whether a particular mode supports
11381 Altivec instructions instead of just checking TARGET_ALTIVEC.
11382 (UNSPEC_VPKUWUM): Likewise.
11383 (UNSPEC_VPKSHSS): Likewise.
11384 (UNSPEC_VPKSWSS): Likewise.
11385 (UNSPEC_VPKUHUS): Likewise.
11386 (UNSPEC_VPKSHUS): Likewise.
11387 (UNSPEC_VPKUWUS): Likewise.
11388 (UNSPEC_VPKSWUS): Likewise.
11389 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
11390 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
11391 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
11392 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
11393 (UNSPEC_VUPKHSB): Likewise.
11394 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
11395 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
11396 (UNSPEC_VUPKHSH): Likewise.
11397 (UNSPEC_VUPKLSB): Likewise.
11398 (UNSPEC_VUPKLSH): Likewise.
11400 (VI_char): Likewise.
11401 (VI_scalar): Likewise.
11402 (VI_unit): Likewise.
11404 (VP_small): Likewise.
11405 (VP_small_lc): Likewise.
11406 (VU_char): Likewise.
11407 (add<mode>3): Likewise.
11408 (altivec_vaddcuw): Likewise.
11409 (altivec_vaddu<VI_char>s): Likewise.
11410 (altivec_vadds<VI_char>s): Likewise.
11411 (sub<mode>3): Likewise.
11412 (altivec_vsubcuw): Likewise.
11413 (altivec_vsubu<VI_char>s): Likewise.
11414 (altivec_vsubs<VI_char>s): Likewise.
11415 (altivec_vavgs<VI_char>): Likewise.
11416 (altivec_vcmpbfp): Likewise.
11417 (altivec_eq<mode>): Likewise.
11418 (altivec_gt<mode>): Likewise.
11419 (altivec_gtu<mode>): Likewise.
11420 (umax<mode>3): Likewise.
11421 (smax<mode>3): Likewise.
11422 (umin<mode>3): Likewise.
11423 (smin<mode>3): Likewise.
11424 (altivec_vpkuhum): Likewise.
11425 (altivec_vpkuwum): Likewise.
11426 (altivec_vpkshss): Likewise.
11427 (altivec_vpkswss): Likewise.
11428 (altivec_vpkuhus): Likewise.
11429 (altivec_vpkshus): Likewise.
11430 (altivec_vpkuwus): Likewise.
11431 (altivec_vpkswus): Likewise.
11432 (altivec_vpks<VI_char>ss): Likewise.
11433 (altivec_vpks<VI_char>us): Likewise.
11434 (altivec_vpku<VI_char>us): Likewise.
11435 (altivec_vpku<VI_char>um): Likewise.
11436 (altivec_vrl<VI_char>): Likewise.
11437 (altivec_vsl<VI_char>): Likewise.
11438 (altivec_vsr<VI_char>): Likewise.
11439 (altivec_vsra<VI_char>): Likewise.
11440 (altivec_vsldoi_<mode>): Likewise.
11441 (altivec_vupkhsb): Likewise.
11442 (altivec_vupkhs<VU_char>): Likewise.
11443 (altivec_vupkls<VU_char>): Likewise.
11444 (altivec_vupkhsh): Likewise.
11445 (altivec_vupklsb): Likewise.
11446 (altivec_vupklsh): Likewise.
11447 (altivec_vcmpequ<VI_char>_p): Likewise.
11448 (altivec_vcmpgts<VI_char>_p): Likewise.
11449 (altivec_vcmpgtu<VI_char>_p): Likewise.
11450 (abs<mode>2): Likewise.
11451 (vec_unpacks_hi_v16qi): Likewise.
11452 (vec_unpacks_hi_v8hi): Likewise.
11453 (vec_unpacks_lo_v16qi): Likewise.
11454 (vec_unpacks_hi_<VP_small_lc>): Likewise.
11455 (vec_unpacks_lo_v8hi): Likewise.
11456 (vec_unpacks_lo_<VP_small_lc>): Likewise.
11457 (vec_pack_trunc_v8h): Likewise.
11458 (vec_pack_trunc_v4si): Likewise.
11459 (vec_pack_trunc_<mode>): Likewise.
11461 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
11463 (vec_vmaxsd): Likewise.
11464 (vec_vmaxud): Likewise.
11465 (vec_vminsd): Likewise.
11466 (vec_vminud): Likewise.
11467 (vec_vpksdss): Likewise.
11468 (vec_vpksdus): Likewise.
11469 (vec_vpkudum): Likewise.
11470 (vec_vpkudus): Likewise.
11471 (vec_vrld): Likewise.
11472 (vec_vsld): Likewise.
11473 (vec_vsrad): Likewise.
11474 (vec_vsrd): Likewise.
11475 (vec_vsubudm): Likewise.
11476 (vec_vupkhsw): Likewise.
11477 (vec_vupklsw): Likewise.
11479 2013-05-29 Jan Hubicka <jh@suse.cz>
11481 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
11482 flags; reorder rest of fields in more consistent way.
11483 (varpool_node): Remove analyzed, finalized and alias.
11484 (cgraph_ndoe): Likewise.
11485 (symtab_alias_ultimate_target): New function.
11486 (cgraph_function_node): Move offline.
11487 (cgraph_reset_node): Declare.
11488 (cgraph_comdat_can_be_unshared_p): Remove.
11489 (varpool_remove_initializer): Declare.
11490 (varpool_first_defined_variable, varpool_next_defined_variable
11491 cgraph_first_defined_function, cgraph_next_defined_function): Update.
11492 (cgraph_function_with_gimple_body_p): Update.
11493 (varpool_all_refs_explicit_p): Update.
11494 (symtab_alias_target): New function.
11495 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
11496 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
11497 (cgraph_function_or_thunk_node): Simplify using
11498 symtab_alias_ultimate_target.
11499 (varpool_variable_node): Likewise.
11500 * cgraph.c (cgraph_create_function_alias): Update.
11501 (cgraph_add_thunk): Update.
11502 (cgraph_remove_node): Update.
11503 (dump_cgraph_node): Do not dump removed flags.
11504 (cgraph_function_body_availability): Update.
11505 (cgraph_propagate_frequency): Update.
11506 (verify_cgraph_node): Check sanity of local flag.
11507 (cgraph_function_node): Move here from cgraph.h; revamp for
11508 cgraph_function_or_thunk_node.
11509 * lto-symtab.c (lto_varpool_replace_node): Update.
11510 (lto_symtab_resolve_can_prevail_p): Update.
11511 (lto_symtab_merge_cgraph_nodes): Update.
11512 * ipa-cp.c (determine_versionability, initialize_node_lattices,
11513 propagate_constants_accross_call, devirtualization_time_bonus,
11514 ipcp_propagate_stage): Update.
11515 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
11516 * ipa-inline-transform.c (clone_inlined_nodes,
11517 preserve_function_body_p): Update.
11518 * ipa-reference.c (propagate): Update.
11519 (write_node_summary_p): Update.
11520 * toplev.c (wrapup_global_declaration_2): Update.
11521 * cgraphunit.c (cgraph_analyze_function): Rename to ...
11522 (analyze_function) ... this one.
11523 (cgraph_process_new_functions): Update.
11524 (cgraph_reset_node): Export.
11525 (cgraph_finalize_function): Update.
11526 (cgraph_add_new_function): Update.
11527 (process_function_and_variable_attributes): Update.
11528 (varpool_finalize_decl): Update.
11529 (symbol_finalized): Remove.
11530 (symbol_finalized_and_needed): Rename to ...
11531 (symbol_defined_and_needed): ... update.
11532 (cgraph_analyze_functions): Update.
11533 (handle_alias_pairs): Update.
11534 (mark_functions_to_output): Update.
11535 (assemble_thunk): Update.
11536 (output_in_order): Update.
11537 (output_weakrefs): Update.
11538 (finalize_compilation_unit): Update.
11539 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
11540 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
11541 input_node, input_varpool_node): Update.
11542 * dbxout.c (dbxout_expand_expr): Update.
11543 * cgraphclones.c (cgraph_clone_node): Update.
11544 (cgraph_copy_node_for_versioning): Update.
11545 (cgraph_materialize_clone): Update.
11546 (cgraph_materialize_all_clones): Update.
11547 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
11548 propagate_pure_const, propagate_nothrow): Update.
11549 * lto-streamer-out.c (lto_output, write_symbol): Update.
11550 * ipa-utils.c (ipa_reverse_postorder): Update.
11551 * ipa-inline.c (can_inline_edge_p): Update.
11552 (update_caller_keys, ipa_inline): Update.
11553 * dwarf2out.c (reference_to_unused,
11554 premark_types_used_by_global_vars_helper): Update.
11555 * tree-eh.c (tree_could_trap_p): Update.
11556 * ipa-split.c (consider_split, execute_split_functions): Update.
11557 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
11558 has_addr_references_p): Update; move ahead in file for better
11560 (process_references): Simplify.
11561 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
11562 bodies are removed.
11563 (cgraph_comdat_can_be_unshared_p): Make static.
11564 (cgraph_externally_visible_p): Update.
11565 (varpool_externally_visible_p): Update.
11566 (function_and_variable_visibility): Update.
11567 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
11568 ipa_tm_mark_force_output_node): Update.
11569 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
11570 estimate_edge_devirt_benefit, inline_generate_summary,
11571 inline_write_summary): Update.
11572 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
11573 * ipa-prop.c (ipa_compute_jump_functions): Update.
11574 (ipa_print_node_params, ipa_prop_read_section,
11575 ipa_update_after_lto_read, read_replacements_section): Update.
11576 * varasm.c (mark_decl_referenced): Update.
11577 (assemble_alias, dump_tm_clone_pairs): Update.
11578 * tree-inline.c (copy_bb): Update.
11579 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
11581 * symtab.c (dump_symtab_base): Print new flags.
11582 (verify_symtab_base): Verify new flags.
11583 (symtab_alias_ultimate_target): New function.
11584 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
11585 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
11587 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
11589 * i386.c (ix86_get_function_versions_dispatcher,
11590 ix86_generate_version_dispatcher_body): Update.
11591 (fold_builtin_cpu): Use varpool_add_new_variable.
11592 * varpool.c (varpool_remove_initializer): Break out from ...
11593 (varpool_remove_node): ... this one.
11594 (dump_varpool_node, varpool_node_for_asm,
11595 cgraph_variable_initializer_availability, varpool_analyze_node,
11596 varpool_assemble_decl, varpool_remove_unreferenced_decls,
11597 varpool_finalize_named_section_flags, varpool_create_variable_alias):
11600 2013-05-29 Jan Hubicka <jh@suse.cz>
11602 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
11604 2013-05-29 Easwaran Raman <eraman@google.com>
11606 PR tree-optimization/57442
11607 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
11608 when control exits the main loop.
11610 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
11612 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
11614 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
11615 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
11616 * rx/t-rx: Add rx100 under multi library matches option for nofpu
11619 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11621 PR tree-optimization/57441
11622 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
11623 Don't limit size of incr_vec to number of candidates.
11625 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
11627 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
11628 and mips16 directories.
11629 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
11630 (MULTILIB_DIRNAMES): Ditto.
11631 (MULTILIB_EXCEPTIONS): Add new exceptions.
11632 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
11633 (MULTILIB_DIRNAMES): Ditto.
11634 (MULTILIB_EXCEPTIONS): Add new exceptions.
11636 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11637 Marcus Shawcroft <marcus.shawcroft@arm.com>
11639 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
11640 SYMBOL_TINY_ABSOLUTE.
11641 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
11642 SYMBOL_TINY_ABSOLUTE.
11643 (aarch64_expand_mov_immediate): Likewise.
11644 (aarch64_classify_symbol): Likewise.
11645 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
11646 Permit SYMBOL_TINY_ABSOLUTE.
11647 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
11649 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11650 Marcus Shawcroft <marcus.shawcroft@arm.com>
11652 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
11653 Refactor if/switch. Replace gcc_assert with if.
11655 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11657 * config/i386/i386.c (initial_ix86_tune_features): Enable
11658 FP Reassociation for AMD bdver1 and bdver2.
11660 2013-05-29 Martin Jambor <mjambor@suse.cz>
11662 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
11663 and IMAGPART_EXPR do not occur within other handled_components.
11665 2013-05-29 Richard Biener <rguenther@suse.de>
11667 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
11668 access on whether the use is in the BB we currently try to
11670 (vect_bb_vectorization_profitable_p): Pass the BB we currently
11671 vectorize to vect_bb_slp_scalar_cost.
11673 2013-05-29 Richard Biener <rguenther@suse.de>
11675 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
11676 computing scalar cost offsetted by stmts that are kept live
11678 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
11679 for computation of scalar cost.
11681 2013-05-28 Steve Ellcey <sellcey@mips.com>
11683 * config/mips/mips-cpus.def (mips32r2): Change processor type.
11685 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11687 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
11688 array notation built-in reduction functions.
11689 * doc/passes.texi (Passes): Added documentation about changes done
11691 * doc/invoke.texi (C Dialect Options): Added documentation about
11692 the -fcilkplus flag.
11693 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
11694 (BUILTINS_DEF): Depend on cilkplus.def.
11695 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
11696 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
11697 * cilkplus.def: New file.
11699 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
11701 PR rtl-optimization/57439
11702 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
11704 2013-05-28 Easwaran Raman <eraman@google.com>
11706 PR tree-optimization/57337
11707 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
11708 (find_insert_point): Correctly identify the insertion point
11709 when two statements with the same UID is compared.
11711 2013-05-28 Richard Biener <rguenther@suse.de>
11713 PR tree-optimization/56787
11714 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
11715 from the list of data references.
11716 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
11718 (vect_analyze_loop_operations): Likewise.
11719 (vect_transform_loop): Remove clobbers.
11721 2013-05-28 Martin Jambor <mjambor@suse.cz>
11723 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
11724 and REALPART_EXPRs have scalar type.
11726 2013-05-28 Richard Biener <rguenther@suse.de>
11728 PR tree-optimization/57411
11729 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
11731 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
11732 virtual operand propagation.
11734 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11736 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
11737 destination register for bmasksi_vis.
11738 (vector_init_bshuffle): Likewise.
11739 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
11741 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11743 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
11744 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
11745 mode if the instruction isn't available in the original mode.
11746 * config/sparc/sparc.opt (mfix-ut699): New option.
11747 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
11748 (divdf3): Turn into expander.
11749 (divdf3_nofix): New insn.
11750 (divdf3_fix): Likewise.
11751 (divsf3): Disable if -mfix-ut699.
11752 (sqrtdf2): Turn into expander.
11753 (sqrtdf2_nofix): New insn.
11754 (sqrtdf2_fix): Likewise.
11755 (sqrtsf2): Disable if -mfix-ut699.
11757 2013-05-27 Richard Biener <rguenther@suse.de>
11759 PR middle-end/57412
11760 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
11761 block for the new loop.
11763 2013-05-27 Richard Biener <rguenther@suse.de>
11765 PR tree-optimization/57343
11766 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
11767 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
11768 (number_of_iterations_cond): Do not build the folded tree.
11770 2013-05-27 Richard Biener <rguenther@suse.de>
11773 PR middle-end/57381
11774 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11775 OEP_CONSTANT_ADDRESS_OF retained.
11777 PR tree-optimization/57417
11778 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
11779 for unchanged base.
11780 (set_ssa_val_to): Compare addresses using
11781 get_addr_base_and_unit_offset.
11783 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
11785 PR rtl-optimization/56833
11786 * postreload.c (move2add_record_mode): New function.
11787 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
11788 (move2add_use_add2_insn): Use move2add_record_sym_value.
11789 (move2add_use_add3_insn): Likewise.
11790 (reload_cse_move2add): Use move2add_valid_value_p and
11791 move2add_record_mode. Invalidate call-clobbered and REG_INC
11792 affected regs by setting reg_mode to VOIDmode.
11793 (move2add_note_store): Don't pretend the inside of a SUBREG is
11794 the actual destination. Invalidate single/leading registers by
11795 setting reg_mode to VOIDmode.
11796 Use move2add_record_sym_value, move2add_valid_value_p and
11797 move2add_record_mode.
11799 2013-05-27 Richard Biener <rguenther@suse.de>
11801 PR tree-optimization/57396
11802 * tree-affine.c (double_int_constant_multiple_p): Properly
11803 return false for val == 0 and div != 0.
11805 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11807 * config/mips/mips.h: Use #elif in preprocessor conditions.
11809 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11812 * config/mips/constraints.md (kl): New constraint.
11813 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
11814 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
11815 constraint for operand 0. Split after CSE for MIPS16. Emit a move
11816 from LO for MIPS16.
11817 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
11819 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11822 * config/mips/mips.c (mips_can_inline_p): New function.
11823 (TARGET_CAN_INLINE_P): Define.
11825 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
11827 * sched-int.h (ds_t, dw_t): Make unsigned int.
11828 Fix documentation that describes how all the ds_t bits are used.
11829 Reserve the last bit for delayed-branch scheduling.
11830 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
11831 (BITS_PER_DEP_WEAK): Fix definition and documentation.
11832 (gen_dep_weak_1): Remove prototype.
11833 * sched-deps.c (get_dep_weak_1): Make static.
11834 * target.def (speculate_insn, needs_block_p, gen_spec_check,
11835 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
11836 * doc/tm.texi: Regenerate.
11837 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
11839 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
11842 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
11844 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
11845 Sandra Loosemore <sandra@codesourcery.com>
11847 * config.gcc (powerpc-*): Allow native for with-cpu.
11849 2013-05-24 Jeff Law <law@redhat.com>
11851 PR tree-optimization/57124
11852 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
11853 conversion feeding a condition if the range has an overflow
11854 if -fstrict-overflow. Add warnings for when we do make the
11857 2013-05-24 Dehao Chen <dehao@google.com>
11859 * tree-cfg.c (locus_discrim_map): Fix the typo.
11860 (locus_discrim_hasher): Likewise.
11861 (locus_discrim_hasher::hash): Likewise.
11862 (locus_discrim_hasher::equal): Likewise.
11864 2013-05-24 Martin Jambor <mjambor@suse.cz>
11866 PR tree-optimization/57294
11867 * cgraph.h (ipa_record_stmt_references): Declare.
11868 * cgraphbuild.c (ipa_record_stmt_references): New function.
11869 (build_cgraph_edges): Use ipa_record_stmt_references.
11870 (rebuild_cgraph_edges): Likewise.
11871 (cgraph_rebuild_references): Likewise.
11872 * ipa-prop.c (ipa_modify_call_arguments): Discard references
11873 associated with the old statement and build references from the
11874 newly built statements.
11875 * ipa-ref.c (ipa_remove_stmt_references): New function.
11876 * ipa-ref.h (ipa_remove_stmt_references): Declare.
11878 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
11880 * lra-constraints.c (emit_spill_move): Use smaller mode for
11882 (check_and_process_move): Consider mem-reg moves for secondary
11884 (curr_insn_transform): Don't lose insns emitted before for
11885 secondary memory moves.
11886 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
11887 reg set up in the current insn.
11889 2013-05-24 Dehao Chen <dehao@google.com>
11891 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
11893 (locus_descrim_hasher::equal): Likewise.
11894 (build_gimple_cfg): New discriminator assignment algorithm.
11895 (make_edges): Likewise.
11896 (next_discriminator_for_locus): Likewise.
11897 (same_line_p): Likewise.
11898 (assign_discriminators): Likewise.
11899 (make_cond_expr_edges): Likewise.
11900 (make_gimple_switch_edges): Likewise.
11901 (make_goto_expr_edges): Likewise.
11902 (make_gimple_asm_edges): Likewise.
11904 2013-05-24 Ian Bolton <ian.bolton@arm.com>
11906 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
11907 X format specifier to only display bottom 16 bits.
11908 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
11909 immediate to match for operand 2, since it will be masked.
11911 2013-05-24 Richard Biener <rguenther@suse.de>
11913 PR tree-optimization/57287
11914 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
11915 all SSA names that occur in abnormal PHIs.
11917 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
11920 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
11921 that index is not negative.
11923 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
11925 PR rtl-optimization/55177
11926 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
11927 (simplify_byte_swapping_operation): New.
11928 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
11929 (simplify_relational_operation_1): Deal with BSWAP.
11931 2013-05-23 Richard Henderson <rth@redhat.com>
11934 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
11935 (ix86_reorg): Call it.
11937 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
11940 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
11941 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
11942 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
11944 2013-05-23 Christian Bruel <christian.bruel@st.com>
11947 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
11949 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11950 Marcus Shawcroft <marcus.shawcroft@arm.com>
11952 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
11953 * config/aarch64/constraints.md (Usa): Remove.
11954 * doc/md.texi (AArch64 Usa): Remove.
11956 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11957 Marcus Shawcroft <marcus.shawcroft@arm.com>
11959 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
11960 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
11961 * config/aarch64/predicates.md (aarch64_const_address): Remove.
11962 (aarch64_mov_operand): Use aarch64_mov_operand_p.
11964 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
11966 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
11967 instruction (AdvSIMD).
11968 * config/aarch64/aarch64-builtins.c
11969 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
11970 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
11972 2013-05-23 Martin Jambor <mjambor@suse.cz>
11974 PR middle-end/57347
11975 * tree.h (contains_bitfld_component_ref_p): Declare.
11976 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
11977 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
11979 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
11980 not access a bit-field. Assert all final offsets are byte-aligned.
11982 2013-05-23 Richard Biener <rguenther@suse.de>
11984 PR tree-optimization/57380
11985 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
11986 least one invariant or re-used load.
11987 * passes.c (init_optimization_passes): Move pass_phiprop before
11990 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
11992 * config/aarch64/aarch64-simd.md
11993 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
11995 2013-05-23 Richard Biener <rguenther@suse.de>
11997 PR middle-end/57381
11998 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11999 OEP_CONSTANT_ADDRESS_OF retained.
12001 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12003 PR middle-end/57344
12004 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12005 don't lower unit. Handle unit not being always BITS_PER_WORD.
12007 2013-05-23 Richard Biener <rguenther@suse.de>
12009 PR rtl-optimization/57341
12010 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12011 instead of true_dependence.
12013 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12015 * bb-reorder.c (branch_threshold): Make const.
12016 (exec_threshold): Ditto.
12018 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12019 Pat Haugen <pthaugen@us.ibm.com>
12020 Peter Bergner <bergner@vnet.ibm.com>
12022 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12023 documentation for the power8 crypto builtins.
12025 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12027 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12028 macros for defining power8 builtin functions.
12029 (BU_P8V_AV_2): Likewise.
12030 (BU_P8V_AV_P): Likewise.
12031 (BU_P8V_VSX_1): Likewise.
12032 (BU_P8V_OVERLOAD_1): Likewise.
12033 (BU_P8V_OVERLOAD_2): Likewise.
12034 (BU_CRYPTO_1): Likewise.
12035 (BU_CRYPTO_2): Likewise.
12036 (BU_CRYPTO_3): Likewise.
12037 (BU_CRYPTO_OVERLOAD_1): Likewise.
12038 (BU_CRYPTO_OVERLOAD_2): Likewise.
12039 (XSCVSPDP): Fix typo, point to the correct instruction.
12040 (VCIPHER): Add power8 crypto builtins.
12041 (VCIPHERLAST): Likewise.
12042 (VNCIPHER): Likewise.
12043 (VNCIPHERLAST): Likewise.
12044 (VPMSUMB): Likewise.
12045 (VPMSUMH): Likewise.
12046 (VPMSUMW): Likewise.
12047 (VPERMXOR_V2DI): Likewise.
12048 (VPERMXOR_V4SI: Likewise.
12049 (VPERMXOR_V8HI: Likewise.
12050 (VPERMXOR_V16QI: Likewise.
12051 (VSHASIGMAW): Likewise.
12052 (VSHASIGMAD): Likewise.
12053 (VPMSUM): Likewise.
12054 (VPERMXOR): Likewise.
12055 (VSHASIGMA): Likewise.
12057 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12058 __CRYPTO__ if the crypto instructions are available.
12059 (altivec_overloaded_builtins): Add support for overloaded power8
12062 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12063 support for power8 crypto builtins.
12064 (builtin_function_type): Likewise.
12065 (altivec_init_builtins): Add support for builtins that take vector
12066 long long (V2DI) arguments.
12068 * config/rs6000/crypto.md: New file, define power8 crypto
12071 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12072 Pat Haugen <pthaugen@us.ibm.com>
12073 Peter Bergner <bergner@vnet.ibm.com>
12075 * doc/invoke.texi (Option Summary): Add power8 options.
12076 (RS/6000 and PowerPC Options): Likewise.
12078 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12079 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12080 wm, wn, wr documentation.
12082 * config/rs6000/constraints.md (wm): New constraint for VSX
12083 registers if direct move instructions are enabled.
12084 (wn): New constraint for no registers.
12085 (wq): New constraint for quad word even GPR registers.
12086 (wr): New constraint if 64-bit instructions are enabled.
12087 (wv): New constraint if power8 vector instructions are enabled.
12088 (wQ): New constraint for quad word memory locations.
12090 * config/rs6000/predicates.md (const_0_to_15_operand): New
12091 constraint for 0..15 for crypto instructions.
12092 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12093 as GPR and floating point registers.
12094 (int_reg_operand): New predicate to match only GPR registers.
12095 (base_reg_operand): New predicate to match base registers.
12096 (quad_int_reg_operand): New predicate to match even GPR registers
12097 for quad memory operations.
12098 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12099 operations in both GPR and VSX registers.
12100 (quad_memory_operand): New predicate for quad memory operations.
12101 (reg_or_indexed_operand): New predicate for direct move support.
12103 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12104 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12105 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12106 (POWERPC_MASKS): Add power8 options.
12107 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12110 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12111 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12113 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12114 (-mpower8-fusion): New power8 options.
12115 (-mpower8-fusion-sign): Likewise.
12116 (-mpower8-vector): Likewise.
12117 (-mcrypto): Likewise.
12118 (-mdirect-move): Likewise.
12119 (-mquad-memory): Likewise.
12121 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12122 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12124 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12125 (rs6000_debug_vector_unit): Add p8_vector.
12126 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12127 definitions. Also print fusion state.
12128 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12129 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12130 (rs6000_option_override_internal): Add support for power8.
12131 (rs6000_common_init_builtins): Add debugging for skipped builtins
12132 if -mdebug=builtin.
12133 (rs6000_adjust_cost): Add power8 support.
12134 (rs6000_issue_rate): Likewise.
12135 (insn_must_be_first_in_group): Likewise.
12136 (insn_must_be_last_in_group): Likewise.
12137 (force_new_group): Likewise.
12138 (rs6000_register_move_cost): Likewise.
12139 (rs6000_opt_masks): Likewise.
12141 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12142 power8 capable assembler, default to power7 options.
12143 (TARGET_DIRECT_MOVE): Likewise.
12144 (TARGET_CRYPTO): Likewise.
12145 (TARGET_P8_VECTOR): Likewise.
12146 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12147 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12148 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12149 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12150 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12151 (TARGET_XSCVDPSPN): Likewise.
12152 (TARGET_XSCVSPDPN): Likewsie.
12153 (TARGET_SYNC_HI_QI): Likewise.
12154 (TARGET_SYNC_TI): Likewise.
12155 (MASK_CRYPTO): Likewise.
12156 (MASK_DIRECT_MOVE): Likewise.
12157 (MASK_P8_FUSION): Likewise.
12158 (MASK_P8_VECTOR): Likewise.
12159 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12160 temporary used by some of the direct move instructions to get two FP
12161 temporary registers does not force creation of a stack frame.
12162 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12163 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12164 that any VSX registers are tieable, even if they are also an
12165 Altivec vector mode.
12166 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12167 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12168 (RS6000_BTM_CRYPTO): Likewise.
12169 (RS6000_BTM_COMMON): Likewise.
12171 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12172 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12173 (enum rs6000_vector): Add power8 vector support.
12175 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12179 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12180 (any_sibcall_could_use_r3): this and handle indirect calls.
12181 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12183 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12185 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12187 2013-05-22 Richard Biener <rguenther@suse.de>
12189 PR middle-end/57349
12190 * profile.c (branch_prob): Do not split blocks that are
12191 abnormally receiving from ECF_RETURNS_TWICE functions.
12193 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12195 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12196 address mode. Move pointer mode initialization to the same place.
12198 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12200 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12201 while it has any effect.
12203 2013-05-21 Easwaran Raman <eraman@google.com>
12205 PR tree-optimization/57322
12206 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12207 UID of the statement added to the BB to be 1.
12209 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12211 PR tree-optimization/57331
12212 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12213 of conversion from pointer type to integral type with integer.
12215 2013-05-21 Martin Jambor <mjambor@suse.cz>
12218 * ipa-prop.c (ipa_read_node_info): Process param_used and
12219 controlled_uses in the same order as when writing.
12221 2013-05-21 Magnus Granberg <baldrick@free.fr>
12224 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12226 2013-05-21 Richard Biener <rguenther@suse.de>
12228 PR tree-optimization/57318
12229 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12230 estimate stmts with side-effects as likely eliminated.
12232 2013-05-21 Richard Biener <rguenther@suse.de>
12234 PR tree-optimization/57330
12235 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12236 preserve the call stmts fntype.
12238 2013-05-21 Richard Biener <rguenther@suse.de>
12240 PR tree-optimization/57303
12241 * tree-ssa-sink.c (statement_sink_location): Improve killing
12242 stmt detection and properly handle self-assignments.
12244 2013-05-21 Christian Bruel <christian.bruel@st.com>
12246 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12247 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12248 registers. Set register size out of the PARALLEL loop.
12250 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12253 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12254 (*fmasf4, *fmasf4_media): New insns.
12256 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12258 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12259 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12260 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12261 (mips_idiv_insns): Update the comments to say that the returned
12262 instruction counts are in units of BASE_INSN_LENGTH.
12263 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12264 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12265 using 2 rather than 4 as the length of indirect MIPS16 and
12266 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12267 length of a NOP. Don't divide MIPS16 lengths by 2.
12268 (mips16_split_long_branches): Assume a branch is long if the
12269 length is greater than 4 rather than 8.
12270 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12271 rather than multiplying them by 2. Multiply instruction counts
12272 by BASE_INSN_LENGTH rather than 4.
12273 (*jump_mips16, tls_get_tp_mips16_<mode>)
12274 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12276 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12278 * config/mips/mips.md (extended_mips16): Remove branch case.
12279 (length): Remove duplicated extended_mips16 test.
12281 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12283 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12285 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12287 * recog.h (Recog_data): Rename to...
12288 (recog_data_d): ...this.
12289 (recog_data): Update accordingly.
12290 * recog.c (recog_data): Likewise.
12291 * reload.c (save_recog_data): Likewise.
12292 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12293 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12295 2013-05-17 Julian Brown <julian@codesourcery.com>
12297 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12298 found in a REG_EQUAL note, invalidate it.
12300 2013-05-17 Easwaran Raman <eraman@google.com>
12302 * tree-ssa-reassoc.c (find_insert_point): New function.
12303 (insert_stmt_after): Likewise.
12304 (get_def_stmt): Likewise.
12305 (ensure_ops_are_available): Likewise.
12306 (not_dominated_by): Likewise.
12307 (rewrite_expr_tree): Do not move statements beyond what is
12308 necessary. Remove call to swap_ops_for_binary_stmt...
12309 (reassociate_bb): ... and move it here.
12310 (build_and_add_sum): Assign UIDs for new statements.
12311 (linearize_expr): Likewise.
12312 (do_reassoc): Renumber gimple statement UIDs.
12314 2013-05-17 Jan Hubicka <jh@suse.cz>
12316 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12318 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12319 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12321 (output_weakrefs): Update.
12323 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12324 Martin Jambor <mjambor@suse.cz>
12326 PR middle-end/57276
12327 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12328 value that corresponds to the given aggval is found in values vector.
12330 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12332 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12333 sse, sse2, sse3, ssse3 and sse4a flags to options.
12335 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12337 * gengtype-state.c: (s_expr_writer): New class, to handle
12338 prettifying of output layout of s-expressions.
12339 (state_writer): New class, to write out gtype.state.
12340 (state_written_type_count): Move this variable into member data of
12342 (s_expr_writer::s_expr_writer): New code: constructor for new class
12343 (state_writer::state_writer(): ditto
12344 (s_expr_writer::write_new_line): New function
12345 (s_expr_writer::write_any_indent): ditto
12346 (s_expr_writer::begin_s_expr): ditto
12347 (s_expr_writer::end_s_expr): ditto
12348 (write_state_fileloc): convert to method of state_writer...
12349 (state_writer:: write_state_fileloc): ...and use methods of
12350 s_expr_writer to write indentation into the gtype.state output file
12351 to visually represent the hierarchical structure of the list
12353 (write_state_fields): ditto, renaming to...
12354 (state_writer::write_state_fields)
12355 (write_state_a_string): ditto, renaming to...
12356 (state_writer::write_state_a_string)
12357 (write_state_string_option): ditto, renaming to...
12358 (state_writer::write_state_string_option)
12359 (write_state_type_option): ditto, renaming to...
12360 (state_writer::write_state_type_option)
12361 (write_state_nested_option): ditto, renaming to...
12362 (state_writer::write_state_nested_option)
12363 (write_state_option): ditto, renaming to...
12364 (state_writer::write_state_option)
12365 (write_state_options): ditto, renaming to...
12366 (state_writer::write_state_options)
12367 (write_state_lang_bitmap): ditto, renaming to...
12368 (state_writer::write_state_lang_bitmap)
12369 (write_state_version): ditto, renaming to...
12370 (state_writer::write_state_version)
12371 (write_state_scalar_type): ditto, renaming to...
12372 (state_writer::write_state_scalar_type)
12373 (write_state_string_type): ditto, renaming to...
12374 (state_writer::write_state_string_type)
12375 (write_state_undefined_type): ditto, renaming to...
12376 (state_writer::write_state_undefined_type)
12377 (write_state_struct_union_type): ditto, renaming to...
12378 (state_writer::write_state_struct_union_type)
12379 (write_state_struct_type): ditto, renaming to...
12380 (state_writer::write_state_struct_type)
12381 (write_state_user_struct_type): ditto, renaming to...
12382 (state_writer::write_state_user_struct_type)
12383 (write_state_lang_struct_type): ditto, renaming to...
12384 (state_writer::write_state_lang_struct_type)
12385 (write_state_param_struct_type): ditto, renaming to...
12386 (state_writer::write_state_param_struct_type)
12387 (write_state_pointer_type): ditto, renaming to...
12388 (state_writer::write_state_pointer_type)
12389 (write_state_array_type): ditto, renaming to...
12390 (state_writer::write_state_array_type)
12391 (write_state_gc_used): ditto, renaming to...
12392 (state_writer::write_state_gc_used)
12393 (write_state_common_type_content): ditto, renaming to...
12394 (state_writer::write_state_common_type_content)
12395 (write_state_type): ditto, renaming to...
12396 (state_writer::write_state_type)
12397 (write_state_pair_list): ditto, renaming to...
12398 (state_writer::write_state_pair_list)
12399 (write_state_pair): ditto, renaming to...
12400 (state_writer::write_state_pair)
12401 (write_state_typedefs): ditto, renaming to...
12402 (state_writer::write_state_typedefs)
12403 (write_state_structures): ditto, renaming to...
12404 (state_writer::write_state_structures)
12405 (write_state_param_structs): ditto, renaming to...
12406 (state_writer::write_state_param_structs)
12407 (write_state_variables): ditto, renaming to...
12408 (state_writer::write_state_variables)
12409 (write_state_srcdir): ditto, renaming to...
12410 (state_writer::write_state_srcdir)
12411 (write_state_files_list): ditto, renaming to...
12412 (state_writer::write_state_files_list)
12413 (write_state_languages): ditto, renaming to...
12414 (state_writer::write_state_languages)
12415 (write_state): create a state_writer instance and use it when
12416 writing out the state file
12418 2013-05-17 Mike Stump <mikestump@comcast.net>
12420 PR rtl-optimization/57304
12421 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
12422 accessing DF_REF_REAL_LOC.
12424 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12426 PR rtl-optimization/57281
12427 PR rtl-optimization/57300
12428 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
12429 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
12430 what the other splitter did if the registers are dead.
12432 2013-05-17 Richard Biener <rguenther@suse.de>
12434 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
12437 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12439 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
12442 2013-05-17 Marek Polacek <polacek@redhat.com>
12444 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
12445 length when doing non-zero store of storing '\0' to '\0'.
12447 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12449 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
12450 vect_external_def oprnd1 with loop_vinfo, try to emit
12451 optional cast, negation and and stmts on the loop preheader
12452 edge instead of into the pattern def seq.
12454 PR tree-optimization/57051
12455 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
12456 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
12458 2013-05-16 Nick Clifton <nickc@redhat.com>
12460 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
12461 (rl78_is_naked_func): New function.
12462 (rl78_expand_prologue): Skip prologue generation for naked functions.
12463 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
12464 * doc/extend.texi (naked): Add RL78 to the list of processors
12465 that supports this attribute.
12467 2013-05-16 Jeff Law <law@redhat.com>
12469 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
12471 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12473 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
12474 cache parameters using detect_caches_amd also for CYRIX,
12475 NSC and TM2 signatures.
12477 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12478 Dzianis Kahanovich <mahatma@eu.by>
12482 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12483 VIA/Centaur processors and determine their cache parameters
12484 using detect_caches_amd.
12486 2013-05-16 Teresa Johnson <tejohnson@google.com>
12488 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
12489 (rtl_verify_edges): New function.
12490 (rtl_verify_bb_insns): Ditto.
12491 (rtl_verify_bb_pointers): Ditto.
12492 (rtl_verify_bb_insn_chain): Ditto.
12493 (rtl_verify_fallthru): Ditto.
12494 (rtl_verify_bb_layout): Ditto.
12495 (rtl_verify_flow_info_1): Outline checks into new functions.
12496 (rtl_verify_flow_info): Ditto.
12498 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
12500 * cfghooks.c (copy_bbs): Add update_dominance argument.
12501 * cfghooks.h (copy_bbs): Update prototype.
12502 * tree-cfg.c (gimple_duplicate_sese_region):
12503 Add update_dominance argument.
12504 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
12505 * tree-ssa-loop-ch.c (copy_loop_headers): Update
12506 gimple_duplicate_sese_region call.
12507 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
12508 Update copy_bbs call.
12509 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
12510 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
12512 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12514 * tree-vectorizer.h (NUM_PATTERNS): Increment.
12515 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
12516 vect_recog_rotate_pattern.
12517 (vect_recog_rotate_pattern): New function.
12519 2013-05-16 Jason Merrill <jason@redhat.com>
12521 * Makefile.in (LLINKER): New variable.
12522 (mostlyclean): Remove link mutex.
12523 * configure.ac: Handle --enable-link-mutex.
12524 * lock-and-run.sh: New script.
12526 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12529 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
12532 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12534 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
12536 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
12538 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
12539 * config/arm/arm.c (next_consecutive_mem): New function.
12540 (gen_movmem_ldrd_strd): Likewise.
12541 * config/arm/arm.md (movmemqi): Update condition and code.
12542 (unaligned_loaddi, unaligned_storedi): New patterns.
12544 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12546 * config.gcc: Obsolete *-*-solaris2.9*.
12547 * doc/install.texi (Specific, *-*-solaris2*): Document it.
12549 2013-05-16 Richard Biener <rguenther@suse.de>
12551 * passes.c (init_optimization_passes): Move pass_parallelize_loops
12552 earlier, after GRAPHITE transforms and IV canonicalization.
12554 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12556 * omp-low.c (extract_omp_for_data): For collapsed loops,
12557 if at least one of the loops is known at compile time to
12558 iterate zero times, set count to 0.
12559 (expand_omp_regimplify_p): New function.
12560 (expand_omp_for_generic): For collapsed loops, if at least
12561 one of the loops isn't known to iterate at least once,
12562 add runtime check with setting count to 0.
12563 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
12564 For unsigned types if it isn't known at compile time that
12565 the loop will iterate at least once, add runtime check to bypass
12566 the whole loop if initial condition isn't true.
12568 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
12570 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
12572 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
12574 PR middle-end/57286
12575 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
12576 transformations to avoid an infinite loop.
12578 2013-05-16 Marek Polacek <polacek@redhat.com>
12580 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
12582 2013-05-15 Leif Ekblad <leif@rdos.net>
12584 * config/i386/i386.c (ix86_decompose_address): Use
12585 DEFAULT_TLS_SEG_REG to access TLS segment register.
12586 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
12587 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
12588 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
12590 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
12593 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
12594 sibling calls to functions that would normally be lazily bound,
12595 unless $gp is call-clobbered.
12597 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
12599 * config/i386/i386.c (ix86_option_override_internal): Update
12600 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
12601 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
12602 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
12603 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
12605 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
12607 2013-05-15 Andreas Schwab <schwab@suse.de>
12609 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
12610 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
12613 2013-05-15 Teresa Johnson <tejohnson@google.com>
12615 * loop-unroll.c (report_unroll_peel): Check decision before
12616 emitting unroll/peel message.
12618 2013-05-15 Teresa Johnson <tejohnson@google.com>
12620 * function.h (has_bb_partition): New rtl_data flag.
12621 (bb_reorder_complete): Ditto.
12622 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
12623 instead of flag_reorder_blocks_and_partition.
12624 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
12625 with some enhancements.
12626 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
12627 * bb-reorder.c (connect_traces): Check for has_bb_partition
12628 instead of flag_reorder_blocks_and_partition.
12629 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
12630 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
12631 verify_hot_cold_block_grouping.
12632 (partition_hot_cold_basic_blocks): Set has_bb_partition.
12634 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12637 * config/arm/predicates.md (call_insn_operand): New predicate.
12638 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
12639 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
12640 if insn is not a tail call.
12641 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
12643 * config/arm/arm.h (enum reg_class): New caller save register class.
12644 (REG_CLASS_NAMES): Likewise.
12645 (REG_CLASS_CONTENTS): Likewise.
12646 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
12649 2013-05-15 Richard Biener <rguenther@suse.de>
12651 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
12652 of MSG_OPTIMIZED_LOCATIONS.
12653 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
12654 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
12656 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
12657 of MSG_OPTIMIZED_LOCATIONS.
12658 (execute_vect_slp): Likewise.
12659 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
12660 (vect_create_cond_for_alias_checks): Likewise.
12661 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
12662 (vect_recog_widen_mult_pattern): Likewise.
12663 (vect_recog_widen_sum_pattern): Likewise.
12664 (vect_recog_over_widening_pattern): Likewise.
12665 (vect_recog_widen_shift_pattern): Likewise.
12666 (vect_recog_vector_vector_shift_pattern): Likewise.
12667 (vect_recog_divmod_pattern): Likewise.
12668 (vect_recog_mixed_size_cond_pattern): Likewise.
12669 (vect_recog_bool_pattern): Likewise.
12670 (vect_pattern_recog_1): Likewise.
12672 2013-05-15 Martin Jambor <mjambor@suse.cz>
12674 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
12675 non-functions to builtin_unreachable.
12676 * ipa-inline-transform.c (inline_call): Do not assert estimates were
12677 correct when new direct edges were discovered.
12679 2013-05-15 Martin Jambor <mjambor@suse.cz>
12681 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
12682 header, print symbol order instead of node uid, print more information
12683 about indirect edge targets.
12684 (ipa_make_edge_direct_to_target): Print symbol order instead of node
12686 (ipa_make_edge_direct_to_target): Likewise.
12687 (remove_described_reference): Likewise.
12688 (propagate_controlled_uses): Likewise.
12689 (ipa_print_node_params): Also print symbol order.
12690 (ipcp_transform_function): Print symbol order instead of node uids.
12691 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
12692 (cgraph_get_create_real_symbol_node): Likewise.
12693 * ipa-cp.c (print_lattice): Likewise.
12694 (print_all_lattices): Likewise.
12695 (determine_versionability): Likewise.
12696 (initialize_node_lattices): Likewise.
12697 (estimate_local_effects): Likewise.
12698 (update_profiling_info): Likewise.
12699 (create_specialized_node): Likewise.
12700 (perhaps_add_new_callers): Likewise.
12701 (decide_about_value): Likewise.
12702 (decide_whether_version_node): Likewise.
12703 (identify_dead_nodes): Likewise.
12704 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
12705 (dump_inline_summary): Likewise.
12706 (estimate_node_size_and_time): Likewise.
12707 (inline_analyze_function): Likewise.
12708 * ipa-inline.c (report_inline_failed_reason): Likewise.
12709 (want_early_inline_function_p): Likewise.
12710 (edge_badness): Likewise.
12711 (update_edge_key): Likewise.
12712 (inline_small_functions): Likewise. Add dumping of order to two other
12714 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
12715 instead of node uids.
12716 (propagate_pure_const): Likewise.
12717 (propagate_pure_const): Likewise.
12718 * ipa-utils.c (dump_cgraph_node_set): Likewise.
12719 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
12720 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
12722 * tree-pretty-print.c (dump_function_header): Likewise.
12723 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
12724 Print symbol order instead of node uids.
12726 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12728 * config/s390/s390.c (s390_register_move_cost): Don't impose the
12729 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
12731 2013-05-15 Richard Biener <rguenther@suse.de>
12733 PR tree-optimization/57275
12734 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
12735 return value for fail to do runtime alias checks for gather loads.
12737 2013-05-15 Jan Hubicka <jh@suse.cz>
12741 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
12742 weakrefs are not external.
12743 (lto_symtab_merge_decls): Fix thinko when dealing with
12744 non-lto_symtab decls.
12745 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
12746 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
12747 * varpool.c (dump_varpool_node): Dump more flags.
12749 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12751 * config/i386/i386.c (processor_alias_table): Add instruction
12752 FSGSBASE for AMD bdver3 architecture.
12754 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12756 * tree.c (warn_deprecated_use): Print file:line using locus color.
12757 * diagnostic.c (diagnostic_report_current_module): Print file:line
12758 and file:line:column using locus color.
12760 2013-05-14 Mike Stump <mikestump@comcast.net>
12762 * gdbinit.in: Add __null.
12764 2013-05-14 Mike Stump <mikestump@comcast.net>
12766 * recog.h: Rename struct recog_data to Recog_data.
12767 * recog.c: Likewise.
12768 * reload.c (can_reload_into): Likewise.
12769 * config/picochip/picochip.c: Likewise.
12771 2013-05-14 Mike Stump <mikestump@comcast.net>
12773 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
12775 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
12777 * resource.h (struct resources): Remove unch_memory member.
12778 (CLEAR_RESOURCE): Don't clear unch_memory.
12779 * resource.c (mark_referenced_resources): Don't set it.
12780 (mark_set_resources): Likewise.
12781 (mark_target_live_regs): Don't clear it.
12782 (init_resource_info): Likewise.
12783 * reorg.c (resource_conflicts_p): Don't compare it.
12784 (redundant_insn): Don't set it.
12786 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
12788 * emit-rtl.c (next_label): Remove unused function.
12789 (skip_consecutive_labels, link_cc0_insns): Move to ...
12790 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
12791 only place where these functions are used, and make them static.
12793 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
12795 * fold-const.c (fold_negate_expr): Handle vectors.
12796 (fold_truth_not_expr): Make it static.
12797 (fold_invert_truthvalue): New static function.
12798 (invert_truthvalue_loc): Handle vectors. Do not call
12799 fold_truth_not_expr directly.
12800 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
12801 <TRUTH_NOT_EXPR>: Do not cast to boolean.
12802 (fold_comparison): Handle vector constants.
12803 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
12804 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
12805 * tree.h (fold_truth_not_expr): Remove declaration.
12807 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
12809 * config/aarch64/aarch64-simd.md
12810 (aarch64_vcond_internal<mode>): Rename to...
12811 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
12812 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
12813 float modes. Clarify all iterator modes.
12814 (vcond<mode><mode>): Use new name for vcond expanders.
12815 (vcond<v_cmp_result><mode>): Likewise.
12816 (vcondu<mode><mode>: Likewise.
12817 * config/aarch64/iterators.md (VDQF_COND): New.
12819 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
12822 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
12823 variable for the shift amount. Check that we shift by non-negative
12826 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
12829 * config/arm/arm.h (EPILOGUE_USES): Only return true
12830 for LR_REGNUM after epilogue_completed.
12832 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
12834 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
12835 is error_mark_node.
12837 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12840 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
12841 and Solaris 11+/x86 with gld.
12842 * configure: Regenerate.
12844 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12846 * expmed.c (expand_shift_1): Canonicalize rotates by
12847 constant bitsize / 2 to bitsize - 1.
12848 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
12849 case ROTATERT>: Likewise.
12852 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12854 * config/i386/i386.md (rotateinv): New code attr.
12855 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12856 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12857 roll $31, %eax, etc.
12859 2013-05-14 Richard Biener <rguenther@suse.de>
12861 PR middle-end/57235
12862 * tree-eh.c (sink_clobbers): Give up for successors with
12863 multiple predecessors and no virtual uses.
12865 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
12867 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
12868 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
12870 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12872 PR middle-end/57251
12873 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
12874 the case when both op0 and op1 have VOIDmode.
12876 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12878 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
12879 in multiply-accumulate mode.
12881 2013-05-13 Guozhi Wei <carrot@google.com>
12883 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
12885 2013-05-13 Kai Tietz <ktietz@redhat.com>
12888 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
12889 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
12890 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
12891 * config/i386/i386.c (ix86_option_override_internal): Likewise.
12892 (ix86_expand_prologue): Likewise.
12893 (ix86_expand_split_stack_prologue): Likewise.
12894 (legitimate_pic_address_disp_p): Likewise.
12895 (legitimize_pic_address): Likewise.
12896 (legitimize_tls_address): Likewise.
12897 (legitimize_pe_coff_symbol): Likewise.
12898 (output_pic_addr_const): Likewise.
12899 (construct_plt_address): Likewise.
12900 (ix86_expand_call): Likewise.
12901 (x86_output_mi_thunk): Likewise.
12902 (x86_function_profiler): Likewise.
12904 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
12906 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
12907 similar switch cases.
12908 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
12909 (aarch64_simd_mov_to_<mode>low): Delete.
12910 (aarch64_simd_mov_to_<mode>high): Delete.
12911 (move_lo_quad_<mode>): Add w<-r alternative.
12912 (aarch64_simd_move_hi_quad_<mode>): Likewise.
12913 (aarch64_simd_mov_from_*): Update type attribute.
12914 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
12917 2013-05-13 Jan Hubicka <jh@suse.cz>
12919 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
12920 * config/i386/i386.c (ix86_compute_frame_layout,
12921 ix86_expand_epilogue, emit_i387_cw_initialization,
12922 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
12923 ix86_local_alignment): Fix use of size/speed predicates.
12925 2013-05-13 Jakub Jelinek <jakub@redhat.com>
12927 PR tree-optimization/45216
12928 PR tree-optimization/57157
12929 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
12930 the (-Y) & (B - 1) variant if OP is |.
12931 * expmed.c (expand_shift_1): For rotations by const0_rtx just
12932 return shifted. Use (-op1) & (prec - 1) as other_amount
12933 instead of prec - op1.
12935 2013-05-13 Martin Jambor <mjambor@suse.cz>
12937 PR middle-end/42371
12938 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
12939 (ipa_constant_data): New type.
12940 (ipa_jump_func): Use ipa_constant_data to hold information about
12941 constant jump functions.
12942 (ipa_get_jf_constant): Adjust to jump function type changes.
12943 (ipa_get_jf_constant_rdesc): New function.
12944 (ipa_param_descriptor): New field controlled_uses.
12945 (ipa_get_controlled_uses): New function.
12946 (ipa_set_controlled_uses): Likewise.
12947 * ipa-ref.h (ipa_find_reference): Declare.
12948 * ipa-prop.c (ipa_cst_ref_desc): New type.
12949 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
12951 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
12952 New parameter cs. Adjust all callers.
12953 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
12954 (remove_described_reference): New function.
12955 (jfunc_rdesc_usable): Likewise.
12956 (try_make_edge_direct_simple_call): Decrement controlled use count,
12957 attempt to remove reference if it hits zero.
12958 (combine_controlled_uses_counters): New function.
12959 (propagate_controlled_uses): Likewise.
12960 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
12961 (ipa_edge_duplication_hook): Duplicate reference descriptions.
12962 (ipa_print_node_params): Print described use counter.
12963 (ipa_write_jump_function): Adjust to jump function type changes.
12964 (ipa_read_jump_function): New parameter CS, pass it to
12965 ipa_set_jf_constant. Adjust caller.
12966 (ipa_write_node_info): Stream controlled use count
12967 (ipa_read_node_info): Likewise.
12968 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
12970 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
12971 count. Remove cloning-added reference if it reaches zero.
12972 * ipa-ref.c (ipa_find_reference): New function.
12974 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12976 * config/i386/i386.c (processor_target_table): Modified default
12977 alignment values for AMD BD and BT architectures.
12979 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
12981 * tree-vect-generic.c (uniform_vector_p): Move ...
12982 * tree.c (uniform_vector_p): ... here.
12983 * tree.h (uniform_vector_p): Declare it.
12984 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
12987 2013-05-13 Jakub Jelinek <jakub@redhat.com>
12989 PR tree-optimization/57230
12990 * tree-ssa-strlen.c (handle_char_store): Record length for
12991 array store from STRING_CST.
12993 PR tree-optimization/57230
12994 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
12997 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
12999 * config/epiphany/epiphany.c (epiphany_init): Check size of
13000 NUM_MODES_FOR_MODE_SWITCHING.
13001 (epiphany_expand_prologue):
13002 Remove CONFIG_REGNUM initial value handling code.
13003 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13004 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13005 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13006 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13007 Don't return 1 for FP_MODE_NONE.
13008 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13009 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13010 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13011 * config/epiphany/epiphany.md (save_config): New pattern.
13013 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13015 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13017 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13019 * config/i386/i386.md (memory): Handle sseishft1.
13020 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13021 (*vec_extractv2di_1): Ditto.
13023 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13025 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13028 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13030 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13031 Add mthumb/march=armv7-a multilib.
13032 Add mthumb/march=armv7-r multilib.
13033 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13035 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13037 * config/v850/t-rtems: Add more multilibs.
13039 2013-05-10 Richard Biener <rguenther@suse.de>
13041 PR tree-optimization/57214
13042 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13043 not propagate from SSA names that occur in abnormal PHI nodes.
13045 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13047 * stor-layout.c (element_precision): New function.
13048 * machmode.h (element_precision): Declare it.
13049 * tree.c (build_minus_one_cst): New function.
13050 (element_precision): Likewise.
13051 * tree.h (build_minus_one_cst): Declare new function.
13052 (element_precision): Likewise.
13053 * fold-const.c (operand_equal_p): Use element_precision.
13054 (fold_binary_loc): Handle vector types.
13055 * convert.c (convert_to_integer): Use element_precision.
13056 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13059 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13061 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13062 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13063 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13064 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13065 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13066 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13067 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13068 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13069 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13070 (Uuw8): New constraints.
13071 (Usb4): Move into alphabetical order.
13072 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13073 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13074 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13075 previously unnamed patterns.
13076 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13077 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13078 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13079 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13080 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13081 of specific lengths.
13083 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13085 * config/i386/i386.md (rotateinv): New code attr.
13086 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13087 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13088 roll $31, %eax, etc.
13090 PR tree-optimization/45216
13091 PR tree-optimization/57157
13092 * tree-ssa-forwprop.c (simplify_rotate): New function.
13093 (ssa_forward_propagate_and_combine): Call it.
13095 2013-05-10 Richard Biener <rguenther@suse.de>
13097 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13098 disable peeling when we version for aliasing.
13099 (vector_alignment_reachable_p): Honor explicit user alignment.
13100 (vect_supportable_dr_alignment): Likewise.
13101 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13102 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13103 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13104 then peeling to arrange for the cost-model check to come first.
13106 2013-05-10 Alan Modra <amodra@gmail.com>
13108 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13109 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13110 * configure: Regenerate.
13112 2013-05-10 Alan Modra <amodra@gmail.com>
13115 * varasm.c (default_elf_select_section): Move !DECL_P check..
13116 (get_named_section): ..to here before calling get_section_name.
13118 (default_section_type_flags): Add DECL_P check.
13119 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13120 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13122 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13124 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13125 When using gen_stack_adjust_str with a register offset, add a
13126 REG_FRAME_RELATED_EXPR note.
13128 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13130 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13131 (*vec_extractv4si_zext_mem): Ditto.
13132 (*vec_extractv2di): Add 0->x and x->x alternatives.
13133 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13134 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13136 2013-05-09 Jason Merrill <jason@redhat.com>
13138 N3639 C++1y VLA support
13139 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13142 * tree.c (build_constructor_va): New.
13143 * tree.h: Declare it.
13145 2013-05-09 Martin Jambor <mjambor@suse.cz>
13148 * gimple-fold.c (canonicalize_constructor_val): Call
13149 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13151 2013-05-09 Jan Hubicka <jh@suse.cz>
13152 Richard Biener <rguenther@suse.de>
13155 * symtab.c (symtab_make_decl_local): Do not add private names.
13157 2013-05-09 Jan Hubicka <jh@suse.cz>
13160 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13161 (unlink_from_assembler_name_hash): Likewise.
13162 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13163 symtab_unregister_node, symtab_initialize_asm_name_hash,
13164 change_decl_assembler_name): Update.
13166 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13168 * config/aarch64/aarch64.md: New movtf split.
13169 (*movtf_aarch64): Update.
13170 (aarch64_movdi_tilow): Handle TF modes and rename to
13171 aarch64_movdi_<mode>low.
13172 (aarch64_movdi_tihigh): Handle TF modes and rename to
13173 aarch64_movdi_<mode>high
13174 (aarch64_movtihigh_di): Handle TF modes and rename to
13175 aarch64_mov<mode>high_di
13176 (aarch64_movtilow_di): Handle TF modes and rename to
13177 aarch64_mov<mode>low_di
13178 (aarch64_movtilow_tilow): Remove spurious whitespace.
13179 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13181 (aarch64_print_operand): Update.
13183 2013-05-09 Alan Modra <amodra@gmail.com>
13185 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13187 * configure: Regenerate.
13189 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13191 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13192 splitter preparation statements.
13193 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13194 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13197 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13199 * gimple-ssa-strength-reduction.c (count_candidates): Change
13200 return value to int.
13201 (analyze_candidates_and_replace): Change type of length to int.
13203 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13205 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13206 (*vec_extract<mode>): Use VI12_128 mode iterator.
13207 (*vec_extract<mode>_mem): Ditto.
13208 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13211 2013-05-08 Diego Novillo <dnovillo@google.com>
13216 2012-08-17 Diego Novillo <dnovillo@google.com>
13219 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13220 * config.in: Regenerate.
13221 * configure: Regenerate.
13222 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13224 2013-05-08 Jan Hubicka <jh@suse.cz>
13227 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13228 * cgraph.h (symtab_node_base): Add unique_name.
13229 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13230 input_overwrite_node, input_varpool_node): Stream unique_name.
13231 * cgraphclones.c (cgraph_create_virtual_clone,
13232 cgraph_function_versioning): Set unique_name.
13233 * ipa.c (function_and_variable_visibility): Set unique_name.
13235 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13237 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13238 (alloc_cand_and_find_basis): Restrict conditional candidate
13239 processing to CAND_MULTs.
13241 2013-05-08 Jan Hubicka <jh@suse.cz>
13244 lto-symtab.c (lto_symtab_symbol_p): New function.
13245 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13246 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13247 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13248 Skip static symbols.
13250 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13252 PR tree-optimization/57200
13253 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13254 Only call inform if the preceding warning_at returns true.
13256 2013-05-07 Han Shen <shenhan@google.com>
13258 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13259 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13260 * common.opt (fstack-protector-strong): New option.
13261 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13262 * doc/invoke.texi (Optimization Options): Document
13263 "-fstack-protector-strong".
13264 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13266 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13268 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13270 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13272 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13273 it is not lhs of the insn.
13274 (setup_reg_equiv): Remove insn having equiv memory which it is not
13276 * lra-constraints.c (process_address): Try to improve generation
13277 code for address base + disp.
13278 (lra_constraints): Make correct the code for checking insn setting
13279 up backward equivalence. Remove insn only if it is in the init
13281 * lra-eliminations.c (update_reg_eliminate): Change return value.
13282 (lra_eliminate): Use the result.
13284 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13286 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13287 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13288 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13289 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13290 register target operands.
13291 (*vec_extractv8hi_sse2): New pattern.
13292 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13293 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13294 (*vec_extract<mode>_mem): New insn and split pattern.
13296 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13298 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13299 (TARGET_ASAN_SHADOW_OFFSET): Define.
13300 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13301 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13303 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13305 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13306 (incr_vec_index): Return -1 if increment not found.
13307 (create_add_on_incoming_edge): Assert if increment not found.
13308 (record_increment): Limit number of increments recorded.
13309 (all_phi_incrs_profitable): Return false if an increment not found.
13310 (replace_profitable_candidates): Don't process increments that were
13312 (analyze_candidates_and_replace): Limit size of incr_vec.
13314 2013-05-07 Richard Biener <rguenther@suse.de>
13316 * calls.c (special_function_p): setjmp-like functions are leaf.
13317 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13318 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13320 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13322 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13323 (aarch64_simd_mov<mode>): New expander.
13324 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13325 (aarch64_simd_mov_to_<mode>high): Likewise.
13326 (aarch64_simd_mov_from_<mode>low): Likewise.
13327 (aarch64_simd_mov_from_<mode>high): Likewise.
13328 (aarch64_dup_lane<mode>): Update.
13329 (aarch64_dup_lanedi): New instruction pattern.
13330 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13331 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13333 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13335 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13336 (replace_mult_candidate): Remove unnecessary argument; remove
13337 unnecessary parameter from call to introduce_cast_before_cand.
13338 (replace_unconditional_candidate): Remove unnecessary parameter
13339 from call to replace_mult_candidate.
13340 (replace_conditional_candidate): Likewise.
13341 (insert_initializers): Use make_temp_ssa_name.
13342 (introduce_cast_before_cand): Remove unnecessary argument; use
13343 make_temp_ssa_name.
13344 (replace_one_candidate): Remove unnecessary argument; remove
13345 unnecessary parameter from calls to introduce_cast_before_cand.
13346 (replace_profitable_candidates): Remove unnecessary parameters
13347 from calls to replace_one_candidate.
13349 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13351 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13352 phi def as possibly hiding a basis for a CAND_ADD whose operands
13353 have been commuted in the analysis.
13354 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13356 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13358 * config/aarch64/aarch64.md
13359 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13360 shift value between 0-4.
13362 2013-05-07 Richard Biener <rguenther@suse.de>
13364 * double-int.h (rshift): New overload.
13365 * double-int.c (rshift): New function.
13366 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13367 (create_reference_ops_from_ref): Remove.
13368 (vn_reference_insert): Use shared ops for constructing the
13369 reference and copy it.
13371 2013-05-07 Richard Biener <rguenther@suse.de>
13373 PR middle-end/57190
13374 * tree-eh.c (sink_clobbers): Properly propagate
13375 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
13377 2013-05-07 Jakub Jelinek <jakub@redhat.com>
13379 PR tree-optimization/57149
13380 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
13381 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
13382 collect_phi_def_edges, execute_late_warn_uninitialized): Use
13383 uninit_undefined_value_p instead of ssa_undefined_value_p.
13386 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
13387 for modifier == EXPAND_INITIALIZER.
13389 2013-05-07 Anton Blanchard <anton@samba.org>
13391 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
13392 for powerpc64 little endian.
13393 * configure: Regenerate.
13395 2013-05-06 Graham Stott <grahams@btinternet.com>
13397 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
13398 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
13399 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
13400 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
13402 2013-05-06 Graham Stott <grahams@btinternet.com>
13404 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
13405 codes which allow non-lvalues.
13407 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
13409 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
13410 components are all 1s.
13411 (integer_minus_onep): New function.
13412 * tree.h (integer_minus_onep): Declare it.
13413 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
13414 integer_minus_onep instead of integer_all_onesp.
13416 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13419 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
13420 variations of these patterns.
13422 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13424 * config/i386/i386.md (isa): Add x64_sse4 member.
13425 (enabled): Handle x64_sse4.
13426 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
13427 instruction for 64bit SSE4_1 targets. Update insn attributes.
13428 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
13429 instruction for SSE4_1 targets. Update insn attributes.
13430 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13431 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
13432 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
13434 (*vec_extractv4si): Rename from *sse4_1_pextrd.
13435 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
13436 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
13438 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13441 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
13443 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
13445 * final.c (do_assembler_dialects): Don't handle curly braces and
13446 vertical bar escaped by % as dialect delimiters.
13447 (output_asm_insn): Print curly braces and vertical bar if escaped
13448 by % and ASSEMBLER_DIALECT defined.
13449 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
13450 * doc/tm.texi: Regenerated.
13452 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13454 * config/mips/mips.c: Include tree-pass.h.
13455 (mips_reorg): Split in pre- and post-dbr_schedule parts.
13456 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
13457 (pass_mips_machine_reorg2): New machine specific pass.
13458 (insert_pass_mips_machine_reorg2): New pass plugin definition.
13459 (mips_option_override): Register the new pass.
13460 * rtl.h (cleanup_barriers): Remove prototype.
13461 (dbr_schedule): Likewise.
13462 * jump.c (cleanup_barriers): Make static.
13463 * reorg.c (dbr_schedule): Likewise.
13465 2013-05-06 Richard Biener <rguenther@suse.de>
13467 PR tree-optimization/57185
13468 * tree-parloops.c (add_field_for_reduction): Handle anonymous
13469 SSA names properly.
13471 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13474 * config/i386/i386.c (add_parameter_dependencies): Add dependence
13475 between "first_arg" and "insn", not "last" and "insn".
13477 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13479 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
13480 (find_candidates_in_block): Re-enable slsr_process_phi.
13481 (create_phi_basis): Fix double counting of candidate adjustment.
13483 2013-05-06 Richard Biener <rguenther@suse.de>
13485 PR middle-end/57147
13486 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
13487 the edge is also fallthru, preserve it and just clear the
13489 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
13490 also complex, preserve that and just clear the fallthru flag.
13491 * tree-inline.c (update_ssa_across_abnormal_edges): Also
13492 update virtual operands.
13494 2013-05-06 Alan Modra <amodra@gmail.com>
13496 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
13497 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
13498 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
13499 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
13500 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13501 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
13503 2013-05-06 Alan Modra <amodra@gmail.com>
13505 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
13506 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
13507 (DEFAULT_ASM_ENDIAN): Define.
13508 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13509 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
13510 Update -K PIC clause from sysv4.h.
13511 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
13512 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
13514 2013-05-06 Alan Modra <amodra@gmail.com>
13516 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
13517 twice for little-endian.
13518 (ashrdi3_no_power, ashrdi3): Support little-endian.
13520 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13523 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
13524 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
13526 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
13527 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
13528 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
13531 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
13532 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13534 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
13535 * config/pa/pa.opt: Make mbig-switch a no-op.
13536 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
13537 (CASE_VECTOR_MODE): Always return SImode.
13538 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
13539 for the !TARGET_BIG_SWITCH case.
13540 * config/pa/pa-linux.h: Likewise.
13541 * config/pa/pa-openbsd.h: Likewise.
13542 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
13543 * config/pa/pa.md (short_jump): Remove define_insn.
13544 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
13545 (casesi0): Remove define_insn.
13546 (type): Remove btable_branch.
13547 (pa_combine_type): Likewise.
13548 (in_nullified_branch_delay): Likewise.
13549 (in_call_delay): Likewise.
13550 (define_delay): Likewise.
13551 (define_insn_reservation "Z3"): Likewise.
13552 (define_insn_reservation "Z4"): Likewise.
13553 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
13554 (pa_adjust_insn_length): Remove adjustment for btable branches.
13555 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
13558 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
13560 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13561 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
13562 Add m->r,x alternatives.
13563 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
13564 splitters using SWI48x mode iterator.
13565 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
13566 TARGET_64BIT. Add m->x alternative.
13567 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
13568 Add o->x alternative. Enable for TARGET_SSE.
13569 (sse_storeq): Remove expander.
13570 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
13571 with memory input operand.
13572 (*vec_extractv2di_1 splitter): New.
13573 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
13574 * config/i386/i386.md (ssevecmodelower): New mode attribute.
13576 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13578 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
13579 (INT_LOWPART): Delete.
13580 (extract_MB): Adjust.
13581 (extract_ME): Adjust.
13582 (print_operand): Adjust.
13584 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13586 * config/rs6000/predicates.md (reg_or_add_cint_operand,
13587 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13588 (reg_or_logical_cint_operand, easy_fp_constant,
13589 logical_const_operand): Delete "CONST_DOUBLE" case.
13590 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
13591 "HOST_BITS_PER_WIDE_INT == 64" test.
13592 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
13593 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
13594 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
13595 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
13597 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
13598 CONST_DOUBLE DImode/VOIDmode case.
13599 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
13600 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13601 CONST_DOUBLE VOIDmode case.
13602 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13603 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
13604 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
13605 Delete CONST_DOUBLE case.
13606 (splitters for mov FMOVE64 const_double): Delete
13607 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13608 "HOST_BITS_PER_WIDE_INT >= 64" test.
13609 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
13611 (mov DI const_double): Delete.
13613 2013-05-04 Jakub Jelinek <jakub@redhat.com>
13615 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
13616 on op shows all bits zero in mode of a lowpart subreg, return zero.
13618 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
13621 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
13622 to save TFmode registers and DImode to save TImode registers for
13623 caller save operations.
13624 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
13625 mark being partially clobbered since they only use the first
13628 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
13629 and TDmode only use the upper 64-bits of each VSX register.
13631 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13633 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
13634 (find_candidates_in_block): Disable slsr_process_phi.
13636 2013-05-03 Guozhi Wei <carrot@google.com>
13638 * coverage.c (coverage_obj_init): Move the construction of gcov
13640 (build_init_ctor): ... here.
13642 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13644 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
13645 (slsr_cand_d): Redefine def_phi.
13646 (stride_status, phi_adjust_status, count_phis_status): New enums.
13647 (find_phi_def): New.
13648 (find_basis_for_base_expr): New.
13649 (find_basis_for_candidate): Handle hidden bases.
13650 (alloc_cand_and_find_basis): Handle phi candidates.
13651 (slsr_process_phi): New.
13652 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
13653 (create_mul_imm_cand): Likewise.
13654 (create_add_ssa_cand): Exclude phi base candidates.
13655 (create_add_imm_cand): Likewise.
13656 (slsr_process_cast): Likewise.
13657 (slsr_process_copy): Likewise.
13658 (find_candidates_in_block): Handle phi candidates.
13659 (dump_candidate): Likewise.
13660 (unconditional_cands): Delete.
13661 (unconditional_cands_with_known_stride_p): Delete.
13662 (phi_dependent_cand_p): New.
13663 (cand_increment): Handle phi-dependent candidates.
13664 (replace_dependent): Delete.
13665 (replace_mult_candidate): New.
13666 (replace_unconditional_candidate): New.
13667 (incr_vec_index): Move to avoid forward reference.
13668 (create_add_on_incoming_edge): New.
13669 (create_phi_basis): New.
13670 (replace_dependents): Delete.
13671 (replace_conditional_candidate): New.
13672 (phi_add_costs): New.
13673 (replace_uncond_cands_and_profitable_phis): New.
13674 (record_increment): Handle phi adjustments.
13675 (record_phi_increments): New.
13676 (record_increments): Handle phi adjustments.
13677 (phi_incr_cost): New.
13678 (lowest_cost_path): Handle phis.
13679 (total_savings): Likewise.
13680 (analyze_increments): Likewise.
13681 (ncd_with_phi): New.
13682 (ncd_of_cand_and_phis): New.
13683 (nearest_common_dominator_for_cands): Handle phi increments.
13684 (all_phi_incrs_profitable): New.
13685 (replace_profitable_candidates): Handle phi-dependent candidates.
13686 (analyze_candidates_and_replace): Likewise.
13688 2013-05-03 Teresa Johnson <tejohnson@google.com>
13691 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
13692 do not exceed REG_BR_PROB_BASE.
13694 2013-05-03 Jeff Law <law@redhat.com>
13696 PR tree-optimization/57144
13697 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
13698 operand of the condition will bit into the new type when eliminating
13699 a cast feeding a condition.
13701 2013-05-03 Jakub Jelinek <jakub@redhat.com>
13703 PR rtl-optimization/57130
13704 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
13705 of COMPARE as in_code to the recursive call if needed.
13707 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
13709 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
13710 (enabled): Handle new members.
13711 * config/i386/sse.md (*vec_concatv2si): Merge from
13712 *vec_concatv2si_sse2 and vec_concatv2si_sse.
13713 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
13715 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
13717 PR tree-optimization/57027
13718 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
13719 for fnms opportunity, check we got the prerequisite kind
13720 of tree / gimple before using accessor functions.
13722 2013-05-03 Richard Biener <rguenther@suse.de>
13724 * double-int.h (lshift): New overload without precision
13725 and arith argument.
13726 (operator *=, operator +=, operator -=): Move ...
13727 * double-int.c (operator *=, operator +=, operator -=): ... here
13728 and implement more efficiently.
13729 (mul_double_with_sign): Remove.
13730 (lshift_double): Adjust to take unsinged shift argument, push
13731 dispatching code to callers.
13732 (mul_double_wide_with_sign): Add early out for callers that
13733 are not interested in high parts or overflow.
13734 (lshift): New function.
13735 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
13736 dispatch code here.
13737 (lrotate, rrotate): Use logical shifts.
13738 * expr.c (get_inner_reference): Use lshift.
13739 * fixed-value.c (do_fixed_divide): Likewise.
13740 * tree-dfa.c (get_ref_base_and_extent): Likewise.
13741 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
13742 (indirect_refs_may_alias_p): Likewise.
13743 (stmt_kills_ref_p_1): Likewise.
13745 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13747 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
13749 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13751 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
13752 scalar form of FABD instruction.
13754 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
13756 * lra-constraints.c (process_alt_operands): Add checking alt
13757 number to choose the best alternative.
13759 2013-05-02 Richard Biener <rguenther@suse.de>
13761 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
13762 bitmap and its handling.
13763 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
13765 2013-05-02 Richard Biener <rguenther@suse.de>
13767 PR middle-end/57140
13768 * tree-inline.c (copy_loops): Properly handle removed loops.
13769 (copy_cfg_body): Mark destination loops for fixup if source
13770 loops needed fixup.
13772 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
13775 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
13776 generating simple_return for naked functions.
13778 2013-05-02 Martin Jambor <mjambor@suse.cz>
13780 PR middle-end/56988
13781 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
13782 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
13784 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
13785 ipa_agg_replacement_value structures.
13786 (known_aggs_to_agg_replacement_list): Likewise.
13787 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
13788 (read_agg_replacement_chain): Likewise.
13789 (ipcp_transform_function): Also check that by_ref flags match.
13791 2013-05-02 Richard Biener <rguenther@suse.de>
13793 * graphds.h (struct graph): Add obstack member.
13794 * graphds.c (new_graph): Initialize obstack and allocate
13796 (add_edge): Allocate edge from the obstack.
13797 (free_graph): Free the obstack instead of all edges and vertices.
13799 2013-05-02 Teresa Johnson <tejohnson@google.com>
13801 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
13803 * cfg.c (update_bb_profile_for_threading): Ditto.
13804 * tree-inline.c (copy_bb): Ditto.
13805 (copy_edges_for_bb): Ditto.
13806 (initialize_cfun): Ditto.
13807 (copy_cfg_body): Ditto.
13808 (expand_call_inline): Ditto.
13809 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
13810 (estimate_node_size_and_time): Ditto.
13811 (inline_merge_summary): Ditto.
13812 * cgraphclones.c (cgraph_clone_edge): Ditto.
13813 (cgraph_clone_node): Ditto.
13814 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13815 (compute_trg_info): Ditto.
13817 2013-05-02 Ian Bolton <ian.bolton@arm.com>
13819 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
13820 S reg when fp attribute set.
13821 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
13823 2013-05-02 Ian Bolton <ian.bolton@arm.com>
13825 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
13827 (*and_one_cmplsi3_compare0_uxtw): Likewise.
13828 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
13829 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
13831 2013-05-02 Richard Biener <rguenther@suse.de>
13833 * tree-scalar-evolution.c (scev_info_hasher): Remove.
13834 (struct instantiate_cache_entry): New type.
13835 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
13836 (struct instantiate_cache_type): New type.
13837 (set_instantiated_value, get_instantiated_value): Remove.
13838 (get_instantiated_value_entry): New function.
13839 (instantiate_scev_name): Use the new cache and adjust.
13840 (instantiate_scev_poly): Adjust.
13841 (instantiate_scev_binary): Likewise.
13842 (instantiate_array_ref): Likewise.
13843 (instantiate_scev_convert): Likewise.
13844 (instantiate_scev_not): Likewise.
13845 (instantiate_scev_3): Likewise.
13846 (instantiate_scev_2): Likewise.
13847 (instantiate_scev_r): Likewise.
13848 (instantiate_scev): Likewise.
13849 (resolve_mixers): Likewise.
13851 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
13854 * lra-constraints.c (best_small_class_operands_num): Remove.
13855 (process_alt_operands): Remove small_class_operands_num. Take
13856 small classes operands into losers and only if the operand is not
13857 matched. Modify debugging output.
13858 (curr_insn_transform): Remove best_small_class_operands_num.
13861 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13863 * config/aarch64/aarch64-builtins.c
13864 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
13865 * config/aarch64/aarch64-simd-builtins.def
13866 (reduc_splus_): Add new modes.
13867 (reduc_uplus_): New.
13868 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
13869 (reduc_uplus_v4sf): Likewise.
13870 (reduc_splus_v4sf): Likewise.
13871 (aarch64_addv<mode>): Likewise.
13872 (reduc_uplus_<mode>): Likewise.
13873 (reduc_splus_<mode>): Likewise.
13874 (aarch64_addvv2di): Likewise.
13875 (reduc_uplus_v2di): Likewise.
13876 (reduc_splus_v2di): Likewise.
13877 (aarch64_addvv2si): Likewise.
13878 (reduc_uplus_v2si): Likewise.
13879 (reduc_splus_v2si): Likewise.
13880 (reduc_<sur>plus_<mode>): New.
13881 (reduc_<sur>plus_v2di): Likewise.
13882 (reduc_<sur>plus_v2si): Likewise.
13883 (reduc_<sur>plus_v4sf): Likewise.
13884 (aarch64_addpv4sf): Likewise.
13885 * config/aarch64/arm_neon.h
13886 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
13887 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
13888 add UNSPEC_SADDV, UNSPEC_UADDV.
13890 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
13892 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13894 * config/aarch64/arm_neon.h
13895 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
13897 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13899 * config/aarch64/aarch64-builtins
13900 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
13902 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13904 * config/aarch64/aarch64-simd-builtins.def
13905 (reduc_smax_): New.
13906 (reduc_smin_): Likewise.
13907 (reduc_umax_): Likewise.
13908 (reduc_umin_): Likewise.
13909 (reduc_smax_nan_): Likewise.
13910 (reduc_smin_nan_): Likewise.
13913 (smax): Update for V2SF, V4SF and V2DF modes.
13916 (smin_nan): Likewise.
13917 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
13918 (<su><maxmin><mode>3): ...This, refactor.
13919 (s<maxmin><mode>3): New.
13920 (<maxmin_uns><mode>3): Likewise.
13921 (reduc_<maxmin_uns>_<mode>): Refactor.
13922 (reduc_<maxmin_uns>_v4sf): Likewise.
13923 (reduc_<maxmin_uns>_v2si): Likewise.
13924 (aarch64_<fmaxmin><mode>: Remove.
13925 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
13927 (vmin<q>_f<32,64>): Likewise.
13928 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13930 (su): Add mappings for smax, smin, umax, umin.
13932 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13933 (FMAXMIN): Rename as...
13934 (FMAXMIN_UNS): ...This.
13936 (fmaxminv): Likewise.
13937 (fmaxmin): Likewise.
13939 (maxmin_uns_op): Likewise.
13941 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13943 * config/aarch64/arm_neon.h
13944 (vac<ge, gt><sd>_f<32, 64>): Rename to...
13945 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
13946 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
13948 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13950 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
13951 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
13953 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13955 * config/aarch64/aarch64-simd.md
13956 (vcond<mode>_internal): Handle special cases for constant masks.
13957 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
13958 (vcondu<mode><mode>): Likewise.
13959 (vcond<v_cmp_result><mode>): New.
13961 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13963 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
13964 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
13965 * config/aarch64/aarch64-simd-builtins.def
13966 (cmeq): Update to BUILTIN_VALLDI.
13971 * config/aarch64/arm_neon.h
13972 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
13973 to builtins or C as appropriate.
13975 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13977 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
13979 (cmhi): Rename to...
13981 * config/aarch64/aarch64-simd.md
13982 (simd_mode): Add SF.
13983 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
13984 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
13985 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
13986 (cstore<mode>_neg): ...This.
13987 * config/aarch64/iterators.md
13989 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
13990 (COMPARISONS): New.
13991 (UCOMPARISONS): Likewise.
13992 (optab): Add missing comparisons.
13998 (VCMP_S): Likewise.
13999 (VCMP_U): Likewise.
14000 (V_cmp_result): Add DF, SF modes.
14001 (v_cmp_result): Likewise.
14003 (vmtype): Likewise.
14004 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14006 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14008 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14009 define_insn to define_insn_and_split.
14010 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14011 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14012 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14013 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14014 (thumb2_negscc): Likewise.
14016 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14018 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14020 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14022 * config/arm/thumb2.md: Remove trailing whitespaces.
14024 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14026 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14027 Use gen_int_mode rather than GEN_INT.
14029 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14031 * value-prof.c (stream_in_histogram_value): Remove the strayed
14034 2013-04-30 Richard Biener <rguenther@suse.de>
14036 PR middle-end/57122
14037 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14039 2013-04-30 Richard Biener <rguenther@suse.de>
14041 PR middle-end/57107
14042 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14044 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14046 PR rtl-optimization/56957
14047 PR rtl-optimization/57105
14048 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14049 variable. Use just INSN_UID for determining whether an insn
14050 should be only disconnected from the insn stream.
14051 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14053 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14055 PR tree-optimization/57104
14056 * tsan.c (instrument_expr): Don't instrument accesses to
14057 DECL_HARD_REGISTER VAR_DECLs.
14059 2013-04-30 Richard Biener <rguenther@suse.de>
14061 * function.h (loops_for_fn): New inline function.
14062 (set_loops_for_fn): Likewise.
14063 * cfgloop.h (place_new_loop): Add struct function parameter.
14064 (get_loop): Likewise.
14065 (get_loops): Likewise.
14066 (number_of_loops): Likewise.
14067 (fel_next): Adjust.
14068 (fel_init): Likewise.
14069 * cfg.c (get_loop_copy): Adjust.
14070 * cfgloop.c (flow_loops_dump): Likewise.
14071 (record_loop_exits): Likewise.
14072 (verify_loop_structure): Likewise.
14073 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14074 (estimate_reg_pressure_cost): Likewise.
14075 (mark_loop_exit_edges): Likewise.
14076 * cfgloopmanip.c (place_new_loop): Likewise.
14077 (add_loop): Likewise.
14078 (duplicate_loop): Likewise.
14079 * graph.c (draw_cfg_nodes): Likewise.
14080 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14081 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14082 (extract_affine_chrec): Likewise.
14083 (build_scop_iteration_domain): Likewise.
14084 * graphite.c (graphite_initialize): Likewise.
14085 * ira-build.c (create_loop_tree_nodes): Likewise.
14086 (more_one_region_p): Likewise.
14087 (rebuild_regno_allocno_maps): Likewise.
14088 (mark_loops_for_removal): Likewise.
14089 (mark_all_loops_for_removal): Likewise.
14090 (remove_unnecessary_regions): Likewise.
14091 (ira_build): Likewise.
14092 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14093 * loop-init.c (fix_loop_structure): Likewise.
14094 (gate_rtl_move_loop_invariants): Likewise.
14095 (gate_rtl_unswitch): Likewise.
14096 (gate_rtl_unroll_and_peel_loops): Likewise.
14097 (rtl_doloop): Likewise.
14098 * lto-streamer-in.c (input_cfg): Likewise.
14099 * lto-streamer-out.c (output_cfg): Likewise.
14100 * modulo-sched.c (sms_schedule): Likewise.
14101 * predict.c (tree_estimate_probability): Likewise.
14102 (tree_estimate_probability_driver): Likewise.
14103 (estimate_loops): Likewise.
14104 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14105 (move_sese_region_to_fn): Likewise.
14106 (debug_loop_num): Likewise.
14107 * tree-chrec.c (chrec_evaluate): Likewise.
14108 (hide_evolution_in_other_loops_than_loop): Likewise.
14109 (chrec_component_in_loop_num): Likewise.
14110 (reset_evolution_in_loop): Likewise.
14111 (evolution_function_is_invariant_rec_p): Likewise.
14112 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14113 * tree-inline.c (copy_loops): Likewise.
14114 (copy_cfg_body): Likewise.
14115 (tree_function_versioning): Likewise.
14116 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14117 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14119 (add_to_evolution_1): Likewise.
14120 (scev_const_prop): Likewise.
14121 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14122 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14123 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14124 (tree_ssa_lim_initialize): Likewise.
14125 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14126 (verify_loop_closed_ssa): Likewise.
14127 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14128 (tree_ssa_loop_im): Likewise.
14129 (tree_ssa_loop_unswitch): Likewise.
14130 (tree_vectorize): Likewise.
14131 (check_data_deps): Likewise.
14132 (tree_ssa_loop_ivcanon): Likewise.
14133 (tree_ssa_loop_bounds): Likewise.
14134 (tree_complete_unroll): Likewise.
14135 (tree_complete_unroll_inner): Likewise.
14136 (tree_parallelize_loops): Likewise.
14137 (tree_ssa_loop_prefetch): Likewise.
14138 (tree_ssa_loop_ivopts): Likewise.
14139 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14140 * tree-vectorizer.c (vectorize_loops): Likewise.
14142 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14144 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14145 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14146 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14147 with EABI_LINK_SPEC.
14149 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14152 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14155 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14158 * lra-constraints.c (process_alt_operands): Discourage a bit more
14159 using memory for pseudos. Print cost dump for alternatives.
14160 Modify cost values for conflicts with early clobbers.
14161 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14163 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14166 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14168 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14170 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14171 from/to S register.
14172 (movdi_aarch64): Support LDR/STR from/to D register.
14174 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14176 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14177 or higher by default.
14179 2013-04-29 Richard Biener <rguenther@suse.de>
14181 PR middle-end/57075
14182 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14183 even if not adding abnormal edges for calls that can make
14186 2013-04-29 Richard Biener <rguenther@suse.de>
14188 PR middle-end/57103
14189 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14191 (move_stmt_r): Remove redundant checking.
14193 2013-04-29 Teresa Johnson <tejohnson@google.com>
14196 * basic-block.h (apply_scale): New function.
14197 (apply_probability): Use apply_scale.
14198 * gimple-streamer-in.c (input_bb): Ditto.
14199 * lto-streamer-in.c (input_cfg): Ditto.
14200 * lto-cgraph.c (merge_profile_summaries): Ditto.
14201 * tree-optimize.c (execute_fixup_cfg): Ditto.
14202 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14203 (copy_edges_for_bb): Ditto.
14204 (copy_cfg_body): Ditto.
14206 2013-04-29 Tom de Vries <tom@codesourcery.com>
14208 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14209 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14210 (tail_merge_optimize): Handle current_loops == NULL.
14212 2013-04-26 Jeff Law <law@redhat.com>
14214 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14215 (simplify_cond_using_ranges): Generalize code to simplify
14216 COND_EXPRs where one argument is a constant and the other
14217 is an SSA_NAME created by an integral type conversion.
14219 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14221 * config/arm/arm.md (store_minmaxsi): Use only when
14222 optimize_insn_for_size_p.
14224 2013-04-29 Christian Bruel <christian.bruel@st.com>
14227 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14229 2013-04-29 Richard Biener <rguenther@suse.de>
14231 PR middle-end/57089
14232 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14233 loop tree make sure to schedule a fixup for the child as well.
14234 (expand_omp_for_generic): Properly add loops.
14235 (expand_omp_for_static_nochunk): Likewise.
14236 (expand_omp_for_static_chunk): Likewise.
14237 (expand_omp_for): For the degenerate case fixup loops.
14238 (expand_omp_sections): Fix default bb placement in loops.
14239 (expand_omp_atomic_pipeline): Properly add loops.
14241 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14243 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14245 2013-04-29 Tom de Vries <tom@codesourcery.com>
14247 * tree-ssa-tail-merge.c: Update header comment.
14249 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14251 * config/aarch64/arm_neon.h
14252 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14253 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14254 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14255 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14256 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14257 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14258 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14259 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14261 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14263 * config/aarch64/aarch64-simd.md
14264 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14265 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14266 fix_trunc, fixuns_trunc.
14267 (ftrunc<VDQF:mode>2): New.
14268 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14269 (fix_trunc_optab): New.
14271 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14273 * config/aarch64/aarch64-builtins.c
14274 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14275 iceilf, lround, iroundf.
14277 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14280 * config/i386/i386.h (enum ix86_tune_indices)
14281 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14282 New, split from X86_TUNE_INTER_UNIT_MOVES.
14283 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14284 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14285 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14286 (TARGET_INTER_UNIT_MOVES): Remove.
14287 * config/i386/i386.c (initial_ix86_tune_features): Update.
14288 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14289 (ix86_expand_convert_uns_didf_sse): Use
14290 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14291 (ix86_expand_vector_init_one_nonzero): Ditto.
14292 (ix86_expand_vector_init_interleave): Ditto.
14293 (inline_secondary_memory_needed): Return true for moves from SSE class
14294 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14295 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14296 * config/i386/constraints.md (Yi, Ym): Depend on
14297 TARGET_INTER_UNIT_MOVES_TO_VEC.
14298 (Yj, Yn): New constraints.
14299 * config/i386/i386.md (*movdi_internal): Change constraints of
14300 operand 1 from Yi to Yj and from Ym to Yn.
14301 (*movsi_internal): Ditto.
14302 (*movdf_internal): Ditto.
14303 (*movsf_internal): Ditto.
14304 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14305 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14306 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14307 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14308 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14309 * config/i386/sse.md (movdi_to_sse): Ditto.
14310 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14311 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14312 TARGET_INTER_UNIT_MOVES.
14313 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14314 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14315 instead of TARGET_INTER_UNIT_MOVES.
14316 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14317 operand 1 from Yi to Yj and from Ym to Yn.
14319 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14321 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14322 (float_truncate_hi_): Likewise.
14323 (float_extend_lo_): Likewise.
14324 (float_truncate_lo_): Likewise.
14325 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14326 (aarch64_float_extend_lo_v2df): Likewise.
14327 (vec_unpacks_hi_v4sf): Likewise.
14328 (aarch64_float_truncate_lo_v2sf): Likewise.
14329 (aarch64_float_truncate_hi_v4sf): Likewise.
14330 (vec_pack_trunc_v2df): Likewise.
14331 (vec_pack_trunc_df): Likewise.
14333 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14335 * config/aarch64/aarch64-builtins.c
14336 (aarch64_fold_builtin): Fold float conversions.
14337 * config/aarch64/aarch64-simd-builtins.def
14338 (floatv2si, floatv4si, floatv2di): New.
14339 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14340 * config/aarch64/aarch64-simd.md
14341 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14342 * config/aarch64/iterators.md (FLOATUORS): New.
14343 (optab): Add float, floatuns.
14344 (su_optab): Likewise.
14346 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14348 * config/aarch64/aarch64-builtins.c
14349 (aarch64_builtin_vectorized_function): Use new names for
14351 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14352 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14353 (fcvtzu): Split as...
14354 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14355 (fcvtas): Split as...
14356 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14357 (fcvtau): Split as...
14358 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14359 (fcvtps): Split as...
14360 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14361 (fcvtpu): Split as...
14362 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14363 (fcvtms): Split as...
14364 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14365 (fcvtmu): Split as...
14366 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14367 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14368 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14369 (lfrintnusf, lfrintnudf): Likewise.
14370 * config/aarch64/aarch64-simd.md
14371 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
14373 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
14374 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
14375 (fcvt_pattern): Likewise.
14377 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14379 * config/aarch64/aarch64-simd.md
14380 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
14381 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
14383 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14385 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
14386 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
14387 (vrnd<a,m,n,p>_f32): Implement using builtins.
14388 (vrnd<i,x><q>_f<32, 64>): New.
14390 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14392 * config/aarch64/aarch64-builtins.c
14393 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
14394 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
14395 (frintz): Rename to...
14397 (frintp): Rename to...
14399 (frintm): Rename to...
14401 (frinti): Rename to...
14402 (nearbyint): ...this.
14403 (frintx): Rename to...
14405 (frinta): Rename to...
14407 * config/aarch64/aarch64-simd.md
14408 (aarch64_frint<frint_suffix><mode>): Delete.
14409 (<frint_pattern><mode>2): Convert to insn.
14410 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
14411 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
14412 (frint_pattern): Likewise.
14413 (frint_suffix): Likewise.
14415 2013-04-29 Richard Biener <rguenther@suse.de>
14417 PR tree-optimization/57081
14418 * loop-init.c: Include tree-flow.h.
14419 (loop_optimizer_finalize): Free number of iteration estimates.
14420 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
14422 2013-04-29 Jakub Jelinek <jakub@redhat.com>
14424 PR tree-optimization/57083
14425 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
14426 non-singleton shift count range, zero extend low_bound for uns case.
14428 * config/i386/predicates.md (general_vector_operand): New predicate.
14429 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
14430 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
14431 if they aren't nonimmediate operands. If their original values
14432 satisfy const_vector_equal_evenodd_p, don't shift them.
14433 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
14434 predicates. For the SSE4.1 case force operands[{1,2}] into registers
14435 if not nonimmediate_operand.
14436 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
14437 instead of register_operand.
14438 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
14440 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
14442 * stor-layout.c (finalize_size_functions): Allocate a structure and
14443 reset cfun before dumping the functions.
14445 2013-04-27 Jakub Jelinek <jakub@redhat.com>
14447 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
14450 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
14451 use xop_pmacsdqh if uns_p.
14452 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
14453 the immediate rotate count.
14455 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
14457 * rtl.h (struct rtx_def): Add comment for field jump.
14458 (LRA_SUBREG_P): New macro.
14459 * recog.c (register_operand): Check LRA_SUBREG_P.
14460 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14462 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14463 (lra_final_code_change): Skip subreg change for operators.
14464 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14465 if there are no operand changes.
14466 * lra-constraints.c (curr_insn_set): New.
14467 (match_reload): Set LRA_SUBREG_P.
14468 (emit_spill_move): Ditto.
14469 (check_and_process_move): Use curr_insn_set. Process only single
14470 set insns. Don't initialize sec_mem_p and change_p.
14471 (simplify_operand_subreg): Use LRA_SUBREG_P.
14472 (reg_in_class_p): New function.
14473 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14474 of #ifdef. Add code to remove cycling.
14475 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14476 non-null disp. Reload inner instead of disp when base and index
14477 are null. Try to put lo_sum into register.
14478 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14479 (check_and_process_move): Move code for move cost check to
14480 simple_move_p. Remove equiv_substitution.
14481 (simple_move_p): New function.
14482 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14483 curr_insn_set. Call check_and_process_move only for single set
14484 insns. Use the new function. Move call of check_and_process_move
14485 after operand equiv substitution and address process.
14487 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14490 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
14491 with nonlocal goto receivers or returns twice calls, ignore
14492 unininitialized values from abnormal edges to nl goto receiver
14493 or returns twice call.
14495 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14497 PR tree-optimization/57051
14498 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
14499 and VEC_RSHIFT_EXPR if shift count is a multiple of element
14502 2013-04-26 Richard Biener <rguenther@suse.de>
14504 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
14505 (expand_omp_taskreg): Likewise. Mark loops for fixup.
14506 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
14507 (fixup_loop_arrays_after_move): New function.
14508 (move_sese_region_to_fn): Properly outline the loop tree parts
14509 of the SESE region.
14511 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
14513 * config/i386/i386.md (type, unit): Fix long lines.
14515 2013-04-26 Richard Biener <rguenther@suse.de>
14517 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
14518 (lto-streamer-out.o): Likewise.
14519 * cfgloop.c (init_loops_structure): Export, add struct function
14520 argument and adjust.
14521 (flow_loops_find): Adjust.
14522 * cfgloop.h (enum loop_estimation): Add EST_LAST.
14523 (init_loops_structure): Declare.
14524 * lto-streamer-in.c: Include cfgloop.h.
14525 (input_cfg): Input the loop tree.
14526 * lto-streamer-out.c: Include cfgloop.h.
14527 (output_cfg): Output the loop tree.
14528 (output_struct_function_base): Do not drop PROP_loops.
14530 2013-03-26 Richard Biener <rguenther@suse.de>
14532 * tree-cfg.c (execute_build_cfg): Build the loop tree.
14533 (pass_build_cfg): Provide PROP_loops.
14534 (move_sese_region_to_fn): Remove loops that are outlined into fn
14536 * tree-inline.c: Include cfgloop.h.
14537 (initialize_cfun): Do not drop PROP_loops.
14538 (copy_loops): New function.
14539 (copy_cfg_body): Copy loop structure.
14540 (tree_function_versioning): Initialize destination loop tree.
14541 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
14542 (pass_parallelize_loops): Do IL verification.
14543 * loop-init.c (loop_optimizer_init): Fixup loops if required.
14544 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
14545 the CFG make sure we fixup loops as well.
14546 * tree-ssa-tail-merge.c: Include cfgloop.h.
14547 (replace_block_by): When merging loop latches mark loops for fixup.
14548 * lto-streamer-out.c (output_struct_function_base): Drop
14549 PROP_loops for now.
14550 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
14551 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
14552 * ipa-split.c: Include cfgloop.h.
14553 (split_function): Add the new return block to the loop tree root.
14554 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
14555 whether we have removed the forwarder block.
14556 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
14557 * cfgloop.h (place_new_loop): Declare.
14558 * cfgloopmanip.c (place_new_loop): Export.
14559 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
14560 (tree-switch-conversion.o): Likewise.
14561 (tree-complex.o): Likewise.
14562 (tree-inline.o): Likewise.
14563 (tree-ssa-tailmerge.o): Likewise.
14564 (ipa-split.o): Likewise.
14565 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
14566 (tree-ssa-copy.o): Likewise.
14567 * tree-switch-conversion.c: Include cfgloop.h
14568 (process_switch): If we emit a bit-test cascade, schedule loops
14570 * tree-complex.c: Include cfgloop.h.
14571 (expand_complex_div_wide): Properly add new basic-blocks to loops.
14572 * asan.c: Include cfgloop.h.
14573 (create_cond_insert_point): Properly add new basic-blocks to
14574 loops, schedule loop fixup.
14575 * cfgloop.c (verify_loop_structure): Check that looks are not
14577 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
14579 (expand_omp_for_generic): Likewise.
14580 (expand_omp_sections): Likewise.
14581 (expand_omp_atomic_pipeline): Schedule loops for fixup.
14582 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
14583 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
14584 is initialized, not when loops are present.
14585 * tree-parloops.c (parallelize_loops): Remove checking here.
14586 * passes.c (init_optimization_passes): Schedule a copy-propagation
14587 pass before complete unrolling of inner loops.
14589 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14591 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
14592 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
14593 (colorize_init): Add argument to _WIN32 version.
14594 * toplev.c: Include diagnostic-color.h.
14595 (process_options): Default to -fdiagnostics-color=auto if
14596 GCC_COLORS env var is in the environment.
14597 * common.opt (fdiagnostics-color=): Add Var and Init.
14598 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
14599 env var is in the environment, the default is auto rather than never.
14601 * diagnostic.h (file_name_as_prefix): Add context argument.
14602 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
14603 the string as locus.
14604 * langhooks.c (lhd_print_error_function): Adjust caller.
14606 2013-04-25 Lawrence Crowl <crowl@google.com>
14608 * var-tracking.c (shared_hash_def::htab):
14609 Change type to hash_table. Update dependent calls and types.
14611 2013-04-25 Lawrence Crowl <crowl@google.com>
14613 * Makefile.in: Update as needed below.
14615 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
14616 Move declaration to after the type's method definitons.
14618 * attribs.c (htab_t scoped_attributes::attribute_hash):
14619 Change type to hash_table. Update dependent calls and types.
14621 * bitmap.c (htab_t bitmap_desc_hash):
14622 Change type to hash_table. Update dependent calls and types.
14624 * cselib.c (htab_t cselib_hash_table):
14625 Change type to hash_table. Update dependent calls and types.
14627 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
14628 (hash_string_slot_node): Move implementation into lto-streamer.h
14629 struct string_slot_hasher.
14630 (eq_string_slot_node): Likewise.
14632 * data-streamer-out.c: Update output_block::string_hash_table
14633 dependent calls and types.
14635 * dwarf2cfi.c (htab_t trace_index):
14636 Change type to hash_table. Update dependent calls and types.
14638 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
14639 Change type to hash_table. Update dependent calls and types.
14640 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
14641 (htab_t optimize_external_refs::map): Likewise.
14642 (htab_t output_comp_unit::extern_map): Likewise.
14643 (htab_t output_comdat_type_unit::extern_map): Likewise.
14644 (htab_t output_macinfo::macinfo_htab): Likewise.
14645 (htab_t optimize_location_lists::htab): Likewise.
14646 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
14648 * except.c (htab_t ehspec_hash_type):
14649 Change type to hash_table. Update dependent calls and types.
14650 (assign_filter_values::ttypes): Likewise.
14651 (assign_filter_values::ehspec): Likewise.
14652 (sjlj_assign_call_site_values::ar_hash): Likewise.
14653 (convert_to_eh_region_ranges::ar_hash): Likewise.
14655 * gcse.c (htab_t pre_ldst_table):
14656 Change type to hash_table. Update dependent calls and types.
14658 * ggc-common.c (htab_t saving_htab):
14659 Change type to hash_table. Update dependent calls and types.
14660 (htab_t loc_hash): Likewise.
14661 (htab_t ptr_hash): Likewise.
14662 (call_count): Rename ggc_call_count.
14663 (call_alloc): Rename ggc_call_alloc.
14664 (loc_descriptor): Rename make_loc_descriptor.
14665 (add_statistics): Rename ggc_add_statistics.
14667 * ggc-common.c (saving_htab):
14668 Change type to hash_table. Update dependent calls and types.
14670 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
14671 (push_gimplify_context): Likewise.
14672 (pop_gimplify_context): Likewise.
14673 (struct gimple_temp_hash_elt): Added.
14674 (struct gimplify_hasher): Likewise.
14675 (struct gimplify_ctx.temp_htab):
14676 Change type to hash_table. Update dependent calls and types.
14678 * gimple-fold.c: Include gimplify-ctx.h.
14680 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
14681 Change type to hash_table. Update dependent calls and types.
14682 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
14683 avoid potential global name collision.
14685 * gimplify.c: Include gimplify-ctx.h.
14686 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
14687 (htab_t gimplify_ctx::temp_htab):
14688 Update dependent calls and types for new type hash_table.
14689 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
14690 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
14692 * gimplify-ctx.h: New.
14693 (struct gimple_temp_hash_elt): Move from gimplify.c.
14694 (class gimplify_hasher): New.
14695 (struct gimplify_ctx): Move from gimple.h.
14696 (htab_t gimplify_ctx::temp_htab):
14697 Change type to hash_table. Update dependent calls and types.
14699 * graphite-clast-to-gimple.c: Include graphite-htab.h.
14700 (htab_t ivs_params::newivs_index):
14701 Change type to hash_table. Update dependent calls and types.
14702 (htab_t ivs_params::params_index): Likewise.
14703 (htab_t print_generated_program::params_index): Likewise.
14704 (htab_t gloog::newivs_index): Likewise.
14705 (htab_t gloog::params_index): Likewise.
14707 * graphite.c: Include graphite-htab.h.
14708 4htab_t graphite_transform_loops::bb_pbb_mapping):
14709 Change type to hash_table. Update dependent calls and types.
14711 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
14712 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
14713 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
14715 * graphite-dependences.c: Include graphite-htab.h.
14716 (loop_is_parallel_p): Change hash table type of parameter.
14718 * graphite-htab.h: New.
14719 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
14720 (extern find_pbb_via_hash): Move from graphite-poly.h.
14721 (extern loop_is_parallel_p): Move from graphite-poly.h.
14722 (extern get_loop_body_pbbs): Move from graphite-poly.h.
14724 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
14725 (extern loop_is_parallel_p): Move to graphite-htab.h.
14726 (extern get_loop_body_pbbs): Move to graphite-htab.h.
14728 * haifa-sched.c (htab_t delay_htab):
14729 Change type to hash_table. Update dependent calls and types.
14730 (htab_t delay_htab_i2): Likewise.
14732 * ira-color.c (htab_t allocno_hard_regs_htab):
14733 Change type to hash_table. Update dependent calls and types.
14735 * ira-costs.c (htab_t cost_classes_htab):
14736 Change type to hash_table. Update dependent calls and types.
14738 * loop-invariant.c (htab_t merge_identical_invariants::eq):
14739 Change type to hash_table. Update dependent calls and types.
14741 * loop-iv.c (htab_t bivs):
14742 Change type to hash_table. Update dependent calls and types.
14744 * loop-unroll.c (htab_t opt_info::insns_to_split):
14745 Change type to hash_table. Update dependent calls and types.
14746 (htab_t opt_info::insns_with_var_to_expand): Likewise.
14748 * lto-streamer.h (struct string_slot): Move from data-streamer.h
14749 (struct string_slot_hasher): New.
14750 (htab_t output_block::string_hash_table):
14751 Change type to hash_table. Update dependent calls and types.
14753 * lto-streamer-in.c (freeing_string_slot_hasher): New.
14754 (htab_t file_name_hash_table):
14755 Change type to hash_table. Update dependent calls and types.
14757 * lto-streamer-out.c: Update output_block::string_hash_table dependent
14760 * lto-streamer.c (htab_t tree_htab):
14761 Change type to hash_table. Update dependent calls and types.
14763 * omp-low.c: Include gimplify-ctx.h.
14765 * passes.c (htab_t name_to_pass_map):
14766 Change type to hash_table. Update dependent calls and types.
14767 (pass_traverse): Rename to passes_pass_traverse.
14769 * plugin.c (htab_t event_tab):
14770 Change type to hash_table. Update dependent calls and types.
14772 * postreload-gcse.c (htab_t expr_table):
14773 Change type to hash_table. Update dependent calls and types.
14774 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
14776 * sese.c (debug_rename_map_1): Make extern.
14777 (htab_t copy_bb_and_scalar_dependences::rename_map):
14778 Change type to hash_table. Update dependent calls and types.
14780 * sese.h (extern debug_rename_map): Move to .c file.
14782 * store-motion.c (htab_t store_motion_mems_table):
14783 Change type to hash_table. Update dependent calls and types.
14785 * trans-mem.c (htab_t tm_new_mem_hash):
14786 Change type to hash_table. Update dependent calls and types.
14788 * tree-browser.c (htab_t TB_up_ht):
14789 Change type to hash_table. Update dependent calls and types.
14791 * tree-cfg.c (htab_t discriminator_per_locus):
14792 Change type to hash_table. Update dependent calls and types.
14794 * tree-complex.c: Include tree-hasher.h
14795 (htab_t complex_variable_components):
14796 Change type to hash_table. Update dependent calls and types.
14798 * tree-eh.c (htab_t finally_tree):
14799 Change type to hash_table. Update dependent calls and types.
14801 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
14802 struct int_tree_hasher.
14803 (extern int_tree_map_eq): Likewise.
14804 (uid_decl_map_hash): Removed.
14805 (extern decl_tree_map_eq): Likewise.
14807 * tree-hasher.h: New.
14808 (struct int_tree_hasher): New.
14809 (typedef int_tree_htab_type): New.
14811 * tree-inline.c: Include gimplify-ctx.h.
14813 * tree-mudflap.c: Include gimplify-ctx.h.
14815 * tree-parloops.c: Include tree-hasher.h.
14816 (htab_t eliminate_local_variables_stmt::decl_address):
14817 Change type to hash_table. Update dependent calls and types.
14818 (htab_t separate_decls_in_region::decl_copies): Likewise.
14820 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
14821 Change type to hash_table. Update dependent calls and types.
14823 * tree-sra.c (candidates):
14824 Change type to hash_table. Update dependent calls and types.
14826 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
14828 (int_tree_map_hash): Likewise.
14830 * tree-ssa-dom.c (htab_t avail_exprs):
14831 Change type to hash_table. Update dependent calls and types.
14833 * tree-ssa-live.c (var_map_base_init::tree_to_index):
14834 Change type to hash_table. Update dependent calls and types.
14836 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
14837 Change type to hash_table. Update dependent calls and types.
14839 * tree-ssa-phiopt.c (seen_ssa_names):
14840 Change type to hash_table. Update dependent calls and types.
14842 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
14843 Change type to hash_table. Update dependent calls and types.
14845 * tree-ssa-uncprop.c (equiv):
14846 Change type to hash_table. Update dependent calls and types.
14848 2013-04-25 Jakub Jelinek <jakub@redhat.com>
14850 PR rtl-optimization/57003
14851 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
14852 call note_stores with kill_clobbered_value callback again after
14853 killing regs_invalidated_by_call.
14855 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14857 * config/aarch64/aarch64-simd.md
14858 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
14859 (aarch64_simd_bsl<mode>): Likewise.
14860 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
14862 2013-04-25 Marek Polacek <polacek@redhat.com>
14864 PR tree-optimization/57066
14865 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
14867 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
14869 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
14871 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14873 * config/aarch64/aarch64-builtins.c
14874 (aarch64_fold_builtin): New.
14875 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
14876 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
14877 * config/aarch64/aarch64-simd-builtins.def (abs): New.
14878 * config/aarch64/arm_neon.h
14879 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
14881 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14882 Tejas Belagod <tejas.belagod@arm.com>
14884 * config/aarch64/aarch64-builtins.c
14885 (aarch64_gimple_fold_builtin): New.
14886 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
14887 * config/aarch64/aarch64-simd-builtins.def (addv): New.
14888 * config/aarch64/aarch64-simd.md (addpv4sf): New.
14889 (addvv4sf): Update.
14890 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
14892 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14894 * config/aarch64/aarch64.md
14895 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
14897 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14899 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
14900 (*ngcsi_uxtw): New pattern.
14902 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14903 Julian Brown <julian@codesourcery.com>
14905 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
14906 (TB_DREG): Add T_V4HF.
14907 (v4hf_UP): New macro.
14908 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14909 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14910 Handle initialisation of V4HF. Adjust initialisation of reinterpret
14912 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14913 (arm_vector_mode_supported_p): Handle V4HF.
14914 (arm_mangle_map): Handle V4HFmode.
14915 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
14916 * config/arm/arm_neon_builtins.def: Add entries for
14917 vcvtv4hfv4sf, vcvtv4sfv4hf.
14918 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
14919 (neon_vcvtv4hfv4sf): Likewise.
14920 * config/arm/neon-gen.ml: Handle half-precision floating point
14922 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
14923 * config/arm/arm_neon.h: Regenerate.
14924 * config/arm/neon.ml (type elts): Add F16.
14925 (type vectype): Add T_float16x4, T_floatHF.
14926 (type vecmode): Add V4HF.
14927 (type features): Add Requires_FP_bit feature.
14928 (elt_width): Handle F16.
14929 (elt_class): Likewise.
14930 (elt_of_class_width): Likewise.
14931 (mode_of_elt): Refactor.
14932 (type_for_elt): Handle F16, fix error messages.
14933 (vectype_size): Handle T_float16x4.
14934 (vcvt_sh): New function.
14935 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
14936 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
14937 (string_of_mode): Handle V4HF.
14938 * doc/arm-neon-intrinsics.texi: Regenerate.
14940 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14942 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
14943 format specifier in 'X' case.
14945 2013-04-25 Alan Modra <amodra@gmail.com>
14948 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
14949 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
14950 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
14951 Repeat for many other rotate/shift and mask patterns using subregs.
14952 Name lshiftrt insns.
14953 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
14954 on WORDS_BIG_ENDIAN.
14956 2013-04-25 Alan Modra <amodra@gmail.com>
14958 * config.gcc: Support little-endian powerpc-linux targets.
14959 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
14960 (LINK_OS_LINUX_SPEC): Define.
14961 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
14962 Preserve MASK_LITTLE_ENDIAN.
14963 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
14964 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
14965 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
14966 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
14967 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
14968 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
14969 Correct fp word order for little-endian. Don't shift toc entries
14970 smaller than a word for little-endian.
14971 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
14972 (bswapdi2 splits): Correct low-part subreg for little-endian.
14973 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
14974 low/high where such is correct only for be.
14975 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
14976 little-endian for -mcall-aixdesc.
14978 2013-04-25 Alan Modra <amodra@gmail.com>
14980 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
14981 replace_equiv_address_nv.
14983 2013-04-25 Alan Modra <amodra@gmail.com>
14985 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
14987 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14990 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14991 * rtl.h (struct rtx_def): ...
14993 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14995 PR rtl-optimizations/57046
14996 * lra-constraints (split_reg): Set up lra_risky_transformations_p
14997 for multi-reg splits.
14999 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15001 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15003 2013-04-24 Sterling Augustine <saugustine@google.com>
15005 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15006 (comp_dir_string, debug_str_dwo_section): New.
15007 (DEBUG_STR_DWO_SECTION): Rename to ...
15008 (DEBUG_DWO_STR_SECTION): ... this.
15009 (DEBUG_NORM_STR_SECTION): Delete.
15010 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15011 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15012 (find_AT_string): Move most logic to ...
15013 (find_AT_string_in_table): ... here. New.
15014 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15015 add_skeleton_AT_string. Delete logic.
15016 (output_skeleton_debug_sections): Remove call to
15017 add_top_level_skeleton_die_attrs.
15018 (add_comp_dir_attribute): Move logic to comp_dir_string.
15019 (dwarf2out_init): Initialize debug_str_dwo_section.
15020 (output_indirect_string): Call find_string_form.
15021 (output_indirect_strings): Rewrite.
15022 (prune_unused_types): Empty skeleton_debug_str_hash.
15023 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15024 (dwarf2out_finish): Call output_indirect_strings.
15026 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15028 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15030 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15032 * rtl.h (struct rtx_def): Add comment for field jump.
15033 (LRA_SUBREG_P): New macro.
15034 * recog.c (register_operand): Check LRA_SUBREG_P.
15035 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15037 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15038 (lra_final_code_change): Skip subreg change for operators.
15039 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15040 if there are no operand changes.
15041 * lra-constraints.c (curr_insn_set): New.
15042 (match_reload): Set LRA_SUBREG_P.
15043 (emit_spill_move): Ditto.
15044 (check_and_process_move): Use curr_insn_set. Process only single
15045 set insns. Don't initialize sec_mem_p and change_p.
15046 (simplify_operand_subreg): Use LRA_SUBREG_P.
15047 (reg_in_class_p): New function.
15048 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15049 of #ifdef. Add code to remove cycling.
15050 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15051 non-null disp. Reload inner instead of disp when base and index
15052 are null. Try to put lo_sum into register.
15053 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15054 (check_and_process_move): Move code for move cost check to
15055 simple_move_p. Remove equiv_substitution.
15056 (simple_move_p): New function.
15057 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15058 curr_insn_set. Call check_and_process_move only for single set
15059 insns. Use the new function. Move call of check_and_process_move
15060 after operand equiv substitution and address process.
15062 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15064 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15065 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15066 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15068 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15070 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15072 2013-04-24 Marek Polacek <polacek@redhat.com>
15074 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15075 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15076 (select_loops_exit_conditions): Likewise.
15077 (number_of_iterations_for_all_loops): Likewise.
15078 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15079 (scev_analysis): Likewise.
15081 2013-04-02 Catherine Moore <clm@codesourcery.com>
15082 Chao-ying Fu <fu@mips.com>
15084 * config/mips/micromips.md (jraddiusp): New pattern.
15085 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15086 instruction if possible.
15088 2013-04-24 Alan Modra <amodra@gmail.com>
15090 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15092 2013-04-24 Julian Brown <julian@codesourcery.com>
15093 Chung-Lin Tang <cltang@codesourcery.com>
15095 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15096 dependency behavior in enumeration type DIE generation. Add TODO note
15097 to comments about future DW_FORM_sdata/udata re-work of related code.
15099 2013-04-23 Lawrence Crowl <crowl@google.com>
15101 * Makefile.in: Update as needed below.
15103 * hash-table.h (class hash_table):
15104 Correct many methods with parameter types compare_type to the correct
15105 value_type. (Correct code was unlikely to notice the change.)
15106 (hash_table::elements_with_deleted) New.
15107 (class hashtable::iterator): New.
15108 (hashtable::begin()): New.
15109 (hashtable::end()): New.
15110 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15112 * statistics.c (statistics_hashes):
15113 Change type to hash_table. Update dependent calls and types.
15115 * tree-into-ssa.c (var_infos):
15116 Change type to hash_table. Update dependent calls and types.
15118 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15119 Change type to hash_table. Update dependent calls and types.
15121 * tree-ssa-loop-im.c (struct mem_ref.refs):
15122 Change type to hash_table. Update dependent calls and types.
15124 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15125 Change type to hash_table. Update dependent calls and types.
15127 * tree-ssa-sccvn.c (vn_tables_s::nary):
15128 Change type to hash_table. Update dependent calls and types.
15129 (vn_tables_s::phis): Likewise.
15130 (vn_tables_s::references): Likewise.
15132 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15133 (vn_reference_eq): Update parameter and return types.
15135 * tree-ssa-structalias.c (pointer_equiv_class_table):
15136 Change type to hash_table. Update dependent calls and types.
15137 (location_equiv_class_table): Likewise.
15139 * tree-vect-data-refs.c: Consequential changes for making
15140 peeling a hash_table.
15142 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15143 (destroy_loop_vec_info): Dependent hash_table update.
15145 * tree-vectorizer.h (peeling_htab):
15146 Change type to hash_table. Update dependent calls and types.
15148 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15150 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15151 to check the register content is equal or not.
15152 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15153 to assign register content record.
15154 * lra-eliminations.c (update_reg_eliminate): Use
15155 lra_update_reg_val_offset to update register content offset.
15156 * lra-int.h (struct lra_reg): Add offset member.
15157 (lra_reg_val_equal_p): New static inline function.
15158 (lra_update_reg_val_offset): New static inline function.
15159 (lra_assign_reg_val): New static inline function.
15160 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15161 to assign register content record.
15162 (initialize_lra_reg_info_element): Initial offset to zero.
15164 2013-04-23 Catherine Moore <clm@codesourcery.com>
15166 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15167 operands. Record compression.
15169 2013-04-23 Xinliang David Li <davidxl@google.com>
15171 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15173 2013-04-23 Richard Biener <rguenther@suse.de>
15175 PR middle-end/57036
15176 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15177 parameter, only add abnormal goto edges from the copied body
15178 if the call could perform abnormal gotos.
15179 (copy_cfg_body): Adjust.
15181 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15183 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15185 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15187 * coretypes.h (gimple_stmt_iterator): Add struct to make
15190 2013-04-23 Richard Biener <rguenther@suse.de>
15192 PR tree-optimization/57026
15193 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15194 from SSA names occuring in abnormal PHI nodes.
15196 2013-04-22 Andi Kleen <ak@linux.intel.com>
15198 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15200 2013-04-22 Andi Kleen <ak@linux.intel.com>
15202 * lto/lto.c (print_lto_report_1): Declare early.
15203 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15205 2013-04-22 Andi Kleen <ak@linux.intel.com>
15207 * common.opt (-flto-report-wpa): Add.
15208 * doc/invoke.texi (-flto-report-wpa): Add.
15209 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15212 2013-04-22 Xinliang David Li <davidxl@google.com>
15214 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15215 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15216 * Makefile.in: New dependency
15218 David Daney <ddaney.cavm@gmail.com>
15220 * configure.ac (gcc_cv_as_micromips_support): Use the
15221 --fatal-warnings option.
15222 * configure: Regenerate.
15224 2013-04-22 Marek Polacek <polacek@redhat.com>
15227 * tsan.c (instrument_expr): Don't instrument expression
15228 in case its size is zero.
15230 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15234 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15236 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15238 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15240 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15241 (gimple_stmt_iterator): New typedef.
15242 * gimple.h (gimple_stmt_iterator): Rename to...
15243 (gimple_stmt_iterator_d): ... This.
15244 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15245 trees be valid for GIMPLE and GENERIC.
15246 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15247 * gimple-fold.c (gimple_fold_call): Call target hook
15248 gimple_fold_builtin.
15249 * hooks.c (hook_bool_gsiptr_false): New.
15250 * hooks.h (hook_bool_gsiptr_false): New.
15251 * target.def (fold_stmt): New.
15252 * doc/tm.texi: Regenerate.
15254 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15257 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15258 a set sp if no stack realignment.
15260 2013-04-22 Nick Clifton <nickc@redhat.com>
15262 * config.gcc (tilegx-linux): Extend extra_objs rather than
15264 (tilepro-linux): Likewise.
15266 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15268 * config/aarch64/aarch64-builtins.c
15270 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15271 (VAR<1-12>): Add MAP parameter.
15272 (BUILTIN_*): Likewise.
15273 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15274 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15275 (aarch64_ushl_n<mode>): Likewise.
15276 (aarch64_sshr_n<mode>): Likewise.
15277 (aarch64_ushr_n<mode>): Likewise.
15278 (aarch64_<maxmin><mode>): Likewise.
15279 (aarch64_sqrt<mode>): Likewise.
15280 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15281 (vshr<q>_n_*): Likewise.
15283 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15285 * config/aarch64/aarch64-builtins.c
15286 (aarch64_simd_builtin_type_mode): Handle SF types.
15288 (BUILTIN_GPF): Define.
15289 (aarch64_init_simd_builtins): Handle SF types.
15290 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15291 (frecps): Likewise.
15292 (frecpx): Likewise.
15293 * config/aarch64/aarch64-simd.md
15294 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15295 (aarch64_frecpe<mode>): New.
15296 (aarch64_frecps<mode>): Likewise.
15297 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15298 (v8type): Add frecp<esx>.
15299 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15300 (aarch64_frecps<mode>): Likewise.
15301 * config/aarch64/iterators.md (FRECP): New.
15302 (frecp_suffix): Likewise.
15303 * config/aarch64/arm_neon.h
15304 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15306 2013-04-22 Christian Bruel <christian.bruel@st.com>
15309 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15310 (REG_CLASS_NAMES): Idem.
15311 (REG_CLASS_CONTENTS): Idem.
15312 (REGCLASS_HAS_FP_REG): Idem.
15313 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15314 (sh_conditional_register_usage): Idem.
15316 2013-04-21 Jeff Law <law@redhat.com>
15318 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15319 (ssa_forward_propagate_and_combine): Use it.
15321 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15323 * lra.c: Update the flow chart diagram.
15325 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15327 PR rtl-optimization/56847
15328 * lra-constraints.c (process_alt_operands): Discourage alternative
15329 with non-matche doffsettable memory constraint fro memory with
15332 2013-04-19 Richard Biener <rguenther@suse.de>
15334 PR tree-optimization/56982
15335 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15337 * gimplify.c (gimplify_call_expr): Notice special calls.
15338 (gimplify_modify_expr): Likewise.
15339 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15340 abnormal control flow receivers.
15341 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15342 in the same way as cfun->has_nonlocal_labels.
15343 (gimple_purge_dead_abnormal_call_edges): Likewise.
15344 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15345 receivers start a basic-block.
15347 2013-04-19 Richard Biener <rguenther@suse.de>
15349 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15351 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15352 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15353 (SLP_TREE_LOAD_PERMUTATION): Add.
15354 (vect_transform_slp_perm_load): Adjust prototype.
15355 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15356 (vect_free_slp_instance): Likewise.
15357 (vect_create_new_slp_node): Likewise.
15358 (vect_supported_slp_permutation_p): Remove.
15359 (vect_slp_rearrange_stmts): Adjust.
15360 (vect_supported_load_permutation_p): Likewise. Inline
15361 vect_supported_slp_permutation_p here.
15362 (vect_analyze_slp_instance): Compute load permutations per
15363 slp node instead of per instance.
15364 (vect_get_slp_defs): Adjust.
15365 (vect_transform_slp_perm_load): Likewise.
15366 (vect_schedule_slp_instance): Remove redundant code.
15367 (vect_schedule_slp): Remove hack for PR56270, add it ...
15368 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15369 CSE loads for SLP. Adjust.
15371 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15373 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
15374 spelling in two comments.
15376 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15379 * config/arm/arm.c (load_multiple_sequence): Require SP
15380 as base register for loads if SP is in the register list.
15382 2013-04-19 Martin Jambor <mjambor@suse.cz>
15384 PR tree-optimization/56718
15385 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
15386 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
15387 and made public. Adjusted all callers.
15388 (ipa_intraprocedural_devirtualization): New function.
15389 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
15390 (ipa_intraprocedural_devirtualization): Likewise.
15391 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
15393 2013-04-19 Richard Biener <rguenther@suse.de>
15395 PR tree-optimization/57000
15396 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
15398 2013-04-19 Terry Guo <terry.guo@arm.com>
15400 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
15402 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
15403 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
15404 (cortex_m4_fmacs): Use new reservations.
15405 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
15407 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
15409 PR rtl-optimization/56999
15410 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
15412 (lra_coalesce): Remove split_origin_bitmap and related code.
15413 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
15414 ranges if necessary.
15416 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
15418 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
15420 (ix86_expand_call): Remove clobbered_registers array and use
15421 x86_64_ms_sysv_extra_clobbered_registers instead.
15422 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
15424 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
15426 * config/i386/i386.md (*call_rex64_ms_sysv): Use
15427 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
15428 (*call_value_rex64_ms_sysv): Ditto.
15430 2013-04-18 Cary Coutant <ccoutant@google.com>
15432 * dwarf2out.c (output_pubnames): Check die_perennial_p of
15433 parent instead of die_mark.
15435 2013-04-18 Diego Novillo <dnovillo@google.com>
15437 * gimple.c (create_gimple_tmp): New.
15438 (get_expr_type): New.
15439 (build_assign): New.
15440 (build_type_cast): New.
15441 * gimple.h (enum ssa_mode): Define.
15442 (gimple_seq_set_location): New.
15443 * asan.c (build_check_stmt): Change some gimple_build_* calls
15444 to use build_assign and build_type_cast.
15446 2013-04-18 Richard Biener <rguenther@suse.de>
15448 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
15449 handle negative step. Remove redundant checks.
15450 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
15451 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
15452 for negative step and grouped loads fail to vectorize.
15454 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15456 * emit-rtl.c (reset_insn_used_flags): New function.
15457 (reset_all_used_flags): Use it.
15458 (verify_insn_sharing): New function.
15459 (verify_rtl_sharing): Fix verification for SEQUENCEs.
15461 2013-04-18 Jakub Jelinek <jakub@redhat.com>
15463 PR tree-optimization/56984
15464 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
15465 and (x >> M) >= N don't register any assertion if N << M is the
15468 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15470 * lower-subreg.c (resolve_simple_move): If called self-recursive,
15471 do not delete_insn insns that have not yet been emitted, only
15472 unlink them with remove_insn.
15473 * df-scan.c (df_insn_delete): Revert r197492.
15475 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
15477 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
15478 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
15480 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15482 * config/arm/arm.md (movsicc_insn): Convert define_insn into
15483 define_insn_and_split.
15484 (and_scc,ior_scc,negscc): Likewise.
15485 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
15487 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15489 * config/arm/arm.c (use_return_insn): Return 0 for targets that
15490 can benefit from using a sequence of LDRD instructions in epilogue
15491 instead of a single LDM instruction.
15493 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
15496 * doc/extend.texi: Fix typo.
15498 2013-04-17 Richard Biener <rguenther@suse.de>
15500 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
15501 (vect_build_slp_tree): ... here.
15502 (vect_build_slp_tree_1): Compute which stmts of the SLP group
15503 match. Remove special-casing of mismatched complex loads.
15504 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
15505 re-try the match with swapped commutative operands.
15506 (vect_supported_load_permutation_p): Remove special-casing of
15507 mismatched complex loads.
15508 (vect_analyze_slp_instance): Adjust.
15510 2013-04-17 Richard Biener <rguenther@suse.de>
15512 PR rtl-optimization/56921
15513 * cfgloop.h (struct loop): Add simple_loop_desc member.
15514 (struct niter_desc): Mark with GTY(()).
15515 (simple_loop_desc): Do not use aux field but simple_loop_desc.
15516 * loop-iv.c (get_simple_loop_desc): Likewise.
15517 (free_simple_loop_desc): Likewise.
15520 2013-04-16 Richard Biener <rguenther@suse.de>
15522 PR rtl-optimization/56921
15523 * loop-init.c (pass_rtl_move_loop_invariants): Add
15524 TODO_do_not_ggc_collect to todo_flags_finish.
15525 (pass_rtl_unswitch): Same.
15526 (pass_rtl_unroll_and_peel_loops): Same.
15527 (pass_rtl_doloop): Same.
15529 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
15531 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
15532 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
15533 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
15535 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
15536 * tree-streamer.c (record_common_node): Adjust reference in comment.
15538 2013-04-17 Terry Guo <terry.guo@arm.com>
15540 * config/arm/cortex-m4.md: Add a new bypass.
15542 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15544 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
15546 (*subs_<optab><mode>_multp2): New pattern.
15547 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15548 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15550 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15552 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
15553 (*subs_mul_imm_<mode>): New pattern.
15555 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
15558 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
15559 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
15560 (vsx_movti_32bit): Change j->wa to O->wa.
15562 2013-04-16 Richard Biener <rguenther@suse.de>
15564 PR rtl-optimization/56921
15565 * loop-init.c (pass_rtl_move_loop_invariants): Add
15566 TODO_do_not_ggc_collect to todo_flags_finish.
15567 (pass_rtl_unswitch): Same.
15568 (pass_rtl_unroll_and_peel_loops): Same.
15569 (pass_rtl_doloop): Same.
15571 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
15573 * config/arm/arm.c (emit_multi_reg_push): New declaration
15574 for an existing function.
15575 (arm_emit_strd_push): New function.
15576 (arm_expand_prologue): Used here.
15577 (arm_emit_ldrd_pop): New function.
15578 (arm_expand_epilogue): Used here.
15579 (arm_get_frame_offsets): Update condition.
15580 (arm_emit_multi_reg_pop): Add a special case for load of a single
15581 register with writeback.
15583 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
15585 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
15588 2013-04-16 Richard Biener <rguenther@suse.de>
15590 PR tree-optimization/56756
15591 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
15592 (first_mem_ref_loc): New.
15593 (execute_sm): Place the load temporarily before a previous
15594 access instead of in the latch edge to ensure its SSA dependencies
15595 are defined at points dominating the load.
15597 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
15599 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
15600 correct fix by moving header and footer insn to the footer of
15601 the merged basic block. Clear BB_END of the merged-away block.
15603 PR middle-end/43631
15604 * emit-rtl.c (make_note_raw): New function.
15605 (link_insn_into_chain): New static inline function.
15606 (add_insn): Use it.
15607 (add_insn_before, add_insn_after): Factor insn chain linking code...
15608 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
15609 using link_insn_into_chain.
15610 (note_outside_basic_block_p): New helper function for emit_note_after
15611 and emit_note_before.
15612 (emit_note_after): Use nobb variant of add_insn_after if the note
15613 should not be contained in a basic block.
15614 (emit_note_before): Use nobb variant of add_insn_before if the note
15615 should not be contained in a basic block.
15616 (emit_note_copy): Use make_note_raw.
15617 (emit_note): Likewise.
15618 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
15619 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
15620 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
15621 the moved barrier the tail of the basic block it follows.
15622 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
15624 2013-04-15 Jakub Jelinek <jakub@redhat.com>
15626 PR tree-optimization/56962
15627 * gimple-ssa-strength-reduction.c (record_increment): Only set
15628 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
15629 either rhs1 or rhs2 is equal to c->base_expr.
15631 2013-04-15 Richard Biener <rguenther@suse.de>
15633 PR tree-optimization/56933
15634 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
15636 (GROUP_READ_WRITE_DEPENDENCE): Remove.
15637 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
15638 * tree-vect-data-refs.c (vect_analyze_group_access): Move
15639 dependence check ...
15640 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
15642 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
15643 GROUP_READ_WRITE_DEPENDENCE.
15645 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15647 * emit-rtl.c (reset_all_used_flags): New function.
15648 (verify_rtl_sharing): Call reset_all_used_flags before and after
15649 performing the checks.
15651 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15653 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
15654 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
15655 * config/arm/constraints.md (De): New constraint.
15656 * config/arm/neon.md (anddi3_neon): Delete.
15657 (neon_vand<mode>): Expand to standard anddi3 pattern.
15658 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
15659 Move earlier in the file.
15660 (neon_inv_logic_op2): Likewise.
15661 (arm_anddi_operand_neon): New predicate.
15663 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15665 * configure.ac (gcc_cv_ld_as_needed): Set
15666 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
15667 Use -z ignore, -z record on *-*-solaris2*.
15668 (HAVE_LD_AS_NEEDED): Update comment.
15669 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
15670 * configure: Regenerate.
15671 * config.in: Regenerate.
15672 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
15673 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
15674 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
15675 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
15676 equivalents. Fix markup.
15677 * doc/tm.texi: Regenerate.
15679 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
15681 * config/i386/i386.opt: New option mstack-protector-guard=.
15682 * config/i386/i386-opts.h: Add enum stack_protector_guard.
15683 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
15684 TARGET_SSP_TLS_GUARD.
15685 * config/i386/i386.c (ix86_option_override_internal): Set
15686 ix86_stack_protector_guard.
15687 * config/i386/i386.md (stack_protect_set): Enable for
15688 TARGET_SSP_TLS_GUARD only.
15689 (stack_protect_set_<mode>): Ditto.
15690 (stack_protect_test): Ditto.
15691 (stack_protect_test_<mode>): Ditto.
15692 * doc/invoke.texi (i386 Option): Document.
15694 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
15697 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
15698 (S_MODES): Set H_MODE bit.
15699 (SF_MODES): Set only S_MODE and SF_MODE bits.
15700 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
15701 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
15702 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
15703 <MODE_FLOAT>: Likewise.
15705 2013-04-15 Joey Ye <joey.ye@arm.com>
15707 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
15709 2013-04-15 Joey Ye <joey.ye@arm.com>
15711 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
15713 (thumb_far_jump_used_p): Count instruction size and set
15716 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
15718 * reorg.c (fill_simple_delay_slots): Reindent block of code.
15719 * resource.c (mark_target_live_regs): Reformat conditional block.
15721 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15723 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
15724 notes, they are emitted only just before final.
15725 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
15727 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15729 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
15730 * cfgrtl.c (delete_insn): Call it here instead.
15731 * lra-spills.c (lra_final_code_change): Use delete_insn.
15732 * haifa-sched.c (sched_remove_insn): Likewise.
15733 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
15734 returning to the nop pool.
15735 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
15736 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
15738 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15740 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
15741 * doc/tm.texi: Regenerated.
15743 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
15745 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
15748 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15750 * df-core.c (df_find_def): Compare register numbers.
15751 (df_find_use): Likewise.
15753 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
15756 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
15757 lra_in_progress for return.
15759 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15761 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
15762 define_insn into define_insn_and_split and emit movsicc patterns.
15764 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15766 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
15768 2013-04-12 Richard Biener <rguenther@suse.de>
15770 * tree-pass.h (TODO_do_not_ggc_collect): New.
15771 * passes.c (execute_one_ipa_transform_pass): Honor
15772 TODO_do_not_ggc_collect.
15773 (execute_one_pass): Likewise.
15776 2013-04-10 Richard Biener <rguenther@suse.de>
15778 * passes.c (init_optimization_passes): Remove reload pass.
15779 * ira.c (do_reload): Merge into ...
15781 (rest_of_handle_reload): Remove.
15782 (pass_reload): Likewise.
15783 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15784 of reload for vzeroupper pass placement.
15786 2013-04-12 Jakub Jelinek <jakub@redhat.com>
15788 PR tree-optimization/56918
15789 PR tree-optimization/56920
15790 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
15791 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
15792 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
15793 use wide_mul_with_sign method.
15795 2013-04-12 Richard Biener <rguenther@suse.de>
15797 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
15798 not be considered a gimple constant.
15800 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
15802 * fold-const.c (const_binop): Handle vector shifts by a scalar.
15803 (fold_binary_loc): Call const_binop also for mixed vector-scalar
15806 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
15807 Jakub Jelinek <jakub@redhat.com>
15809 * opts.c: Include diagnostic-color.h.
15810 (common_handle_option): Handle OPT_fdiagnostics_color_.
15811 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
15812 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
15813 (diagnostic-color.o): New.
15814 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
15815 (diagnostic_color_rule): New enum.
15816 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
15817 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
15818 the location string.
15819 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
15820 either NULL, or color kind.
15821 * diagnostic-color.c: New file.
15822 * diagnostic-color.h: New file.
15823 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
15825 * doc/invoke.texi (-fdiagnostics-color): Document.
15826 * pretty-print.h (pp_show_color): Define.
15827 (struct pretty_print_info): Add show_color field.
15828 * diagnostic.c: Include diagnostic-color.h.
15829 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
15830 macros. Colorize error:, warning: etc. strings and also the location
15832 (diagnostic_show_locus): Colorize the caret line.
15833 * pretty-print.c: Include diagnostic-color.h.
15834 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
15835 inside of %< %> quotes or quoted through q format modifier.
15837 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15839 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
15841 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15843 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
15844 code in CC_NZ mode.
15845 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
15848 2013-04-11 Marek Polacek <polacek@redhat.com>
15850 PR tree-optimization/48184
15851 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
15853 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
15855 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
15856 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
15857 (skip_simple_arithmetic): Tidy up.
15858 * tree.h (skip_simple_constant_arithmetic): Declare.
15860 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15862 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
15864 2013-04-11 Richard Biener <rguenther@suse.de>
15866 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15867 generate vector constants.
15869 2013-04-11 Richard Biener <rguenther@suse.de>
15871 PR tree-optimization/56878
15872 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
15873 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
15875 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
15876 Prefer to align the DR with the most invariant base address.
15878 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15880 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
15883 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
15885 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
15886 floating-point vector comparisons against 0.
15888 2013-04-11 Jakub Jelinek <jakub@redhat.com>
15890 PR tree-optimization/56899
15891 * fold-const.c (extract_muldiv_1): Apply distributive law
15892 only if TYPE_OVERFLOW_WRAPS (ctype).
15894 2013-04-11 Bin Cheng <bin.cheng@arm.com>
15897 * ira-costs.c (scan_one_insn): Check whether the source rtx of
15898 loading has side effect.
15900 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
15902 * config/sparc/sparc.c: Include tree-pass.h.
15903 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
15904 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
15905 head of file. Change return type. Split off gate function.
15906 (sparc_gate_work_around_errata): New function.
15907 (pass_work_around_errata): New pass definition.
15908 (insert_pass_work_around_errata) New pass insert definition to
15909 insert pass_work_around_errata just after delayed-branch scheduling.
15910 (sparc_option_override): Insert the pass.
15911 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
15913 2013-04-10 David S. Miller <davem@davemloft.net>
15915 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
15916 or -mcpu=hypersparc.
15918 * target.def (cstore_mode): New hook.
15919 * target.h: Include insn-codes.h
15920 * targhooks.c: Likewise.
15921 (default_cstore_mode): New function.
15922 * targhooks.h: Declare it.
15923 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
15924 * doc/tm.texi: Rebuild.
15925 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
15926 target hook, rather than inspecting the insn_data.
15927 * config/sparc/sparc.c (sparc_cstore_mode): New function.
15928 (TARGET_CSTORE_MODE): Redefine.
15929 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
15931 * config/sparc/predicates.md (cstore_result_operand): New special
15933 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
15934 Use it for operand 0.
15935 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
15936 (*snesi_special): Likewise.
15937 (*snesi_zero): Likewise.
15938 (*seqsi_zero): Likewise.
15939 (*sltu_insn): Likewise.
15940 (*sgeu_insn): Likewise.
15941 (*seqdi_special): Make operand 0 and comparison operation be of
15943 (*snedi_special): Likewise.
15944 (*snedi_special_vis3): Likewise.
15945 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
15946 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
15947 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
15948 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
15949 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
15950 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
15951 (*sltu_extend_sp64): Likewise.
15952 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
15953 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
15954 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
15955 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
15956 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
15958 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
15960 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
15961 (aarch64_start_file): Use the new function.
15963 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15964 Jason Merrill <jason@redhat.com>
15966 * common.opt: Add -gdwarf.
15967 * opts.c (common_handle_option): Handle it.
15968 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
15970 2013-04-10 Richard Biener <rguenther@suse.de>
15972 * passes.c (execute_todo): Do not call ggc_collect conditional here.
15973 (execute_one_ipa_transform_pass): But unconditionally here.
15974 (execute_one_pass): And here.
15975 (init_optimization_passes): Remove reload pass.
15976 * tree-pass.h (TODO_ggc_collect): Remove.
15977 (pass_reload): Likewise.
15978 * ira.c (do_reload): Merge into ...
15980 (rest_of_handle_reload): Remove.
15981 (pass_reload): Likewise.
15982 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15983 of reload for vzeroupper pass placement.
15984 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
15985 and todo_flags_finish of all passes.
15987 2013-04-10 Richard Biener <rguenther@suse.de>
15989 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
15990 first_const_oprnd field, rename first_def_type to first_op_type.
15991 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
15992 (vect_get_and_check_slp_defs): Always use the type of the
15993 operand. Allow mixed vect_external_def, vect_constant_def types.
15994 (vect_get_constant_vectors): Handle mixed vect_external_def,
15995 vect_constant_def types.
15997 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
15999 PR tree-optimization/55524
16000 * tree-ssa-math-opts.c
16001 (convert_mult_to_fma): Don't use an fms construct
16002 when we don't have an fms operation, but fnma, and it looks
16003 likely that we'll be able to use the latter.
16005 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16007 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16009 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16010 inline fail caused by overwritable functions.
16012 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16014 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16015 unnecessary bits in the constant power of two case.
16017 2013-04-10 Richard Biener <rguenther@suse.de>
16019 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16020 broken code swapping operands.
16021 (vect_build_slp_tree): Do not compute load permutations here.
16022 (vect_analyze_slp_instance): Compute load permutations here,
16023 after building the SLP tree.
16025 2013-04-09 Christian Bruel <christian.bruel@st.com>
16027 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16028 of next/prev_real_insn.
16030 2013-04-09 Jan Hubicka <jh@suse.cz>
16032 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16033 Drop aliased parameter.
16034 (function_and_variable_visibility): Do not handle alias pairs.
16035 * cgraph.c (varpool_externally_visible_p): Update prototype.
16036 * varpool.c (varpool_add_new_variable): Update.
16038 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16040 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16042 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16044 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16046 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16048 2013-04-09 Marek Polacek <polacek@redhat.com>
16050 PR tree-optimization/48762
16051 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16053 2013-04-09 Richard Biener <rguenther@suse.de>
16055 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16057 (vect_build_slp_tree): Likewise.
16058 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16059 calculating the cost of a SLP instance.
16060 (vect_analyze_slp_instance): Use it from here, after building
16063 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16065 PR middle-end/56883
16066 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16067 expand_omp_for_static_chunk): Use simple_p = true in
16068 force_gimple_operand_gsi calls when assigning to addressable decls.
16070 2013-04-09 Jeff Law <law@redhat.com>
16072 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16073 when the boolean was created by converting a wider object which
16074 had a boolean range.
16076 2013-04-09 Richard Biener <rguenther@suse.de>
16078 * tree-vectorizer.h (slp_void_p): Remove.
16079 (slp_tree): Typedef before _slp_tree declaration.
16080 (struct _slp_tree): Use a vector of slp_tree as children.
16081 (vect_get_place_in_interleaving_chain): Remove.
16082 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16084 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16086 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16087 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16088 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16089 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16090 Use slp_node instead of slp_void_p and adjust.
16092 2013-04-09 Richard Biener <rguenther@suse.de>
16094 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16095 work that is not necessary.
16097 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16099 PR tree-optimization/56854
16100 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16101 forward into clobber stmts if it would change MEM_REF lhs into
16104 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16106 * tree.c (type_hash_lookup, type_hash_add): Make static.
16107 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16109 2013-04-09 Richard Biener <rguenther@suse.de>
16111 * tree.h (unsave_expr_now): Remove.
16112 * tree-inline.c (mark_local_for_remap_r): Remove.
16113 (unsave_expr_1): Likewise.
16114 (unsave_r): Likewise.
16115 (unsave_expr_now): Likewise.
16116 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16117 (propagate_tree_value): Likewise.
16119 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16121 * doc/rtl.texi (sequence): Rewrite documentation to match the
16122 current use of SEQUENCE rtl objects.
16123 * rtl.def (SEQUENCE): Likewise.
16125 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16126 Update documentation.
16127 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16128 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16130 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16132 2013-04-08 Teresa Johnson <tejohnson@google.com>
16134 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16135 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16137 (estimate_edge_size_and_time): Add comment to suggest using rounding
16139 (estimate_node_size_and_time): Ditto.
16140 (remap_edge_change_prob): Use helper rounding divide methods.
16141 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16142 (gimple_mod_pow2_value_transform): Ditto.
16143 (gimple_mod_subtract_transform): Ditto.
16144 (gimple_ic_transform): Ditto.
16145 (gimple_stringops_transform): Ditto.
16146 * stmt.c (conditional_probability): Ditto.
16147 (emit_case_dispatch_table): Ditto.
16148 * lto-cgraph.c (merge_profile_summaries): Ditto.
16149 * tree-optimize.c (execute_fixup_cfg): Ditto.
16150 * cfgcleanup.c (try_forward_edges): Ditto.
16151 * cfgloopmanip.c (scale_loop_profile): Ditto.
16153 (duplicate_loop_to_header_edge): Ditto.
16154 (lv_adjust_loop_entry_edge): Ditto.
16155 * tree-vect-loop.c (vect_transform_loop): Ditto.
16156 * profile.c (compute_branch_probabilities): Ditto.
16157 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16158 * lto-streamer-in.c (input_cfg): Ditto.
16159 * gimple-streamer-in.c (input_bb): Ditto.
16160 * ipa-cp.c (update_profiling_info): Ditto.
16161 (update_specialized_profile): Ditto.
16162 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16163 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16165 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16166 (compute_trg_info): Ditto.
16167 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16168 (purge_dead_edges): Ditto.
16169 * loop-unswitch.c (unswitch_loop): Ditto.
16170 * cgraphclones.c (cgraph_clone_edge): Ditto.
16171 (cgraph_clone_node): Ditto.
16172 * tree-inline.c (copy_bb): Ditto.
16173 (copy_edges_for_bb): Ditto.
16174 (initialize_cfun): Ditto.
16175 (copy_cfg_body): Ditto.
16176 (expand_call_inline): Ditto.
16178 2013-04-08 Kai Tietz <ktietz@redhat.com>
16180 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16181 TARGET_CYGWIN64 by TARGET_64BIT.
16183 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16185 * config/epiphany/epiphany.md (GPR_1): New constant.
16186 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16187 * config/epiphany/epiphany.c (gen_compare_reg):
16188 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16189 is already in place.
16190 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16191 Don't require being called during rtl expansion; If y operlaps r0,
16193 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16194 (epiphany_expand_epilogue): Likewise.
16196 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16197 Don't use CC_FPmode for ORDERED / UNORDERED.
16198 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16200 * config/epiphany/constraints.md (CnL): New constraint.
16201 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16202 * config/epiphany/predicates.md (add_operand): Allow 1024.
16204 * config/epiphany/epiphany.md (logical_op): New code iterator.
16205 (op_mnc): New code attribute.
16206 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16207 (mov_f+1, mov_f+2): New peephole2 patterns.
16209 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16210 (cstoresi4): Also allow re-use of zero result when doing a NE
16211 comparison to a non-zero operand.
16212 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16214 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16215 Use gen_addsi3_i / gen_subsi3_i.
16217 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16221 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16222 contain anything but clobbers, at most one __builtin_stack_restore,
16223 optionally debug stmts and final resx, and if it has at least one
16224 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16225 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16226 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16227 which isn't defaut definition, remove them.
16228 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16229 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16230 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16231 with MEM_REF LHS with SSA_NAME address.
16233 2013-04-08 Jeff Law <law@redhat.com>
16235 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16237 2013-04-08 Richard Biener <rguenther@suse.de>
16239 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16241 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16242 determined vector type.
16243 (vect_analyze_data_refs): Likewise.
16244 (vect_get_new_vect_var): Adjust.
16245 (vect_create_destination_var): Preserve SSA name versions.
16246 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16247 not dump anything here.
16249 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16251 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16252 Add member lr_slot_known.
16253 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16255 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16256 Remove code that sets lr_slot_offset according to what a previous
16257 version of epiphany_emit_save_restore used to do.
16258 (epiphany_emit_save_restore): When doing an lr save or restore,
16259 set/verify lr_slot_known and lr_slot_offset.
16261 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16264 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16267 2013-04-08 Richard Biener <rguenther@suse.de>
16269 * alias.c (find_base_term): Fix thinko in previous change.
16271 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16273 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16274 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16275 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16276 if possible to compute val.
16277 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16278 For QImode integers don't require anything about precision. Use
16279 const_with_all_bytes_same to find out if the constant doesn't have
16280 repeated bytes in it.
16282 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16284 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16287 2013-04-08 Marek Polacek <polacek@redhat.com>
16289 PR rtl-optimization/48182
16290 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16293 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16296 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16297 nuses, make sure we have a label.
16299 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16302 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16303 (rs6000_emit_swdiv_low_precision): Remove.
16304 (rs6000_emit_swdiv): Rewrite to handle between one and four
16305 iterations of Newton-Raphson generally; modify required number of
16306 iterations for some cases.
16307 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16309 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16311 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16312 set-but-unused variable.
16314 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16315 basic blocks of released function bodies garbage-collectable.
16317 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16318 (struct rtl_opt_pass): Add TODO_df_finish.
16320 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16322 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16324 * config/arm/constraints.md (q): New constraint.
16325 * config/arm/ldrdstrd.md: New file.
16326 * config/arm/arm.md (ldrdstrd.md) New include.
16327 (arm_movdi): Use "q" instead of "r" constraint
16328 for double-word memory access.
16329 (movdf_soft_insn): Likewise.
16330 * config/arm/vfp.md (movdi_vfp): Likewise.
16331 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16332 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16333 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16334 (mem_ok_for_ldrd_strd): Likewise.
16335 (output_move_double): Update assertion.
16337 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16339 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16341 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16343 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16344 define_insn_and_split.
16345 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16347 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16349 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16350 define_insn_and_split.
16351 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16352 (shiftsi3_compare): New pattern.
16353 (rrx): New pattern.
16354 * config/arm/unspecs.md (UNSPEC_RRX): New.
16356 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16358 * config/arm/arm.md (negdi_extendsidi): New pattern.
16359 (negdi_zero_extendsidi): Likewise.
16361 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16363 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16364 define_insn_and_split.
16365 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16366 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16368 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16370 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16371 define_insn_and_split.
16372 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
16373 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
16375 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16377 * config/arm/arm.md (subsi3_carryin): New pattern.
16378 (subsi3_carryin_const): Likewise.
16379 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
16380 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
16382 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16384 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
16386 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16388 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
16389 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
16391 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16393 * config/arm/arm.c (arm_expand_builtin): Change fcode
16394 type to unsigned int.
16396 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16398 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
16400 2013-04-04 Ian Lance Taylor <iant@google.com>
16402 * doc/standards.texi (Standards): The Go frontend supports the Go 1
16405 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
16407 PR middle-end/56729
16408 * df-scan.c (df_insn_delete): Disable failing assert.
16410 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16412 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
16413 New function prototype.
16414 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
16415 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
16416 (arm_builtin_vectorized_function): New function.
16418 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16420 * config/arm/arm_neon_builtins.def: New file.
16421 * config/arm/arm.c (neon_builtin_data): Move contents to
16422 arm_neon_builtins.def.
16423 (enum arm_builtins): Include neon builtin definitions.
16424 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
16425 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
16427 2013-04-04 Marek Polacek <polacek@redhat.com>
16429 PR tree-optimization/48186
16430 * predict.c (maybe_hot_frequency_p): Return false if
16431 HOT_BB_FREQUENCY_FRACTION is 0.
16432 (cgraph_maybe_hot_edge_p): Likewise.
16434 2013-04-04 Richard Biener <rguenther@suse.de>
16436 PR tree-optimization/56826
16437 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
16440 2013-04-04 Richard Biener <rguenther@suse.de>
16442 PR tree-optimization/56213
16443 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
16444 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
16446 2013-04-04 Richard Biener <rguenther@suse.de>
16448 PR tree-optimization/56837
16449 * tree-loop-distribution.c (classify_partition): For non-zero
16450 values require that the value has the same precision as its
16451 mode to be useful as memset value.
16453 2013-04-03 Nick Clifton <nickc@redhat.com>
16455 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
16456 (fmssf4): Use fmsf.s on E3V5 architectures.
16457 (fnmasf4): Use fnmaf.s on E3V5 architectures.
16458 (fnmssf4): Use fnmsf.s on E3V5 architectures.
16460 2013-04-03 Jeff Law <law@redhat.com>
16462 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
16463 (lra-eliminations.o): Likewise.
16465 2013-04-03 Teresa Johnson <tejohnson@google.com>
16467 * gcov-io.c (compute_working_sets): Moved most of body of old
16468 compute_working_sets here from profile.c.
16469 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
16470 (gcov_working_set_t): Moved typedef here from basic-block.h
16471 (compute_working_set): Declare.
16472 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
16473 (get_working_sets): Renamed from compute_working_set,
16474 replace most of body with call to new compute_working_sets.
16475 (get_exec_counts): Replace call to compute_working_sets
16476 to get_working_sets.
16477 * profile.h (get_working_sets): Renamed from compute_working_set.
16478 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
16479 to get_working_sets.
16480 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
16481 * gcov-dump.c (dump_working_sets): New function.
16483 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
16485 * hwint.c (sext_hwi, zext_hwi): New functions.
16486 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
16487 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
16488 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
16489 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
16490 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
16491 (sext_hwi, zext_hwi): New functions.
16493 2013-04-03 Jeff Law <law@redhat.com>
16495 PR tree-optimization/56799
16496 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
16497 back test for widening conversion erroneously dropped in prior change.
16499 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16502 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
16503 instead of next_real_insn.
16505 2013-04-03 Marek Polacek <polacek@redhat.com>
16508 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
16510 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16513 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
16515 (thumb1_output_casesi): Likewise.
16516 (thumb2_output_casesi): Likewise.
16518 2013-04-03 Richard Biener <rguenther@suse.de>
16520 PR tree-optimization/56817
16521 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
16523 (tree_unroll_loops_completely_1): ... new function to manually
16524 walk the loop tree, properly defering outer loops of unrolled
16525 loops to later iterations.
16527 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16529 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
16530 (vectorizable_load): Likewise.
16531 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
16532 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
16534 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16536 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
16539 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16541 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
16543 2013-04-03 Bin Cheng <bin.cheng@arm.com>
16545 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
16547 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16549 PR tree-optimization/56790
16550 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
16553 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16555 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
16557 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
16558 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
16561 2013-04-03 Jakub Jelinek <jakub@redhat.com>
16564 * tree.h (force_folding_builtin_constant_p): New decl.
16565 * builtins.c (force_folding_builtin_constant_p): New variable.
16566 (fold_builtin_constant_p): Fold immediately also if
16567 force_folding_builtin_constant_p.
16569 2013-04-03 Richard Biener <rguenther@suse.de>
16571 PR tree-optimization/56812
16572 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
16573 DRs of the same interleaving chain are independent.
16575 2013-04-02 Jason Merrill <jason@redhat.com>
16577 * gdbinit.in (pbb): Use debug fn.
16579 2013-04-02 Lawrence Crowl <crowl@google.com>
16581 * sese.h (struct ivtype_map_elt_s): Remove unused.
16582 (extern debug_ivtype_map): Remove unused.
16583 (extern eq_ivtype_map_elts): Remove unused.
16584 * sese.c (debug_ivtype_map): Removed unused.
16585 (debug_ivtype_map_1): Removed unused.
16586 (debug_ivtype_elt): Remove unused.
16587 (eq_ivtype_map_elts): Remove unused.
16590 2013-04-02 Kai Tietz <ktietz@redhat.com>
16593 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
16594 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
16595 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
16597 (legitimize_pe_coff_symbol): Likewise.
16598 (is_imported_p): New helper-function.
16599 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
16600 for Windows x64 targets.
16601 (ix86_expand_prologue): Optimize for pe-coff targets.
16602 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
16603 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
16604 medium/large code-model.
16605 (legitimize_pic_address): Likewise.
16606 (legitimize_tls_address): Likewise.
16607 (ix86_expand_call): Likewise.
16608 (x86_output_mi_thunk): Likewise.
16609 (get_dllimport_decl): Add new beimport argument.
16610 (construct_plt_address): Don't assert for x64 pe-coff targets.
16611 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
16613 (SYMBOL_FLAG_STUBVAR): New macro.
16614 (SYMBOL_REF_STUBVAR_P): Likewise.
16615 * config/i386/winnt.c (stub_list): New structure.
16616 (stub_head): New local variable.
16617 (i386_pe_record_stub): New function.
16618 (i386_pe_file_end): Emit refptr-stubs.
16620 2013-04-02 Jakub Jelinek <jakub@redhat.com>
16622 PR rtl-optimization/56745
16623 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
16624 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
16627 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
16628 and both of them are MEM_REFs, just compare first argument for
16629 equality and attempt to deal even with differing offsets.
16632 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
16633 of gimple_clobber_p to be MEM_REF.
16634 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
16635 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
16636 after gimplification.
16637 * asan.c (get_mem_ref_of_assignment): Don't instrument
16638 gimple_clobber_p stmts.
16639 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
16640 gimple_clobber_p stmt if they have MEM_REF lhs and
16641 are dead because of another gimple_clobber_p stmt.
16642 * tree-ssa-live.c (clear_unused_block_pointer): Treat
16643 gimple_clobber_p stmts like debug stmts.
16644 (remove_unused_locals): Remove clobbers with MEM_REF lhs
16645 that refer to unused VAR_DECLs or uninitialized values.
16646 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
16647 gimple_clobber_p stmts if they refer to removed parameters.
16648 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
16651 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
16653 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
16654 using SWI48 mode attribute.
16656 2013-04-02 Wei Mi <wmi@google.com>
16658 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
16659 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
16660 *<rotate_insn><mode>3_mask in i386.md.
16662 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
16664 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
16666 2013-04-02 Richard Biener <rguenther@suse.de>
16668 PR tree-optimization/56778
16669 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
16670 Runtime alias tests are not supported for gather loads.
16671 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
16672 stmts referenced from SSA operands before updating SSA form.
16674 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
16675 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16677 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
16678 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
16679 * config/arm/cortex-a53.md: New file.
16680 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
16681 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
16682 * config/arm/arm.c (arm_issue_rate): Likewise.
16683 * config/arm/arm-tune.md: Regenerate
16684 * config/arm/arm-tables.opt: Regenerate.
16685 * config/arm/arm-cores.def: Add cortex-a53.
16687 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
16689 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
16692 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16694 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
16695 scalar load/store operations using B/H registers.
16696 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
16698 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16700 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
16702 * config/aarch64/aarch64.c
16703 (aarch64_simd_scalar_immediate_valid_for_move): New.
16704 * config/aarch64/aarch64-protos.h
16705 (aarch64_simd_scalar_immediate_valid_for_move): New.
16706 * config/aarch64/constraints.md (Dh, Dq): New.
16707 * config/aarch64/iterators.md (hq): New.
16709 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
16711 * reorg.c (get_branch_condition): Deal with conditional returns.
16712 (fill_simple_delay_slots): Remove dead code dealing with jumps.
16714 2013-04-01 Wei Mi <wmi@google.com>
16716 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
16717 Truncate operand 2 using %b asm operand modifier.
16718 (*<shift_insn><mode>3_mask): Ditto.
16719 (*<rotate_insn><mode>3_mask): Ditto.
16721 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
16723 PR middle-end/56798
16724 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
16726 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
16728 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
16730 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
16732 2013-03-30 Lawrence Crowl <crowl@google.com>
16734 * dse.c (clear_alias_sets): Remove never set.
16735 (disqualified_clear_alias_sets): Remove never set.
16736 (clear_alias_mode_pool): Remove never set.
16737 (dse_step0): Remove condition that is never true.
16738 (canon_address): Remove condition that is never true.
16739 (dse_step7): Remove condition that is never true.
16740 (rest_of_handle_dse): Remove condition that is never true.
16741 (rest_of_handle_dse::did_global): Remove never read from above.
16742 (dse_step2_spill): Remove never called from above.
16743 (dse_step5_spill): Remove never called from above.
16745 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
16747 * doc/md.texi (Standard Names) <casesi>: Update documentation for
16748 JUMP_TABLE_DATA changes.
16749 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
16750 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
16751 (Insns) <jump_table_data>: New entry.
16752 * doc/tm.texi: Regenerate.
16754 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
16756 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
16757 for table jump at the end of a basic block using tablejump_p.
16758 * targhooks.c (default_invalid_within_doloop): Likewise.
16759 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
16760 target hook implementation that is identical to the default hook.
16761 (rs6000_invalid_within_doloop): Remove.
16763 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
16764 unused variable from tablejump_p call.
16766 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
16767 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
16768 (INSN_DELETED_P): Likewise.
16769 (emit_jump_table_data): New prototype.
16770 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
16771 after 4th as unused.
16772 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
16773 * sched-vis.c (print_insn): Likewise.
16774 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
16775 insn for compatibility with back ends that use next_active_insn to
16776 identify jump table data.
16777 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
16778 (remove_insn): Likewise.
16779 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
16781 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
16782 (emit_jump_table_data): New function.
16784 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
16785 basic block, a JUMP_TABLE_DATA never is.
16786 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
16787 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
16788 off from code handling real insns.
16789 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
16790 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
16791 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
16792 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
16793 is not a NONDEBUG_INSN_P.
16794 * ira-costs.c (scan_one_insn): Likewise.
16795 * jump.c (mark_all_labels): Likewise.
16796 (mark_jump_label_1): Likewise.
16797 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16798 * lra.c (get_insn_freq): Expect all insns reaching here to be in
16800 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
16801 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
16802 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
16803 JUMP_TABLE_DATA_P insns.
16804 (calculate_elim_costs_all_insns): Likewise.
16805 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
16806 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
16807 (delete_output_reload): Code style fixups.
16808 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
16809 insn flags on this non-insn.
16810 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
16811 as scheduling barriers, for pre-change compatibility.
16812 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
16813 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
16815 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
16816 redundant JUMP_TABLE_DATA_P test.
16817 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
16818 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16819 (frv_for_each_packet): Likewise.
16820 * config/i386/i386.c (min_insn_size): Likewise.
16821 (ix86_avoid_jump_mispredicts): Likewise.
16822 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16823 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
16824 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16825 (mips16_insn_length): Robustify.
16826 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
16827 (mips16_split_long_branches): Likewise.
16828 * config/pa/pa.c (pa_combine_instructions): Likewise.
16829 * config/rs6000/rs6000.c (get_next_active_insn): Treat
16830 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
16831 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
16832 as contributing to pool range lengths.
16833 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
16834 Remove redundant JUMP_TABLE_DATA_P test.
16835 (sh_loop_align): Likewise.
16836 (split_branches): Likewise.
16837 (sh_insn_length_adjustment): Likewise.
16838 * config/spu/spu.c (get_branch_target): Likewise.
16840 2013-03-29 Jan Hubicka <jh@suse.cz>
16842 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
16843 gcov streaming; stream hot bb threshold to ltrans.
16844 * predict.c (get_hot_bb_threshold): Break out from ....
16845 (maybe_hot_count_p): ... here.
16846 (set_hot_bb_threshold): New function.
16847 * lto-section-in.c (lto_section_name): Add profile.
16848 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
16849 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
16850 and data-streamer.h
16851 (histogram_entry): New structure.
16852 (histogram, histogram_pool): New global vars.
16853 (histogram_hash): New structure.
16854 (histogram_hash::hash): New method.
16855 (histogram_hash::equal): Likewise.
16856 (account_time_size): New function.
16857 (cmp_counts): New function.
16858 (dump_histogram): New function.
16859 (ipa_profile_generate_summary): New function.
16860 (ipa_profile_write_summary): New function.
16861 (ipa_profile_read_summary): New function.
16862 (ipa_profile): Decide on threshold.
16863 (pass_ipa_profile): Add ipa_profile_write_summary and
16864 ipa_profile_read_summary.
16865 * Makefile.in (ipa.o): Update dependencies.
16866 * lto-streamer.h (LTO_section_ipa_profile): New section.
16868 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
16870 * tree.h (VAR_P): New.
16872 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
16875 * doc/invoke.texi ([-fwhole-program]): Fix typo.
16877 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
16879 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
16880 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16881 (control_flow_insn_p): Likewise.
16882 * cfgrtl.c (duplicate_insn_chain): Likewise.
16883 * final.c (get_attr_length_1): Likewise.
16884 (shorten_branches): Likewise.
16885 (final_scan_insn): Likewise.
16886 * function.c (instantiate_virtual_regs): Likewise.
16887 * gcse.c (insert_insn_end_basic_block): Likewise.
16888 * ira-costs.c (scan_one_insn): Likewise.
16889 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16890 * lra.c (check_rtl): Likewise.
16891 * reload1.c (elimination_costs_in_insn): Likewise.
16892 * reorg.c (follow_jumps): Likewise.
16894 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
16895 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16896 (thumb_far_jump_used_p): Likewise.
16897 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
16898 (workaround_speculation): Likewise.
16899 (add_sched_insns_for_speculation): Likewise.
16900 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
16901 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16902 (frv_for_each_packet): Likewise.
16903 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
16904 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
16905 (final_emit_insn_group_barriers): Likewise.
16906 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16907 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16908 (mips16_insn_length): Likewise.
16909 * config/pa/pa.c (pa_reorg): Likewise.
16910 (pa_combine_instructions): Likewise.
16911 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
16912 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
16913 (sh_reorg): Likewise.
16914 (split_branches): Likewise.
16915 * config/spu/spu.c (get_branch_target): Likewise.
16917 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
16920 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
16922 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
16923 Fix declaration name.
16925 2013-03-28 Lawrence Crowl <crowl@google.com>
16927 * graphds.h (struct graph.indicies): Remove unused.
16928 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
16929 (SCOP_ORIGINAL_PDDRS): Remove unused.
16930 * sese.h (extern insert_loop_close_phis): Removed unused.
16931 (extern insert_guard_phis): Removed unused.
16932 (extern ivtype_map_elt_info): Removed unused.
16933 (new_ivtype_map_elt): Removed unused.
16934 * sese.c (ivtype_map_elt_info): Removed unused.
16936 2013-03-28 Lawrence Crowl <crowl@google.com>
16938 * Makefile.in: Add several missing include dependences.
16940 (test-dump.o): New. This object is not added to any executable,
16941 but is present for ad-hoc testing.
16943 (debug (const bitmap_head_def &)): New.
16944 (debug (const bitmap_head_def *)): New.
16946 (extern debug (const bitmap_head_def &)): New.
16947 (extern debug (const bitmap_head_def *)): New.
16949 (debug (edge_def &)): New.
16950 (debug (edge_def *)): New.
16952 (debug (basic_block_def &)): New.
16953 (debug (basic_block_def *)): New.
16955 (dump_node (const_tree, int, FILE *)): Correct source file.
16957 (debug (die_struct &)): New.
16958 (debug (die_struct *)): New.
16960 (extern debug (die_struct &)): New.
16961 (extern debug (die_struct *)): New.
16962 * gimple-pretty-print.c
16963 (debug (gimple_statement_d &)): New.
16964 (debug (gimple_statement_d *)): New.
16965 * gimple-pretty-print.h
16966 (extern debug (gimple_statement_d &)): New.
16967 (extern debug (gimple_statement_d *)): New.
16969 (debug (ira_allocno_copy &)): New.
16970 (debug (ira_allocno_copy *)): New.
16971 (debug (ira_allocno &)): New.
16972 (debug (ira_allocno *)): New.
16974 (extern debug (ira_allocno_copy &)): New.
16975 (extern debug (ira_allocno_copy *)): New.
16976 (extern debug (ira_allocno &)): New.
16977 (extern debug (ira_allocno *)): New.
16979 (debug (live_range &)): New.
16980 (debug (live_range *)): New.
16982 (debug (lra_live_range &)): New.
16983 (debug (lra_live_range *)): New.
16985 (debug (lra_live_range &)): New.
16986 (debug (lra_live_range *)): New.
16988 (debug (omega_pb_d &)): New.
16989 (debug (omega_pb_d *)): New.
16991 (extern debug (omega_pb_d &)): New.
16992 (extern debug (omega_pb_d *)): New.
16994 (debug (const rtx_def &)): New.
16995 (debug (const rtx_def *)): New.
16997 (debug_tree (tree): Move within file.
16998 (debug_raw (const tree_node &)): New.
16999 (debug_raw (const tree_node *)): New.
17000 (dump_tree_via_hooks (const tree_node *, int)): New.
17001 (debug (const tree_node &)): New.
17002 (debug (const tree_node *)): New.
17003 (debug_verbose (const tree_node &)): New.
17004 (debug_verbose (const tree_node *)): New.
17005 (debug_head (const tree_node &)): New.
17006 (debug_head (const tree_node *)): New.
17007 (debug_body (const tree_node &)): New.
17008 (debug_body (const tree_node *)): New.
17009 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17010 (debug (vec<tree, va_gc> &)): New.
17011 (debug (vec<tree, va_gc> *)): New.
17013 (extern debug (const rtx_def &)): New.
17014 (extern debug (const rtx_def *)): New.
17016 (debug_raw (simple_bitmap_def &)): New.
17017 (debug_raw (simple_bitmap_def *)): New.
17018 (debug (simple_bitmap_def &)): New.
17019 (debug (simple_bitmap_def *)): New.
17021 (extern debug (simple_bitmap_def &)): New.
17022 (extern debug (simple_bitmap_def *)): New.
17023 (extern debug_raw (simple_bitmap_def &)): New.
17024 (extern debug_raw (simple_bitmap_def *)): New.
17026 (debug (vinsn_def &)): New.
17027 (debug (vinsn_def *)): New.
17028 (debug_verbose (vinsn_def &)): New.
17029 (debug_verbose (vinsn_def *)): New.
17030 (debug (expr_def &)): New.
17031 (debug (expr_def *)): New.
17032 (debug_verbose (expr_def &)): New.
17033 (debug_verbose (expr_def *)): New.
17034 (debug (vec<rtx> &)): New.
17035 (debug (vec<rtx> *)): New.
17037 (extern debug (vinsn_def &)): New.
17038 (extern debug (vinsn_def *)): New.
17039 (extern debug_verbose (vinsn_def &)): New.
17040 (extern debug_verbose (vinsn_def *)): New.
17041 (extern debug (expr_def &)): New.
17042 (extern debug (expr_def *)): New.
17043 (extern debug_verbose (expr_def &)): New.
17044 (extern debug_verbose (expr_def *)): New.
17045 (extern debug (vec<rtx> &)): New.
17046 (extern debug (vec<rtx> *)): New.
17048 (_list_iter_cond_expr): Make inline instead of static.
17050 (debug (sreal &)): New.
17051 (debug (sreal *)): New.
17053 (extern debug (sreal &)): New.
17054 (extern debug (sreal *)): New.
17056 (extern debug_raw (const tree_node &)): New.
17057 (extern debug_raw (const tree_node *)): New.
17058 (extern debug (const tree_node &)): New.
17059 (extern debug (const tree_node *)): New.
17060 (extern debug_verbose (const tree_node &)): New.
17061 (extern debug_verbose (const tree_node *)): New.
17062 (extern debug_head (const tree_node &)): New.
17063 (extern debug_head (const tree_node *)): New.
17064 (extern debug_body (const tree_node &)): New.
17065 (extern debug_body (const tree_node *)): New.
17066 (extern debug (vec<tree, va_gc> &)): New.
17067 (extern debug (vec<tree, va_gc> *)): New.
17069 (debug (struct loop &)): New.
17070 (debug (struct loop *)): New.
17071 (debug_verbose (struct loop &)): New.
17072 (debug_verbose (struct loop *)): New.
17073 * tree-dump.c: Add header dependence.
17075 (extern debug (struct loop &)): New.
17076 (extern debug (struct loop *)): New.
17077 (extern debug_verbose (struct loop &)): New.
17078 (extern debug_verbose (struct loop *)): New.
17080 (debug (data_reference &)): New.
17081 (debug (data_reference *)): New.
17082 (debug (vec<data_reference_p> &)): New.
17083 (debug (vec<data_reference_p> *)): New.
17084 (debug (vec<ddr_p> &)): New.
17085 (debug (vec<ddr_p> *)): New.
17087 (extern debug (data_reference &)): New.
17088 (extern debug (data_reference *)): New.
17089 (extern debug (vec<data_reference_p> &)): New.
17090 (extern debug (vec<data_reference_p> *)): New.
17091 (extern debug (vec<ddr_p> &)): New.
17092 (extern debug (vec<ddr_p> *)): New.
17094 (debug (pt_solution &)): New.
17095 (debug (pt_solution *)): New.
17097 (extern debug (pt_solution &)): New.
17098 (extern debug (pt_solution *)): New.
17100 (debug (_var_map &)): New.
17101 (debug (_var_map *)): New.
17102 (debug (tree_live_info_d &)): New.
17103 (debug (tree_live_info_d *)): New.
17105 (extern debug (_var_map &)): New.
17106 (extern debug (_var_map *)): New.
17107 (extern debug (tree_live_info_d &)): New.
17108 (extern debug (tree_live_info_d *)): New.
17110 2013-03-28 Jan Hubicka <jh@suse.cz>
17112 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17114 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17116 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17117 record only when desired or required.
17119 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17121 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17122 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17124 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17126 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17127 (*andsi3_compare0_uxtw): New pattern.
17128 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17129 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17131 2013-03-28 Jan Hubicka <jh@suse.cz>
17133 * data-streamer-in.c (streamer_read_gcov_count): New function.
17134 * gimple-streamer-out.c: Include value-prof.h.
17135 (output_gimple_stmt): Output histogram.
17136 (output_bb): Use streamer_write_gcov_count.
17137 * value-prof.c: Include data-streamer.h
17138 (dump_histogram_value): Add HIST_TYPE_MAX.
17139 (stream_out_histogram_value): New function.
17140 (stream_in_histogram_value): New function.
17141 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17142 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17143 * data-streamer-out.c (streamer_write_gcov_count): New function.
17144 (streamer_write_gcov_count_stream): New function.
17145 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17146 (lto_output_node): Likewise.
17147 (input_node, input_edge): Likewise.
17148 * lto-streamer-out.c (output_cfg): Update streaming.
17149 * lto-streamer-in.c (input_cfg): Likewise.
17150 * data-streamer.h (streamer_write_gcov_count,
17151 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17152 * gimple-streamer-in.c: Include value-prof.h
17153 (input_gimple_stmt): Input histograms.
17154 (input_bb): Update profile streaming.
17156 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17158 * genmodes.c (emit_max_int): New function.
17159 (emit_insn_modes_h): Added call to emit_max_function.
17160 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17162 * machmode.def: Fixed comment.
17164 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17166 * combine.c (try_combine): Removed useless assert.
17167 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17169 2013-03-28 Marek Polacek <polacek@redhat.com>
17170 Richard Biener <rguenther@suse.de>
17172 PR tree-optimization/56695
17173 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17174 build signed result of a vector comparison.
17175 * tree-cfg.c (verify_gimple_comparison): Check that a result
17176 of a vector comparison has signed type.
17178 2013-03-28 Richard Biener <rguenther@suse.de>
17180 PR tree-optimization/37021
17181 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17182 do not restrict gaps between groups.
17183 * tree-vect-stmts.c (vectorizable_load): Properly account for
17184 a gap between groups.
17186 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17188 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17189 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17192 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17194 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17195 * doc/extend.texi (Named Address Spaces): Ditto.
17196 (Variable Attributes): Ditto.
17198 2013-03-27 Kai Tietz <ktietz@redhat.com>
17200 * config.build: Add support for cygwin x64 target.
17201 * config.gcc: Likewise.
17202 * config.host: Likewise.
17203 * configure.ac: Likewise
17204 * configure: Regenerated.
17206 2013-03-27 Kai Tietz <ktietz@redhat.com>
17208 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17209 * config/i386/t-cygwin-w64: New file.
17210 * config/i386/cygwin-w64.h: New file.
17211 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17212 and add support for x64-cygwin target.
17213 (CPP_SPEC): Likewise.
17214 (CXX_WRAP_SPEC_LIST): Undefine before define.
17215 (LIBGCJ_SONAME): Use 15 as version.
17217 2013-03-27 Richard Biener <rguenther@suse.de>
17219 PR tree-optimization/56716
17220 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17221 dumping for ref nodes.
17223 2013-03-27 Martin Jambor <mjambor@suse.cz>
17225 PR tree-optimization/55334
17226 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17227 restricted pointers to arrays.
17229 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17231 * Makefile.in (.SUFFIXES): Add .cc.
17232 (.c.o): Apply same recipe for implicit rule .cc.o.
17234 2013-03-27 Richard Biener <rguenther@suse.de>
17236 PR tree-optimization/37021
17237 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17238 REALPART/IMAGPART_EXPRs around the supported refs.
17239 * tree-ssa-structalias.c (find_func_aliases): Assume that
17240 floating-point values are not used to transfer pointers.
17242 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17244 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17245 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17246 * doc/tm.texi: Regenerate.
17247 * targhooks.h (default_has_ifunc_p): New.
17248 * targhooks.c (default_has_ifunc_p): Ditto.
17249 * config/linux-protos.h: New file.
17250 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17251 hook for linux which disables support of indirect functions in android.
17252 * config/linux-android.c: New file.
17253 * config/t-linux-android.c: Ditto.
17254 * config.gcc: Added new object file linux-android.o.
17255 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17256 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17257 * varasm.c (do_assemble_alias): Likewise.
17258 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17259 doesn't support indirect functions.
17260 * configure: Regenerate.
17262 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17265 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17266 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17269 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17271 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17273 2013-03-27 Terry Guo <terry.guo@arm.com>
17275 * config/arm/arm-cores.def: Added core cortex-r7.
17276 * config/arm/arm-tune.md: Regenerated.
17277 * config/arm/arm-tables.opt: Regenerated.
17278 * doc/invoke.texi: Added entry for core cortex-r7.
17280 2013-03-27 Walter Lee <walt@tilera.com>
17282 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17283 double-decrement of next_scratch_regno.
17285 2013-03-27 Walter Lee <walt@tilera.com>
17287 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17289 (insn_v1mulus): Ditto.
17290 (insn_v2muls): Ditto.
17292 2013-03-27 Walter Lee <walt@tilera.com>
17294 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17295 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17297 2013-03-27 Walter Lee <walt@tilera.com>
17299 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17300 (*sibcall_value): Ditto.
17302 2013-03-27 Walter Lee <walt@tilera.com>
17304 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17305 (insn_mnz_v8qi): ... this ...
17306 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17308 (insn_v<n>mnz): Replaced by ...
17309 (insn_v1mnz): ... this ...
17310 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17312 (insn_mz_<mode>): Replaced by ...
17313 (insn_mz_v8qi): ... this ...
17314 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17316 (insn_v<n>mz): Replaced by ...
17317 (insn_v1mz): ... this ...
17318 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17321 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17323 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17325 2013-03-26 Roland McGrath <mcgrathr@google.com>
17327 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17328 than fprintf with a non-constant, non-format string.
17330 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17332 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17333 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17334 operand 0 predicate.
17335 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17336 attribute. Use general_x64nomem_operand as operand 1 predicate.
17337 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17338 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17339 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17340 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17341 (mov<mode>_insv_1): Remove expander. Merge insn with
17342 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17343 Use general_x64nomem_operand as operand 1 predicate.
17344 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17345 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17346 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17347 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17348 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17349 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17350 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17351 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17352 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17353 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17354 (general_x64nomem_operand): Ditto.
17356 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17358 * config/rtems.opt: Add -pthread option.
17360 2013-03-26 Richard Biener <rguenther@suse.de>
17362 * alias.c (find_base_term): Avoid redundant and not used recursion.
17363 (base_alias_check): Get the initial base term from the caller.
17364 (true_dependence_1): Compute and pass base terms to base_alias_check.
17365 (write_dependence_p): Likewise.
17366 (may_alias_p): Likewise.
17368 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17370 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17371 PC-relative load in SI modes and above only.
17373 2013-03-26 Xinyu Qi <xyqi@marvell.com>
17375 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
17376 * config/arm/iwmmxt.md (WCGR0): Update.
17377 (WCGR1, WCGR2, WCGR3): Likewise.
17379 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17381 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
17382 Use x64 and nox64 isa attributes.
17384 2013-03-26 Richard Biener <rguenther@suse.de>
17386 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
17387 alignment computations and rely on get_object_alignment_1
17388 for the !TYPE_P case.
17389 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
17391 2013-03-26 Walter Lee <walt@tilera.com>
17393 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
17394 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
17396 2013-03-25 Jeff Law <law@redhat.com>
17398 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
17399 check for INTEGRAL_TYPE_P that was missing due to checking in
17400 wrong version of prior patch.
17402 2013-03-25 Walter Lee <walt@tilera.com>
17404 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
17405 TILEGX_INSN_SHUFFLEBYTES1.
17406 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
17408 (tilegx_builtins): Ditto.
17409 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
17411 2013-03-25 Walter Lee <walt@tilera.com>
17413 * config/tilegx/tilegx.md (floatsisf2): New pattern.
17414 (floatunssisf2): New pattern.
17415 (floatsidf2): New pattern.
17416 (floatunssidf2): New pattern.
17418 2013-03-25 Walter Lee <walt@tilera.com>
17420 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
17421 tests for constraint J, K, N, P.
17423 2013-03-25 Walter Lee <walt@tilera.com>
17425 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
17426 Use indirect/pcrel encoding.
17427 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
17430 2013-03-25 Steve Ellcey <sellcey@mips.com>
17432 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
17433 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
17434 * config/mips/mips.c (mips_option_override): Set IMADD default.
17435 * config/mips/mips.h (PTF_AVOID_IMADD): New.
17436 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
17437 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
17438 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
17439 * doc/invoke.texi (-mimadd/-mno-imadd): New.
17441 2013-03-25 Jeff Law <law@redhat.com>
17443 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
17444 slightly to avoid creating and folding useless trees. Simplify
17445 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
17447 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
17449 * config/i386/i386.md (*zero_extendsidi2): Merge with
17450 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
17451 * config/i386/predicates.md (x86_64_zext_operand): Rename from
17452 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
17453 targets. Clarify comment.
17455 2013-03-25 Martin Jambor <mjambor@suse.cz>
17457 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
17458 pass-through jump functions differently.
17459 (ipa_read_jump_function): Likewise. Also use setter functions to set
17462 2013-03-25 Martin Jambor <mjambor@suse.cz>
17464 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
17465 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
17467 (ipa_get_indirect_edge_target): New function.
17468 (devirtualization_time_bonus): New parameter known_aggs, pass it to
17469 ipa_get_indirect_edge_target. Update all callers.
17470 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
17471 ipa_get_indirect_edge_target_1 instead of calling
17472 ipa_get_indirect_edge_target.
17473 (create_specialized_node): Pass aggvlas to
17474 ipcp_discover_new_direct_edges.
17476 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17478 * config/arm/arm.md (f_sels, f_seld): New types.
17479 (*cmov<mode>): New pattern.
17480 * config/arm/predicates.md (arm_vsel_comparison_operator): New
17483 2013-03-25 Kai Tietz <ktietz@redhat.com>
17485 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
17486 POSIX-printf for mingw-hosted builds.
17488 2013-03-25 Richard Biener <rguenther@suse.de>
17490 PR middle-end/56694
17491 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
17492 must-not-throw stmt location.
17494 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17496 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
17497 Emit load-acquire versions when acq is true.
17498 (arm_emit_store_exclusive): Add rel parameter.
17499 Emit store-release versions when rel is true.
17500 (arm_split_compare_and_swap): Use acquire-release instructions
17502 of barriers when appropriate.
17503 (arm_split_atomic_op): Likewise.
17504 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
17505 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
17506 (VUNSPEC_SLX): Likewise.
17507 (VUNSPEC_LDA): Likewise.
17508 (VUNSPEC_STL): Likewise.
17509 * config/arm/sync.md (atomic_load<mode>): New pattern.
17510 (atomic_store<mode>): Likewise.
17511 (arm_load_acquire_exclusive<mode>): Likewise.
17512 (arm_load_acquire_exclusivesi): Likewise.
17513 (arm_load_acquire_exclusivedi): Likewise.
17514 (arm_store_release_exclusive<mode>): Likewise.
17516 2013-03-25 Catherine Moore <clm@codesourcery.com>
17518 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
17519 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
17520 * config/mip/predicates.md (lwsp_swsp_operand,
17521 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
17522 sb16_operand, db4_operand, db7_operand, ib3_operand,
17523 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
17524 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
17525 andi16_operand): New predicates.
17526 * config/mips/mips.md (compression): New attribute.
17527 (enabled): New attribute.
17528 (length): Consider compression in computing length.
17529 (shift_compression): New code attribute.
17530 (*add<mode>3): New operands. Record compression.
17531 (sub<mode>3): Likewise.
17532 (one_cmpl<mode>2): Likewise.
17533 (*and<mode>3): Likewise.
17534 (*ior<mode>3): Likewise.
17535 (unnamed pattern for xor): Likewise.
17536 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
17537 (*<optab><mode>3): Likewise.
17538 (*mov<mode>_internal: Likewise.
17539 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
17540 (mips_unsigned_immediate_p): New.
17541 (umips_lwsp_swsp_address_p): New.
17542 (m16_based_address_p): New.
17543 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
17544 (mips_unsigned_immediate_p): New prototype.
17545 (lwsp_swsp_address_p): New prototype.
17546 (m16_based_address_p): New prototype.
17547 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
17548 (mips_signed_immediate_p): New function.
17549 (m16_based_address_p): New function.
17550 (lwsp_swsp_address_p): New function.
17551 (mips_print_operand_punctuation): Recognize short delay slot insns
17552 for microMIPS.add<mode>3"
17554 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17557 * config/arm/iterators.md (v_cmp_result): New mode attribute.
17558 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
17560 2013-03-25 Richard Biener <rguenther@suse.de>
17562 PR tree-optimization/56689
17563 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
17566 2013-03-25 Richard Biener <rguenther@suse.de>
17568 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
17570 (memory_references): Likewise.
17571 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
17572 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
17573 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
17574 (gather_mem_refs_in_loops): Fold into ...
17575 (analyze_memory_references): ... this. Move initialization
17576 to tree_ssa_lim_initialize.
17577 (fill_always_executed_in): Rename to ...
17578 (fill_always_executed_in_1): ... this.
17579 (fill_always_executed_in): Move contains_call computation to
17580 this new function from ...
17581 (tree_ssa_lim_initialize): ... here.
17582 (tree_ssa_lim): Call fill_always_executed_in.
17584 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17586 * postreload.c (reload_combine): Fix code detecting returns.
17588 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17590 * function.c (emit_use_return_register_into_block): On cc0 targets,
17591 do not emit the sequence between cc0 setter and user.
17593 2013-03-25 Kai Tietz <ktietz@redhat.com>
17595 * config/i386/predicates.md (local_symbolic_operand): Interpret
17596 dll-imported symbols as none-local.
17598 2013-03-25 Richard Biener <rguenther@suse.de>
17600 * tree-ssa-loop-im.c (struct depend): Remove.
17601 (struct lim_aux_data): Make depends a vec of gimples.
17602 (free_lim_aux_data): Adjust.
17603 (add_dependency): Likewise.
17604 (set_level): Likewise.
17606 2013-03-25 Richard Biener <rguenther@suse.de>
17608 PR middle-end/56434
17609 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
17610 the pointer returned by calls with ECF_MALLOC set.
17612 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17614 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
17616 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17618 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
17619 using MMXMODE mode iterator.
17620 (*move<mode>_internal): Merge with *movv2sf_internal and
17621 *movv2sf_internal_rex64 using MMXMODE mode iterator.
17623 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17625 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
17626 (record_last_mem_set_info): Likewise.
17628 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
17629 of XNEWVEC followed by memset.
17630 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
17632 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17634 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
17635 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
17636 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
17637 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
17638 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
17639 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
17640 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
17641 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
17642 BARRIER_P instead of GET_CODE.
17644 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17646 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
17647 inaccuracy in the probing code.
17649 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
17650 (ctrapdi4): Likewise.
17652 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17654 * calls.c (expand_call): Add missing guard to code handling return
17655 of non-BLKmode structures in MSB.
17656 * function.c (expand_function_end): Likewise.
17658 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17660 * combine.c (try_combine): Adjust comment. Do not add the set of
17661 insn #0 if the destination indirectly is set or dies in insn #2.
17662 Tidy up code to distribute a new note.
17664 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17666 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
17667 also for alternatives 16 and 17.
17669 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17671 * config/i386/sse.md (*mov<mode>_internal): Merge with
17672 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
17673 Emit insn template depending on type attribute. Use
17674 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17675 movd instead of movq mnemonic for interunit moves. Rewrite mode
17676 attribute calculation. Remove unit attribute calculation.
17677 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
17678 Set prefix_data16 attribute for DImode ssemov types.
17679 Use Ym instead of y for SSE-MMX conversion alternatives.
17680 Reorder operand constraints.
17682 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
17684 * df.h (df_insn_delete): Adjust prototype.
17685 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
17686 and let it decide whether mark the basic block dirty.
17687 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
17688 * df-scan.c (df_insn_info_delete): New helper function, split
17689 off from df_insn_delete.
17690 (df_scan_free_bb_info): Use it.
17691 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
17693 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
17694 that the insn is actually an insn and it has a non-NULL basic block.
17695 Do not mark basic block dirty if only deleting a DEBUG_INSN.
17697 2013-03-22 Richard Biener <rguenther@suse.de>
17699 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
17701 (mem_ref_alloc): Do not allocate them.
17702 (refs_independent_p): Do not query or maintain a cache.
17704 2013-03-22 Richard Biener <rguenther@suse.de>
17706 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
17707 (gather_mem_refs_in_loops): Do not compute it.
17708 (analyze_memory_references): Do not allocate it.
17709 (tree_ssa_lim_finalize): Do not free it.
17710 (for_all_locs_in_loop): Do not query all_refs_in_loop.
17712 2013-03-22 Richard Biener <rguenther@suse.de>
17714 * is-a.h (as_a): Use gcc_checking_assert.
17716 2013-03-22 Ian Bolton <ian.bolton@arm.com>
17718 * config/aarch64/aarch64.c (aarch64_print_operand): New
17719 format specifier for printing a constant in hex.
17720 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
17721 format specifier for printing second operand.
17723 2013-03-22 Richard Biener <rguenther@suse.de>
17725 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
17727 (gather_mem_refs_in_loops): Perform store accumulation here.
17728 (create_vop_ref_mapping_loop): Remove.
17729 (create_vop_ref_mapping): Likewise.
17730 (analyze_memory_references): Initialize refs_stored_in_loop.
17731 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
17732 (record_indep_loop): Remove.
17733 (record_dep_loop): New function.
17734 (ref_indep_loop_p_1): Adjust to only walk over references
17735 in the loop, not its subloops.
17736 (ref_indep_loop_p): Rename to ...
17737 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
17738 maintaining a more fine-grained cache.
17739 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
17740 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
17742 2013-03-22 Richard Biener <rguenther@suse.de>
17744 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
17745 (struct mem_ref): Make accesses_in_loop a vec of a vec of
17746 aggregate mem_ref_loc.
17747 (free_mem_ref_locs): Inline into ...
17748 (memref_free): ... this and adjust.
17749 (mem_ref_alloc): Adjust.
17750 (mem_ref_locs_alloc): Remove.
17751 (record_mem_ref_loc): Adjust.
17752 (get_all_locs_in_loop): Rewrite into ...
17753 (for_all_locs_in_loop): ... this iterator.
17754 (rewrite_mem_ref_loc): New functor.
17755 (rewrite_mem_refs): Use for_all_locs_in_loop.
17756 (sm_set_flag_if_changed): New functor.
17757 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
17758 (ref_always_accessed): New functor.
17759 (ref_always_accessed_p): Use for_all_locs_in_loop.
17761 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17763 * tree-pass.h (PROP_gimple_lvec): New.
17764 * passes.c (dump_properties): Handle PROP_gimple_lvec.
17765 (init_optimization_passes): Move pass_lower_vector.
17766 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
17768 (pass_lower_vector): Provide PROP_gimple_lvec.
17769 (pass_lower_vector_ssa): Likewise.
17770 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
17772 2013-03-21 Mark Wielaard <mjw@redhat.com>
17774 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
17776 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17778 * config/i386/i386.md (*movdi_internal): Disparage slightly
17779 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
17780 conversion alternatives.
17782 2013-03-21 Jakub Jelinek <jakub@redhat.com>
17784 PR middle-end/48087
17785 * diagnostic.def (DK_WERROR): New kind.
17786 * diagnostic.h (werrorcount): Define.
17787 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
17788 promoted to DK_ERROR, increment DK_WERROR counter instead of
17790 * toplev.c (toplev_main): Call print_ignored_options even if
17791 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
17792 even if just werrorcount is non-zero.
17795 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
17797 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
17798 (string_cst_pool_decl): New function.
17799 (optimize_one_addr_into_implicit_ptr): New function.
17800 (resolve_addr_in_expr): Optimize DWARF location expression
17801 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
17802 which doesn't live in memory, but has DW_AT_location or
17803 DW_AT_const_value, or refers to a string literal, into
17804 DW_OP_GNU_implicit_pointer.
17805 (optimize_location_into_implicit_ptr): New function.
17806 (resolve_addr): If removing DW_AT_location of a variable because
17807 it was DW_OP_addr of address of the variable, but the variable doesn't
17808 live in memory, try to emit const value attribute for the initializer.
17810 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17812 * tree.h (VECTOR_TYPE_P): New macro.
17813 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
17814 TYPE_MODE): Use it.
17815 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
17816 VEC_COND_EXPR cannot be lvalues.
17817 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
17819 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17821 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
17822 Restrict the transformation to equal modes.
17824 2013-03-21 Richard Biener <rguenther@suse.de>
17826 PR tree-optimization/39326
17827 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
17828 (MEM_ANALYZABLE): Adjust.
17829 (record_mem_ref_loc): Move bitmap ops ...
17830 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
17831 unanalyzable refs, do not record locations for it.
17832 (analyze_memory_references): Allocate ref zero as shared
17834 (refs_independent_p): Do not test for unanalyzed mems here.
17835 (ref_indep_loop_p_1): Special-case disambiguation against
17836 the unanalyzed ref.
17837 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
17839 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
17841 * config/arm/arm-protos.h (tune_params): Add
17842 prefer_neon_for_64bits field.
17843 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
17844 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
17845 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
17846 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
17847 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
17848 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
17849 (arm_option_override): Handle -mneon-for-64bits new option.
17850 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
17851 (prefer_neon_for_64bits): Declare new variable.
17852 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
17853 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
17854 (arch_enabled): Handle new arch types. Remove support for onlya8
17856 (one_cmpldi2): Use new arch names.
17857 (zero_extend<mode>di2, extend<mode>di2): Ditto.
17858 * config/arm/arm.opt (mneon-for-64bits): Add option.
17859 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
17860 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
17861 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
17863 * doc/invoke.texi (-mneon-for-64bits): Document.
17865 2013-03-21 Richard Biener <rguenther@suse.de>
17867 PR tree-optimization/39326
17868 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
17869 (sort_bbs_in_loop_postorder_cmp): New function.
17870 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
17872 2013-03-21 Richard Biener <rguenther@suse.de>
17874 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
17875 (vect_insert_into_interleaving_chain): Likewise.
17876 (vect_drs_dependent_in_basic_block): Inline ...
17877 (vect_slp_analyze_data_ref_dependence): ... here. New function,
17879 (vect_analyze_data_ref_dependence): ... here. Simplify.
17880 (vect_check_interleaving): Simplify.
17881 (vect_analyze_data_ref_dependences): Likewise. Split out ...
17882 (vect_slp_analyze_data_ref_dependences): ... this new function.
17883 (dr_group_sort_cmp): New function.
17884 (vect_analyze_data_ref_accesses): Compute data-reference groups
17885 here instead of in vect_analyze_data_ref_dependence. Use
17886 a more efficient algorithm.
17887 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
17888 vect_slp_analyze_data_ref_dependences. Call
17889 vect_analyze_data_ref_accesses earlier.
17890 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
17891 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
17892 (vect_slp_analyze_data_ref_dependences): New prototype.
17894 2013-03-21 Richard Biener <rguenther@suse.de>
17896 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
17897 ref is stored in the loop.
17898 (find_refs_for_sm): Walk only over all stores.
17899 (store_motion_loop): Allocate from lim_bitmap_obstack.
17900 (store_motion): Likewise.
17902 2013-03-21 Richard Biener <rguenther@suse.de>
17904 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
17905 Update virtual SSA form.
17907 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17909 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
17910 * configure: Regenerate.
17911 * config.in: Regenerate.
17912 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
17913 if !HAVE_LD_EH_FRAME_CIEV3.
17915 2013-03-21 Richard Biener <rguenther@suse.de>
17917 * tree-cfg.c (verify_expr_no_block): New function.
17918 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
17919 nor DECL_VALUE_EXPR have locations with associated blocks.
17920 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
17921 (clear_unused_block_pointer): Remove code dealing with
17922 blocks in DECL_DEBUG_EXPR locations.
17924 2013-03-21 Richard Biener <rguenther@suse.de>
17926 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
17927 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
17928 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
17929 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
17930 instead of DECL_DEBUG_EXPR_IS_FROM.
17931 * gimplify.c (gimplify_modify_expr): Likewise.
17932 * tree-cfg.c (verify_expr_location_1): Likewise.
17933 * tree-complex.c (create_one_component_var): Likewise.
17934 * tree-sra.c (create_access_replacement): Likewise.
17935 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
17936 (clear_unused_block_pointer): Likewise.
17937 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
17938 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
17939 * var-tracking.c (var_debug_decl): Likewise.
17940 (track_expr_p): Likewise.
17941 * tree-inline.c (add_local_variables): Likewise. Set
17942 DECL_HAS_DEBUG_EXPR_P after copying it.
17943 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
17944 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
17946 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17949 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
17950 * configure: Regenerate.
17951 * config.in: Regenerate.
17952 * config/i386/i386.md (*movdf_internal): Use
17953 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17954 movd instead of movq mnemonic for interunit moves.
17955 (*movdi_internal): Ditto.
17957 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17959 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
17960 (abd<mode>_3): New pattern.
17961 (aba<mode>_3): New pattern.
17962 (fabd<mode>_3): New pattern.
17964 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17966 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
17967 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
17968 occurrence of REGISTER_PREFIX as its empty string.
17970 2013-03-20 Jeff Law <law@redhat.com>
17972 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
17973 addititional equivalences for equality comparisons between an SSA_NAME
17974 and a constant where the SSA_NAME was set from a widening conversion.
17976 2013-03-20 Walter Lee <walt@tilera.com>
17978 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
17980 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17982 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
17983 depending on type attribute.
17984 (*movti_internal): Ditto.
17985 (*movtf_internal): Ditto.
17986 (*movxf_internal): Ditto.
17987 (*movdf_internal): Ditto.
17988 (*movsf_internal): Ditto.
17990 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17992 * config/i386/i386.md (*movti_internal): Set prefix attribute to
17993 maybe_vex for sselog1 and ssemov types.
17994 (*movdi_internal): Reorder operand constraints.
17995 (*movsi_internal): Ditto. Set prefix attribute to
17996 maybe_vex for sselog1 and ssemov types.
17997 (*movtf_internal): Set prefix attribute to maybe_vex
17998 for sselog1 and ssemov types.
17999 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18000 DImode ssemov types. Reorder operand constraints.
18001 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18002 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18003 attribute for SImode ssemov types. Reorder operand constraints.
18005 2013-03-20 Martin Jambor <mjambor@suse.cz>
18007 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18008 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18010 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18012 * config/rs6000/predicates.md (indexed_address, update_address_mem
18013 update_indexed_address_mem): New predicates.
18014 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18015 attribute for load/store instructions.
18016 * config/rs6000/dfp.md (movsd_store): Likewise.
18017 (movsd_load): Likewise.
18018 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18019 (unnamed HI->DI extend define_insn): Likewise.
18020 (unnamed SI->DI extend define_insn): Likewise.
18021 (unnamed QI->SI extend define_insn): Likewise.
18022 (unnamed QI->HI extend define_insn): Likewise.
18023 (unnamed HI->SI extend define_insn): Likewise.
18024 (unnamed HI->SI extend define_insn): Likewise.
18025 (extendsfdf2_fpr): Likewise.
18026 (movsi_internal1): Likewise.
18027 (movsi_internal1_single): Likewise.
18028 (movhi_internal): Likewise.
18029 (movqi_internal): Likewise.
18030 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18031 attribute for load/store instructions.
18032 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18034 (mov<mode>_softfloat): Likewise.
18035 (mov<mode>_hardfloat32): Likewise.
18036 (mov<mode>_hardfloat64): Likewise.
18037 (mov<mode>_softfloat64): Likewise.
18038 (movdi_internal32): Likewise.
18039 (movdi_internal64): Likewise.
18040 (probe_stack_<mode>): Likewise.
18042 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18044 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18045 floating point, and decimal floating point to reload iterator.
18047 * config/rs6000/constraints.md (wl constraint): New constraints to
18048 return FLOAT_REGS if certain options are used to reduce the number
18049 of separate patterns that exist in the file.
18050 (wx constraint): Likewise.
18051 (wz constraint): Likewise.
18053 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18054 -mdebug=reg, print wg, wl, wx, and wz constraints.
18055 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18056 Initialize the reload functions for 64-bit binary/decimal floating
18058 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18059 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18060 create the buffer on the stack to overcome not having a 32-bit
18062 (rs6000_emit_move): Likewise.
18063 (rs6000_secondary_memory_needed_rtx): Likewise.
18064 (rs6000_alloc_sdmode_stack_slot): Likewise.
18065 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18066 via xxlxor, just like DFmode 0.0.
18068 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18069 define as 1 if we are running on a power7 or newer.
18070 (enum r6000_reg_class_enum): Add new constraints.
18072 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18073 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18074 with other moves by using conditional constraits (wg). Use LFIWZX
18075 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18076 (movsd splitter): Likewise.
18077 (movsd_hardfloat): Likewise.
18078 (movsd_softfloat): Likewise.
18080 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18081 binary and decimal floating point moves.
18082 (fmove_ok): New attributes to combine binary and decimal floating
18083 point moves, and to combine power6x (mfpgpr) moves along normal
18085 (real_value_to_target): Likewise.
18086 (f32_lr): Likewise.
18087 (f32_lm): Likewise.
18088 (f32_li): Likewise.
18089 (f32_sr): Likewise.
18090 (f32_sm): Likewise.
18091 (f32_si): Likewise.
18092 (movsf): Combine binary and decimal floating point moves. Combine
18093 power6x (mfpgpr) moves with other moves by using conditional
18094 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18095 (mov<mode> for SFmode/SDmode); Likewise.
18096 (SFmode/SDmode splitters): Likewise.
18097 (movsf_hardfloat): Likewise.
18098 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18099 (movsf_softfloat): Likewise.
18100 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18102 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18103 wx and wz constraints.
18105 * config/rs6000/constraints.md (wg constraint): New constraint to
18106 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18108 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18111 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18112 -mdebug=reg, print wg, wl, wx, and wz constraints.
18113 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18114 Initialize the reload functions for 64-bit binary/decimal floating
18116 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18117 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18118 create the buffer on the stack to overcome not having a 32-bit
18120 (rs6000_emit_move): Likewise.
18121 (rs6000_secondary_memory_needed_rtx): Likewise.
18122 (rs6000_alloc_sdmode_stack_slot): Likewise.
18123 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18124 via xxlxor, just like DFmode 0.0.
18126 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18127 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18128 with other moves by using conditional constraits (wg). Use LFIWZX
18129 and STFIWX for loading SDmode on power7.
18130 (movdd splitters): Likewise.
18131 (movdd_hardfloat32): Likewise.
18132 (movdd_softfloat32): Likewise.
18133 (movdd_hardfloat64_mfpgpr): Likewise.
18134 (movdd_hardfloat64): Likewise.
18135 (movdd_softfloat64): Likewise.
18137 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18138 64-bit binary and decimal floating point moves.
18139 (FMOVE64X): Likewise.
18140 (movdf): Combine 64-bit binary and decimal floating point moves.
18141 Combine power6x (mfpgpr) moves with other moves by using
18142 conditional constraits (wg).
18143 (mov<mode> for DFmode/DDmode): Likewise.
18144 (DFmode/DDmode splitters): Likewise.
18145 (movdf_hardfloat32): Likewise.
18146 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18147 (movdf_softfloat32): Likewise.
18148 (movdf_hardfloat64_mfpgpr): Likewise.
18149 (movdf_hardfloat64): Likewise.
18150 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18151 (movdf_softfloat64): Likewise.
18152 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18153 (reload_<mode>_load): Move to later in the file so they aren't in
18154 the middle of the floating point move insns.
18155 (reload_<mode>_store): Likewise.
18157 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18160 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18161 constraint if -mdebug=reg.
18162 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18163 Enable using dd reload support if needed.
18165 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18166 binary and decimal floating point moves in rs6000.md.
18167 (movtd_internal): Likewise.
18169 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18170 decimal floating point moves.
18172 (movtf_internal): Likewise.
18173 (mov<mode>_internal, TDmode/TFmode): Likewise.
18174 (movtf_softfloat): Likewise.
18175 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18177 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18178 movdi_internal64, using wg constraint for move direct operations.
18179 (movdi_internal64): Likewise.
18181 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18182 MODES_TIEABLE_P for selected modes. Print the numerical value of
18183 the various virtual registers. Use GPR/FPR first/last values,
18184 instead of hard coding the register numbers. Print which modes
18185 have reload functions registered.
18186 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18187 settings before/after setting cpu, target and subtarget settings.
18188 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18189 and for secondary reload failures in rs6000_secondary_reload_inner.
18190 (rs6000_secondary_reload_fail): Likewise.
18191 (rs6000_secondary_reload_inner): Likewise.
18193 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18194 macros for first/last GPR and FPR registers.
18195 (LAST_GPR_REGNO): Likewise.
18196 (FIRST_FPR_REGNO): Likewise.
18197 (LAST_FPR_REGNO): Likewise.
18199 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18200 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18201 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18202 (vcond<mode><mode>): Likewise.
18203 (vcondu<mode><mode>): Likewise.
18204 (vector_gtu<mode>): Likewise.
18205 (vector_gte<mode>): Likewise.
18206 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18207 to prevent the compiler from converting DImode operations to TImode.
18208 (ior<mode>3): Likewise.
18209 (and<mode>3): Likewise.
18210 (one_cmpl<mode>2): Likewise.
18211 (nor<mode>3): Likewise.
18212 (andc<mode>3): Likewise.
18214 * config/rs6000/constraints.md (wt constraint): New constraint
18215 that returns VSX_REGS if TImode is allowed in VSX registers.
18217 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18218 constant under VSX.
18220 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18221 similar to TImode, but it is restricted to being in the GPRs.
18223 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18224 TImode to occupy a single VSX register.
18226 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18227 -mvsx-timode for power7/power8.
18228 (power7 cpu): Likewise.
18229 (power8 cpu): Likewise.
18231 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18232 sure that TFmode/TDmode take up two registers if they are ever
18233 allowed in the upper VSX registers.
18234 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18236 (rs6000_init_hard_regno_mode_ok): Likewise.
18237 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18238 constraint. Print if LRA is turned on.
18239 (rs6000_option_override_internal): Give an error if -mvsx-timode
18240 and VSX is not enabled.
18241 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18242 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18243 to reg+offset addressing. Use PTImode when checking offset
18244 addresses for validity.
18245 (reg_offset_addressing_ok_p): Likewise.
18246 (rs6000_legitimate_offset_address_p): Likewise.
18247 (rs6000_legitimize_address): Likewise.
18248 (rs6000_legitimize_reload_address): Likewise.
18249 (rs6000_legitimate_address_p): Likewise.
18250 (rs6000_eliminate_indexed_memrefs): Likewise.
18251 (rs6000_emit_move): Likewise.
18252 (rs6000_secondary_reload): Likewise.
18253 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18254 reloads to fpr registers to continue to use reg+offset addressing,
18255 but 64-bit reloads to altivec registers need reg+reg addressing.
18256 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18257 it. Treat LO_SUM like a PLUS operation.
18258 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18259 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18260 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18261 registers to share a register with a smaller sized type, since VSX
18262 puts scalars in the upper 64-bits.
18263 (print_operand): Add support for PTImode.
18264 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18265 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18266 registers, but don't have arithmetic support.
18267 (rs6000_memory_move_cost): Add test for VSX.
18268 (rs6000_opt_masks): Add -mvsx-timode.
18270 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18273 (VSr): Use wt constraint for TImode.
18274 (VSv): Drop TImode support.
18275 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18276 (vsx_movti_64bit): Likewise.
18277 (vsx_movti_32bit): Likewise.
18278 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18279 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18280 one '?' on the appropriate output constraint. Do not allow TImode
18281 logical operations on 32-bit systems.
18282 (vsx_ior<mode>3): Likewise.
18283 (vsx_xor<mode>3): Likewise.
18284 (vsx_one_cmpl<mode>2): Likewise.
18285 (vsx_nor<mode>3): Likewise.
18286 (vsx_andc<mode>3): Likewise.
18287 (vsx_concat_<mode>): Likewise.
18288 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18290 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18291 OPTION_MASK_VSX_TIMODE.
18292 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18293 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18295 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18296 (TI2 iterator): New iterator for TImode, PTImode.
18297 (wd mode attribute): Add values for vector types.
18298 (movti_string): Replace TI move operations with operations for TImode
18299 and PTImode. Add support for TImode being allowed in VSX registers.
18300 (mov<mode>_string, TImode/PTImode): Likewise.
18301 (movti_ppc64): Likewise.
18302 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18303 (TI mode splitters): Likewise.
18305 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18308 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18310 PR tree-optimization/56355
18311 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18312 Also handle integers with undefined overflow.
18314 2013-03-20 Catherine Moore <clm@codesourcery.com>
18315 Maciej W. Rozycki <macro@codesourcery.com>
18316 Tom de Vries <tom@codesourcery.com>
18317 Nathan Sidwell <nathan@codesourcery.com>
18318 Iain Sandoe <iain@codesourcery.com>
18319 Nathan Froyd <froydnj@codesourcery.com>
18320 Chao-ying Fu <fu@mips.com>
18322 * doc/extend.texi: (micromips, nomicromips, nocompression):
18323 Document new function attributes.
18324 * doc/invoke.texi (minterlink-compressed, mmicromips,
18325 m14k, m14ke, m14kec): Document new options.
18326 (minterlink-mips16): Update documentation.
18327 * doc/md.texi (ZC, ZD): Document new constraints.
18328 * configure.ac (gcc_cv_as_micromips): Check if linker
18329 supports the .set micromips directive.
18330 * configure: Regenerate.
18331 * config.in: Regenerate.
18332 * config/mips/mips-tables.opt: Regenerate.
18333 * config/mips/micromips.md: New file.
18334 * constraints.md (ZC, ZD): New constraints.
18335 * config/mips/predicates.md (movep_src_register): New predicate.
18336 (movep_src_operand): New predicate.
18337 (non_volatile_mem_operand): New predicate.
18338 * config/mips/mips.md (multimem): New type.
18339 (length): Differentiate between 17-bit and 18-bit branch offsets.
18340 (MOVEP1, MOVEP2): New mode iterator.
18341 (mov_<load>l): Use ZC constraint.
18342 (mov_<load>r): Likewise.
18343 (mov_<store>l): Likewise.
18344 (mov_<store>r): Likewise.
18345 (*branch_equality<mode>_inverted): Add microMIPS support.
18346 (*branch_equality<mode>): Likewise.
18347 (*jump_absolute): Likewise.
18348 (indirect_jump_<mode>): Likewise.
18349 (tablejump_<mode>): Likewise.
18350 (<optab>_internal): Likewise.
18351 (sibcall_internal): Likewise.
18352 (sibcall_value_internal): Likewise.
18353 (prefetch): Use constraint ZD.
18354 * config/mips/mips.opt (minterlink-compressed): New option.
18355 (minterlink-mips16): Now an alias for minterlink-compressed.
18356 (mmicromips): New option.
18357 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18358 (compare_and_swap_12): Likewise.
18359 (sync_add<mode>): Likewise.
18360 (sync_<optab>_12): Likewise.
18361 (sync_old_<optab>_12): Likewise.
18362 (sync_new_<optab>_12): Likewise.
18363 (sync_nand_12): Likewise.
18364 (sync_old_nand_12): Likewise.
18365 (sync_new_nand_12): Likewise.
18366 (sync_sub<mode>): Likewise.
18367 (sync_old_add<mode>): Likewise.
18368 (sync_old_sub<mode>): Likewise.
18369 (sync_new_add<mode>): Likewise.
18370 (sync_new_sub<mode>): Likewise.
18371 (sync_<optab><mode>): Likewise.
18372 (sync_old_<optab><mode>): Likewise.
18373 (sync_new_<optab><mode>): Likewise.
18374 (sync_nand<mode>): Likewise.
18375 (sync_old_nand<mode>): Likewise.
18376 (sync_new_nand<mode>): Likewise.
18377 (sync_lock_test_and_set<mode>): Likewise.
18378 (test_and_set_12): Likewise.
18379 (atomic_compare_and_swap<mode>): Likewise.
18380 (atomic_exchange<mode>_llsc): Likewise.
18381 (atomic_fetch_add<mode>_llsc): Likewise.
18382 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18383 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18384 (umips_save_restore_pattern_p): Likewise.
18385 (umips_load_store_pair_p): Likewise.
18386 (umips_output_load_store_pair): Likewise.
18387 (umips_movep_target_p): Likewise.
18388 (umips_12bit_offset_address_p): Likewise.
18389 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18390 (mips_base_mips16): Rename this...
18391 (mips_base_compression_flags): ...to this. Update all uses.
18392 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18393 (mips_mips16_decl_p): Delete.
18394 (mips_nomips16_decl_p): Delete.
18395 (mips_get_compress_on_flags): New function.
18396 (mips_get_compress_off_flags): New function.
18397 (mips_get_compress_mode): New function.
18398 (mips_get_compress_on_name): New function.
18399 (mips_get_compress_off_name): New function.
18400 (mips_insert_attributes): Support multiple compression types.
18401 (mips_merge_decl_attributes): Likewise.
18402 (umips_12bit_offset_address_p): New function.
18403 (mips_start_function_definition): Emit .set micromips directive.
18404 (mips_call_may_need_jalx_p): New function.
18405 (mips_function_ok_for_sibcall): Add microMIPS support.
18406 (mips_print_operand_punctuation): Support short delay slots and
18408 (umips_swm_mask, umips_swm_encoding): New.
18409 (umips_build_save_restore): New function.
18410 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18411 (was_mips16_p): Remove.
18412 (old_compression_mode): New.
18413 (mips_set_compression_mode): New function.
18414 (mips_set_current_function): Add microMIPS support.
18415 (mips_option_override): Likewise.
18416 (umips_save_restore_pattern_p): New function.
18417 (umips_output_save_restore): New function.
18418 (umips_load_store_pair_p_1): New function.
18419 (umips_load_store_pair_p): New function.
18420 (umips_output_load_store_pair_1): New function.
18421 (umips_output_load_store_pair): New function.
18422 (umips_movep_target_p) New function.
18423 (mips_prepare_pch_save): Add microMIPS support.
18424 * config/mips/mips.h (TARGET_COMPRESSION): New.
18425 (TARGET_CPU_CPP_BUILTINS): Update macro
18426 to use new compression flags and to support microMIPS.
18427 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18428 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18429 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18430 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18431 (ASM_SPEC): Support mmicromips and mno-micromips.
18432 (M16STORE_REG_P): New macro.
18433 (MIPS_CALL): Support TARGET_MICROMIPS.
18434 (MICROMIPS_J): New macro.
18435 (mips_base_mips16): Rename this...
18436 (mips_base_compression_flags): ...to this.
18437 (UMIPS_12BIT_OFFSET_P): New macro.
18438 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18439 (MULTILIB_DIRNAMES): Likewise.
18440 2013-03-20 Richard Biener <rguenther@suse.de>
18442 PR tree-optimization/56661
18443 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
18444 the result does not have to be distinct.
18446 2013-03-20 Richard Biener <rguenther@suse.de>
18448 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
18451 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18452 Steven Bosscher <steven@gcc.gnu.org>
18454 PR rtl-optimization/56605
18455 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
18457 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18460 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
18461 that require movd instead of movq.
18463 2013-03-20 Richard Biener <rguenther@suse.de>
18465 * tree-ssa-structalias.c (struct variable_info): Add pointer
18466 to the first field of an aggregate with sub-vars. Make
18467 this and the pointer to the next subfield its ID.
18468 (vi_next): New function.
18469 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
18470 storedanything_id, integer_id): Increment by one.
18471 (new_var_info, get_call_vi, lookup_call_clobber_vi,
18472 get_call_clobber_vi): Adjust.
18473 (solution_set_expand): Simplify and speedup.
18474 (solution_set_add): Inline into ...
18475 (set_union_with_increment): ... this. Adjust accordingly.
18476 (do_sd_constraint): Likewise.
18477 (do_ds_constraint): Likewise.
18478 (do_complex_constraint): Simplify.
18479 (build_pred_graph): Adjust.
18480 (solve_graph): Likewise. Simplify and speedup.
18481 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
18482 get_constraint_for_component_ref, get_constraint_for_1,
18483 first_vi_for_offset, first_or_preceding_vi_for_offset,
18484 create_function_info_for, create_variable_info_for_1,
18485 create_variable_info_for, intra_create_variable_infos): Adjust.
18486 (init_base_vars): Push NULL for ID zero.
18487 (compute_points_to_sets): Adjust.
18489 2013-03-20 Richard Biener <rguenther@suse.de>
18491 * cfgloop.c (verify_loop_structure): Streamline and avoid
18492 ICEing on corrupt loop tree.
18493 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
18496 2013-03-20 Richard Biener <rguenther@suse.de>
18498 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
18499 check whether an SSA update is needed.
18501 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
18503 * config/mips/constraints.md (T): Rename to...
18507 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
18508 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
18510 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18512 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
18513 (*subsi3_carryin_uxtw): Likewise.
18515 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18517 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
18518 (*rorsi3_insn_uxtw): Likewise.
18520 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18522 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
18523 (*extrsi5_insn_uxtw): Likewise.
18525 2013-03-19 Richard Biener <rguenther@suse.de>
18527 PR tree-optimization/56273
18528 * passes.c (init_optimization_passes): Move second VRP after DOM.
18530 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
18532 * config/i386/i386.md (*movti_internal): Merge from
18533 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
18534 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
18535 nox64 isa attributes.
18537 2013-03-18 Richard Biener <rguenther@suse.de>
18539 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
18541 (merge_node_constraints): Likewise.
18542 (build_succ_graph): Likewise.
18543 (valid_graph_edge): Inline into single caller.
18544 (unify_nodes): Likewise. Use bitmap_set_bit return value
18546 (scc_visit): Fix formatting and variable use.
18547 (do_sd_constraint): Use gcc_checking_assert.
18548 (do_ds_constraint): Likewise.
18549 (do_complex_constraint): Likewise.
18550 (condense_visit): Likewise. Cleanup.
18551 (dump_pred_graph): New function.
18552 (perform_var_substitution): Dump the pred-graph before
18553 variable substitution.
18554 (find_equivalent_node): Use gcc_checking_assert.
18555 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
18557 2013-03-18 Richard Biener <rguenther@suse.de>
18559 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
18560 Remove cond_expr_stmt_list argument and do not gimplify the
18562 (vect_loop_versioning): Adjust.
18563 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
18564 Cleanup to use less temporaries.
18565 (vect_create_data_ref_ptr): Cleanup.
18567 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18569 PR tree-optimization/56635
18570 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
18571 require types_compatible_p types.
18573 2013-03-18 Nick Clifton <nickc@redhat.com>
18575 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
18576 spurious backslash.
18578 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
18579 Add missing line to comment describing function.
18581 2013-03-18 Richard Biener <rguenther@suse.de>
18583 PR tree-optimization/56210
18584 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18585 Handle string / character search functions.
18586 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
18588 2013-03-18 Richard Biener <rguenther@suse.de>
18590 PR middle-end/56483
18591 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
18592 and implement properly.
18593 * gimple.h (gimple_cond_single_var_p): Remove.
18595 2013-03-18 Richard Biener <rguenther@suse.de>
18597 * tree-data-ref.h (find_data_references_in_loop): Declare.
18598 * tree-data-ref.c (get_references_in_stmt): Use a stack
18599 vector pre-allocated in the callers.
18600 (find_data_references_in_stmt): Adjust.
18601 (graphite_find_data_references_in_stmt): Likewise.
18602 (create_rdg_vertices): Likewise.
18603 (find_data_references_in_loop): Export.
18604 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
18605 Compute dependences here...
18606 (vect_analyze_data_refs): ...not here. When we encounter
18607 a non-vectorizable data reference in basic-block vectorization
18608 truncate the data reference vector. Do not bother to
18609 fixup data-dependence information for gather loads.
18610 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
18611 of data references, as reported.
18613 2013-03-18 Richard Biener <rguenther@suse.de>
18615 PR tree-optimization/3713
18616 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
18617 has_constants and expr.
18618 (stmt_has_constants): Properly valueize SSA names when deciding
18619 whether the stmt has constants.
18621 2013-03-18 Richard Biener <rguenther@suse.de>
18623 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
18624 whole function when there is nothing to do.
18625 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
18626 * tree-vectorizer.c (vectorize_loops): Update virtual and
18627 loop-closed SSA once.
18628 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
18630 2013-03-18 Richard Biener <rguenther@suse.de>
18632 PR middle-end/56113
18633 * domwalk.c (bb_postorder): New global static.
18634 (cmp_bb_postorder): New function.
18635 (walk_dominator_tree): Replace scheme imposing an order for
18636 visiting dominator sons by one sorting them at the time they
18637 are pushed on the stack.
18639 2013-03-18 Richard Biener <rguenther@suse.de>
18641 PR tree-optimization/39326
18642 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
18643 (struct mem_ref): Replace mem member with ao_ref typed member.
18644 (MEM_ANALYZABLE): Adjust.
18645 (memref_eq): Likewise.
18646 (mem_ref_alloc): Likewise.
18647 (gather_mem_refs_stmt): Likewise.
18648 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
18649 (execute_sm_if_changed_flag_set): Adjust.
18650 (execute_sm): Likewise.
18651 (ref_always_accessed_p): Likewise.
18652 (refs_independent_p): Likewise.
18653 (can_sm_ref_p): Likewise.
18655 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18658 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
18659 return 1 even for !unsignedp.
18661 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18663 * config/i386/i386.md (isa): Add x64 and nox64.
18664 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
18665 (*pushtf): Enable *roF alternative for x64 isa only.
18666 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
18667 mode attribute of integer alternatives to DImode for TARGET_64BIT.
18668 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
18669 (*movtf_internal): Merge from *movtf_internal_rex64 and
18670 *movtf_internal_sse. Use x64 and nox64 isa attributes.
18671 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
18672 nox64 isa attributes.
18673 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
18674 nox64 isa attributes.
18675 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
18677 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18679 * config/alpha/alpha.c (TARGET_LRA_P): New define.
18681 2013-03-17 Jakub Jelinek <jakub@redhat.com>
18684 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
18685 class names. Remove trailing comma after "ALL_REGS".
18687 2013-03-16 Jan Hubicka <jh@suse.cz>
18689 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
18690 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
18691 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
18692 of cgraph_get_create_node.
18693 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
18695 2013-03-16 Jason Merrill <jason@redhat.com>
18698 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
18699 with DW_AT_default_value.
18701 2013-03-16 Jakub Jelinek <jakub@redhat.com>
18703 * BASE-VER: Set to 4.9.0.
18705 2013-03-14 Andi Kleen <ak@linux.intel.com>
18708 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
18709 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
18710 Document _x* TSX intrinsics.
18712 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
18713 David Holsgrove <david.holsgrove@xilinx.com>
18715 * configure.ac: Add MicroBlaze TLS support detection.
18716 * configure: Regenerate.
18717 * config/microblaze/microblaze-protos.h
18718 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
18719 symbol_mentioned_p, label_mentioned_p): Add prototypes.
18720 * config/microblaze/microblaze.c (microblaze_address_type): Add
18721 ADDRESS_TLS and tls_reloc address types.
18722 (microblaze_address_info): Add tls_reloc.
18723 (TARGET_HAVE_TLS): Define.
18724 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
18725 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
18726 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
18727 load_tls_operand, microblaze_call_tls_get_addr,
18728 microblaze_legitimize_tls_address): New functions.
18729 (microblaze_classify_unspec): Handle UNSPEC_TLS.
18730 (get_base_reg): Use microblaze_tls_symbol_p.
18731 (microblaze_classify_address): Handle TLS.
18732 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
18733 label_mentioned_p and microblaze_tls_referenced_p.
18734 (microblaze_legitimize_address): Handle TLS.
18735 (microblaze_address_insns): Handle ADDRESS_TLS.
18736 (pic_address_needs_scratch): Handle TLS.
18737 (print_operand_address): Handle TLS.
18738 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
18739 (microblaze_expand_move): Handle TLS.
18740 (microblaze_legitimate_constant_p): Check
18741 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
18742 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
18743 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
18744 (PIC_OFFSET_TABLE_REGNUM): Set.
18745 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
18746 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
18747 (addsi3, movsi_internal2, movdf_internal): Update constraints
18748 * config/microblaze/predicates.md (arith_plus_operand): Define
18749 (move_operand): Redefine as move_src_operand,
18750 check microblaze_tls_referenced_p.
18752 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18754 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
18755 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
18757 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18759 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
18762 2013-03-14 Jakub Jelinek <jakub@redhat.com>
18764 PR tree-optimization/53265
18765 * common.opt (Waggressive-loop-optimizations): New option.
18766 * tree-ssa-loop-niter.c: Include tree-pass.h.
18767 (do_warn_aggressive_loop_optimizations): New function.
18768 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
18769 if number_of_latch_executions returned constant.
18770 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
18771 early. If number_of_latch_executions returned constant, set
18772 nb_iterations_upper_bound back to it.
18773 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
18775 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
18776 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
18778 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
18779 (MULTILIB_OSDIRNAMES): Set.
18780 * genmultilib: If defaultosdirname doesn't start with :: , set
18781 defaultosdirname2 instead, clear it and emit two . multilib_raw
18782 entries instead of just one.
18784 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
18786 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
18787 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
18788 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
18789 (SUBTARGET_OVERRIDE_OPTIONS): New.
18791 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
18794 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
18795 (musermode): Convert to Var(TARGET_USERMODE).
18796 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
18797 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
18798 * config/sh/sh.c (sh_option_override): Use
18799 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
18800 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
18802 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
18804 (udivsi3_i4_single, divsi3_i4_single): Use
18805 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
18807 2013-03-13 Dave Korn <dave.korn.cygwin@....>
18809 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
18812 2013-03-13 Richard Biener <rguenther@suse.de>
18814 PR tree-optimization/56608
18815 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
18816 calls when vectorizing basic-blocks.
18818 2013-03-13 Jakub Jelinek <jakub@redhat.com>
18821 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
18824 2013-03-12 Jakub Jelinek <jakub@redhat.com>
18826 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
18828 2013-03-11 Jan Hubicka <jh@suse.cz>
18831 * lto-streamer-out.c (output_symbol_p): Skip references from
18832 constructors of external variables.
18834 2013-03-11 Jan Hubicka <jh@suse.cz>
18836 PR middle-end/56571
18837 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
18839 * emit-rtl.c (verify_rtx_sharing): Likewise.
18840 (copy_insn_1): Likewise.
18841 * rtl.c (copy_rtx): Likewise.
18843 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
18846 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
18847 output_operand_lossage message.
18849 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
18852 * arm.c (shift_op): Validate RTL pattern on the fly.
18853 (arm_print_operand, case 'S'): Don't use shift_operator to validate
18856 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18859 * config/pa/pa.md (call_value): Check for calls to powf and direct to
18860 new call patterns that clobber %fr12.
18861 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
18862 split and postreload patterns.
18863 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
18864 registers %fr12 and %fr12R as call used.
18866 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
18868 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
18869 (canon_address, record_store, replace_read, check_mem_read_rtx,
18870 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
18871 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
18872 rest_of_handle_dse): Likewise.
18874 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
18876 PR middle-end/56524
18877 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
18879 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
18880 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
18881 (save_optabs_if_changed): Replace with...
18882 (init_tree_optimization_optabs): ...this.
18883 * optabs.c (save_optabs_if_changed): Rename to...
18884 (init_tree_optimization_optabs): ...this. Take the optimization node
18885 as argument. Do nothing if the base optabs are already correct.
18886 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
18887 to recompute optabs.
18888 * function.h (function): Remove optabs field.
18889 * function.c (invoke_set_current_function_hook): Call
18890 init_tree_optimization_optabs. Use the result to initialize
18893 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
18895 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
18896 if GTMA_HAS_NO_INSTRUMENTATION.
18897 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
18898 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
18899 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
18900 * gimple-pretty-print.c (dump_gimple_transaction): Handle
18901 GTMA_HAS_NO_INSTRUMENTATION.
18903 2013-03-08 Jakub Jelinek <jakub@redhat.com>
18905 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
18908 2013-03-08 Marek Polacek <polacek@redhat.com>
18909 Jakub Jelinek <jakub@redhat.com>
18911 PR tree-optimization/56478
18912 * predict.c (is_comparison_with_loop_invariant_p): Change the
18913 type of loop_step to tree.
18914 (predict_loops): Adjust.
18915 (predict_iv_comparison): Perform the computations on double_ints.
18917 2013-03-08 Richard Biener <rguenther@suse.de>
18919 PR tree-optimization/56570
18920 * tree-cfg.c (verify_expr_location_1): Verify locations for
18922 * tree-sra.c (create_access_replacement): Strip locations
18923 from DECL_DEBUG_EXPRs.
18925 2013-03-08 Richard Biener <rguenther@suse.de>
18927 * tree-inline.c (expand_call_inline): Do not associate
18928 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
18929 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
18931 2013-03-08 Richard Biener <rguenther@suse.de>
18933 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
18934 or block changes with -Og. Fix for location / block encoding
18935 changes and PHI arguments with locations.
18937 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
18939 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
18941 (struct output_info): Likewise.
18942 (register_overhead): Remove bad gcc_assert.
18943 (bitmap_find_bit): If there is only a single bitmap element, do not
18944 count a miss as a search.
18945 (print_statistics): Update for counter type changes.
18946 (dump_bitmap_statistics): Likewise. Print headers such that they
18947 are properly lined up with the printed counters.
18949 2013-03-07 Jakub Jelinek <jakub@redhat.com>
18951 PR tree-optimization/56559
18952 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
18953 check that it has only a single use.
18955 2013-03-07 Richard Biener <rguenther@suse.de>
18957 * doc/invoke.texi (fwhole-program): Discourage use in combination
18960 2013-03-06 Jakub Jelinek <jakub@redhat.com>
18962 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
18964 PR tree-optimization/56539
18965 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
18966 instead of GSI_CONTINUE_LINKING as last argument to
18967 force_gimple_operand_gsi. Adjust function comment.
18969 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
18972 PR middle-end/56548
18973 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
18974 promoted mode, convert the result back to the original mode.
18976 2013-03-06 Richard Biener <rguenther@suse.de>
18978 PR middle-end/56294
18979 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
18980 (insert_updated_phi_nodes_compare_uids): New function.
18981 (update_ssa): Sort symbols_to_rename after UID before
18982 traversing it to insert PHI nodes.
18984 2013-03-06 Richard Biener <rguenther@suse.de>
18986 PR middle-end/50494
18987 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
18988 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
18991 2013-02-13 Richard Biener <rguenther@suse.de>
18994 * varasm.c (output_constant_def_1): Get the decl representing
18995 the constant as argument.
18996 (output_constant_def): Wrap output_constant_def_1.
18997 (make_decl_rtl): Use output_constant_def_1 with the decl
18998 representing the constant.
18999 (build_constant_desc): Optionally re-use a decl already
19000 representing the constant.
19001 (tree_output_constant_def): Adjust.
19003 2013-03-06 Joey Ye <joey.ye@arm.com>
19006 * gcc.c (convert_white_space): New function.
19007 (main): Handles white space in function name.
19009 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19012 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19013 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19014 to SH_DIV_CALL_TABLE for TARGET_SH2.
19015 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19017 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19018 call-table options.
19020 2013-03-05 Sterling Augustine <saugustine@google.com>
19021 Cary Coutant <ccoutant@google.com>
19024 * dwarf2out.c (resolve_addr): Don't call
19025 remove_loc_list_addr_table_entries a second time for the same
19028 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19031 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19032 (avoid_complex_debug_insns): New function.
19033 (expand_debug_locations): Call it.
19035 PR rtl-optimization/56484
19036 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19037 lifetimes of hard registers on small register class machines.
19039 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19041 * config/microblaze/microblaze-protos.h: Rename
19042 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19043 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19045 (microblaze_fast_interrupt_function_p): New function.
19046 (microblaze_is_interrupt_handler): Rename to
19047 microblaze_is_interrupt_variant and add fast_interrupt check.
19048 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19049 (save_restore_insns): Likewise.
19050 (compute_frame_size): Likewise.
19051 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19052 (microblaze_globalize_label): Likewise.
19053 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19054 * config/microblaze/microblaze.md: Use wrapper
19055 microblaze_is_interrupt_variant.
19057 2013-03-05 Kai Tietz <ktietz@redhat.com>
19059 * sdbout.c (sdbout_one_type): Switch to current function's section
19060 supporting cold/hot.
19062 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19064 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19067 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19069 PR middle-end/56461
19070 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19071 if VALGRIND_GET_VBITS is defined, temporarily make object
19072 memory all defined, and restore previous valgrind addressability
19073 and definability afterwards. Free this_object at the end.
19075 PR middle-end/56461
19076 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19077 right before calling lra_create_live_ranges, also call it
19078 when clearing live_p. Only call lra_clear_live_ranges
19079 at the end if live_p.
19081 PR middle-end/56461
19082 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19084 2013-03-05 Richard Biener <rguenther@suse.de>
19086 PR tree-optimization/56521
19087 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19090 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19093 * except.h (remove_unreachable_eh_regions): New prototype.
19094 * except.c (remove_eh_handler_splicer): New function, split out
19095 of remove_eh_handler.
19096 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19097 warning about running it on many EH regions one at a time.
19098 (remove_unreachable_eh_regions_worker): New function, walk the
19099 EH tree in depth-first order and remove non-marked regions.
19100 (remove_unreachable_eh_regions): New function.
19101 * tree-eh.c (mark_reachable_handlers): New function, split out
19102 from remove_unreachable_handlers.
19103 (remove_unreachable_handlers): Use mark_reachable_handlers and
19104 remove_unreachable_eh_regions.
19105 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19106 and remove_unreachable_eh_regions.
19108 2013-03-05 Richard Biener <rguenther@suse.de>
19110 PR middle-end/56525
19111 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19112 not freeing them until the end.
19114 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19116 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19118 2013-03-05 Richard Biener <rguenther@suse.de>
19120 PR tree-optimization/56270
19121 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19122 of loads after scheduling an SLP instance.
19124 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19126 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19128 (check_gcc_parallelize): Run guality.exp as a separate job from
19129 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19130 struct-layout-1.exp with stackalign.exp.
19132 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19134 PR middle-end/56461
19135 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19136 load_index sbitmap even if some bit in it isn't set.
19138 PR middle-end/56461
19139 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19140 (discover_iteration_bound_by_body_walk): Change queues to
19141 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19142 spelling in comment. Call safe_push on queues[bound_index] directly.
19143 Release queues[queue_index] in every iteration unconditionally.
19144 Release bounds vector.
19146 PR middle-end/56461
19147 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19148 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19149 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19152 2013-03-05 Richard Biener <rguenther@suse.de>
19155 * tree-inline.c (remap_blocks_to_null): New function.
19156 (expand_call_inline): When expanding a call stmt without
19157 an associated block inline remap all callee blocks to NULL.
19159 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19161 PR rtl-optimization/56494
19162 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19163 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19164 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19166 PR middle-end/56461
19167 * sel-sched-ir.c (free_sched_pools): Release
19168 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19169 if succs_info_pool.max_top isn't -1.
19172 * opts.c (opts_obstack, opts_concat): Moved to...
19173 * opts-common.c (opts_obstack, opts_concat): ... here.
19175 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19177 PR middle-end/56461
19178 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19180 2013-03-04 Martin Jambor <mjambor@suse.cz>
19182 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19183 all appropriate places.
19185 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19187 PR tree-optimization/56424
19188 * ipa-split.c (split_function): Do not set the RSO flag if result is
19189 not by reference and its type is a register type.
19191 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19193 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19194 (microblaze_legitimate_pic_operand): Likewise
19195 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19196 new function microblaze_legitimate_pic_operand
19197 * config/microblaze/microblaze-protos.h
19198 (microblaze_legitimate_pic_operand): Declare.
19200 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19202 * config/microblaze/predicates.md (call_insn_simple_operand):
19203 New predicate for supported rtx code types.
19204 * config/microblaze/microblaze.md (call_internal1): Use
19205 call_insn_simple_operand predicate.
19207 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19209 PR middle-end/56461
19210 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19211 partitions.ordered_remove.
19213 PR middle-end/56461
19214 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19215 vec_oprnds0.create (1) for modifier == NONE.
19217 PR middle-end/56461
19218 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19219 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19220 vec_oprnds1 right before pushing anything to it for
19223 PR middle-end/56461
19224 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19225 set nbbs to 0 instead of having separate code path.
19226 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19227 instead of false as last argument if returning NULL.
19229 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19231 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19232 the attribute is now called "target" instead of "option".
19233 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19234 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19235 attribute/pragma name for TARGET_OPTION_VALID_P and
19236 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19237 * doc/tm.texi: Regenerated.
19239 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19241 * config/microblaze/microblaze.c:
19242 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19243 * config/microblaze/microblaze.h: Add -mxl-reorder to
19245 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19246 instructions emitted if TARGET_REORDER.
19247 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19248 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19251 2013-03-01 Xinliang David Li <davidxl@google.com>
19253 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19256 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19258 PR middle-end/56461
19259 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19260 vector even when returning true. Fix up function comment formatting.
19262 PR middle-end/56461
19263 * ira-build.c (ira_loop_nodes_count): New variable.
19264 (create_loop_tree_nodes): Initialize it.
19265 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19267 PR middle-end/56461
19268 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19269 method on dr_chain and result_chain.
19270 * tree-vect-stmts.c (vectorizable_store): Only call
19271 result_chain.create if j == 0.
19273 PR middle-end/56461
19274 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19275 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19276 before overwriting it.
19278 2013-03-01 Tobias Burnus <burnus@net-b.de>
19280 * doc/extended.texi (C Extensions): Change order in @menu
19282 (Other MIPS Built-in Functions): Move last MIPS entry before
19283 "picoChip Built-in Functions".
19284 (SH Built-in Functions): Move after RX Built-in Functions.
19285 * doc/gcc.texi (Introduction): Change order in @menu
19287 * doc/md.texi (Constraints): Ditto.
19288 * gty.texi (Type Information): Ditto.
19289 (User-provided marking routines for template types): Make
19291 * doc/invoke.texi (AArch64 Options): Move before
19292 "Adapteva Epiphany Options".
19294 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19295 Jakub Jelinek <jakub@redhat.com>
19298 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19299 no_address_safety_analysis attribute.
19300 * doc/extend.texi (no_address_safety_attribute): Rename to
19301 no_sanitize_address attribute, mention no_address_safety_analysis
19302 attribute as deprecated alias.
19304 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19306 PR middle-end/56461
19307 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19308 type to vec<vec<tree> > *.
19309 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19310 to be vec<tree> instead of vec<tree> *, set vec_defs
19311 to vNULL and call vec_defs.create (number_of_vects), adjust other
19313 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19314 vectorizable_condition): Adjust vect_get_slp_defs callers.
19316 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19318 * config/aarch64/aarch64.c
19319 (aarch64_float_const_representable): Remove unused variable.
19321 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19323 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19325 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19327 * config/aarch64/aarch64-builtins.c
19328 (aarch64_init_simd_builtins): Make static.
19330 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19332 * config/aarch64/aarch64.c
19333 (aarch64_simd_make_constant): Make static.
19335 2013-02-28 Martin Jambor <mjambor@suse.cz>
19337 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19338 with no initialization to the RHS of debug statements.
19340 2013-02-28 Martin Jambor <mjambor@suse.cz>
19342 PR tree-optimization/56294
19343 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19345 (get_access_replacement): Do not call create_access_replacement.
19346 Assert a replacement exists.
19347 (get_repl_default_def_ssa_name): Create the replacement declaration
19350 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19352 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19353 final_end_function.
19355 2013-02-28 Marek Polacek <polacek@redhat.com>
19357 PR rtl-optimization/56466
19358 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19359 if we're changing a loop.
19360 (peel_loops_completely): Likewise.
19362 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19365 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19367 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19370 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19371 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19372 INTX_FTYPE_FX, FX_FTYPE_INTX.
19373 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
19375 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19377 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
19378 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
19379 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19380 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
19381 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
19382 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
19383 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
19384 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
19385 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
19386 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
19387 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
19388 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
19389 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
19390 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
19391 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
19392 (avrxmega6): Increase max flash segments from 5 to 6.
19393 * config/avr/t-multilib: Regenerate.
19394 * config/avr/avr-tables.opt: Regenerate.
19395 * doc/avr-mmcu.texi: Regenerate.
19397 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19399 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
19400 (avr_device_to_arch): Rename to avr_device_to_ld.
19401 (avr_device_to_as): New prototype.
19402 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
19403 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
19404 * config/avr/driver-avr.c (avr_device_to_as): New.
19405 (avr_device_to_arch): Rename to avr_device_to_ld.
19407 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19409 PR middle-end/56461
19410 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
19411 method on dr_chain and result_chain.
19413 PR middle-end/56461
19414 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
19415 pointer_set_destroy on not_executed_last_iteration.
19417 PR middle-end/56461
19418 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
19420 PR middle-end/56461
19421 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
19422 FOR_EACH_DEFINED_FUNCTION when freeing state.
19424 PR middle-end/56461
19425 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
19427 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
19430 PR middle-end/56461
19431 * ipa-cp.c (decide_whether_version_node): Call vec_free on
19432 known_aggs[i].items and release known_aggs vector.
19434 PR middle-end/56461
19435 * ipa-reference.c (propagate): Free node_info even for alias nodes.
19437 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19439 * config/microblaze/microblaze.c (microblaze_emit_compare):
19440 Use xor for EQ/NE comparisions.
19441 * config/microblaze/microblaze.md (cstoresf4): Add constraints
19442 (cbranchsf4): Adjust operator to comparison_operator.
19444 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19446 PR middle-end/56461
19447 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
19449 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
19450 vec_safe_push, always update *slot.
19451 (redirect_edge_var_map_clear): Use vec_free.
19452 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
19453 (free_var_map_entry): Use vec_free.
19454 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
19455 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
19457 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
19459 PR middle-end/45472
19460 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
19461 when the may_trap_p bit of the exprs being merged differs.
19462 Reorder tests for speculativeness in the logical and operator.
19464 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19466 * incpath.c (add_standard_paths): Use reconcat instead of concat
19467 where appropriate and avoid leaking memory.
19469 * opts.h: Include obstack.h.
19470 (opts_concat): New prototype.
19471 (opts_obstack): New declaration.
19472 * opts.c (opts_concat): New function.
19473 (opts_obstack): New variable.
19474 (init_options_struct): Call gcc_init_obstack on opts_obstack.
19475 (finish_options): Use opts_concat instead of concat
19476 and XOBNEWVEC instead of XNEWVEC.
19477 * opts-common.c (generate_canonical_option, decode_cmdline_option,
19478 generate_option): Likewise.
19479 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
19480 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
19483 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
19484 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
19486 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19488 PR middle-end/56461
19489 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
19491 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19493 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
19494 (arm_block_move_unaligned_straight): Likewise.
19495 (arm_adjust_block_mem): Likewise.
19497 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19500 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
19501 temp, cond and label.
19502 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
19505 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
19506 * config/c6x/c6x.h (dbx_register_map): Update declaration.
19509 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
19510 of prologue/epilogue functions.
19513 * config/tilegx/tilegx.c (tilegx_expand_prologue):
19514 Remove unused variable cfa_offset.
19515 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
19518 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
19519 type promotion to unsigned.
19522 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
19523 for HOST_WIDE_INT of 32 bit / same size as int.
19524 (arm_block_move_unaligned_straight): Likewise.
19525 (arm_adjust_block_mem): Likewise.
19528 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
19531 2013-02-26 Marek Polacek <polacek@redhat.com>
19533 PR tree-optimization/56426
19534 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
19536 2013-02-26 Richard Biener <rguenther@suse.de>
19539 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
19540 unused variable loops.
19542 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19544 PR tree-optimization/56448
19545 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
19546 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
19547 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
19548 later operands of the references, or even first operand for
19549 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
19551 PR tree-optimization/56443
19552 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
19553 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
19554 to type_for_mode langhook.
19556 2013-02-25 Matt Turner <mattst88@gmail.com>
19558 * doc/invoke.texi: Document r4700.
19560 2013-02-25 Richard Biener <rguenther@suse.de>
19562 PR tree-optimization/56175
19563 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
19565 (simplify_bitwise_binary): ... here. Also guard the conversion
19566 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
19568 2013-02-25 Catherine Moore <clm@codesourcery.com>
19571 2013-02-24 Catherine Moore <clm@codesourcery.com>
19572 Maciej W. Rozycki <macro@codesourcery.com>
19573 Tom de Vries <tom@codesourcery.com>
19574 Nathan Sidwell <nathan@codesourcery.com>
19575 Iain Sandoe <iain@codesourcery.com>
19576 Nathan Froyd <froydnj@codesourcery.com>
19577 Chao-ying Fu <fu@mips.com>
19579 * doc/extend.texi: (micromips, nomicromips, nocompression):
19580 Document new function attributes.
19581 * doc/invoke.texi (minterlink-compressed, mmicromips,
19582 m14k, m14ke, m14kec): Document new options.
19583 (minterlink-mips16): Update documentation.
19584 * doc/md.texi (ZC, ZD): Document new constraints.
19585 * configure.ac (gcc_cv_as_micromips): Check if linker
19586 supports the .set micromips directive.
19587 * configure: Regenerate.
19588 * config.in: Regenerate.
19589 * config/mips/mips-tables.opt: Regenerate.
19590 * config/mips/micromips.md: New file.
19591 * constraints.md (ZC, AD): New constraints.
19592 * config/mips/predicates.md (movep_src_register): New predicate.
19593 (movep_src_operand): New predicate.
19594 (non_volatile_mem_operand): New predicate.
19595 * config/mips/mips.md (multimem): New type.
19596 (length): Differentiate between 17-bit and 18-bit branch offsets.
19597 (MOVEP1, MOVEP2): New mode iterator.
19598 (mov_<load>l): Use ZC constraint.
19599 (mov_<load>r): Likewise.
19600 (mov_<store>l): Likewise.
19601 (mov_<store>r): Likewise.
19602 (*branch_equality<mode>_inverted): Add microMIPS support.
19603 (*branch_equality<mode>): Likewise.
19604 (*jump_absolute): Likewise.
19605 (indirect_jump_<mode>): Likewise.
19606 (tablejump_<mode>): Likewise.
19607 (<optab>_internal): Likewise.
19608 (sibcall_internal): Likewise.
19609 (sibcall_value_internal): Likewise.
19610 (prefetch): Use constraint ZD.
19611 * config/mips/mips.opt (minterlink-compressed): New option.
19612 (minterlink-mips16): Now an alias for minterlink-compressed.
19613 (mmicromips): New option.
19614 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19615 (compare_and_swap_12): Likewise.
19616 (sync_add<mode>): Likewise.
19617 (sync_<optab>_12): Likewise.
19618 (sync_old_<optab>_12): Likewise.
19619 (sync_new_<optab>_12): Likewise.
19620 (sync_nand_12): Likewise.
19621 (sync_old_nand_12): Likewise.
19622 (sync_new_nand_12): Likewise.
19623 (sync_sub<mode>): Likewise.
19624 (sync_old_add<mode>): Likewise.
19625 (sync_old_sub<mode>): Likewise.
19626 (sync_new_add<mode>): Likewise.
19627 (sync_new_sub<mode>): Likewise.
19628 (sync_<optab><mode>): Likewise.
19629 (sync_old_<optab><mode>): Likewise.
19630 (sync_new_<optab><mode>): Likewise.
19631 (sync_nand<mode>): Likewise.
19632 (sync_old_nand<mode>): Likewise.
19633 (sync_new_nand<mode>): Likewise.
19634 (sync_lock_test_and_set<mode>): Likewise.
19635 (test_and_set_12): Likewise.
19636 (atomic_compare_and_swap<mode>): Likewise.
19637 (atomic_exchange<mode>_llsc): Likewise.
19638 (atomic_fetch_add<mode>_llsc): Likewise.
19639 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19640 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19641 (umips_save_restore_pattern_p): Likewise.
19642 (umips_load_store_pair_p): Likewise.
19643 (umips_output_load_store_pair): Likewise.
19644 (umips_movep_target_p): Likewise.
19645 (umips_12bit_offset_address_p): Likewise.
19646 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19647 (mips_base_mips16): Rename this...
19648 (mips_base_compression_flags): ...to this. Update all uses.
19649 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19650 (mips_mips16_decl_p): Delete.
19651 (mips_nomips16_decl_p): Delete.
19652 (mips_get_compress_on_flags): New function.
19653 (mips_get_compress_off_flags): New function.
19654 (mips_get_compress_mode): New function.
19655 (mips_get_compress_on_name): New function.
19656 (mips_get_compress_off_name): New function.
19657 (mips_insert_attributes): Support multiple compression types.
19658 (mips_merge_decl_attributes): Likewise.
19659 (umips_12bit_offset_address_p): New function.
19660 (mips_start_function_definition): Emit .set micromips directive.
19661 (mips_call_may_need_jalx_p): New function.
19662 (mips_function_ok_for_sibcall): Add microMIPS support.
19663 (mips_print_operand_punctuation): Support short delay slots and
19665 (umips_swm_mask, umips_swm_encoding): New.
19666 (umips_build_save_restore): New function.
19667 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19668 (was_mips16_p): Remove.
19669 (old_compression_mode): New.
19670 (mips_set_compression_mode): New function.
19671 (mips_set_current_function): Add microMIPS support.
19672 (mips_option_override): Likewise.
19673 (umips_save_restore_pattern_p): New function.
19674 (umips_output_save_restore): New function.
19675 (umips_load_store_pair_p_1): New function.
19676 (umips_load_store_pair_p): New function.
19677 (umips_output_load_store_pair_1): New function.
19678 (umips_output_load_store_pair): New function.
19679 (umips_movep_target_p) New function.
19680 (mips_prepare_pch_save): Add microMIPS support.
19681 * config/mips/mips.h (TARGET_COMPRESSION): New.
19682 (TARGET_CPU_CPP_BUILTINS): Update macro
19683 to use new compression flags and to support microMIPS.
19684 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19685 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19686 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19687 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19688 (ASM_SPEC): Support mmicromips and mno-micromips.
19689 (M16STORE_REG_P): New macro.
19690 (MIPS_CALL): Support TARGET_MICROMIPS.
19691 (MICROMIPS_J): New macro.
19692 (mips_base_mips16): Rename this...
19693 (mips_base_compression_flags): ...to this.
19694 (UMIPS_12BIT_OFFSET_P): New macro.
19695 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19696 (MULTILIB_DIRNAMES): Likewise.
19698 2013-02-25 Tom de Vries <tom@codesourcery.com>
19700 PR rtl-optimization/56131
19701 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
19702 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
19703 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
19705 2013-02-25 Tobias Burnus <burnus@net-b.de>
19707 * doc/invoke.texi (-fsanitize=): Move from optimization
19708 to debugging options.
19710 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19712 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
19714 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19715 Alexander Monakov <amonakov@ispras.ru>
19717 PR middle-end/56077
19718 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
19719 flush pending lists also on non-jumps. Adjust comment.
19721 2013-02-24 Catherine Moore <clm@codesourcery.com>
19722 Maciej W. Rozycki <macro@codesourcery.com>
19723 Tom de Vries <tom@codesourcery.com>
19724 Nathan Sidwell <nathan@codesourcery.com>
19725 Iain Sandoe <iain@codesourcery.com>
19726 Nathan Froyd <froydnj@codesourcery.com>
19727 Chao-ying Fu <fu@mips.com>
19729 * doc/extend.texi: (micromips, nomicromips, nocompression):
19730 Document new function attributes.
19731 * doc/invoke.texi (minterlink-compressed, mmicromips,
19732 m14k, m14ke, m14kec): Document new options.
19733 (minterlink-mips16): Update documentation.
19734 * doc/md.texi (ZC, ZD): Document new constraints.
19735 * configure.ac (gcc_cv_as_micromips): Check if linker
19736 supports the .set micromips directive.
19737 * configure: Regenerate.
19738 * config.in: Regenerate.
19739 * config/mips/mips-tables.opt: Regenerate.
19740 * config/mips/micromips.md: New file.
19741 * constraints.md (ZC, AD): New constraints.
19742 * config/mips/predicates.md (movep_src_register): New predicate.
19743 (movep_src_operand): New predicate.
19744 (non_volatile_mem_operand): New predicate.
19745 * config/mips/mips.md (multimem): New type.
19746 (length): Differentiate between 17-bit and 18-bit branch offsets.
19747 (MOVEP1, MOVEP2): New mode iterator.
19748 (mov_<load>l): Use ZC constraint.
19749 (mov_<load>r): Likewise.
19750 (mov_<store>l): Likewise.
19751 (mov_<store>r): Likewise.
19752 (*branch_equality<mode>_inverted): Add microMIPS support.
19753 (*branch_equality<mode>): Likewise.
19754 (*jump_absolute): Likewise.
19755 (indirect_jump_<mode>): Likewise.
19756 (tablejump_<mode>): Likewise.
19757 (<optab>_internal): Likewise.
19758 (sibcall_internal): Likewise.
19759 (sibcall_value_internal): Likewise.
19760 (prefetch): Use constraint ZD.
19761 * config/mips/mips.opt (minterlink-compressed): New option.
19762 (minterlink-mips16): Now an alias for minterlink-compressed.
19763 (mmicromips): New option.
19764 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19765 (compare_and_swap_12): Likewise.
19766 (sync_add<mode>): Likewise.
19767 (sync_<optab>_12): Likewise.
19768 (sync_old_<optab>_12): Likewise.
19769 (sync_new_<optab>_12): Likewise.
19770 (sync_nand_12): Likewise.
19771 (sync_old_nand_12): Likewise.
19772 (sync_new_nand_12): Likewise.
19773 (sync_sub<mode>): Likewise.
19774 (sync_old_add<mode>): Likewise.
19775 (sync_old_sub<mode>): Likewise.
19776 (sync_new_add<mode>): Likewise.
19777 (sync_new_sub<mode>): Likewise.
19778 (sync_<optab><mode>): Likewise.
19779 (sync_old_<optab><mode>): Likewise.
19780 (sync_new_<optab><mode>): Likewise.
19781 (sync_nand<mode>): Likewise.
19782 (sync_old_nand<mode>): Likewise.
19783 (sync_new_nand<mode>): Likewise.
19784 (sync_lock_test_and_set<mode>): Likewise.
19785 (test_and_set_12): Likewise.
19786 (atomic_compare_and_swap<mode>): Likewise.
19787 (atomic_exchange<mode>_llsc): Likewise.
19788 (atomic_fetch_add<mode>_llsc): Likewise.
19789 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19790 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19791 (umips_save_restore_pattern_p): Likewise.
19792 (umips_load_store_pair_p): Likewise.
19793 (umips_output_load_store_pair): Likewise.
19794 (umips_movep_target_p): Likewise.
19795 (umips_12bit_offset_address_p): Likewise.
19796 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19797 (mips_base_mips16): Rename this...
19798 (mips_base_compression_flags): ...to this. Update all uses.
19799 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19800 (mips_mips16_decl_p): Delete.
19801 (mips_nomips16_decl_p): Delete.
19802 (mips_get_compress_on_flags): New function.
19803 (mips_get_compress_off_flags): New function.
19804 (mips_get_compress_mode): New function.
19805 (mips_get_compress_on_name): New function.
19806 (mips_get_compress_off_name): New function.
19807 (mips_insert_attributes): Support multiple compression types.
19808 (mips_merge_decl_attributes): Likewise.
19809 (umips_12bit_offset_address_p): New function.
19810 (mips_start_function_definition): Emit .set micromips directive.
19811 (mips_call_may_need_jalx_p): New function.
19812 (mips_function_ok_for_sibcall): Add microMIPS support.
19813 (mips_print_operand_punctuation): Support short delay slots and
19815 (umips_swm_mask, umips_swm_encoding): New.
19816 (umips_build_save_restore): New function.
19817 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19818 (was_mips16_p): Remove.
19819 (old_compression_mode): New.
19820 (mips_set_compression_mode): New function.
19821 (mips_set_current_function): Add microMIPS support.
19822 (mips_option_override): Likewise.
19823 (umips_save_restore_pattern_p): New function.
19824 (umips_output_save_restore): New function.
19825 (umips_load_store_pair_p_1): New function.
19826 (umips_load_store_pair_p): New function.
19827 (umips_output_load_store_pair_1): New function.
19828 (umips_output_load_store_pair): New function.
19829 (umips_movep_target_p) New function.
19830 (mips_prepare_pch_save): Add microMIPS support.
19831 * config/mips/mips.h (TARGET_COMPRESSION): New.
19832 (TARGET_CPU_CPP_BUILTINS): Update macro
19833 to use new compression flags and to support microMIPS.
19834 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19835 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19836 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19837 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19838 (ASM_SPEC): Support mmicromips and mno-micromips.
19839 (M16STORE_REG_P): New macro.
19840 (MIPS_CALL): Support TARGET_MICROMIPS.
19841 (MICROMIPS_J): New macro.
19842 (mips_base_mips16): Rename this...
19843 (mips_base_compression_flags): ...to this.
19844 (UMIPS_12BIT_OFFSET_P): New macro.
19845 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19846 (MULTILIB_DIRNAMES): Likewise.
19848 2013-02-24 Jakub Jelinek <jakub@redhat.com>
19851 * target-globals.c (save_target_globals): For init_reg_sets and
19852 target_reinit remporarily set this_fn_optabs to this_target_optabs.
19854 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
19856 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
19857 * config/aarch64/t-aarch64
19858 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
19860 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
19862 PR inline-asm/56148
19863 * lra-constraints.c (process_alt_operands): Reload operand
19864 conflicting with earlier clobber only if no more other conflicting
19867 2013-02-22 Jakub Jelinek <jakub@redhat.com>
19870 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
19871 if not linking a shared library.
19873 2013-02-22 Seth LaForge <sethml@google.com>
19875 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
19877 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
19879 * config/arm/arm.md (split for extendsidi): Update condition.
19880 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
19881 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
19882 (qhs_zextenddi_cstr): Likewise.
19884 2013-02-21 Jakub Jelinek <jakub@redhat.com>
19886 PR middle-end/56420
19887 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
19888 avoid signed wrapping.
19889 (expand_mult): Handle properly multiplication by
19890 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
19891 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
19892 in the compiler if coeff is HOST_WIDE_INT_MIN.
19893 (expand_divmod): Don't make ext_op1 static, change it's type to
19894 uhwi. Avoid undefined behavior in -INTVAL (op1).
19896 PR rtl-optimization/50339
19897 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
19899 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
19900 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
19901 into splitting_ashiftrt field.
19902 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
19904 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
19907 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
19909 PR middle-end/56108
19910 * trans-mem.c (execute_tm_mark): Do not expand transactions that
19911 are sure to go irrevocable.
19913 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
19915 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
19916 scalars are valid operands.
19918 2013-02-21 Martin Jambor <mjambor@suse.cz>
19920 PR tree-optimization/56310
19921 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
19922 only matching indices and non-negative final offsets.
19923 (intersect_aggregates_with_edge): Pass src_idx to
19924 agg_replacements_to_vector. Pass src_idx insstead of index to
19925 intersect_with_agg_replacements.
19927 2013-02-21 Martin Jambor <mjambor@suse.cz>
19929 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
19930 instead of hard-wired defaults.
19932 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
19934 * doc/invoke.texi (MIPS Options): Update documentation of the
19935 floating-point multiply-accumulate instruction restrictions.
19937 2013-02-21 Kostya Serebryany <kcc@google.com>
19939 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
19940 asan_shadow_offset on x86_64 linux.
19942 2013-02-21 Richard Biener <rguenther@suse.de>
19944 PR tree-optimization/56415
19946 2013-02-11 Richard Biener <rguenther@suse.de>
19948 PR tree-optimization/56273
19949 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19952 2013-02-21 Jakub Jelinek <jakub@redhat.com>
19955 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
19958 PR inline-asm/56405
19959 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
19960 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
19962 2013-02-20 Jan Hubicka <jh@suse.cz>
19964 PR tree-optimization/56265
19965 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
19966 when target is referenced for first time.
19968 2013-02-20 Richard Biener <rguenther@suse.de>
19970 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
19971 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
19972 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
19973 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
19974 not return anything.
19975 (rename_ssa_copies): Do not remove unused locals.
19976 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
19977 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
19978 * passes.c (execute_function_todo): Do not schedule unused locals
19979 removal if cleanup_tree_cfg did something.
19980 * tree-ssa-live.c (remove_unused_locals): Dump statistics
19981 about the number of removed locals.
19983 2013-02-20 Richard Biener <rguenther@suse.de>
19985 PR tree-optimization/56398
19986 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
19988 2013-02-20 Martin Jambor <mjambor@suse.cz>
19990 PR tree-optimization/55334
19991 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
19992 restricted pointers to arrays.
19994 2013-02-20 Richard Biener <rguenther@suse.de>
19995 Jakub Jelinek <jakub@redhat.com>
19997 PR tree-optimization/56396
19998 * tree-ssa-ccp.c (n_const_val): New static variable.
19999 (get_value): Return NULL for SSA names we don't have a lattice
20001 (ccp_initialize): Initialize n_const_val.
20002 * tree-ssa-copy.c (n_copy_of): New static variable.
20003 (init_copy_prop): Initialize n_copy_of.
20004 (get_value): Return NULL_TREE for SSA names we don't have a
20007 2013-02-20 Martin Jambor <mjambor@suse.cz>
20009 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20011 2013-02-20 Richard Biener <rguenther@suse.de>
20013 * genpreds.c (write_lookup_constraint): Do not compare first
20014 letter of the constraint again.
20016 2013-02-20 Richard Biener <rguenther@suse.de>
20018 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20020 (get_use_iv_cost): Terminate hashtable walk when coming across
20023 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20025 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20026 reassociation for avx2 targets.
20028 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20030 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20031 Add version check for v8.10.a to enable microblaze_has_clz
20032 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20033 version and TARGET_PATTERN_COMPARE check
20034 * config/microblaze/microblaze.md: New clzsi2 instruction
20036 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20038 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20039 function before branching.
20041 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20043 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20045 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20047 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20049 PR middle-end/55889
20050 * sel-sched.c: Include ira.h.
20051 (implicit_clobber_conflict_p): New function.
20052 (moveup_expr): Use it.
20053 * Makefile.in (sel-sched.o): Depend on ira.h.
20055 2013-02-19 Richard Biener <rguenther@suse.de>
20057 PR tree-optimization/56384
20058 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20059 (vn_hash_type): Split out from ...
20060 (vn_hash_constant_with_type): ... here.
20061 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20062 (vn_phi_eq): Compare types from vn_phi_s structure.
20063 (vn_phi_lookup): Populate vn_phi_s type.
20064 (vn_phi_insert): Likewise.
20066 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20068 PR tree-optimization/56350
20069 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20070 if haven't found reduction or nested cycle operand, rather than
20071 asserting we must find it.
20073 PR tree-optimization/56381
20074 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20077 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20078 Jakub Jelinek <jakub@redhat.com>
20081 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20082 (swap_optab_enable): Same.
20083 (init_all_optabs): Use argument instead of global.
20084 * tree.h (struct tree_optimization_option): New field target_optabs.
20085 * expr.h (init_all_optabs): Add argument to prototype.
20086 (TREE_OPTIMIZATION_OPTABS): New.
20087 (save_optabs_if_changed): Protoize.
20088 * optabs.h: Declare this_fn_optabs.
20089 * optabs.c (save_optabs_if_changed): New.
20090 Declare this_fn_optabs.
20091 (init_optabs): Add argument to init_all_optabs() call.
20092 * function.c (invoke_set_current_function_hook): Handle per
20094 * function.h (struct function): New field optabs.
20095 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20096 optimization_current_node has changed.
20097 * target-globals.h (save_target_globals_default_opts): Protoize.
20098 * target-globals.c (save_target_globals_default_opts): New.
20100 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20103 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20104 registers %fr12 and %fr12R as call used.
20107 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20108 and HImode, require all displacements to be an integer multiple of
20110 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20111 only allow QImode and HImode when reload is in progress and strict is
20112 true. Likewise for symbolic addresses. Use base14_operand to check
20113 displacements in REG+BASE addresses.
20115 2013-02-18 Richard Biener <rguenther@suse.de>
20117 PR tree-optimization/56366
20118 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20119 handle sign-conversion of outer-loop initial induction value.
20121 2013-02-18 Richard Biener <rguenther@suse.de>
20123 PR middle-end/56349
20124 * cfghooks.c (merge_blocks): If we merge a latch into another
20125 block adjust references to it.
20126 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20127 (verify_loop_structure): Verify that a recorded latch is in fact
20130 2013-02-18 Richard Biener <rguenther@suse.de>
20132 PR tree-optimization/56321
20133 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20134 order SSA name release and virtual operand unlinking.
20136 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20138 * config/microblaze/microblaze.md (save_stack_block): Define.
20139 (restore_stack_block): Likewise.
20141 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20143 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20144 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20145 * config/microblaze/microblaze.c (microblaze_option_override):
20146 Bail out early for PIC modes when target does not support PIC.
20148 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20150 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20151 Replace with a microblaze version.
20152 (microblaze_trampoline_init): Adapt for microblaze.
20153 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20156 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20157 Dodji Seketeli <dodji@redhat.com>
20160 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20161 (instrument_mem_region_access): Do not forget to always put
20162 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20163 0) statement, even for cases where either 'base' or 'base + len'
20164 are not instrumented -- because they have been previously
20165 instrumented. Simplify the logic by putting all the statements
20166 instrument 'base + len' inside a sequence, and then insert that
20167 sequence right before the current insertion point. Then, to
20168 instrument 'base + len', just get an iterator on that statement.
20169 And do not forget to update the pointer to iterator the function
20170 received as argument.
20172 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20174 PR rtl-optimization/56348
20175 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20177 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20179 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20180 (clean_graph_dump_file): Pass base to start_graph_dump.
20182 2013-02-14 Richard Henderson <rth@redhat.com>
20185 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20187 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20189 * collect2-aix.h: Define F_LOADONLY.
20191 2013-02-14 Richard Biener <rguenther@suse.de>
20194 * varasm.c (output_constant_def_1): Get the decl representing
20195 the constant as argument.
20196 (output_constant_def): Wrap output_constant_def_1.
20197 (make_decl_rtl): Use output_constant_def_1 with the decl
20198 representing the constant.
20199 (build_constant_desc): Optionally re-use a decl already
20200 representing the constant.
20201 (tree_output_constant_def): Adjust.
20203 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20206 * asan.c (instrument_builtin_call): Really put the length of the
20207 second source argument into src1_len.
20209 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20211 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20212 argument. If it is false, don't create edge from then_bb to
20214 (insert_if_then_before_iter): Pass true to it.
20215 (build_check_stmt): Pass false to it.
20216 (transform_statements): Flush hash table only on extended basic
20217 block boundaries, rather than at the beginning of every bb.
20218 Don't flush hash table on nonfreeing_call_p calls.
20219 * tree-flow.h (nonfreeing_call_p): New prototype.
20220 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20222 2013-02-13 David S. Miller <davem@davemloft.net>
20224 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20226 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20229 * ira.c (max_regno_before_ira): Move from ...
20231 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20234 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20236 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20238 2013-02-13 Richard Biener <rguenther@suse.de>
20241 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20242 globals in MEM_REFs.
20244 2013-02-13 Richard Biener <rguenther@suse.de>
20246 * loop-init.c (loop_optimizer_init): Clear loop state when
20247 re-initializing preserved loops.
20248 * loop-unswitch.c (unswitch_single_loop): Return whether
20249 we unswitched the loop. Do not verify loop state here.
20250 (unswitch_loops): When we unswitched a loop discover new loops.
20252 2013-02-13 Kostya Serebryany <kcc@google.com>
20254 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20256 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20258 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20260 Avoid instrumenting duplicated memory access in the same basic block
20261 * Makefile.in (asan.o): Add new dependency on hash-table.h
20262 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20263 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20264 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20265 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20266 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20267 (get_mem_ref_of_assignment): New functions.
20268 (get_mem_refs_of_builtin_call): Extract from
20269 instrument_builtin_call and tweak a little bit to make it fit with
20271 (instrument_builtin_call): Use the new
20272 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20273 of is_gimple_builtin_call.
20274 (instrument_derefs, instrument_mem_region_access): Insert the
20275 instrumented memory reference into the hash table.
20276 (maybe_instrument_assignment): Renamed instrument_assignment into
20277 this, and change it to advance the iterator when instrumentation
20278 actually happened and return true in that case. This makes it
20279 homogeneous with maybe_instrument_assignment, and thus give a
20280 chance to callers to be more 'regular'.
20281 (transform_statements): Clear the memory reference hash table
20282 whenever we enter a new BB, when we cross a function call, or when
20283 we are done transforming statements. Use
20284 maybe_instrument_assignment instead of instrumentation. No more
20285 need to special case maybe_instrument_assignment and advance the
20286 iterator after calling it; it's now handled just like
20287 maybe_instrument_call. Update comment.
20289 2013-02-13 Richard Biener <rguenther@suse.de>
20291 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20292 Fix loop discovery code.
20294 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20296 PR inline-asm/56148
20297 * lra-constraints.c (process_alt_operands): Match early clobber
20298 operand with itself. Check conflicts with earlyclobber only if
20299 the operand is not reloaded. Prefer to reload conflicting operand
20300 if earlyclobber and matching operands are the same.
20302 2013-02-12 Richard Biener <rguenther@suse.de>
20305 * lto-streamer-out.c (write_symbol): Do not output symbols
20306 for hard register variables.
20308 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20311 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20312 (umulsidi3_insn, mulsidi3_insn): New insns.
20314 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20316 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20317 (struct tune_params): Add vec_costs field.
20318 * config/arm/arm.c (arm_builtin_vectorization_cost)
20319 (arm_add_stmt_cost): New functions.
20320 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20321 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20322 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20323 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20324 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20325 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20326 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20328 2013-02-12 Richard Biener <rguenther@suse.de>
20331 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20332 decls again if possible.
20334 2013-02-12 Richard Biener <rguenther@suse.de>
20336 PR middle-end/56288
20337 * tree-ssa.c (verify_ssa_name): Fix check, move
20338 SSA_NAME_IN_FREE_LIST check up.
20340 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20341 Steven Bosscher <steven@gcc.gnu.org>
20343 PR rtl-optimization/56151
20344 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20345 equal to op0 or op1, and last_insn pattern is CODE operation
20346 with MEM dest and one of the operands matches that MEM.
20348 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20350 * doc/extend.texi: Document Function Multiversioning and "default"
20351 parameter string to target attribute.
20352 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20353 target attribute parameter is "default".
20354 (ix86_compare_version_priority): Remove checks for target attribute.
20355 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20356 Remove check for target attribute equal to NULL. Add assert.
20357 (ix86_generate_version_dispatcher_body): Change error to sorry.
20359 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20360 Jack Howarth <howarth@bromo.med.uc.edu>
20361 Patrick Marlier <patrick.marlier@gmail.com>
20364 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20365 define ENDFILE_SPEC as TM_DESTRUCTOR.
20366 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20368 2013-02-11 Alexander Potapenko <glider@google.com>
20369 Jack Howarth <howarth@bromo.med.uc.edu>
20370 Jakub Jelinek <jakub@redhat.com>
20373 * config/darwin.c (cdtor_record): Rename ctor_record.
20374 (sort_cdtor_records): Rename sort_ctor_records.
20375 (finalize_dtors): New routine to sort destructors by
20376 priority before use in assemble_integer.
20377 (machopic_asm_out_destructor): Use finalize_dtors if needed.
20379 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
20381 PR rtl-optimization/56275
20382 * simplify-rtx.c (avoid_constant_pool_reference): Check that
20383 offset is non-negative and less than cmode size before
20384 calling simplify_subreg.
20386 2013-02-11 Richard Biener <rguenther@suse.de>
20388 PR tree-optimization/56264
20389 * cfgloop.h (fix_loop_structure): Adjust prototype.
20390 * loop-init.c (fix_loop_structure): Return the number of
20391 newly discovered loops.
20392 * tree-cfgcleanup.c (repair_loop_structures): When new loops
20393 are discovered, do a full loop-closed SSA rewrite.
20395 2013-02-11 Richard Biener <rguenther@suse.de>
20397 PR tree-optimization/56273
20398 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20400 (check_array_ref): Fix missing newline in dumps.
20401 (search_for_addr_array): Likewise.
20403 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
20405 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
20407 2013-02-09 Jakub Jelinek <jakub@redhat.com>
20410 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
20412 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
20414 PR rtl-optimization/56246
20415 * lra-constraints.c (simplify_operand_subreg): Try to reuse
20417 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
20418 constraints are satisfied.
20420 2013-02-08 Jeff Law <law@redhat.com>
20423 * emit-rtl.c (reg_is_parm_p): New function.
20424 * regs.h (reg_is_parm_p): New prototype.
20425 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
20426 callee-clobbered registers.
20428 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
20431 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
20432 If there is no implicit builtin declaration, just return NULL.
20434 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20436 * config/i386/sse.md (FMAMODEM): New mode iterator.
20437 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
20438 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
20440 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20442 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
20443 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
20444 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
20446 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20448 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
20449 (microblaze*-*-elf): Likewise.
20450 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
20452 * config/microblaze/microblaze-c.c: Add builtin defines for
20453 _LITTLE_ENDIAN and _BIG_ENDIAN.
20454 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
20455 add to TARGET_DEFAULT flags.
20456 Expand ASM_SPEC and LINK_SPEC.
20457 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
20458 * config/microblaze/microblaze.md: Update extendsidi2 and
20459 movdi_internal instructions to use low-order / high-order reg
20461 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
20462 options and inversemask / mask of LITTLE_ENDIAN.
20463 * config/microblaze/t-microblaze: Expand multilib options to
20464 include mlittle-endian (le) and update exceptions patterns.
20466 2013-02-08 Jakub Jelinek <jakub@redhat.com>
20468 PR rtl-optimization/56195
20469 * lra-constraints.c (get_reload_reg): Don't reuse regs
20470 if they have smaller mode than requested, if they have
20471 wider mode than requested, try to return a SUBREG.
20473 PR tree-optimization/56250
20474 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
20475 if type is unsigned and code isn't MULT_EXPR.
20477 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20479 PR tree-optimization/56064
20480 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
20481 bits according to mode.
20482 * fixed-value.h (fixed_from_double_int)
20483 (const_fixed_from_double_int): Adjust comments.
20485 2013-02-08 Richard Biener <rguenther@suse.de>
20488 * lto-streamer.h (struct data_in): Remove current_file, current_line
20489 and current_col members.
20490 * lto-streamer-out.c (lto_output_location): Stream changed bits
20491 en-block for efficiency.
20492 * lto-streamer-in.c (clear_line_info): Remove.
20493 (lto_input_location): Cache current file, line and column
20494 globally via local statics. Read changed bits en-block.
20495 (input_function): Do not call clear_line_info.
20496 (lto_read_body): Likewise.
20497 (lto_input_toplevel_asms): Likewise.
20499 2013-02-08 Michael Matz <matz@suse.de>
20501 PR tree-optimization/52448
20502 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
20503 (nt_call_phase): New static.
20504 (add_or_mark_expr): Only mark accesses with newer phase than any
20506 (nonfreeing_call_p): New.
20507 (nt_init_block): Update nt_call_phase, mark blocks as visited.
20508 (nt_fini_block): Keep blocks marked as visited.
20509 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
20511 2013-02-08 Richard Biener <rguenther@suse.de>
20513 * ira.c (ira): Free broken dominator information.
20515 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20517 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
20519 2013-02-08 Marek Polacek <polacek@redhat.com>
20521 * cfgloop.c (verify_loop_structure): Add more checking of headers.
20523 2013-02-08 Richard Biener <rguenther@suse.de>
20525 PR middle-end/56181
20526 * cfgloop.h (flow_loops_find): Adjust.
20527 (bb_loop_header_p): Declare.
20528 * cfgloop.c (bb_loop_header_p): New function split out from ...
20529 (flow_loops_find): ... here. Adjust function signature,
20530 support incremental loop structure update.
20531 (verify_loop_structure): Cleanup. Verify a loop is a loop.
20532 * cfgloopmanip.c (fix_loop_structure): Move ...
20533 * loop-init.c (fix_loop_structure): ... here.
20534 (apply_loop_flags): Split out from ...
20535 (loop_optimizer_init): ... here.
20536 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
20537 in incremental mode, only remove dead loops here.
20539 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20542 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
20543 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
20544 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
20545 (*round<mode>3.libgcc): New insns for fixed-modes.
20546 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
20547 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
20548 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
20549 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
20550 implementations. Define to __builtin_avr_absFX,
20551 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
20552 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
20553 __builtin_avr_countlsFX, respectively.
20554 * config/avr/avr-c.c (target.h): Include it.
20555 (enum avr_builtin_id): New enum.
20556 (avr_resolve_overloaded_builtin): New static function.
20557 (avr_register_target_pragmas): Use it to set
20558 targetm.resolve_overloaded_builtin.
20559 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
20560 tree nodes used by DEF_BUILTIN.
20561 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
20562 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
20563 <AVR_BUILTIN_xxBITS>: Same.
20565 2013-02-08 Richard Biener <rguenther@suse.de>
20567 * cfgloop.c (verify_loop_structure): Properly handle
20568 a loop exiting to another loop header.
20569 * ira-int.h (ira_loops): Remove.
20570 * ira.c (ira_loops): Remove.
20571 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
20572 (do_reload): Use loop_optimizer_finalize.
20573 * ira-build.c (create_loop_tree_nodes): Use get_loops and
20574 number_of_loops to access the loop tree.
20575 (more_one_region_p): Likewise.
20576 (finish_loop_tree_nodes): Likewise.
20577 (rebuild_regno_allocno_maps): Likewise.
20578 (mark_loops_for_removal): Likewise.
20579 (mark_all_loops_for_removal): Likewise.
20580 (remove_unnecessary_regions): Likewise.
20581 (ira_build): Likewise.
20582 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20584 2013-02-08 Richard Biener <rguenther@suse.de>
20586 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
20587 * ipa-pure-const.c (analyze_function): Avoid calling
20588 mark_irreducible_loops twice.
20589 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
20591 2013-02-07 David S. Miller <davem@davemloft.net>
20593 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
20595 * var-tracking.c (vt_add_function_parameter): Test the presence of
20596 HAVE_window_save properly and do not remap argument registers when
20597 we have a leaf function.
20599 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
20602 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
20604 * config/i386/i386.c (ix86_print_operand): Ditto.
20606 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20608 * lra-constraints.c (process_alt_operands): Fix recently added comment.
20610 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20612 PR rtl-optimization/56225
20613 * lra-constraints.c (process_alt_operands): Check that reload hard
20614 reg can hold value for strict_low_part.
20616 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20619 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
20620 dwarf2out_end_function.
20621 (in_first_function_p, maybe_at_text_label_p,
20622 first_loclabel_num_not_at_text_label): New variables.
20623 (dwarf2out_var_location): In the first function find out
20624 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
20625 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
20628 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
20630 PR rtl-optimization/56178
20631 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
20632 SUBREG of a register. Tidy up related block of code.
20633 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
20634 note if the source is a register or a SUBREG of a register.
20636 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20639 * config/rs6000/rs6000.md (ptrm): New mode attr.
20640 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
20641 call_value_indirect_aix<pttrsize>,
20642 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
20645 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20647 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
20648 if -bnortl. Convert to strcmp and strncmp.
20650 2013-02-07 Alan Modra <amodra@gmail.com>
20653 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
20654 addresses won't wrap when offsetting.
20655 (rs6000_secondary_reload): Provide secondary reloads needed for
20656 wrapping LO_SUM addresses.
20658 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
20660 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
20661 MACH, just __MACH__.
20663 2013-02-06 Richard Biener <rguenther@suse.de>
20665 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
20666 instead of calling fix_loop_structure.
20668 2013-02-06 Jakub Jelinek <jakub@redhat.com>
20670 PR middle-end/56217
20671 * omp-low.c (use_pointer_for_field): Return false if
20672 lower_send_shared_vars doesn't generate any copy-out code.
20674 2013-02-06 Tom de Vries <tom@codesourcery.com>
20676 PR rtl-optimization/56131
20677 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
20678 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
20679 of the label is NULL. Add comment.
20681 2013-02-05 Jakub Jelinek <jakub@redhat.com>
20683 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
20686 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
20687 (STATIC_LIBTSAN_LIBS): Likewise.
20688 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
20689 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
20690 is defined, don't add anything else beyond that.
20691 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
20692 (LINK_COMMAND_SPEC): Use them.
20694 PR tree-optimization/56205
20695 * tree-stdarg.c (check_all_va_list_escapes): Return true if
20696 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
20697 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
20699 2013-02-05 Richard Biener <rguenther@suse.de>
20701 PR tree-optimization/53342
20702 PR tree-optimization/53185
20703 * tree-vectorizer.h (vect_check_strided_load): Remove.
20704 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
20705 not disallow peeling for vectorized strided loads.
20706 (vect_check_strided_load): Make static and simplify.
20707 (vect_analyze_data_refs): Adjust.
20708 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
20709 correctly when vectorizing strided loads.
20711 2013-02-05 Richard Biener <rguenther@suse.de>
20713 * doc/install.texi: Refer to ISL, not PPL.
20715 2013-02-05 Jan Hubicka <jh@suse.cz>
20717 PR tree-optimization/55789
20718 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
20720 2013-02-05 Jan Hubicka <jh@suse.cz>
20722 PR tree-optimization/55789
20723 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
20724 the dead call anyway.
20726 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
20729 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
20731 2013-02-04 Alexander Potapenko <glider@google.com>
20732 Jack Howarth <howarth@bromo.med.uc.edu>
20733 Jakub Jelinek <jakub@redhat.com>
20736 * config/darwin.c (sort_ctor_records): Stabilized qsort
20737 on constructor priority by using original position.
20738 (finalize_ctors): New routine to sort constructors by
20739 priority before use in assemble_integer.
20740 (machopic_asm_out_constructor): Use finalize_ctors if needed.
20742 2013-02-04 Jakub Jelinek <jakub@redhat.com>
20745 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
20746 about visibility on artificial decls.
20747 * config/sol2.c (solaris_assemble_visibility): Likewise.
20749 2013-02-04 Kai Tietz <ktietz@redhat.com>
20752 * config/i386/i386.c (function_value_ms_64): Add additional valtype
20753 argument and improve checking of return-argument types for 16-byte
20755 (ix86_function_value_1): Add additional valtype argument on call
20756 of function_value_64.
20757 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
20758 handling infunction_value_64 function.
20760 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
20762 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
20764 2013-02-04 Richard Biener <rguenther@suse.de>
20766 PR tree-optimization/56188
20767 * tree-ssa-structalias.c (label_visit): Consider case with
20768 initially non-empty points-to set.
20769 (perform_var_substitution): Dump node mapping and clean up.
20771 2013-02-04 Richard Guenther <rguenther@suse.de>
20774 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
20775 node prevail as last resort.
20776 (lto_symtab_merge_decls): Remove guard on LTRANS here.
20777 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
20779 2013-02-04 Richard Biener <rguenther@suse.de>
20781 PR tree-optimization/56113
20782 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
20784 (equiv_class_lookup_or_add): ... this.
20785 (label_visit): Adjust and fix error in previous patch.
20786 (perform_var_substitution): Adjust.
20788 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
20790 * config/sh/divtab.c: Fix formatting and comments throughout the file.
20791 * config/sh/sh4-300.md: Likewise.
20792 * config/sh/sh4a.md: Likewise.
20793 * config/sh/constraints.md: Likewise.
20794 * config/sh/sh.md: Likewise.
20795 * config/sh/netbsd-elf.h: Likewise.
20796 * config/sh/predicates.md: Likewise.
20797 * config/sh/sh-protos.h: Likewise.
20798 * config/sh/ushmedia.h: Likewise.
20799 * config/sh/linux.h: Likewise.
20800 * config/sh/sh.c: Likewise.
20801 * config/sh/superh.h: Likewise.
20802 * config/sh/elf.h: Likewise.
20803 * config/sh/sh4.md: Likewise.
20804 * config/sh/sh.h: Likewise.
20806 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20808 * config/pa/constraints.md: Adjust unused letters. Change "T"
20809 constraint to match_test floating_point_store_memory_operand().
20810 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
20811 (base14_operand): New.
20812 (floating_point_store_memory_operand): New.
20813 (integer_store_memory_operand): Revise to use base14_operand and
20814 reg_plus_base_memory_operand.
20815 (move_dest_operand): Allow symbolic_memory_operands.
20816 (symbolic_memory_operand): Check for LO_SOM.
20817 (symbolic_operand): Change default case to break.
20818 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
20819 CONST_DOUBLE values to be reloaded by putting them into memory when
20820 the destination is a floating point register.
20821 (movdf): Remove code to handle CONST_DOUBLE.
20823 (reload_indf_r1): New.
20824 (reload_insf_r1): New.
20825 Consistently use "Q" and "T" constraints with integer and floating
20826 point move instructions, respectively.
20827 (movdi): Remove FAIL.
20828 Change predicate for source operand unamed DImode move from
20829 general_operand to move_src_operand.
20830 (umulsidi3): Change predicate for destination operand to
20832 Likewise for similar unamed patterns.
20833 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
20834 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
20835 (hppa_legitimize_address): Simplify mask calculation.
20836 (pa_emit_move_sequence): Revised handling of secondary reloads from
20837 REG+D addresses for floating point loads and stores. Directly handle
20838 loading CONST0_RTX (mode) to a floating point register.
20839 (pa_secondary_reload): Handle reloading DF and SFmode constant values
20840 to floating point registers. Don't restrict secondary reloads to
20841 floating point registers to integer modes. Revise some comments and
20843 (TARGET_LEGITIMATE_ADDRESS_P): Define.
20844 (pa_legitimate_address_p): New.
20845 (pa_legitimize_reload_address): New.
20846 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
20847 (STRICT_REG_OK_FOR_BASE_P): New.
20848 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
20849 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
20851 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
20852 Andrew Dixie <andrewd@gentrack.com>
20854 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
20857 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
20859 * expmed.c (extract_bit_field_1): Pass the full width of the
20860 structure to get_best_reg_extraction_insn.
20862 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
20865 * configure.ac (use_cxa_atexit): Add AIX.
20866 * configure: Regenerate.
20868 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
20870 2013-02-01 Jakub Jelinek <jakub@redhat.com>
20873 * final.c (need_profile_function): New variable.
20874 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
20875 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
20876 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
20877 notes, targetm.asm_out.function_prologue doesn't emit anything,
20878 HAVE_prologue and profiler should be emitted before prologue,
20879 set need_profile_function instead of emitting it.
20880 (final_scan_insn): If need_profile_function, emit
20881 profile_function on the first NOTE_INSN_BASIC_BLOCK or
20882 NOTE_INSN_FUNCTION_BEG note.
20884 2013-02-01 Richard Henderson <rth@redhat.com>
20886 * config/rs6000/rs6000.md (smulditi3): New.
20889 * config/alpha/alpha.md (umulditi3): New.
20891 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
20893 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
20894 (ASM_OUTPUT_ALIGNED_LOCAL): New.
20896 2013-02-01 Richard Biener <rguenther@suse.de>
20898 PR tree-optimization/56113
20899 * tree-ssa-structalias.c (label_visit): Reduce work for
20900 single-predecessor nodes.
20902 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
20904 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
20905 range isn't testing for zero.
20907 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
20909 PR middle-end/56113
20910 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
20912 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
20913 Nick Clifton <nickc@redhat.com>
20915 * config/v850/constraints.md (Q): Define as a memory constraint.
20916 * config/v850/predicates.md (label_ref_operand): New predicate.
20917 (e3v5_shift_operand): New predicate.
20918 (ior_operator): New predicate.
20919 * config/v850/t-v850: Add e3v5 multilib.
20920 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
20921 (v850_gen_movdi): Prototype.
20922 * config/v850/v850.c: Add support for e3v5 architecture.
20923 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
20925 (construct_save_jarl): Add e3v5 long JARL support.
20926 (v850_adjust_insn_length): New function. Adjust length of call
20927 insns when using e3v5 instructions.
20928 (v850_gen_movdi): New function: Generate instructions to move a
20930 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
20931 (CPP_SPEC): Define __v850e3v5__ as appropriate.
20932 (TARGET_USE_FPU): Enable for e3v5.
20933 (CONST_OK_FOR_W): New macro.
20934 (ADJUST_INSN_LENGTH): Define.
20935 * config/v850/v850.md (UNSPEC_LOOP): Define.
20936 (attr cpu): Add v850e3v5.
20937 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
20938 (movdi): New pattern.
20939 (movdi_internal): New pattern.
20940 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
20941 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
20942 (cstoresf4): Likewise.
20943 (cstoredf4): Likewise.
20944 (insv): New pattern.
20945 (rotlso3_a): New pattern.
20946 (rotlsi3_b): New pattern
20947 (rotlsi3_v850e3v5): New pattern.
20948 (doloop_begin): New pattern.
20949 (fix_loop_counter): New pattern.
20950 (doloop_end): New pattern.
20951 (branch_normal): Add e3v5 long branch support.
20952 (branch_invert): Likewise.
20953 (branch_z_normal): Likewise.
20954 (branch_z_invert): Likewise.
20955 (branch_nz_normal): Likewise.
20956 (branch_nz_invert): Likewise.
20957 (call_internal_short): Add e3v5 register-indirect JARL support.
20958 (call_internal_long): Likewise.
20959 (call_value_internal_short): Likewise.
20960 (call_value_internal_long): Likewise.
20961 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
20962 (mloop): New option.
20963 * config.gcc: Add support for configuring v840e3v5 target.
20964 * doc/invoke.texi: Document new v850 specific command line options.
20966 2013-01-31 Paul Koning <ni1d@arrl.net>
20970 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
20971 children if parent is a class.
20972 (prune_unused_types_prune): Don't add DW_AT_declaration.
20974 2013-01-31 Richard Biener <rguenther@suse.de>
20976 PR tree-optimization/56157
20977 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
20978 match up operand with SLP child.
20980 2013-01-31 Jason Merrill <jason@redhat.com>
20983 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
20984 parameters the first time.
20985 (gen_scheduled_generic_parms_dies): Check completeness here.
20987 2013-01-31 Richard Biener <rguenther@suse.de>
20989 PR middle-end/53073
20990 * common.opt (faggressive-loop-optimizations): New flag,
20991 enabled by default.
20992 * doc/invoke.texi (faggressive-loop-optimizations): Document.
20993 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
20994 infer_loop_bounds_from_undefined by it.
20996 2013-01-31 Richard Biener <rguenther@suse.de>
20998 PR tree-optimization/56150
20999 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21000 visit virtual operands.
21001 (find_uses_to_rename_bb): Likewise.
21003 2013-01-31 Richard Biener <rguenther@suse.de>
21005 PR tree-optimization/56150
21006 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21007 mixed store non-store stmts.
21009 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21012 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21013 LIBASAN_EARLY_SPEC is defined.
21014 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21015 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21017 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21020 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21021 invalid args instead of ICEing on it.
21022 (ix86_valid_target_attribute_tree): Return error_mark_node if
21023 ix86_valid_target_attribute_inner_p failed.
21024 (ix86_valid_target_attribute_p): Return false only if
21025 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21026 target("default") attribute.
21027 (sorted_attr_string): Change argument from const char * to tree,
21028 merge in all target attribute arguments rather than just one.
21029 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21030 instead of free. Avoid using strcat.
21031 (ix86_mangle_function_version_assembler_name): Mangle
21032 target("default") as if no target attribute is present. Adjust
21033 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21034 instead of xmalloc and XDELETEVEC instead of free.
21035 (ix86_function_versions): Don't return true if one of the decls
21036 doesn't have target attribute. If they don't and one of the decls
21037 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21038 sorted_attr_string caller. Use XDELETEVEC instead of free.
21039 (ix86_supports_function_versions): Remove.
21040 (make_name): Fix up formatting.
21041 (make_dispatcher_decl): Remove resolver_name and its initialization.
21042 Avoid leaking memory.
21043 (is_function_default_version): Return true if there is
21044 target("default") attribute rather than no target attribute at all.
21045 (make_resolver_func): Avoid leaking memory.
21046 (ix86_generate_version_dispatcher_body): Likewise.
21047 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21048 * target.def (supports_function_versions): Remove.
21049 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21050 * doc/tm.texi: Regenerated.
21052 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21054 PR rtl-optimization/56144
21055 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21056 for values with side effects.
21058 2013-01-30 Richard Biener <rguenther@suse.de>
21060 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21061 (sparseset_pop): Likewise.
21062 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21063 to be able to use quick_push in the worker loop.
21065 2013-01-30 Marek Polacek <polacek@redhat.com>
21067 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21069 2013-01-30 Richard Biener <rguenther@suse.de>
21072 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21074 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21076 PR tree-optimization/56064
21077 * fixed-value.c (fixed_from_double_int): New function.
21078 * fixed-value.h (fixed_from_double_int): New prototype.
21079 (const_fixed_from_double_int): New static inline function.
21080 * fold-const.c (native_interpret_fixed): New static function.
21081 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21082 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21083 (native_encode_fixed): New static function.
21084 (native_encode_expr) <FIXED_CST>: Use it.
21085 (native_interpret_int): Move double_int worker code to...
21086 * double-int.c (double_int::from_buffer): ...this new static method.
21087 * double-int.h (double_int::from_buffer): Prototype it.
21089 2013-01-30 Richard Biener <rguenther@suse.de>
21091 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21092 New pointer-map and obstack.
21093 (init_alias_vars): Allocate pointer-map and obstack.
21094 (delete_points_to_sets): Free them.
21095 (find_what_var_points_to): Cache result.
21096 (find_what_p_points_to): Adjust for changed interface of
21097 find_what_var_points_to.
21098 (compute_points_to_sets): Likewise.
21099 (ipa_pta_execute): Likewise.
21101 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21103 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21104 * configure: Regenerate.
21105 * config.in: Regenerate.
21106 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21107 #nobits/#progbits if supported.
21109 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21112 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21113 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21114 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21116 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21118 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21119 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21121 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21123 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21124 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21126 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21128 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21130 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21131 * config/arm/cortex-a7.md: New bypasses using
21132 arm_mac_accumulator_is_result.
21134 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21136 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21137 (cortex_a7_neon_mla): Likewise.
21138 (cortex_a7_fpfmad): New reservation.
21139 (cortex_a7_fpmacs): Use ffmas and update required units.
21140 (cortex_a7_fpmuld): Update required units and latency.
21141 (cortex_a7_fpmacd): Likewise.
21142 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21143 (cortex_a7_neon). Likewise.
21144 (bypass) Update participating units.
21146 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21148 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21149 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21151 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21152 (vfp_fmul): Use ffmad.
21153 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21154 (cortex_r4_fmacd): Use ffmad.
21155 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21156 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21157 (cortex_a9_fmacd): Use ffmad.
21158 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21159 (cortex_a8_vfp_macd): Use ffmad.
21160 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21161 (cortex_a5_fpmacd): Use ffmad.
21162 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21163 (cortex_a15_vfp_macd): Use ffmad.
21164 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21166 2013-01-29 Jason Merrill <jason@redhat.com>
21169 * varasm.c (default_assemble_visibility): Don't warn about
21170 visibility on artificial decls.
21172 2013-01-29 Richard Biener <rguenther@suse.de>
21174 PR tree-optimization/56113
21175 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21177 (label_visit): Free duplicate bitmaps and record the leader instead.
21178 (perform_var_substitution): Adjust.
21180 2013-01-29 Richard Biener <rguenther@suse.de>
21182 PR tree-optimization/55270
21183 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21184 the CFG, schedule loops for fixup.
21186 2013-01-29 Nick Clifton <nickc@redhat.com>
21188 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21191 2013-01-28 Leif Ekblad <leif@rdos.net>
21193 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21194 * config/i386/i386.h (TARGET_RDOS): New macro.
21195 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21196 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21197 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21198 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21199 DEFAULT_LARGE_SECTION_THRESHOLD.
21200 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21201 * config/i386/rdos.h: New file.
21202 * config/i386/rdos64.h: New file.
21204 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21207 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21210 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21212 PR rtl-optimization/56117
21213 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21214 call cselib_lookup_from_insn on the MEM before calling
21215 add_insn_mem_dependence.
21217 2013-01-28 Richard Biener <rguenther@suse.de>
21219 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21220 to a stmt that didn't have one.
21221 (copy_phis_for_bb): Likewise for PHI arguments.
21222 (copy_debug_stmt): Likewise for debug stmts.
21224 2013-01-28 Richard Biener <rguenther@suse.de>
21226 PR tree-optimization/56034
21227 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21228 (partition_builtin_p): Adjust.
21229 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21230 it is the last partition.
21231 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21232 up the vertex for the definition.
21233 (classify_partition): Classify whether a partition is a
21234 PKIND_REDUCTION, thus has uses outside of the loop.
21235 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21236 Merge all PKIND_REDUCTION partitions into the last partition.
21237 (tree_loop_distribution): Seed partitions from reductions as well.
21239 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21241 PR tree-optimization/56125
21242 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21243 pow(x,c) into sqrt(x) * powi(x, n/2) or
21244 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21245 optimizing for size.
21246 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21247 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21250 PR tree-optimization/56094
21251 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21252 to UNKNOWN_LOCATION while gimplifying expr.
21254 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21257 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21258 operand 0 in movabs insn template for -masm=intel asm alternative.
21259 (*movabs<mode>_2): Ditto for operand 1.
21261 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21264 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21267 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21269 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21272 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21274 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21275 Undef to avoid warning.
21277 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21279 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21280 * configure: Regenerate.
21282 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21284 PR tree-optimization/56098
21285 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21286 for stmts with volatile ops.
21287 (cond_store_replacement): Don't optimize if assign has volatile ops.
21288 (cond_if_else_store_replacement_1): Don't optimize if either
21289 then_assign or else_assign have volatile ops.
21290 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21293 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21295 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21297 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21299 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21300 missing ':' in asm example.
21302 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21304 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21305 entries into lane and laneq entries.
21306 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21307 Remove AdvSIMD scalar modes.
21308 (aarch64_sq<r>dmulh_laneq<mode>): New.
21309 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21311 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21312 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21313 * config/aarch64/iterators.md (VCOND): New.
21316 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21319 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21320 Add NULL LIBNAME argument to existing definitions.
21321 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21322 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21323 * config/avr/avr.c (DEF_BUILTIN): Same.
21324 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21325 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21326 implementation is available (DECL_ASSEMBLER_NAME is set).
21327 (avr_fold_absfx): New static function.
21328 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21329 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21330 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21331 AVR_BUILTIN_ABSLLK.
21332 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21333 (abshk, absk, abslk, absllk): Provide as static inline functions.
21335 2013-01-25 Marek Polacek <polacek@redhat.com>
21337 PR tree-optimization/56035
21338 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21340 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21342 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21343 (*movtf_internal_rex64): Add (!o,C) alternative
21344 (*movxf_internal_rex64): Ditto.
21345 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21347 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21349 * doc/invoke.texi: fix typo.
21350 * doc/objc.texi: fix typo.
21352 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21354 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21355 for the first two alternatives.
21357 2013-01-24 Diego Novillo <dnovillo@google.com>
21359 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21360 (ggc-zone.o): Remove.
21361 * configure.ac: Remove option --with-gc.
21362 * configure: Re-generate.
21363 * doc/install.texi: Remove documentation for --with-gc.
21364 * gengtype.c (write_enum_defn): Remove. Update all users.
21365 (write_Types_process_field): Remove generation of gt_e_* argument.
21366 (output_type_enum): Remove. Update all users.
21367 (write_enum_defn): Remove. Update all users.
21368 (enum alloc_zone): Remove. Update all users.
21369 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21370 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21371 Update all callers.
21372 (struct ptr_data): Remove field TYPE. Update all users.
21373 (gt_pch_note_object): Remove argument TYPE. Update all users.
21374 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
21376 * ggc-none.c (ggc_alloc_typed_stat): Remove.
21377 (struct alloc_zone): Remove.
21378 (ggc_internal_alloc_zone_stat): Remove.
21379 (ggc_internal_cleared_alloc_zone_stat): Remove.
21380 * ggc-page.c (ggc_alloc_typed_stat): Remove.
21381 (ggc_pch_count_object): Remove last argument. Update all users.
21382 (ggc_pch_alloc_object): Remove last argument. Update all users.
21383 (struct alloc_zone): Remove.
21384 * ggc-zone.c: Remove.
21385 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
21386 (struct alloc_zone): Remove.
21387 (ggc_alloc_typed_stat): Remove.
21388 (ggc_alloc_typed): Remove.
21389 (ggc_splay_alloc): Remove first argument.
21390 (rtl_zone): Remove. Update all users.
21391 (tree_zone): Remove. Update all users.
21392 (tree_id_zone): Remove. Update all users.
21393 (ggc_internal_zone_alloc_stat): Remove. Update all users.
21394 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
21395 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
21396 * tree-ssanames.c: Remove references to zone allocator in comments.
21398 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
21400 * config/avr/avr.c (avr_out_fract): Make register numbers that
21401 might be outside of source operand signed.
21403 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
21405 * config/i386/constraints.md (Yf): New constraint.
21406 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
21407 of f constraint to conditionaly disable x87 register preferences.
21408 (*movdf_internal): Ditto.
21409 (*movsf_internal): Ditto.
21411 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
21413 PR inline-asm/55934
21414 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
21415 that have operands with impossible constraints.
21416 Add a FIXME for a speed-up opportunity.
21417 * lra-constraints.c (process_alt_operands): Verify that a class
21418 selected from constraints on asms is valid for the operand mode.
21419 (curr_insn_transform): Remove incorrect comment.
21421 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
21423 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
21424 TOC operand is a valid symbol ref in the constant pool.
21426 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21428 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
21430 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
21433 * config/avr/stdfix.h: New file.
21434 * t-avr (stdfix-gcc.h): New rule to build it.
21435 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
21437 2013-01-23 Kostya Serebryany <kcc@google.com>
21439 * config/darwin.h: remove dependency on
21440 CoreFoundation (asan on Mac OS).
21442 2013-01-23 Jakub Jelinek <jakub@redhat.com>
21445 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
21446 instead of cmpdi_operand for first comparison operand.
21447 Don't assert that comparison operands aren't both constants.
21449 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21451 * doc/install.texi (Downloading the Source): Update references to
21452 downloading separate components.
21454 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21456 * doc/extend.texi (__int128): Improve grammar.
21458 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
21461 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
21462 alternative to (o,r).
21463 (*movdi_internal_rex64): Remove (!o,n) alternative.
21464 (DImode immediate->memory splitter): Remove.
21465 (DImode immediate->memory peephole2): Remove.
21466 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
21467 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
21468 alternative to (!o,*r).
21469 (*movtf_internal_sse): New pattern.
21470 (*movxf_internal_rex64): New pattern.
21471 (*movxf_internal): Disable for TARGET_64BIT.
21472 (*movdf_internal_rex64): Remove (!o,F) alternative.
21474 2013-01-22 Jakub Jelinek <jakub@redhat.com>
21476 PR middle-end/56074
21477 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
21478 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
21479 * tree-vect-loop-manip.c (find_loop_location): Also ignore
21480 stmt locations where LOCATION_LOCUS of the stmt location is
21481 UNKNOWN_LOCATION or BUILTINS_LOCATION.
21484 * config/i386/i386.md (UNSPEC_STOS): New.
21485 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
21486 *strsetqi_1): Add UNSPEC_STOS.
21488 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
21491 * doc/invoke.texi: Remove left over -Wsynth example.
21493 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21495 PR tree-optimization/56051
21496 * fold-const.c (fold_binary_loc): Don't fold
21497 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
21498 a narrowing conversion, or widening conversion from signed
21501 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21503 PR rtl-optimization/56023
21504 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
21505 dependent on debug instruction.
21507 2013-01-21 Martin Jambor <mjambor@suse.cz>
21509 PR middle-end/56022
21510 * function.c (allocate_struct_function): Call
21511 invoke_set_current_function_hook earlier.
21513 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21515 * reload1.c (init_reload): Only initialize reload_obstack
21516 during the first call.
21518 2013-01-21 Marek Polacek <polacek@redhat.com>
21520 * cfgloop.c (verify_loop_structure): Fix up grammar.
21522 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
21524 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
21525 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
21527 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21530 * config/arm/marvell-pj4.md: Update copyright year.
21531 Fix up use of alu to alu_reg and simple_alu_imm.
21533 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21535 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
21537 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
21540 * lra-constraints.c (curr_insn_transform): Don't reuse original
21541 insn for secondary memory move when memory mode should be different.
21543 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21545 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
21546 atomic_storedi_1): New patterns.
21548 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21550 btver2 pipeline descriptions.
21551 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
21553 * config/i386/i386.md (btver2_decode): New type attributes.
21554 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
21556 * config/i386/btver2.md: New file describing btver2 pipelines.
21558 2013-01-19 Andrew Pinski <apinski@cavium.com>
21560 PR tree-optimization/52631
21561 * tree-ssa-sccvn (visit_use): Before looking up the original
21562 statement, try looking up the simplified expression.
21564 2013-01-19 Anthony Green <green@moxielogic.com>
21566 * config/moxie/moxie.c (moxie_expand_prologue): Set
21567 current_function_static_stack_size.
21569 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21571 PR tree-optimization/56029
21572 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
21573 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
21575 2013-01-18 Sharad Singhai <singhai@google.com>
21577 PR tree-optimization/55995
21578 * dumpfile.c (dump_loc): Print location only if available.
21579 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
21581 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
21584 * lra-constraints.c (curr_insn_transform): Reuse original insn for
21585 secondary memory move.
21586 (inherit_reload_reg): Use rclass instead of cl for
21587 check_secondary_memory_needed_p.
21589 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21591 PR middle-end/56015
21592 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
21593 the case where writing real complex part of target modifies op1.
21595 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
21597 * config/aarch64/aarch64-simd.md
21598 (aarch64_vcond_internal<mode>): Handle unordered cases.
21599 * config/aarch64/iterators.md (v_cmp_result): New.
21601 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
21602 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21604 * config/arm/marvell-pj4.md: New file.
21605 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
21606 * config/arm/arm.md (generic_sched): Add marvell_pj4.
21607 (generic_vfp): Likewise.
21608 * config/arm/arm-cores.def: Add marvell-pj4.
21609 * config/arm/arm-tune.md: Regenerate.
21610 * config/arm/arm-tables.opt: Regenerate.
21611 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
21612 * doc/invoke.texi: Document marvell-pj4.
21614 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
21616 * config/aarch64/arm_neon.h: Map scalar types to standard types.
21618 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21623 * var-tracking.c (negative_power_of_two_p): New.
21624 (global_get_addr_cache, local_get_addr_cache): New.
21625 (get_addr_from_global_cache, get_addr_from_local_cache): New.
21626 (vt_canonicalize_addr): Rewrite using the above. Adjust the
21628 (vt_stack_offset_p): Remove.
21629 (vt_canon_true_dep): Always canonicalize loc's address.
21630 (clobber_overlapping_mems): Make sure we have a MEM.
21631 (local_get_addr_clear_given_value): New.
21632 (val_reset): Clear local cached entries.
21633 (compute_bb_dataflow): Create and release the local cache.
21634 Disable duplicate MEMs clobbering.
21635 (emit_notes_in_bb): Clobber MEMs likewise.
21636 (vt_emit_notes): Create and release the local cache.
21637 (vt_initialize, vt_finalize): Create and release the global
21638 cache, respectively.
21639 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
21641 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21643 PR libmudflap/53359
21644 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
21645 not found in the symtab.
21647 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21650 PR rtl-optimization/55547
21651 PR rtl-optimization/53827
21654 * alias.c (offset_overlap_p): New, factored out of...
21655 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
21656 the conservative special case for symbolic constants. Don't
21657 adjust zero sizes on alignment.
21659 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
21661 PR rtl-optimization/52573
21662 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
21663 REG_UNUSED for the same register.
21665 2013-01-17 Richard Biener <rguenther@suse.de>
21666 Marek Polacek <polacek@redhat.com>
21668 PR rtl-optimization/55833
21669 * loop-unswitch.c (unswitch_loops): Move loop verification...
21670 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
21671 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
21672 Set it to true when we're removing a loop from hierarchy tree in
21673 an irreducible region.
21674 (fix_bb_placements): Adjust caller.
21675 (fix_loop_placements): Likewise.
21677 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
21679 * config/avr/builtins.def (DEF_BUILTIN): Factor out
21680 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
21681 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
21682 Remove ID. Adjust comments.
21683 * config/avr/avr-c.c (avr_builtin_name): Remove.
21684 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
21685 * config/avr/avr.c (avr_tolower): New static function.
21686 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
21687 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
21688 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
21691 2013-01-17 Jan Hubicka <jh@suse.cz>
21693 PR tree-optimization/55273
21694 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
21696 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
21699 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
21700 store through atomic_store<mode>_1.
21701 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
21703 2013-01-17 Martin Jambor <mjambor@suse.cz>
21705 PR tree-optimizations/55264
21706 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
21707 for virtual methods.
21708 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
21709 virtual methods before inlining is over.
21710 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
21712 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
21715 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
21717 PR rtl-optimization/56005
21718 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
21719 pending reads for prefetch.
21721 2013-01-16 Ian Bolton <ian.bolton@arm.com>
21723 * config/aarch64/aarch64.md
21724 (*cstoresi_neg_uxtw): New pattern.
21725 (*cmovsi_insn_uxtw): New pattern.
21726 (*<optab>si3_uxtw): New pattern.
21727 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
21728 (*<optab>si3_insn_uxtw): New pattern.
21729 (*bswapsi2_uxtw): New pattern.
21731 2013-01-16 Richard Biener <rguenther@suse.de>
21733 * tree-inline.c (tree_function_versioning): Remove set but
21734 never used variable.
21736 2013-01-16 Richard Biener <rguenther@suse.de>
21738 PR tree-optimization/55964
21739 * tree-flow.h (rename_variables_in_loop): Remove.
21740 (rename_variables_in_bb): Likewise.
21741 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
21742 (copy_loop_before): Adjust and delete update-ssa status.
21743 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
21744 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
21745 (rename_variables_in_loop): Remove.
21746 (slpeel_update_phis_for_duplicate_loop): Likewise.
21747 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
21748 use available cfg machinery instead of duplicating it.
21749 Update PHI nodes and perform poor-mans SSA update here.
21750 (slpeel_tree_peel_loop_to_edge): Adjust.
21752 2013-01-16 Richard Biener <rguenther@suse.de>
21754 PR tree-optimization/54767
21755 PR tree-optimization/53465
21756 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
21757 (vrp_visit_phi_node): For PHI arguments coming via backedges
21758 drop all symbolical range information.
21759 (execute_vrp): Compute backedges.
21761 2013-01-16 Richard Biener <rguenther@suse.de>
21763 * doc/install.texi: Update CLooG and ISL requirements to
21766 2013-01-16 Christian Bruel <christian.bruel@st.com>
21769 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
21770 (broken_move): Handle UNSPECV_SP_SWITCH_B.
21771 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
21773 2013-01-16 DJ Delorie <dj@redhat.com>
21775 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
21776 (UNSPECV_SP_SWITCH_E): New.
21777 (sp_switch_1): Change to an unspec.
21778 (sp_switch_2): Change to an unspec. Don't use post-inc when we
21781 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
21783 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
21784 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
21785 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
21786 (expand_mem_thread_fence): Ditto.
21787 (expand_mem_signal_fence): Ditto.
21788 (expand_atomic_load): Ditto.
21789 (expand_atomic_store): Ditto.
21791 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
21793 PR rtl-optimization/55547
21794 PR rtl-optimization/53827
21797 * alias.c (memrefs_conflict_p): Set sizes to negative after
21800 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21803 * function.c (thread_prologue_and_epilogue_insns): Always
21804 add crtl->drap_reg to set_up_by_prologue.set, even if
21805 stack_realign_drap is false.
21807 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
21809 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
21810 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
21811 *call): Fix indention.
21813 2013-01-15 Tom de Vries <tom@codesourcery.com>
21816 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
21819 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
21821 PR rtl-optimization/55153
21822 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
21824 2013-01-15 Martin Jambor <mjambor@suse.cz>
21826 PR tree-optimization/55920
21827 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
21828 accesses as grp_to_be_debug_replaced.
21830 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21832 PR tree-optimization/55920
21833 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
21834 there is non-useless type conversion needed from debug rhs to lhs,
21835 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
21837 2013-01-15 Joseph Myers <joseph@codesourcery.com>
21838 Mikael Pettersson <mikpe@it.uu.se>
21841 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
21843 (ASM_OUTPUT_CASE_LABEL): Remove.
21844 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
21845 * final.c (shorten_branches): Update alignment of labels before
21846 jump tables if CASE_VECTOR_SHORTEN_MODE.
21848 2013-01-15 Richard Biener <rguenther@suse.de>
21851 * system.h: Do not include gmp.h for building host tools.
21853 2013-01-15 Richard Biener <rguenther@suse.de>
21855 PR middle-end/55882
21856 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
21857 account for bitpos when computing alignment.
21859 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
21861 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
21862 (ix86_target_macros_internal): Likewise.
21864 * config/i386/i386.c (m_CORE2I7): Removed.
21865 (m_CORE_HASWELL): New macro.
21866 (m_CORE_ALL): Likewise.
21867 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
21868 (initial_ix86_arch_features): Likewise.
21869 (processor_target_table): Initializations for Core avx2.
21870 (cpu_names): New names "core-avx2".
21871 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
21872 PROCESSOR_CORE_HASWELL.
21873 (ix86_issue_rate): New case.
21874 (ia32_multipass_dfa_lookahead): Likewise.
21875 (ix86_sched_init_global): Likewise.
21877 * config/i386/i386.h (TARGET_HASWELL): New macro.
21878 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
21879 (processor_type): New PROCESSOR_HASWELL.
21881 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21883 PR tree-optimization/55955
21884 * tree-vect-loop.c (vectorizable_reduction): Give up early on
21885 *SHIFT_EXPR and *ROTATE_EXPR codes.
21887 PR tree-optimization/48766
21888 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
21889 -ftrapv disable -fwrapv.
21891 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21894 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
21895 etc. to 1 and not to __flash.
21896 Use LL suffix for __INT24_MAX__ with -mint8.
21897 Use ULL suffix for __UINT24_MAX__ with -mint8.
21899 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21901 * config/avr/avr-arch.h
21902 (struct base_arch_s): Use typedef avr_arch_t instead.
21903 (struct arch_info_s): Use typedef avr_arch_info_t instead.
21904 (struct mcu_type_s): Use typedef avr_mcu_t instead.
21905 * config/avr/avr.c: Same.
21906 * config/avr/avr-devices.c: Same.
21907 * config/avr/driver-avr.c: Same.
21908 * config/avr/gen-avr-mmcu-texi.c: Same.
21909 * config/avr/avr-mcus.def: Adjust comment.
21911 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
21913 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
21914 * config/aarch64/iterators.md (VALLDI): New.
21916 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
21917 Andi Kleen <ak@linux.intel.com>
21920 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
21921 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
21924 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21926 * config/avr/avr-stdint.h: Remove trailing blanks.
21927 * config/avr/avr-log.h: Same.
21928 * config/avr/avr-arch.h: Same.
21929 * config/avr/avr-devices.c: Same.
21930 * config/avr/avr-dimode.md: Same.
21931 * config/avr/predicates.md: Same.
21932 * config/avr/avr-c.c: Same. And fix typo.
21934 * config/avr/avr-protos.h: Same. And:
21935 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
21936 (init_cumulative_args): Rename to avr_init_cumulative_args.
21937 (expand_prologue): Rename to avr_expand_prologue.
21938 (expand_epilogue): Rename to avr_expand_epilogue.
21939 (adjust_insn_length): Rename to avr_adjust_insn_length.
21940 (notice_update_cc): Rename to avr_notice_update_cc.
21941 (final_prescan_insn): Rename to avr_final_prescan_insn.
21942 * config/avr/avr.c: Same.
21943 * config/avr/avr.h: Same.
21944 * config/avr/avr.md: Remove trailing blanks.
21945 (prologue): Use avr_expand_prologue.
21946 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
21948 2013-01-14 Richard Biener <rguenther@suse.de>
21950 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
21951 verify_location, collect_subblocks): New functions.
21952 (verify_gimple_in_cfg): Verify that locations only reference
21953 BLOCKs in the functions BLOCK tree.
21955 2013-01-14 Richard Biener <rguenther@suse.de>
21957 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
21959 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
21961 (insert_out_of_ssa_copy_on_edge): Likewise.
21962 (rewrite_close_phi_out_of_ssa): Likewise.
21963 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
21965 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
21966 propagated constants.
21967 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
21970 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21972 * config/avr/avr-modes.def: Add GPL copyright notice.
21974 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
21976 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
21977 MEMMODEL_MASK to determine memory model.
21978 (atomic_store<mode>): Ditto from operands[2].
21979 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
21981 2013-01-13 Jakub Jelinek <jakub@redhat.com>
21984 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
21985 (fold_gimple_assign): Don't call unshare_expr here.
21986 (fold_ctor_reference): Call unshare_expr.
21988 2013-01-13 Terry Guo <terry.guo@arm.com>
21990 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
21991 * doc/fragments.texi: Document MULTILIB_REUSE.
21992 * gcc.c (multilib_reuse): New internal spec.
21993 (set_multilib_dir): Also search multilib from multilib_reuse.
21994 * genmultilib (tmpmultilib3): Refactor code.
21995 (tmpmultilib4): Ditto.
21996 (multilib_reuse): New multilib argument.
21998 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22000 * Makefile.in: Update copyright.
22002 2013-01-12 Tom de Vries <tom@codesourcery.com>
22004 PR middle-end/55890
22005 * calls.c (expand_call): Check if arg_nr is valid.
22007 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22009 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22010 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22011 documentation. Add missing '__' in front of
22012 __builtin_ia32_packssdw256.
22014 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22017 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22018 NO_REGS for larl operands.
22019 (s390_reload_larl_operand): Use s390_load_address instead of
22022 2013-01-11 Richard Biener <rguenther@suse.de>
22024 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22025 (verify_node_sharing): ... here.
22026 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22028 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22030 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22031 Substitute TREECHECKING.
22032 * configure: Regenerate.
22033 * Makefile.in (TREECHECKING): New.
22035 2013-01-11 Richard Guenther <rguenther@suse.de>
22037 PR tree-optimization/44061
22038 * tree-vrp.c (extract_range_basic): Compute zero as
22039 value-range for __builtin_constant_p of function parameters.
22041 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22043 Update copyright years.
22045 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22047 PR rtl-optimization/55672
22048 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22049 const to be eliminable.
22051 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22053 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22054 * configure: Regenerate.
22056 2013-01-10 Richard Biener <rguenther@suse.de>
22058 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22060 2013-01-10 Richard Biener <rguenther@suse.de>
22063 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22064 locations for virtual PHI arguments.
22065 (rewrite_update_phi_arguments): Likewise.
22067 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22069 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22072 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22074 PR tree-optimization/55921
22075 * tree-complex.c (expand_complex_asm): New function.
22076 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22078 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22081 * config/s390/s390.c (s390_symref_operand_p)
22082 (s390_loadrelative_operand_p): Merge the two functions.
22083 (s390_check_qrst_address, print_operand_address): Add parameters
22084 to s390_loadrelative_operand_p invokation.
22085 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22086 (s390_reload_larl_operand, s390_secondary_reload): Use
22087 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22088 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22090 2013-01-09 Mike Stump <mikestump@comcast.net>
22092 * dse.c (record_store): Remove unnecessary assert.
22094 2013-01-09 Jan Hubicka <jh@suse.cz>
22096 PR tree-optimization/55569
22097 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22098 * cfgloop.h (scale_loop_profile): Likewise.
22100 2013-01-09 Jan Hubicka <jh@suse.cz>
22103 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22105 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22107 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22109 PR middle-end/55114
22110 * expr.h (maybe_emit_group_store): Declare.
22111 * expr.c (maybe_emit_group_store): New function.
22112 * builtins.c (expand_builtin_int_roundingfn): Call it.
22113 (expand_builtin_int_roundingfn_2): Likewise.
22115 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22117 PR rtl-optimization/55829
22118 * lra-constraints.c (match_reload): Add code for absent output.
22119 (curr_insn_transform): Add code for reloads of matched inputs
22122 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22124 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22125 attribute of movddup insn to DF.
22126 (*vec_interleave_lowv2df): Ditto.
22127 (vec_dupv2df): Ditto.
22129 2013-01-09 Jan Hubicka <jh@suse.cz>
22131 PR tree-optimiation/55875
22132 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22133 EVERY_ITERATION parameter.
22134 (number_of_iterations_exit): Check if exit is executed every iteration.
22135 (idx_infer_loop_bounds): Similarly here.
22136 (n_of_executions_at_most): Simplify
22137 to only test for cases where statement is dominated by the
22138 particular bound; handle correctly the "postdominance" test.
22139 (scev_probably_wraps_p): Use max loop iterations info
22140 as a global bound first.
22142 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22143 Nick Clifton <nickc@redhat.com>
22145 * config/v850/v850.md (cbranchsf4): New pattern.
22146 (cstoresf4): New pattern.
22147 (cbranchdf4): New pattern.
22148 (cstoredf4): New pattern.
22149 (movsicc): Disallow floating point comparisons.
22150 (cmpsf_le_insn): Fix order of operators.
22151 (cmpsf_lt_insn): Likewise.
22152 (cmpsf_eq_insn): Likewise.
22153 (cmpdf_le_insn): Likewise.
22154 (cmpdf_lt_insn): Likewise.
22155 (cmpdf_eq_insn): Likewise.
22156 (cmpsf_ge_insn): Use LE comparison.
22157 (cmpdf_ge_insn): Likewise.
22158 (cmpsf_gt_insn): Use LT comparison.
22159 (cmpdf_gt_insn): Likewise.
22160 (cmpsf_ne_insn): Delete pattern.
22161 (cmpdf_ne_insn): Delete pattern.
22162 * config/v850/v850.c (v850_gen_float_compare): Use
22163 gen_cmpdf_eq_insn for NE comparison.
22164 (v850_float_z_comparison_operator)
22165 (v850_float_nz_comparison_operator): Move from here ...
22166 * config/v850/predicates.md: ... to here. Move GT and GE
22167 comparisons into v850_float_z_comparison_operator.
22168 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22170 (v850_float_nz_comparison_operator): Likewise.
22172 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22174 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22175 with calls to gen_insvsi/gen_insvdi.
22177 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22179 * config/i386/i386.c (initial_ix86_tune_features): Set up
22180 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22182 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22183 Jakub Jelinek <jakub@redhat.com>
22185 PR tree-optimization/48189
22186 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22187 If nitercst is 0, don't predict the exit edge.
22189 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22191 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22192 in asm_fprintf with reg_names.
22193 (aarch64_print_operand_address): Likewise.
22194 (aarch64_return_addr): Likewise.
22195 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22197 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22199 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22200 (INT_U6_BITS): Likewise.
22201 * config/pa/predicates.md (uint6_operand): New predicate.
22202 (shift5_operand, shift6_operand): Likewise.
22203 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22205 (lshrdi3): Use shift6_operand.
22206 (shrpsi4, shrpdi4): New insn patterns.
22207 (extzv): Delete expander.
22208 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22209 predicates in unamed zero extract patterns. Tighten common constraint.
22210 (extv): Delete expander.
22211 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22212 predicates in unamed sign extract patterns. Tighten common constraint.
22213 (insv): Delete expander.
22214 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22215 predicates in unamed insert patterns. Tighten common constraint.
22216 Change uint32_operand predicate to uint6_operand predicate in unamed
22217 DImode pattern to insert constant values of type 1...1xxxx.
22219 2013-01-04 Jan Hubicka <jh@suse.cz>
22221 PR tree-optimization/55823
22222 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22225 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22226 Uros Bizjak <ubizjak@gmail.com>
22228 PR rtl-optimization/55845
22229 * df-problems.c (can_move_insns_across): Stop scanning at
22230 volatile_insn_p source instruction or give up if
22231 across_from .. across_to range contains any volatile_insn_p
22234 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22236 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22237 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22239 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22240 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22242 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22245 * asan.c (asan_clear_shadow): New function.
22246 (asan_emit_stack_protection): Use it.
22248 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22250 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22251 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22252 with tab instead of space.
22254 2013-01-08 Nick Clifton <nickc@redhat.com>
22256 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22257 register bank 0 at the start of an interrupt handler.
22258 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22261 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22263 * config/aarch64/aarch64-simd.md
22264 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22265 (aarch64_simd_bsl): Likewise.
22266 (aarch64_vcond_internal<mode>): Likewise.
22267 (vcond<mode><mode>): Likewise.
22268 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22269 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22271 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22273 * config/aarch64/aarch64-builtins.c
22274 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22276 2013-01-08 Martin Jambor <mjambor@suse.cz>
22279 * tree-sra.c (analyze_access_subtree): Return true also after
22280 potentially creating a debug-only replacement.
22282 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22284 PR middle-end/55890
22285 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22287 PR tree-optimization/54120
22288 * tree-vrp.c (range_fits_type_p): Don't allow
22289 src_precision < precision from signed vr to unsigned_p
22290 if vr->min or vr->max is negative.
22291 (simplify_float_conversion_using_ranges): Test can_float_p
22292 against CODE_FOR_nothing.
22294 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22295 Richard Biener <rguenther@suse.de>
22297 PR middle-end/55851
22298 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22299 types instead of just INTEGER_TYPE types.
22301 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22303 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22306 2013-01-07 Steve Ellcey <sellcey@mips.com>
22309 * config/mips/mips.opt: Change mad to mmad to match documentation.
22311 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22314 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22315 .progmemx.data now.
22317 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22320 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22321 (avr_addrspace_t): Add .section_name field.
22322 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22324 (avr_addrspace): Same. Initialize .section_name. Remove last
22325 NULL entry. Put __memx into .progmemx.data.
22326 (progmem_section_prefix): Remove.
22327 (avr_asm_init_sections): No need to initialize progmem_section.
22328 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22329 section name prefix.
22330 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22331 retrieve the progmem section.
22332 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22333 boundary to run over avr_addrspace[].
22334 (avr_register_target_pragmas): Ditto.
22336 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22338 * varasm.c (output_constant_def_contents): For asan_protect_global
22339 protected strings, adjust DECL_ALIGN if needed, before testing for
22341 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22342 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22344 (output_object_block): For asan protected decls, emit asan padding
22345 after their contents.
22346 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22347 (asan_finish_file): Test it here instead.
22349 2013-01-07 Nick Clifton <nickc@redhat.com>
22350 Matthias Klose <doko@debian.org>
22351 Doug Kwan <dougkwan@google.com>
22352 H.J. Lu <hongjiu.lu@intel.com>
22355 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22357 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22359 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22361 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22363 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22365 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22368 * doc/install.texi (Cross-Compiler-Specific Options): Document
22371 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
22373 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
22374 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
22375 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
22376 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
22377 vqmovun_high_s64): Fix source operand number and update copyright.
22379 2013-01-07 Richard Biener <rguenther@suse.de>
22381 PR middle-end/55890
22382 * gimple.h (gimple_call_builtin_p): New overload.
22383 * gimple.c (validate_call): New function.
22384 (gimple_call_builtin_p): Likewise.
22385 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22386 Use gimple_call_builtin_p.
22387 (find_func_clobbers): Likewise.
22388 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
22389 (strlen_optimize_stmt): Likewise.
22391 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22393 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
22394 (vld1q_dup_*): Likewise.
22395 (vld1_*): Likewise.
22396 (vld1q_*): Likewise.
22397 (vld1_lane_*): Likewise.
22398 (vld1q_lane_*): Likewise.
22400 2013-01-07 Richard Biener <rguenther@suse.de>
22402 * lto-streamer.h (LTO_minor_version): Bump to 2.
22404 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22406 * config/aarch64/aarch64-protos.h
22407 (aarch64_const_double_zero_rtx_p): Rename to...
22408 (aarch64_float_const_zero_rtx_p): ...this.
22409 (aarch64_float_const_representable_p): New.
22410 (aarch64_output_simd_mov_immediate): Likewise.
22411 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
22412 move immediate case.
22413 * config/aarch64/aarch64.c
22414 (aarch64_const_double_zero_rtx_p): Rename to...
22415 (aarch64_float_const_zero_rtx_p): ...this.
22416 (aarch64_print_operand): Allow printing of new constants.
22417 (aarch64_valid_floating_const): New.
22418 (aarch64_legitimate_constant_p): Check for valid floating-point
22420 (aarch64_simd_valid_immediate): Likewise.
22421 (aarch64_vect_float_const_representable_p): New.
22422 (aarch64_float_const_representable_p): Likewise.
22423 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
22424 (aarch64_output_simd_mov_immediate): New.
22425 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
22426 (*movdf_aarch64): Likewise.
22427 * config/aarch64/constraints.md (Ufc): New.
22428 (Y): call aarch64_float_const_zero_rtx.
22429 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
22431 2013-01-07 Richard Biener <rguenther@suse.de>
22433 PR tree-optimization/55888
22434 PR tree-optimization/55862
22435 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
22436 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
22437 not if it is contained therein.
22439 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22441 * config/avr/t-avr: Typo.
22443 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22446 * config/avr/t-avr: Don't automatically rebuild
22447 $(srcdir)/config/avr/t-multilib
22448 $(srcdir)/config/avr/avr-tables.opt
22449 $(srcdir)/doc/avr-mmcu.texi
22450 (avr-mcus): New phony target to build them on request.
22451 (s-avr-mlib, s-avr-mmcu-texi): Remove.
22452 * avr/avr-mcus.def: Adjust comments.
22454 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
22456 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
22458 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22460 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
22462 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22464 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
22466 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
22468 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
22469 to generate profiling.
22470 * config/rs6000/aix64.h (LIB_SPEC): Same.
22472 2013-01-04 Andrew Pinski <apinski@cavium.com>
22474 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
22476 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
22478 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
22480 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
22482 (ix86_expand_move): Ditto.
22483 (ix86_zero_extend_to_Pmode): Ditto.
22484 (ix86_expand_call): Ditto.
22485 (ix86_expand_special_args_builtin): Ditto.
22486 (ix86_expand_builtin): Ditto.
22488 2013-01-04 Richard Biener <rguenther@suse.de>
22490 PR tree-optimization/55862
22491 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
22492 translating them through PHI nodes.
22494 2013-01-04 Martin Jambor <mjambor@suse.cz>
22496 PR tree-optimization/55755
22497 * tree-sra.c (sra_modify_assign): Do not check that an access has no
22498 children when trying to avoid producing a VIEW_CONVERT_EXPR.
22500 2013-01-04 Marek Polacek <polacek@redhat.com>
22502 PR middle-end/55859
22503 * opts.c (default_options_optimization): Clarify error message.
22505 2013-01-04 Richard Biener <rguenther@suse.de>
22507 PR middle-end/55863
22508 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
22511 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22514 * config/pa/pa.md (movsi): Revert previous change.
22515 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
22518 2013-01-03 Richard Henderson <rth@redhat.com>
22520 * config/i386/i386.c (ix86_expand_move): Always assign to op1
22521 after eliminating TLS symbols.
22523 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
22526 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
22527 * graphite-poly.c (debug_gmp_value): Likewise.
22529 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
22532 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
22533 selected code model, define __code_mode_small__, __code_model_medium__,
22534 __code_model_large__, __code_model_32__ or __code_model_kernel__.
22535 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
22536 xchg temporary register with %k. Declare temporary register as
22538 [__x86_64__]: For medium and large code models, preserve %rbx register.
22540 2013-01-03 Richard Biener <rguenther@suse.de>
22542 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
22543 (dump_subscript): Adjust.
22544 (finalize_ddr_dependent): Do not dump redundant info.
22545 (analyze_siv_subscript): Adjust.
22546 (subscript_dependence_tester): Likewise.
22547 (compute_affine_dependence): Likewise.
22549 2013-01-03 Richard Biener <rguenther@suse.de>
22552 2013-01-03 Richard Biener <rguenther@suse.de>
22554 PR tree-optimization/55857
22555 * tree-vect-stmts.c (vectorizable_load): Do not setup
22556 re-alignment for invariant loads.
22558 2013-01-02 Richard Biener <rguenther@suse.de>
22560 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22561 invariant load do not generate a vector load from the scalar location.
22563 2013-01-03 Richard Biener <rguenther@suse.de>
22565 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
22566 for not vectorizing.
22567 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
22568 not build INDIRECT_REFs, call get_name once only.
22569 (vect_create_data_ref_ptr): Likewise. Dump base object kind
22570 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
22572 2013-01-03 Richard Biener <rguenther@suse.de>
22574 PR tree-optimization/55857
22575 * tree-vect-stmts.c (vectorizable_load): Do not setup
22576 re-alignment for invariant loads.
22578 2013-01-03 Richard Biener <rguenther@suse.de>
22581 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
22582 prefer a built-in decl.
22584 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22586 * gcc.c (process_command): Update copyright notice dates.
22587 * gcov.c (print_version): Likewise.
22588 * gcov-dump.c (print_version): Likewise.
22590 PR rtl-optimization/55838
22591 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
22592 iv0.step, iv1.step and step.
22594 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22595 Marc Glisse <marc.glisse@inria.fr>
22597 PR tree-optimization/55832
22598 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
22599 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
22600 integer_{one,zero}_node.
22602 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22605 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
22606 * var-tracking.c (reverse_op): Don't add reverse ops to
22607 VALUEs that have already
22608 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
22610 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
22612 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
22614 2013-01-02 Teresa Johnson <tejohnson@google.com>
22616 * dumpfile.c (dump_loc): Print filename with location.
22617 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
22618 new location_t parameter to emit complete unroll message with
22619 new dump framework.
22620 (canonicalize_loop_induction_variables): Compute loops location
22621 and pass to try_unroll_loop_completely.
22622 * loop-unroll.c (report_unroll_peel): New function.
22623 (peel_loops_completely): Use new dump format with location
22624 for main dumpfile message, and invoke report_unroll_peel on success.
22625 (decide_unrolling_and_peeling): Ditto.
22626 (decide_peel_once_rolling): Remove old dumpfile message subsumed
22627 by report_unroll_peel.
22628 (decide_peel_completely): Ditto.
22629 (decide_unroll_constant_iterations): Ditto.
22630 (decide_unroll_runtime_iterations): Ditto.
22631 (decide_peel_simple): Ditto.
22632 (decide_unroll_stupid): Ditto.
22633 * cfgloop.c (get_loop_location): New function.
22634 * cfgloop.h (get_loop_location): Declare.
22636 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22638 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
22641 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22643 PR middle-end/55198
22644 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
22645 BLKmode objects when EXPAND_MEMORY is specified.
22647 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22649 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
22651 (fold_builtin_cpu): Do not share cpu model decls across statements.
22653 2013-01-02 Jason Merrill <jason@redhat.com>
22656 * tree.c (build_array_type_1): Revert earlier change.
22658 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
22660 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
22662 * config/aarch64/aarch64-tune.md: Re-generate.
22664 2013-01-02 Richard Biener <rguenther@suse.de>
22666 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22667 invariant load do not generate a vector load from the scalar location.
22669 2013-01-02 Richard Biener <rguenther@suse.de>
22672 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
22673 * configure: Regenerate.
22675 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22677 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
22678 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
22679 (expand_builtin_int_roundingfn_2): Keep the original target around
22680 for the fallback case.
22682 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22684 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
22685 to be clear for sign changes.
22687 2013-01-01 Jan Hubicka <jh@suse.cz>
22689 * ipa-inline-analysis.c: Fix formatting.
22691 2013-01-01 Jakub Jelinek <jakub@redhat.com>
22693 PR tree-optimization/55831
22694 * tree-vect-loop.c (get_initial_def_for_induction): Use
22695 gsi_after_labels instead of gsi_start_bb.
22697 Copyright (C) 2013 Free Software Foundation, Inc.
22699 Copying and distribution of this file, with or without modification,
22700 are permitted in any medium without royalty provided the copyright
22701 notice and this notice are preserved.