1 2013-09-24 Xinliang David Li <davidxl@google.com>
3 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
4 Check max peel iterations parameter.
5 * param.def: New parameter.
6 * doc/invoke.texi: Document New parameter.
8 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
10 * gimple-pretty-print.c: Various whitespace tweaks.
11 * tree-core.h: Likewise.
12 * tree-pretty-print.c: Likewise.
13 * tree-ssa-alias.c: Likewise.
14 * tree-ssa-copy.c: Likewise.
15 * tree-ssanames.c: Likewise.
16 * tree-ssanames.h: Likewise.
17 * tree-vrp.c: Likewise.
19 2013-09-24 Alan Modra <amodra@gmail.com>
23 * stmt.c (expand_asm_operands): Call expand_expr with
24 EXPAND_MEMORY for output operands that disallow regs. Don't
25 use EXPAND_WRITE on inout operands.
27 2013-09-24 Richard Biener <rguenther@suse.de>
30 * tree.c (reference_alias_ptr_type): Move ...
31 * alias.c (reference_alias_ptr_type): ... here and implement
32 in terms of the new reference_alias_ptr_type_1.
33 (ref_all_alias_ptr_type_p): New helper.
34 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
35 use ref_all_alias_ptr_type_p.
36 (get_deref_alias_set): Add flag_strict_aliasing check here.
37 (reference_alias_ptr_type_1): New function, split out from ...
38 (get_alias_set): ... here.
39 (alias_ptr_types_compatible_p): New function.
40 * alias.h (reference_alias_ptr_type): Declare.
41 (alias_ptr_types_compatible_p): Likewise.
42 * tree.h (reference_alias_ptr_type): Remove.
43 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
44 to compare MEM_REF alias types.
46 2013-09-24 Richard Biener <rguenther@suse.de>
48 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
50 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
52 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
53 reload helper function arrays into a single array reg_addr.
54 (reload_fpr_gpr): Likewise.
55 (reload_gpr_vsx): Likewise.
56 (reload_vsx_gpr): Likewise.
57 (struct rs6000_reg_addr): Likewise.
59 (rs6000_debug_reg_global): Change rs6000_vector_reload,
60 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
61 (rs6000_init_hard_regno_mode_ok): Likewise.
62 (rs6000_secondary_reload_direct_move): Likewise.
63 (rs6000_secondary_reload): Likewise.
65 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
66 constraints: wu, ww, and wy. Repurpose wv constraint added during
67 power8 changes. Put wg constraint in alphabetical order.
69 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
70 for future work to add ISA 2.07 VSX single precision support.
71 (-mvsx-scalar-double): Change default from -1 to 1, update
72 documentation comment.
73 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
74 (-mupper-regs-df): New debug switch to control whether DF values
75 can go in the traditional Altivec registers.
76 (-mupper-regs-sf): New debug switch to control whether SF values
77 can go in the traditional Altivec registers.
79 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
81 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
82 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
83 Add new constraints, wu/ww/wy. Repurpose wv constraint.
84 (rs6000_debug_legitimate_address_p): Print if we are running
85 before, during, or after reload.
86 (rs6000_secondary_reload): Add a comment.
87 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
89 * config/rs6000/constraints.md (wa constraint): Sort w<x>
90 constraints. Update documentation string.
91 (wd constraint): Likewise.
92 (wf constraint): Likewise.
93 (wg constraint): Likewise.
94 (wn constraint): Likewise.
95 (ws constraint): Likewise.
96 (wt constraint): Likewise.
97 (wx constraint): Likewise.
98 (wz constraint): Likewise.
99 (wu constraint): New constraint for ISA 2.07 SFmode scalar
101 (ww constraint): Likewise.
102 (wy constraint): Likewise.
103 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
104 the previous submissions.
105 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
107 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
109 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
110 (REG_BR_PROB): Say that the probability is stored in an int_list.
111 * reg-notes.def: Update commentary to mention INT_LIST.
112 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
114 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
115 * rtlanal.c (int_reg_note_p): New function.
116 (alloc_reg_note): Assert that the note does not have an int argument.
117 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
118 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
119 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
120 rather than an INSN_LIST. Handle INT_LIST.
121 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
122 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
123 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
124 Manipulate them as ints rather than rtxes.
125 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
126 * regmove.c (copy_src_to_dest): Likewise.
127 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
129 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
130 into the cases that need it.
131 * config/arm/arm.c (arm_unwind_emit): Likewise.
133 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
134 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
135 * loop-doloop.c (add_test, doloop_modify): Likewise.
136 * loop-unswitch.c (compare_and_jump_seq): Likewise.
137 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
138 * predict.c (combine_predictions_for_insn): Likewise.
139 * print-rtl.c (print_rtx): Handle INT_LIST.
140 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
141 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
142 * config/arm/arm.c (emit_unlikely_jump): Likewise.
143 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
144 (ix86_split_fp_branch, predict_jump): Likewise.
145 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
146 * config/sh/sh.c (expand_cbranchsi4): Likewise.
147 * config/spu/spu.c (ea_load_store_inline): Likewise.
149 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
150 value of a REG_BR_PROB note.
151 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
152 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
153 * emit-rtl.c (try_split): Likewise.
154 * predict.c (br_prob_note_reliable_p): Likewise.
155 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
156 * reorg.c (mostly_true_jump): Likewise.
157 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
158 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
159 * config/i386/i386.c (ix86_print_operand): Likewise.
160 * config/ia64/ia64.c (ia64_print_operand): Likewise.
161 * config/mmix/mmix.c (mmix_print_operand): Likewise.
162 * config/rs6000/rs6000.c (output_cbranch): Likewise.
163 * config/s390/s390.c (s390_expand_tbegin): Likewise.
164 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
165 * config/sparc/sparc.c (output_cbranch): Likewise.
166 * config/spu/spu.c (get_branch_target): Likewise.
167 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
168 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
170 2013-09-23 Jan Hubicka <jh@suse.cz>
172 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
174 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
175 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
177 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
178 (gimple_fold_call): Dump inconsistent devirtualizations; add
179 sanity check for type based devirtualizations.
180 * ipa-prop.c: Include ipa-utils.h
181 (ipa_intraprocedural_devirtualization): Add sanity check.
182 (try_make_edge_direct_virtual_call): Likewise.
184 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
186 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
187 assignment statements.
189 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
191 * gimple-pretty-print.c (dump_ssaname_info): New function.
192 (dump_gimple_phi): Call it.
193 (pp_gimple_stmt_1): Likewise.
194 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
195 (range_info_def): Declare.
196 * tree-pretty-print.c (pp_double_int): New function.
197 (dump_generic_node): Call it.
198 * tree-pretty-print.h (pp_double_int): Declare.
199 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
200 * tree-ssanames.h (range_info_def): New structure.
201 (value_range_type): Move definition here.
202 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
204 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
206 (set_range_info): New function.
207 (get_range_info): Likewise.
208 (duplicate_ssa_name_range_info): Likewise.
209 (duplicate_ssa_name_fn): Check pointer type and call
210 duplicate_ssa_name_range_info.
211 * tree-ssa-copy.c (fini_copy_prop): Likewise.
212 * tree-vrp.c (value_range_type): Remove definition, now in
214 (vrp_finalize): Call set_range_info to update value range of
216 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
217 (SSA_NAME_RANGE_INFO): New macro.
219 2013-09-23 Richard Biener <rguenther@suse.de>
221 PR tree-optimization/58464
222 * tree-ssa-pre.c (phi_trans_lookup): Remove.
223 (phi_trans_add): Change to add conditionally on being not
225 (phi_translate_1): Remove recursion detection here.
226 (phi_translate): Pre-seed the cache with NULL to catch
227 recursion here in a more generic way.
228 (bitmap_find_leader): Adjust comment.
229 (get_representative_for): Dump value-numbers.
230 (create_expression_by_pieces): Likewise.
231 (insert_into_preds_of_block): Likewise.
233 2013-09-23 Christian Bruel <christian.bruel@st.com>
236 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
237 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
239 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
242 2013-09-20 Renlin Li <renlin.li@arm.com>
244 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
245 (aarch64_expand_epilogue): Likewise.
246 (aarch64_legitimize_reload_address): Likewise.
248 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
250 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
251 NULL_TREE before pushing them onto the vector. Likewise for labels.
253 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
255 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
256 comparison operators when -fno-trapping-math is in effect.
257 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
258 comparison operators in TFmode and assert that unsupported operators
260 (ia64_print_operand): Likewise.
262 2013-09-21 Jan Hubicka <jh@suse.cz>
264 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
265 (sse_typeless_stores): Enable for core
266 (sse_load0_by_pxor): Likewise.
267 (four_jump_limit): Disable for core.
268 (pad_returns): Likewise.
269 (avoid_vector_decode): Likewise.
270 (fuse_cmp_and_branch): Enable for cores.
271 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
273 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
276 * config/pa/pa.c (pa_option_override): Disable auto increment and
277 decrement instructions until reload is completed.
279 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
280 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
281 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
283 2013-09-20 DJ Delorie <dj@redhat.com>
284 Nick Clifton <nickc@redhat.com>
286 * config/rl78/rl78.c: Various whitespace and comment tweaks.
287 (need_to_save): Save bank 0 on interrupts.
288 (characterize_address): Strip far address wrappers.
289 (rl78_as_legitimate_address): Likewise.
290 (transcode_memory_rtx): Likewise.
291 (rl78_peep_movhi_p): Disable this peephole after devirt.
292 (rl78_propogate_register_origins): Forget all origins when a
294 * config/rl78/rl78-virt.md: Various whitespace tweaks.
295 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
297 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
298 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
299 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
301 * config/rl78/constraints.md: Various whitespace and paren tweaks.
303 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
305 * config/pa/pa.md: In "scc" insn patterns, change output template to
306 handle const0_rtx in reg_or_0_operand operands.
308 2013-09-20 Martin Husemann <martin@NetBSD.org>
311 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
312 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
314 2013-09-20 Richard Biener <rguenther@suse.de>
317 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
318 remembering SSA name version and block index.
319 (new_scev_info_str): Adjust.
320 (hash_scev_info): Likewise. Also hash the block index.
321 (eq_scev_info): Adjust.
322 (find_var_scev_info): Likewise.
323 (struct instantiate_cache_entry): Remove.
324 (struct instantiate_cache_type): Use a htab to map name, block
326 (instantiate_cache_type::~instantiate_cache_type): Adjust.
327 (get_instantiated_value_entry): Likewise.
328 (hash_idx_scev_info, eq_idx_scev_info): New functions.
329 (instantiate_scev_name): Adjust.
331 2013-09-20 Jeff Law <law@redhat.com>
333 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
335 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
337 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
338 Call aarch64_simd_expand_args to update op[argc].
340 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
342 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
345 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
347 * gengtype.c (file_rules): Added rule for *.cc files.
348 (get_output_file_with_visibility): Give fatal message when no
351 2013-09-20 Renlin Li <renlin.li@arm.com>
353 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
354 (aarch64_expand_epilogue): Likewise.
355 (aarch64_legitimize_reload_address): Likewise.
357 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
360 * expr.c (expand_assignment): Remove misalignp code path.
362 2013-09-20 Marek Polacek <polacek@redhat.com>
365 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
366 TYPE_PRECISION. Add asserts.
368 2013-09-20 Richard Biener <rguenther@suse.de>
370 PR tree-optimization/58453
371 * tree-loop-distribution.c (distribute_loop): Apply the cost
372 model for -ftree-loop-distribute-patterns, too.
374 2013-09-20 Richard Biener <rguenther@suse.de>
377 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
378 make type comparison less strict.
380 2013-09-20 Alan Modra <amodra@gmail.com>
382 * configure: Regenerate.
383 * aclocal.m4: Regenerate.
385 2013-09-20 Marek Polacek <polacek@redhat.com>
388 * doc/extend.texi: Document that attribute used is meant to be used
389 on variables with static storage duration.
391 2013-09-19 Jakub Jelinek <jakub@redhat.com>
393 PR tree-optimization/58472
394 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
395 simd_lane_access set inv_p = false.
396 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
397 the simduid magic VAR_DECL.
399 2013-09-19 Jan Hubicka <jh@suse.cz>
401 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
403 2013-09-17 Jeff Law <law@redhat.com>
405 * tree-ssa-dom.c (record_temporary_equivalences): New function
406 split out of dom_opt_dom_walker::after_dom_children.
407 (dom_opt_dom_walker::thread_across_edge): Move common code
408 in here from dom_opt_dom_walker::after_dom_children.
409 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
411 2013-09-19 Jan Hubicka <jh@suse.cz>
413 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
414 (TARGET_GENERIC): Use PROCESOR_GENERIC
415 (enum processor_type): Unify generic32 and 64.
416 * i386.md (cpu): Likewise.
417 * x86-tune.def (use_leave): Enable for generic32.
418 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
419 * athlon.md: Change generic64 to generic in all occurences.
420 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
421 (ix86_target_macros_internal): Likewise.
422 * driver-i386.c (host_detect_local_cpu): Likewise.
423 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
424 (generic_memcpy, generic_memset, generic_cost): This one.
425 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
426 (m_GENERIC32, m_GENERIC64): Remove.
427 (m_GENERIC): Turn into one flag.
428 (processor_target): Unify generic tunnings.
429 (ix86_option_override_internal): Replace generic32/64 by generic.
430 (ix86_issue_rate): Likewise.
431 (ix86_adjust_cost): Likewise.
433 2013-09-19 Jan Hubicka <jh@suse.cz>
435 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
438 2013-09-19 Jakub Jelinek <jakub@redhat.com>
440 * omp-low.c (expand_omp_sections): Always pass len - 1 to
441 GOMP_sections_start, even if !exit_reachable.
443 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
445 * lra-constraints.c (need_for_all_save_p): Use macro
446 HARD_REGNO_CALL_PART_CLOBBERED.
447 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
448 set up pseudo conflict hard regs.
450 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
453 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
456 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
458 PR rtl-optimization/58438
459 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
460 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
461 from previous subpasses.
463 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
465 * arm.c (arm_get_frame_offsets): Validate architecture supports
466 LDRD/STRD before accepting the tuning preference.
467 (arm_expand_prologue): Likewise.
468 (arm_expand_epilogue): Likewise.
470 2013-09-18 Richard Biener <rguenther@suse.de>
472 PR tree-optimization/58417
473 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
474 have chrecs with symbols defined in the loop as operands.
475 (chrec_fold_multiply): Likewise.
476 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
477 parameters before folding binary operations.
478 (struct instantiate_cache_entry_hasher): Remove.
479 (struct instantiate_cache_type): Use a pointer-map.
480 (instantiate_cache_type::instantiate_cache_type): New function.
481 (instantiate_cache_type::get): Likewise.
482 (instantiate_cache_type::set): Likewise.
483 (instantiate_cache_type::~instantiate_cache_type): Adjust.
484 (get_instantiated_value_entry): Likewise.
485 (global_cache): New global.
486 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
487 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
488 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
489 (instantiate_scev_name): Adjust.
490 (instantiate_scev): Construct global instead of local cache.
491 (resolve_mixers): Likewise.
493 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
494 Paolo Carlini <paolo.carlini@oracle.com>
497 * doc/implement-cxx.texi: Fix references to the C++ standards.
499 2013-09-18 Jakub Jelinek <jakub@redhat.com>
501 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
502 * tree-vect-data-refs.c (vect_analyze_data_refs): For
503 simd_lane_access drs, update also DR_ALIGNED_TO.
505 2013-09-18 Marek Polacek <polacek@redhat.com>
508 * doc/extend.texi: Document no_sanitize_undefined attribute.
509 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
510 no_sanitize_undefined attribute.
512 2013-09-18 Nick Clifton <nickc@redhat.com>
514 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
515 (ASM_DECLARE_FUNCTION_NAME): Define.
517 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
519 * compare-elim.c (find_comparison_dom_walker): New class
520 (find_comparisons_in_bb): Rename to
521 find_comparison_dom_walker::before_dom_children
522 (find_comparisons): Adjust
523 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
525 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
526 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
527 (init_walk_dominator_tree): Remove declaration.
528 (fini_walk_dominator_tree): Remove declaration.
529 * fwprop.c (single_def_use_dom_walker): New class
530 (single_def_use_enter_block): Convert to
531 single_def_use_dom_walker::before_dom_children.
532 (single_def_use_leave_block): Convert to
533 single_def_use_dom_walker::after_dom_children.
534 (build_single_def_use_links): Adjust.
535 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
537 (find_candidates_in_block): Convert to
538 find_candidates_dom_walker::before_dom_children.
539 (execute_strength_reduction): Adjust.
540 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
541 (sese_dom_walker): New class.
542 (sese_dom_walker::sese_dom_walker): New constructor.
543 (sese_dom_walker::~sese_dom_walker): New destructor.
544 (build_sese_conditions_before): Convert to
545 sese_dom_walker::before_dom_children.
546 (build_sese_conditions_after): Convert to
547 sese_dom_walker::after_dom_children.
548 (build_poly_scop): Adjust
549 * tree-into-ssa.c (rewrite_dom_walker): New class
550 (rewrite_enter_block): Convert to
551 rewrite_dom_walker::before_dom_children.
552 (rewrite_leave_block): Convert to
553 rewrite_dom_walker::after_dom_children.
554 (rewrite_update_dom_walker): New class.
555 (rewrite_update_enter_block): Convert to
556 rewrite_update_dom_walker::before_dom_children.
557 (rewrite_update_leave_block): Convert to
558 rewrite_update_dom_walker::after_dom_children.
559 (rewrite_blocks, rewrite_into_ssa): Adjust.
560 (mark_def_dom_walker): New class.
561 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
562 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
563 (mark_def_sites_blocks): Convert to
564 mark_def_dom_walker::before_dom_children.
565 (mark_def_site_blocks): Remove.
566 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
567 (tree_ssa_dominator_optimize): Adjust.
568 (dom_thread_across_edge): Convert to method
569 dom_opt_dom_walker::thread_across_edge.
570 (dom_opt_enter_block): Convert to member function
571 dom_opt_dom_walker::before_dom_children.
572 (dom_opt_leave_block): Convert to member function
573 dom_opt_dom_walker::after_dom_children.
574 * tree-ssa-dse.c (dse_dom_walker): New class.
575 (dse_enter_block): Convert to member function
576 dse_dom_walker::before_dom_children.
577 (tree_ssa_dse): Adjust.
578 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
579 (determine_invariantness_stmt): Convert to method
580 invariantness_dom_walker::before_dom_children.
581 (determine_invariantness): Remove
582 (move_computations_dom_walker): New class.
583 (move_computations_stmt): Convert to method
584 move_computations_dom_walker::before_dom_children.
585 (move_computations, tree_ssa_lim): Adjust.
586 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
587 (nt_init_block): Make method
588 notrappping_dom_walker::before_dom_children.
589 (nt_fini_block): Make
590 method nontrapping_dom_walker::after_dom_children.
591 (get_non_trapping): Adjust.
592 * tree-ssa-pre.c (eliminate_dom_walker): New class.
593 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
594 (eliminate_leave_block): Make method.
595 eliminate_dom_walker::after_dom_children.
597 * tree-ssa-strlen.c (strlen_dom_walker): New class.
598 (strlen_enter_block): Make method
599 strlen_dom_walker::before_dom_children.
600 (strlen_leave_block): Make
601 method strlen_dom_walker::after_dom_children.
602 (tree_ssa_strlen): Adjust.
603 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
604 (tree_ssa_uncprop): Adjust.
605 (uncprop_leave_block): Make method
606 uncprop_dom_walker::after_dom_children.
607 (uncprop_leave_block): Make method
608 uncprop_dom_walker::before_dom_children.
610 2013-09-18 Bin Cheng <bin.cheng@arm.com>
612 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
613 before branch in same basic block. Check both src and dest of
616 2013-09-17 Nick Clifton <nickc@redhat.com>
618 * config/rl78/rl78-real.md (bf): New pattern.
620 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
621 (rl78_print_operand): Do not put a # before a %B.
622 * config/rl78/rl78.opt: Tweak doc strings.
624 2013-09-17 DJ Delorie <dj@redhat.com>
626 * config/rl78/constraints.md (Wcv): Allow up to $r31.
627 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
628 (rl78_option_override): Likewise, if -mallregs.
629 (is_virtual_register): Likewise.
630 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
631 (REGNO_OK_FOR_BASE_P): Likewise.
632 * config/rl78/rl78.opt (-mallregs): New.
634 2013-09-17 Nick Clifton <nickc@redhat.com>
636 * config/rl78/rl78.c (need_to_save): Change return type to bool.
637 For interrupt functions: save all call clobbered registers if the
638 interrupt handler is not a leaf function.
639 (rl78_expand_prologue): Always recompute the frame information.
640 For interrupt functions: only select bank 0 if one of the bank 0
641 registers is going to be psuhed.
643 2013-09-17 DJ Delorie <dj@redhat.com>
645 * config/rl78/constraints.md: For each W* constraint, rename to C*
646 and create a W* constraint that checks for an optional ES: prefix
648 * config/rl78/rl78.md (UNS_ES_ADDR): New.
649 (es_addr): New. Used to wrap far addresses.
650 * config/rl78/rl78-protos.h (rl78_es_addr): New.
652 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
653 wrapped far addresses.
654 (rl78_print_operand_1): Unwrap far addresses before processing.
655 (rl78_lo16): Wrap far addresses in unspecs.
658 (insn_ok_now): Check for not-yet-wrapped far addresses.
659 (transcode_memory_rtx): Properly re-wrap far addresses.
661 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
663 * config/sparc/t-rtems: Add leon3 multilibs.
665 2013-09-17 Cong Hou <congh@google.com>
667 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
668 when checking the dot production pattern. The type of rhs operand
669 of multiply is now checked correctly.
671 2013-09-17 Jeff Law <law@redhat.com>
673 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
674 edge implied equivalences into successor phis.
675 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
676 here from tree-ssa-threadedge.c.
677 (mark_threaded_blocks): When threading through a joiner, if both
678 successors of the joiner's clone reach the same block, verify the
679 PHI arguments are equal. If not, cancel the jump threading request.
680 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
681 tree-ssa-threadupdate.c
682 (thread_across_edge): Don't check PHI argument equality when
683 threading through joiner block here.
685 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
687 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
688 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
689 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
690 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
691 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
692 (has_undefined_value_p): New. Pass dependant parts of
693 ssa_undefined_value_p.
694 (uninit_undefined_value_p): Use has_undefined_value_p.
695 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
696 make_pass_early_warn_uninitialized): Move from tree-ssa.c
697 * tree-ssa.h: Adjust prototypes
699 2013-09-17 Jan Hubicka <jh@suse.cz>
702 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
703 * ipa-inline.c (can_inline_edge_p): Do not downgrade
704 FUNCTION_NOT_OPTIMIZED.
705 * ipa-inline-analysis.c (compute_inline_parameters): Function
706 not optimized is not inlinable unless it is alwaysinline.
707 (inline_analyze_function): Force calls in not optimized
708 function not inlinable.
710 2013-09-17 Jan Hubicka <jh@suse.cz>
713 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
715 * ipa.c (function_and_variable_visibility): Likewise.
716 * ipa-profile.c (ipa_profile): Likewise.
718 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
721 * cgraph.c (cgraph_function_body_availability): Check for ifunc
722 attribute, and don't inline the resolver in this case.
724 2013-09-17 Teresa Johnson <tejohnson@google.com>
726 * coverage.c (get_coverage_counts): Add missing newline.
728 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
730 PR tree-optimization/58088
731 * fold-const.c (mask_with_trailing_zeros): New function.
732 (fold_binary_loc): Make sure we don't recurse infinitely
733 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
734 Use mask_with_trailing_zeros where appropriate.
736 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
738 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
739 of 'prev' var to get better distance estimation.
741 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
743 * tree-inline.h (struct copy_body_data): Add transform_parameter.
744 * tree-inline.c (is_parameter_of): New predicate.
745 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
746 a parameter has been remapped.
747 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
748 (optimize_inline_calls): Initialize transform_parameter.
749 (copy_gimple_seq_and_replace_locals): Likewise.
750 (tree_function_versioning): Likewise.
751 (maybe_inline_call_in_expr): Likewise.
753 2013-09-17 Nick Clifton <nickc@redhat.com>
755 * config/msp430/msp430-protos.h: Add prototypes for new functions.
756 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
758 (is_attr_func): New function.
759 (msp430_is_interrupt_func): New function.
760 (is_naked_func): New function.
761 (is_reentrant_func): New function.
762 (is_critical_func): New function.
763 (msp430_start_function): Add annotations for function attributes.
764 (msp430_attr): New function.
765 (msp430_attribute_table): New.
766 (msp430_function_section): New function.
767 (TARGET_ASM_FUNCTION_SECTION): Define.
768 (msp430_builtin): New enum.
769 (msp430_init_builtins): New function.
770 (msp430_builtin_devl): New function.
771 (msp430_expand_builtin): New function.
772 (TARGET_INIT_BUILTINS): Define.
773 (TARGET_EXPAND_BUILTINS): Define.
774 (TARGET_BUILTIN_DECL): Define.
775 (msp430_expand_prologue): Add support for naked, interrupt,
776 critical and reentrant functions.
777 (msp430_expand_epilogue): Likewise.
778 (msp430_print_operand): Handle 'O' character.
779 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
781 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
782 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
783 (pushm): Use a 'n' rather than an 'i' constraint.
784 (msp_return): Add generation of the interrupt return instruction.
785 (disable_interrupts): New pattern.
786 (enable_interrupts): New pattern.
787 (push_intr_state): New pattern.
788 (pop_intr_state): New pattern.
789 (bic_SR): New pattern.
790 (bis_SR): New pattern.
791 * doc/extend.texi: Document MSP430 function attributes and builtin
794 2013-09-17 Richard Biener <rguenther@suse.de>
796 PR tree-optimization/58432
797 * tree-loop-distribution.c (tree_loop_distribution): Also
798 scan PHIs for outside loop uses and seed a partition from them.
800 2013-09-17 Bin Cheng <bin.cheng@arm.com>
802 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
803 (restructure_reference): Call backtrace_base_for_ref.
805 2013-09-17 Alan Modra <amodra@gmail.com>
808 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
810 2013-09-16 DJ Delorie <dj@redhat.com>
812 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
813 vregs location for RL78/G10.
814 (rl78_expand_prologue): Avoid SEL on G10.
815 (rl78_expand_epilogue): Likewise.
816 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
817 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
818 __RL78_G10__ when appropriate.
819 (ASM_SPEC): Pass -mg10 along to the assembler.
820 * config/rl78/rl78.md (sel_rb): Disable for G10.
821 * config/rl78/rl78.opt: Add -mg10 option.
822 * config/rl78/t-rl78: Add -mg10 multilib.
824 2013-09-16 Xinliang David Li <davidxl@google.com>
826 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
827 * omp-low.c (omp_max_vf): Ditto.
828 (expand_omp_simd): Ditto.
829 * tree-vectorizer.c (vectorize_loops): Ditto.
830 (gate_vect_slp): Ditto.
831 (gate_increase_alignment): Ditto.
832 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
833 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
834 (gate_tree_vectorize): Name change.
835 (tree_vectorize): Ditto.
836 (pass_vectorize::gate): Call new function.
837 (pass_vectorize::execute): Ditto.
838 opts.c: O3 default setting change.
839 (finish_options): Check new flag.
840 * doc/invoke.texi: Document new flags.
841 * common.opt: New flags.
843 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
845 * doc/tm.texi.in (Cond Exec Macros): Remove node.
846 (Condition Code): Don't reference it.
847 * doc/tm.texi: Regenerate.
849 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
852 * lra-constraints.c (undo_optional_reloads): Consider all optional
853 reload even if it did not get a hard reg.
855 2013-09-16 Teresa Johnson <tejohnson@google.com>
857 * dumpfile.c (dump_loc): Remove newline emission.
858 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
859 emission to dump_printf_loc calls where missing.
860 (vect_mark_for_runtime_alias_test): Ditto.
861 (vect_analyze_data_ref_dependence): Ditto.
862 (vect_analyze_data_ref_dependences): Ditto.
863 (vect_slp_analyze_data_ref_dependence): Ditto.
864 (vect_slp_analyze_data_ref_dependences): Ditto.
865 (vect_compute_data_ref_alignment): Ditto.
866 (vect_update_misalignment_for_peel): Ditto.
867 (vect_verify_datarefs_alignment): Ditto.
868 (vector_alignment_reachable_p): Ditto.
869 (vect_get_data_access_cost): Ditto.
870 (vect_enhance_data_refs_alignment): Ditto.
871 (vect_find_same_alignment_drs): Ditto.
872 (vect_analyze_data_refs_alignment): Ditto.
873 (vect_analyze_group_access): Ditto.
874 (vect_analyze_data_ref_access): Ditto.
875 (vect_analyze_data_ref_accesses): Ditto.
876 (vect_prune_runtime_alias_test_list): Ditto.
877 (vect_analyze_data_refs): Ditto.
878 (vect_create_addr_base_for_vector_ref): Ditto.
879 (vect_create_data_ref_ptr): Ditto.
880 (vect_grouped_store_supported): Ditto.
881 (vect_grouped_load_supported): Ditto.
882 * value-prof.c (check_counter): Ditto.
883 (check_ic_target): Ditto.
884 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
885 (vect_recog_widen_mult_pattern): Ditto.
886 (vect_recog_widen_sum_pattern): Ditto.
887 (vect_recog_over_widening_pattern): Ditto.
888 (vect_recog_widen_shift_pattern): Ditto.
889 (vect_recog_rotate_pattern): Ditto.
890 (vect_recog_vector_vector_shift_pattern): Ditto.
891 (vect_recog_divmod_pattern): Ditto.
892 (vect_recog_mixed_size_cond_pattern): Ditto.
893 (vect_recog_bool_pattern): Ditto.
894 (vect_pattern_recog_1): Ditto.
895 (vect_pattern_recog): Ditto.
896 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
897 (vect_is_simple_iv_evolution): Ditto.
898 (vect_analyze_scalar_cycles_1): Ditto.
899 (vect_get_loop_niters): Ditto.
900 (vect_analyze_loop_1): Ditto.
901 (vect_analyze_loop_form): Ditto.
902 (vect_analyze_loop_operations): Ditto.
903 (vect_analyze_loop_2): Ditto.
904 (vect_analyze_loop): Ditto.
905 (report_vect_op): Ditto.
906 (vect_is_slp_reduction): Ditto.
907 (vect_is_simple_reduction_1): Ditto.
908 (vect_get_known_peeling_cost): Ditto.
909 (vect_estimate_min_profitable_iters): Ditto.
910 (vect_model_reduction_cost): Ditto.
911 (vect_model_induction_cost): Ditto.
912 (get_initial_def_for_induction): Ditto.
913 (vect_create_epilog_for_reduction): Ditto.
914 (vectorizable_reduction): Ditto.
915 (vectorizable_induction): Ditto.
916 (vectorizable_live_operation): Ditto.
917 (vect_loop_kill_debug_uses): Ditto.
918 (vect_transform_loop): Ditto.
919 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
920 (vect_stmt_relevant_p): Ditto.
921 (process_use): Ditto.
922 (vect_mark_stmts_to_be_vectorized): Ditto.
923 (vect_model_simple_cost): Ditto.
924 (vect_model_promotion_demotion_cost): Ditto.
925 (vect_model_store_cost): Ditto.
926 (vect_get_store_cost): Ditto.
927 (vect_model_load_cost): Ditto.
928 (vect_get_load_cost): Ditto.
929 (vect_init_vector_1): Ditto.
930 (vect_get_vec_def_for_operand): Ditto.
931 (vect_finish_stmt_generation): Ditto.
932 (vectorizable_call): Ditto.
933 (vectorizable_conversion): Ditto.
934 (vectorizable_assignment): Ditto.
935 (vectorizable_shift): Ditto.
936 (vectorizable_operation): Ditto.
937 (vectorizable_store): Ditto.
938 (vectorizable_load): Ditto.
939 (vectorizable_condition): Ditto.
940 (vect_analyze_stmt): Ditto.
941 (vect_transform_stmt): Ditto.
942 (vect_is_simple_use): Ditto.
943 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
944 (vect_can_advance_ivs_p): Ditto.
945 (vect_update_ivs_after_vectorizer): Ditto.
946 (vect_do_peeling_for_loop_bound): Ditto.
947 (vect_gen_niters_for_prolog_loop): Ditto.
948 (vect_update_inits_of_drs): Ditto.
949 (vect_create_cond_for_alias_checks): Ditto.
950 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
951 (vect_build_slp_tree_1): Ditto.
952 (vect_supported_load_permutation_p): Ditto.
953 (vect_analyze_slp_instance): Ditto.
954 (vect_analyze_slp): Ditto.
955 (vect_make_slp_decision): Ditto.
956 (vect_detect_hybrid_slp): Ditto.
957 (vect_bb_vectorization_profitable_p): Ditto.
958 (vect_slp_analyze_bb_1): Ditto.
959 (vect_update_slp_costs_according_to_vf): Ditto.
960 (vect_get_mask_element): Ditto.
961 (vect_transform_slp_perm_load): Ditto.
962 (vect_schedule_slp_instance): Ditto.
963 (vect_schedule_slp): Ditto.
964 (vect_slp_transform_bb): Ditto.
965 * profile.c (read_profile_edge_counts): Ditto.
966 (compute_branch_probabilities): Ditto.
967 * coverage.c (get_coverage_counts): Ditto.
969 2013-09-16 Diego Novillo <dnovillo@google.com>
971 * tree-core.h: Add missing comment lines from refactoring
974 2013-09-16 Jan Hubicka <jh@suse.cz>
976 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
977 abstract functions; for static functions check the presence
980 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
982 * config/aarch64/aarch64-simd-builtins.def (fma): New.
983 * config/aarch64/aarch64-simd.md
984 (aarch64_mla_elt<mode>): New.
985 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
986 (aarch64_mls_elt<mode>): Likewise.
987 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
988 (aarch64_fma4_elt<mode>): Likewise.
989 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
990 (aarch64_fma4_elt_to_128v2df): Likewise.
991 (aarch64_fma4_elt_to_64df): Likewise.
992 (fnma<mode>4): Likewise.
993 (aarch64_fnma4_elt<mode>): Likewise.
994 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
995 (aarch64_fnma4_elt_to_128v2df): Likewise.
996 (aarch64_fnma4_elt_to_64df): Likewise.
997 * config/aarch64/iterators.md (VDQSF): New.
998 * config/aarch64/arm_neon.h
999 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1000 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1002 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1004 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1005 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1006 (aarch64_mul3_elt_to_128df): Likewise.
1007 (aarch64_mul3_elt_to_64v2df): Likewise.
1008 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1010 (VMUL_CHANGE_NLANES) Likewise.
1013 * config/aarch64/arm_neon.h
1014 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1016 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1018 * config/aarch64/arm_neon.h
1019 (vcvtx_high_f32_f64): Fix parameters.
1021 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1022 Uros Bizjak <ubizjak@gmail.com>
1024 * config/alpha.c: Include tree-ssanames.h.
1026 2013-09-16 Richard Biener <rguenther@suse.de>
1028 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1029 (dot_rdg_1): Handle control_dd.
1030 (create_edge_for_control_dependence): New function.
1031 (create_rdg_edges): Add control dependences if asked for.
1032 (build_rdg): Likewise.
1033 (generate_loops_for_partition): If there are not necessary
1034 control stmts remove all their dependencies.
1035 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1036 (distribute_loop): Pass on control dependences.
1037 (tree_loop_distribution): Compute control dependences and remove
1038 restriction on number of loop nodes.
1040 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1042 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1045 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1047 * cse.c (try_const_anchors): Punt on CC modes.
1049 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1051 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1053 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1056 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1059 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1062 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1064 (darwin_objc1_section): Likewise.
1065 (darwin_file_end): Emit Image Info section when required.
1067 2013-09-14 Jan Hubicka <jh@suse.cz>
1069 * tree-into-ssa.c (gate_into_ssa): New.
1070 (pass_data_build_ssa): Use it.
1071 * cgraph.h (expand_thunk): Update prototype.
1072 * cgraphunit.c (analyze_function): Expand thunks early.
1073 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1074 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1075 set lowered flag; do not add new function.
1076 (assemble_thunks_and_aliases): Update.
1077 * tree-ssa.c (gate_init_datastructures): New gate.
1078 (pass_data_init_datastructures): Use it.
1080 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1083 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1084 xmm register set described in the psABI.
1086 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1088 * dwarf2out.c (should_emit_struct_debug): Add check
1089 for type_decl variable is not NULL.
1091 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1093 * config.gcc: Use new winnt-c.c target hooks
1094 * config/t-winnt: New file
1095 * config/winnt-c.c: New file
1096 * doc/tm.texi.in: Document new hook
1097 * doc/tm.texi: Regenerated
1099 2013-09-13 Jan Hubicka <jh@suse.cz>
1102 * ipa-inline.c (check_callers): New function.
1103 (check_caller_edge): Remove.
1104 (want_inline_function_to_all_callers_p): Also permit alises that are
1106 (inline_to_all_callers): Terminate the walk when devirtualization
1107 introduce new calls.
1109 2013-09-13 Jan Hubicka <jh@suse.cz>
1111 * ipa-inline-analysis.c (struct growth_data): Add node.
1112 (do_estimate_growth_1): Fix detection of recursion.
1114 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1116 PR tree-optimization/58392
1117 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1118 to avoid shadowing of outer loop variable. If
1119 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1120 replace_by_duplicate_decl simduid of loops that have it set and
1121 set dest_cfun->has_simduid_loops and/or
1122 dest_cfun->has_force_vect_loops.
1123 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1124 instead of maybe_lookup_decl.
1125 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1126 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1127 call. Copy over force_vect and copy and remap simduid. Set
1128 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1129 (copy_cfg_body): Remove blocks_to_copy argument. Use
1130 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1131 caller. Don't set cfun->has_simduid_loops and/or
1132 cfun->has_force_vect_loops here.
1133 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1135 (expand_call_inline, tree_function_versioning): Adjust copy_body
1138 2013-09-13 Martin Jambor <mjambor@suse.cz>
1141 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1142 the assert if the edge was a speculative one.
1144 2013-09-13 Richard Biener <rguenther@suse.de>
1146 * tree-data-ref.h (known_dependences_p): Move here ...
1147 * tree-loop-distribution.c (known_dependences_p): ... from here.
1148 (dump_rdg_component, debug_rdg_component): Remove.
1150 (generate_loops_for_partition): Use gimple_uid instead of
1151 relying on matching stmt visit order.
1152 (rdg_build_partitions): Take starting stmt vector.
1153 (ldist_gen): Merge into ...
1154 (distribute_loop): ... this function. Do not compute starting
1156 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1158 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1160 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1161 Set type attribute correctly. Set predicable_short_it attribute.
1162 (cmpsi_shiftsi): Remove %? from output template.
1164 2013-09-13 Richard Biener <rguenther@suse.de>
1166 * tree-loop-distribution.c (struct rdg_component,
1167 rdg_defs_used_in_other_loops_p, free_rdg_components,
1168 rdg_build_components): Remove.
1169 (stmts_from_loop): Do not record virtual PHIs.
1170 (generate_loops_for_partition): Skip virtual PHIs.
1171 (build_rdg_partition_for_component): Rename to ...
1172 (build_rdg_partition_for_vertex): ... this and adjust.
1173 (rdg_build_partitions): Take a vector of starting vertices
1174 instead of components. Remove unnecessary leftover handling.
1175 (ldist_gen): Do not build components or record other stores.
1176 (distribute_loop): Do not distribute loops containing stmts
1179 2013-09-13 Christian Bruel <christian.bruel@st.com>
1182 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1184 2013-09-13 Kai Tietz <ktietz@redhat.com>
1186 * config.gcc: Separate cases for mingw and cygwin targets,
1187 and add 64-bit cygwin target case.
1189 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1190 dll-export inline-functions.
1191 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1194 2013-09-13 Jeff Law <law@redhat.com>
1198 2013-09-06 Jeff Law <law@redhat.com>
1200 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1201 edge implied equivalences into successor phis.
1203 2013-09-12 DJ Delorie <dj@redhat.com>
1205 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1208 2013-09-12 Brooks Moses <bmoses@google.com>
1211 * Makefile.in: Do not install driver binaries in $(target)/bin.
1213 2013-09-12 DJ Delorie <dj@redhat.com>
1215 * config/rl78/rl78.opt (mrelax): New.
1216 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1217 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1219 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1221 (rl78_expand_epilogue): Likewise.
1223 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1226 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1227 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1228 pointer elimination with using remove_reg_equal_offset_note.
1230 2013-09-12 DJ Delorie <dj@redhat.com>
1232 * config/msp430/: New port.
1233 * config.gcc (msp430): Added.
1234 * doc/invoke.texi: Document MSP430 options.
1235 * doc/install.texi: Document msp430-elf
1236 * doc/md.texi: Document msp430-elf
1237 * doc/contrib.texi: Document msp430-elf
1239 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1242 2013-09-12 Martin Jambor <mjambor@suse.cz>
1245 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1246 reference descriptor is NULL.
1247 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1250 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1252 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1253 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1254 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1255 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1256 make_temp_ssa_name): move to tree-ssanames.h
1257 * tree-ssa-alias.h: Move prototype.
1258 * tree-ssa.h: Include tree-ssanames.h.
1259 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1260 * tree-ssanames.h: New. Move items from tree-flow*.h
1261 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1263 2013-09-12 Richard Biener <rguenther@suse.de>
1265 PR tree-optimization/58404
1266 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1267 propagate non-invariant addresses into dereferences wrapped
1268 in component references.
1270 2013-09-12 Richard Biener <rguenther@suse.de>
1272 PR tree-optimization/58402
1273 * passes.def: Move pass_late_warn_uninitialized later.
1275 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1277 * tree-ssa.h: New. Move content from tree-flow.h and
1279 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1280 Move prototypes belonging to tree-ssa.c.
1281 * tree-flow-inline.h (redirect_edge_var_map_def,
1282 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1284 * gimple.h: Adjust prototypes.
1285 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1287 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1288 * tree.h: Move prototype to tree-ssa.h.
1289 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1290 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1291 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1292 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1293 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1294 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1295 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1296 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1297 graphite-blocking.c, graphite-clast-to-gimple.c,
1298 graphite-dependences.c, graphite-interchange.c,
1299 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1300 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1301 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1302 ipa-reference.c, ipa-split.c, ipa-utils.c,
1303 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1304 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1305 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1306 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1307 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1308 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1309 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1310 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1311 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1312 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1313 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1314 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1315 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1316 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1317 value-prof.c, var-tracking.c,
1318 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1320 2013-09-12 Richard Biener <rguenther@suse.de>
1322 PR tree-optimization/58396
1323 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1324 (build_rdg): Take a loop-nest parameter, fix memleaks.
1325 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1327 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1329 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1332 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1334 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1335 The FMA instruction names should have a 'v' prefix.
1337 2013-09-12 Richard Biener <rguenther@suse.de>
1339 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1340 (dot_rdg): Use popen instead of system in optional code.
1341 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1342 (already_processed_vertex_p): Adjust.
1343 (has_anti_or_output_dependence, predecessor_has_mem_write,
1344 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1345 rdg_flag_uses): Remove.
1346 (rdg_flag_vertex): Simplify.
1347 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1349 (build_rdg_partition_for_component): Process the first vertex
1350 of a component only.
1351 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1353 2013-09-12 Alan Modra <amodra@gmail.com>
1355 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1357 2013-09-11 DJ Delorie <dj@redhat.com>
1358 Nick Clifton <nickc@redhat.com>
1360 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1361 (rl78_stack_based_mem): New.
1362 * config/rl78/constraints.md (Iv08): New.
1371 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1373 (movsi): Change from expand to insn-and-split.
1374 (ashrsi3): Clobber AX.
1378 * config/rl78/rl78.md (CC_REG): Fix.
1379 (addsi3): Allow memory and immediate operands.
1380 (addsi3_internal): Split into...
1381 (addsi3_internal_virt): ...new, and ...
1382 (addsi3_internal_real): ...new.
1384 (subsi3_internal_virt): New.
1385 (subsi3_internal_real): New.
1386 (mulsi3): Add memory operand.
1387 (mulsi3_rl78): Likewise.
1388 (mulsi3_g13): Likewise.
1389 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1390 (cbranchqi4_real): Add more constraint options.
1391 (cbranchhi4_real): Expand pattern.
1392 (cbranchhi4_real_signed): New.
1393 (cbranchhi4_real_inverted): New.
1394 (cbranchsi4_real_lt): New.
1395 (cbranchsi4_real_ge): New.
1396 (cbranchsi4_real_signed): New.
1397 (cbranchsi4_real): New.
1399 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1401 (lshrsi3_virt): Likewise.
1402 (ashlsi3_virt): Likewise.
1403 (cbranchqi4_virt_signed): New.
1404 (cbranchhi4_virt_signed): New.
1405 (cbranchsi4_virt): New.
1406 * config/rl78/rl78.c: Whitespace fixes throughout.
1407 (move_elim_pass): New.
1408 (pass_data_rl78_move_elim): New.
1409 (pass_rl78_move_elim): New.
1410 (make_pass_rl78_move_elim): New.
1411 (rl78_devirt_info): Run devirt earlier.
1412 (rl78_move_elim_info): New.
1413 (rl78_asm_file_start): Register it.
1414 (rl78_split_movsi): New.
1415 (rl78_as_legitimate_address): Allow virtual base registers when
1417 (rl78_addr_space_convert): Remove spurious debug stuff.
1418 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1419 (rl78_print_operand): More cases for not printing '#'.
1420 (rl78_expand_compare): Remove most of the logic.
1421 (content_memory): New.
1422 (clear_content_memory): New.
1423 (get_content_index): New.
1424 (get_content_name): New.
1425 (display_content_memory): New.
1426 (update_content): New.
1427 (record_content): New.
1428 (already_contains): New.
1429 (insn_ok_now): Re-recog insns with virtual registers.
1430 (add_postponed_content_update): New.
1431 (process_postponed_content_update): New.
1432 (gen_and_emit_move): New.
1433 (transcode_memory_rtx): Record new location content.
1434 Use gen_and_emit_move.
1435 (force_into_acc): New.
1436 (move_to_acc): Use gen_and_emit_move.
1437 (move_from_acc): Likewise.
1438 (move_acc_to_reg): Likewise.
1439 (move_to_x): Likewise.
1440 (move_to_hl): Likewise.
1441 (move_to_de): Likewise.
1442 (rl78_alloc_physical_registers_op1): Record location content.
1443 (has_constraint): New.
1444 (rl78_alloc_physical_registers_op2): Record location content.
1446 (rl78_alloc_physical_registers_ro1): Likewise.
1447 (rl78_alloc_physical_registers_cmp): Likewise.
1448 (rl78_alloc_physical_registers_umul): Likewise.
1449 (rl78_alloc_address_registers_macax): New.
1450 (rl78_alloc_physical_registers): Initialize and set location
1451 content memory as needed.
1452 (rl78_reorg): Make sure split2 is called.
1453 (rl78_rtx_costs): New.
1455 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1457 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1458 for (not (neg ...)) and (neg (not ...)) cases.
1460 2013-09-11 Richard Biener <rguenther@suse.de>
1463 * passes.def: Split critical edges before late uninit warning passes.
1464 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1466 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1468 PR tree-optimization/58385
1469 * fold-const.c (build_range_check): If both low and high are NULL,
1470 use omit_one_operand_loc to preserve exp side-effects.
1472 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1474 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1476 2013-09-11 Richard Biener <rguenther@suse.de>
1478 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1479 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1480 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1481 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1482 stmts_from_loop, known_dependences_p, build_empty_rdg,
1483 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1484 * tree-loop-distribution.c: ... here.
1485 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1486 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1487 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1488 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1489 * tree-loop-distribution.c: ... here.
1490 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1491 (struct partition_s): Add loops member.
1492 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1493 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1494 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1496 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1497 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1498 Sergey Lega <sergey.s.lega@intel.com>
1499 Anna Tikhonova <anna.tikhonova@intel.com>
1500 Ilya Tocar <ilya.tocar@intel.com>
1501 Andrey Turetskiy <andrey.turetskiy@intel.com>
1502 Ilya Verbin <ilya.verbin@intel.com>
1503 Kirill Yukhin <kirill.yukhin@intel.com>
1504 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1506 * config/i386/constraints.md (k): New.
1508 * config/i386/i386.c (const regclass_map): Add new mask registers.
1509 (dbx_register_map): Ditto.
1510 (dbx64_register_map): Ditto.
1511 (svr4_dbx_register_map): Ditto.
1512 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1514 (ix86_preferred_reload_class): Disable constants for mask registers.
1515 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1516 (ix86_hard_regno_mode_ok): Support new mask registers.
1517 (x86_order_regs_for_local_alloc): Ditto.
1518 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1519 (FIXED_REGISTERS): Add new mask registers.
1520 (CALL_USED_REGISTERS): Ditto.
1521 (REG_ALLOC_ORDER): Ditto.
1522 (VALID_MASK_REG_MODE): New.
1523 (FIRST_MASK_REG): Ditto.
1524 (LAST_MASK_REG): Ditto.
1525 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1526 (MAYBE_MASK_CLASS_P): New.
1527 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1528 (REG_CLASS_CONTENTS): Ditto.
1529 (MASK_REGNO_P): New.
1530 (ANY_MASK_REG_P): Ditto.
1531 (HI_REGISTER_NAMES): Add new mask registers.
1532 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1533 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1535 (attribute "type"): Add mskmov, msklog.
1536 (attribute "length_immediate"): Support them.
1537 (attribute "memory"): Ditto.
1538 (attribute "prefix_0f"): Ditto.
1539 (*movhi_internal): Support new mask registers.
1540 (*movqi_internal): Ditto.
1541 (define_split): Split out clobber pattern is a logic
1542 insn on mask registers.
1543 (*k<logic><mode>): New.
1544 (*andhi_1): Extend to support mask regs.
1545 (*andqi_1): Extend to support mask regs.
1547 (define_split): Split and-not to and and not if operands
1549 (*<code><mode>_1): Separate HI mode to new pattern...
1550 (*<code>hi_1): This.
1551 (*<code>qi_1): Extend to support mask regs.
1553 (kortestzhi): Ditto.
1554 (kortestchi): Ditto.
1556 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1557 (*one_cmplhi2_1): ...Here, now with mask registers support.
1558 (*one_cmplqi2_1): Support new mask registers.
1559 (HI/QImode arithmetics splitter): Don't split if mask registers
1561 (HI/QImode not splitter): Ditto.
1562 * config/i386/predicated.md (mask_reg_operand): New.
1563 (general_reg_operand): Ditto.
1565 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1567 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1568 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1570 2013-09-10 Jeff Law <law@redhat.com>
1572 PR tree-optimization/58380
1573 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1574 of threading through a buried loop header.
1576 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1577 return value for single successor case.
1579 2013-09-10 Jan Hubicka <jh@suse.cz>
1581 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1583 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1586 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1587 support conditional execution.
1588 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
1590 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
1592 * lra.c (lra): Clear lra_optional_reload_pseudos before every
1594 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
1595 Check destination too to check move insn.
1596 (undo_optional_reloads): Add check that the original peudo did not
1597 changed its allocation and the optional reload was inherited on last
1598 inheritance pass. Break loop after deciding to keep optional reload.
1599 (lra_undo_inheritance): Add check that inherited pseudo still in
1602 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
1604 * config/aarch64/aarch64.md (generic_sched): New.
1605 * config/aarch64/aarch64-generic.md (load): Make conditional
1606 on generic_sched attribute.
1607 (nonload): Likewise.
1609 2013-09-10 Jan Hubicka <jh@suse.cz>
1611 * lto-cgraph.c: Include ipa-utils.h.
1612 (compute_ltrans_boundary): Also add possible targets into the boundary.
1614 2013-09-10 Jan Hubicka <jh@suse.cz>
1616 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
1617 VAR_DECL of vtable rather than full expression.
1619 2013-09-10 Jan Hubicka <jh@suse.cz>
1620 Paolo Carlini <paolo.carlini@oracle.com>
1622 * cgraphunit.c (analyze_functions): Save input_location, set it
1623 to UNKNOWN_LOCATION and restore it at the end.
1625 2013-09-10 Martin Jambor <mjambor@suse.cz>
1627 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
1628 represented by a thunk.
1630 2013-09-10 Jeff Law <law@redhat.com>
1632 PR tree-optimization/58343
1633 * tree-ssa-threadupdate.c (thread_block): Identify and disable
1634 jump threading requests through loop headers buried in the middle
1635 of a jump threading path.
1637 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
1638 in return value/type.
1640 2013-09-10 Jakub Jelinek <jakub@redhat.com>
1642 PR rtl-optimization/58365
1643 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
1644 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
1647 2013-09-10 Richard Biener <rguenther@suse.de>
1649 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
1650 * tree-data-ref.c (create_rdg_vertices): Collect all data
1651 references, signal failure to the caller, use data-ref API.
1652 (build_rdg): Compute data references only once. Maintain lifetime
1653 of data references and data dependences from within RDG.
1654 (free_rdg): Free dependence relations.
1655 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
1656 inventing extra dependences.
1657 (distribute_loop): Update for RDG API changes.
1659 2013-09-10 Kai Tietz <ktietz@redhat.com>
1661 * doc/invoke.texi (fms-extensions): Document changed
1662 behavior for ms-abi targets.
1663 * config/i386/i386.c (ix86_option_override_internal):
1664 Set default value of option -fms-extension for ms-abi targets.
1666 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1668 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
1670 2013-09-10 Alan Modra <amodra@gmail.com>
1673 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
1675 2013-09-10 Alan Modra <amodra@gmail.com>
1677 * config/rs6000/predicates.md (add_cint_operand): New.
1678 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
1679 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
1680 using add_cint_operand.
1681 (largetoc_high_plus_aix): Likewise.
1683 2013-09-09 Jakub Jelinek <jakub@redhat.com>
1685 PR tree-optimization/58364
1686 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
1687 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
1688 the current range can't be an unconditional true or false.
1690 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
1692 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
1694 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
1696 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
1698 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
1701 * doc/invoke.texi (-Wdelete-incomplete): Document it.
1703 2013-09-09 Ian Bolton <ian.bolton@arm.com>
1705 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
1706 NO_REGS for immediate that can't be moved directly into FP_REGS.
1708 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1710 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
1711 comparison with negated operand.
1712 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
1715 2013-09-09 Richard Biener <rguenther@suse.de>
1718 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
1719 of a subloop record all its block as affecting loop-closed SSA form.
1721 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1723 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
1724 of an rtx/bitpos pair.
1725 (store_fixed_bit_field): Update accordingly.
1727 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1729 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
1731 * builtins.c (expand_errno_check): Likewise.
1732 * dwarf2cfi.c (init_return_column_size): Likewise.
1733 * except.c (sjlj_mark_call_sites): Likewise.
1734 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
1735 * lra-constraints.c (emit_inc): Likewise.
1736 * ree.c (combine_set_extension): Likewise.
1737 * regmove.c (fixup_match_2): Likewise.
1738 * reload1.c (inc_for_reload): Likewise.
1740 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1742 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
1743 (force_to_mode, simplify_shift_const_1, simplify_comparison):
1744 Use gen_int_mode with the mode of the associated simplify_* call.
1745 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
1746 * expmed.c (expand_shift_1): Likewise.
1747 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1748 * loop-iv.c (iv_number_of_iterations): Likewise.
1749 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1750 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
1752 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1754 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
1755 of the associated expand_* call.
1756 (asan_emit_stack_protection): Likewise.
1757 * builtins.c (round_trampoline_addr): Likewise.
1758 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
1759 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
1760 (emit_store_flag): Likewise.
1761 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
1762 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
1764 * function.c (instantiate_virtual_regs_in_insn): Likewise.
1765 * ifcvt.c (noce_try_store_flag_constants): Likewise.
1766 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
1767 * modulo-sched.c (generate_prolog_epilog): Likewise.
1768 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
1769 (expand_ctz, expand_ffs, expand_unop): Likewise.
1771 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1773 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
1774 of the associated gen_rtx_* call.
1775 * caller-save.c (init_caller_save): Likewise.
1776 * combine.c (find_split_point, make_extraction): Likewise.
1777 (make_compound_operation): Likewise.
1778 * dwarf2out.c (mem_loc_descriptor): Likewise.
1779 * explow.c (plus_constant, probe_stack_range): Likewise.
1780 * expmed.c (expand_mult_const): Likewise.
1781 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
1782 * reload1.c (init_reload): Likewise.
1783 * valtrack.c (cleanup_auto_inc_dec): Likewise.
1784 * var-tracking.c (adjust_mems): Likewise.
1785 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
1786 rather than gen_rtx_fmt_ee.
1788 2013-09-09 Jan Hubicka <jh@suse.cz>
1791 * value-prof.c (gimple_ic): Copy also abnormal edges.
1793 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1795 * asan.c (asan_shadow_cst): Use gen_int_mode.
1797 2013-09-08 Jan Hubicka <jh@suse.cz>
1799 * ipa-profile.c: Add toplevel comment.
1800 (ipa_propagate_frequency_1): Be more conservative when profile is read.
1801 (contains_hot_call_p): New function.
1802 (ipa_propagate_frequency): Set frequencies based on counts when
1804 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
1805 profile; do not tamper with profile after inlining if it is read.
1807 2013-09-08 Jan Hubicka <jh@suse.cz>
1809 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
1812 2013-09-08 Jan Hubicka <jh@suse.cz>
1814 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
1817 2013-09-08 Jeff Law <law@redhat.com>
1820 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
1823 2013-09-08 Andi Kleen <ak@linux.intel.com>
1825 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
1827 2013-09-08 Jan Hubicka <jh@suse.cz>
1829 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
1831 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
1833 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
1834 for non-debug insns.
1835 * lra.c (new_insn_reg): Take the containing insn as a parameter.
1836 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
1837 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
1840 2013-09-08 Jan Hubicka <jh@suse.cz>
1842 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
1843 targets and devirtualize to BUILT_IN_UNREACHABLE.
1844 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
1845 * ipa.c (walk_polymorphic_call_targets): New function.
1846 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
1847 functions; use the new timevar.
1848 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
1849 was removed from the program.
1850 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
1851 not consider it in the walk when its vtable is dead.
1852 (possible_polymorphic_call_targets_1): Pass anonymous flag to
1854 (devirt_variable_node_removal_hook): New function.
1855 (possible_polymorphic_call_targets): Also register
1856 devirt_variable_node_removal_hook.
1857 (ipa_devirt): Do not do non-speculative devirtualization.
1858 (gate_ipa_devirt): One execute if devirtualizing speculatively.
1860 2013-09-08 Jan Hubicka <jh@suse.cz>
1862 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
1863 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
1864 varpool_remove_variable_insertion_hook): Declare.
1865 * varpool.c (varpool_node_hook_list): New structure.
1866 (first_varpool_node_removal_hook,
1867 first_varpool_variable_insertion_hook): New variables.
1868 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
1869 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
1870 varpool_remove_variable_insertion_hook,
1871 varpool_call_variable_insertion_hooks): New functions.
1872 (varpool_remove_node): Use it.
1874 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
1877 * diagnostic.c (diagnostic_build_prefix): When s.file is
1878 "<built-in>" don't output line and column numbers.
1880 2013-09-06 Jan Hubicka <jh@suse.cz>
1882 * cgraphunit.c (expand_thunk): Get body before touching arguments.
1883 * lto-streamer-out.c: Stream thunks, too.
1884 * lto-streamer-in.c (input_function): Pop cfun here
1885 (lto_read_body): Instead of here.
1887 2013-09-06 Caroline Tice <cmtice@google.com>
1889 * doc/install.texi: Add documentation for the --enable-vtable-verify
1890 and the --disable-libvtv configure options.
1892 2013-09-06 Jeff Law <law@redhat.com>
1894 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1895 edge implied equivalences into successor phis.
1897 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
1899 * resource.c (mark_referenced_resources): Handle COND_EXEC.
1901 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
1903 * resource.c (mark_target_live_regs): Compute resources taking
1904 into account if a call is predicated or not.
1906 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1908 * toplev.c (output_stack_usage): Be prepared for suffixes created by
1909 the compiler in the function names.
1911 2013-09-06 Jan Hubicka <jh@suse.cz>
1914 * ipa-inline.c (has_caller_p): New function.
1915 (want_inline_function_to_all_callers_p): Use it.
1916 (sum_callers, inline_to_all_callers): Break out from ...
1917 (ipa_inline): ... here.
1919 2013-09-06 Jan Hubicka <jh@suse.cz>
1921 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
1922 only when AVX is enabled.
1924 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1926 * config/aarch64/aarch64.md
1927 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
1928 is fpsimd_<load/store>2.
1929 (load_pair<mode>): Likewise.
1930 (store_pair<mode>): Likewise.
1932 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1934 * config/arm/types.md (type): Add "mrs" type.
1935 * config/aarch64/aarch64.md
1936 (aarch64_load_tp_hard): Make type "mrs".
1938 (load_tp_hard): Make type "mrs".
1939 * config/arm/cortex-a15.md: Update with new attributes.
1940 * config/arm/cortex-a5.md: Update with new attributes.
1941 * config/arm/cortex-a53.md: Update with new attributes.
1942 * config/arm/cortex-a7.md: Update with new attributes.
1943 * config/arm/cortex-a8.md: Update with new attributes.
1944 * config/arm/cortex-a9.md: Update with new attributes.
1945 * config/arm/cortex-m4.md: Update with new attributes.
1946 * config/arm/cortex-r4.md: Update with new attributes.
1947 * config/arm/fa526.md: Update with new attributes.
1948 * config/arm/fa606te.md: Update with new attributes.
1949 * config/arm/fa626te.md: Update with new attributes.
1950 * config/arm/fa726te.md: Update with new attributes.
1952 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1954 * config/aarch64/aarch64.md
1955 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
1956 (*movtf_aarch64): Likewise.
1958 (thumb1_movdi_insn): Use "multiple" for type where more than one
1959 instruction is used for a move.
1960 (*arm32_movhf): Likewise.
1961 (*thumb_movdf_insn): Likewise.
1963 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1965 * config/arm/types.md (type): Rename fcpys to fmov.
1967 (*arm_movsi_vfp): Rename type fcpys as fmov.
1968 (*thumb2_movsi_vfp): Likewise
1969 (*movhf_vfp_neon): Likewise
1970 (*movhf_vfp): Likewise
1971 (*movsf_vfp): Likewise
1972 (*thumb2_movsf_vfp): Likewise
1973 (*movsfcc_vfp): Likewise
1974 (*thumb2_movsfcc_vfp): Likewise
1975 * config/aarch64/aarch64-simd.md
1976 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
1977 * config/aarch64/aarch64.md
1978 (*movsi_aarch64): Replace type mov_reg with fmovs.
1979 (*movdi_aarch64): Likewise
1980 (*movsf_aarch64): Likewise
1981 (*movdf_aarch64): Likewise
1983 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
1984 * config/arm/iwmmxt.md
1985 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
1986 * config/arm/arm1020e.md: Update with new attributes.
1987 * config/arm/cortex-a15-neon.md: Update with new attributes.
1988 * config/arm/cortex-a5.md: Update with new attributes.
1989 * config/arm/cortex-a53.md: Update with new attributes.
1990 * config/arm/cortex-a7.md: Update with new attributes.
1991 * config/arm/cortex-a8-neon.md: Update with new attributes.
1992 * config/arm/cortex-a9.md: Update with new attributes.
1993 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1994 * config/arm/cortex-r4f.md: Update with new attributes.
1995 * config/arm/marvell-pj4.md: Update with new attributes.
1996 * config/arm/vfp11.md: Update with new attributes.
1998 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2000 * config/aarch64/aarch64.md
2001 (*madd<mode>): Fix type attribute.
2002 (*maddsi_uxtw): Likewise.
2003 (*msub<mode>): Likewise.
2004 (*msubsi_uxtw): Likewise.
2005 (<su_optab>maddsidi4): Likewise.
2006 (<su_optab>msubsidi4): Likewise.
2008 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2010 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2011 * config/arm/arm.md (core_cycles): Remove fdiv.
2012 * config/arm/vfp.md:
2013 (*sqrtsf2_vfp): Update for attribute changes.
2014 (*sqrtdf2_vfp): Likewise.
2015 * config/aarch64/aarch64.md:
2016 (sqrt<mode>2): Update for attribute changes.
2017 * config/arm/arm1020e.md: Update with new attributes.
2018 * config/arm/cortex-a15-neon.md: Update with new attributes.
2019 * config/arm/cortex-a5.md: Update with new attributes.
2020 * config/arm/cortex-a53.md: Update with new attributes.
2021 * config/arm/cortex-a7.md: Update with new attributes.
2022 * config/arm/cortex-a8-neon.md: Update with new attributes.
2023 * config/arm/cortex-a9.md: Update with new attributes.
2024 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2025 * config/arm/cortex-r4f.md: Update with new attributes.
2026 * config/arm/marvell-pj4.md: Update with new attributes.
2027 * config/arm/vfp11.md: Update with new attributes.
2029 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2031 * config/arm/types.md
2032 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2033 * config/aarch64/aarch64.md
2034 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2036 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2037 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2038 (float<GPI:mode><GPF:mode>2): Likewise.
2040 (*truncsisf2_vfp): Update with new attributes.
2041 (*truncsidf2_vfp): Likewise.
2042 (fixuns_truncsfsi2): Likewise.
2043 (fixuns_truncdfsi2): Likewise.
2044 (*floatsisf2_vfp): Likewise.
2045 (*floatsidf2_vfp): Likewise.
2046 (floatunssisf2): Likewise.
2047 (floatunssidf2): Likewise.
2048 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2049 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2050 * config/arm/arm1020e.md: Update with new attributes.
2051 * config/arm/cortex-a15-neon.md: Update with new attributes.
2052 * config/arm/cortex-a5.md: Update with new attributes.
2053 * config/arm/cortex-a53.md: Update with new attributes.
2054 * config/arm/cortex-a7.md: Update with new attributes.
2055 * config/arm/cortex-a8-neon.md: Update with new attributes.
2056 * config/arm/cortex-a9.md: Update with new attributes.
2057 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2058 * config/arm/cortex-r4f.md: Update with new attributes.
2059 * config/arm/marvell-pj4.md: Update with new attributes.
2060 * config/arm/vfp11.md: Update with new attributes.
2062 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2064 * config/aarch64/arm_neon.h
2065 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2066 (vqtbx<1,2,3,4><q>_s8): Likewise.
2068 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2070 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2071 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2072 (add<mode>3): Add type attribute.
2073 (add<mode>3): Likewise.
2074 (usadd<mode>3): Likewise.
2075 (ssadd<mode>3): Likewise.
2076 (sub<mode>3): Likewise.
2077 (sub<mode>3): Likewise.
2078 (ussub<mode>3): Likewise.
2079 (sssub<mode>3): Likewise.
2080 (ssmulsa3): Likewise.
2081 (usmulusa3): Likewise.
2082 (arm_usatsihi): Likewise.
2084 (*movdi_vfp): Add types for all instructions.
2085 (*movdi_vfp_cortexa8): Likewise.
2086 (*movhf_vfp_neon): Likewise.
2087 (*movhf_vfp): Likewise.
2088 (*movdf_vfp): Likewise.
2089 (*thumb2_movdf_vfp): Likewise.
2090 (*thumb2_movdfcc_vfp): Likewise.
2091 * config/arm/arm.md: Add type attribute to all insn patterns.
2092 (*thumb1_adddi3): Add type attribute.
2093 (*arm_adddi3): Likewise.
2094 (*adddi_sesidi_di): Likewise.
2095 (*adddi_zesidi_di): Likewise.
2096 (*thumb1_addsi3): Likewise.
2097 (addsi3_compare0): Likewise.
2098 (*addsi3_compare0_scratch): Likewise.
2099 (*compare_negsi_si): Likewise.
2100 (cmpsi2_addneg): Likewise.
2101 (*addsi3_carryin_<optab>): Likewise.
2102 (*addsi3_carryin_alt2_<optab>): Likewise.
2103 (*addsi3_carryin_clobercc_<optab>): Likewise.
2104 (*subsi3_carryin): Likewise.
2105 (*subsi3_carryin_const): Likewise.
2106 (*subsi3_carryin_compare): Likewise.
2107 (*subsi3_carryin_compare_const): Likewise.
2108 (*arm_subdi3): Likewise.
2109 (*thumb_subdi3): Likewise.
2110 (*subdi_di_zesidi): Likewise.
2111 (*subdi_di_sesidi): Likewise.
2112 (*subdi_zesidi_di): Likewise.
2113 (*subdi_sesidi_di): Likewise.
2114 (*subdi_zesidi_ze): Likewise.
2115 (thumb1_subsi3_insn): Likewise.
2116 (*arm_subsi3_insn): Likewise.
2117 (*anddi3_insn): Likewise.
2118 (*anddi_zesidi_di): Likewise.
2119 (*anddi_sesdi_di): Likewise.
2120 (*ne_zeroextracts): Likewise.
2121 (*ne_zeroextracts): Likewise.
2122 (*ite_ne_zeroextr): Likewise.
2123 (*ite_ne_zeroextr): Likewise.
2124 (*anddi_notdi_di): Likewise.
2125 (*anddi_notzesidi): Likewise.
2126 (*anddi_notsesidi): Likewise.
2127 (andsi_notsi_si): Likewise.
2128 (thumb1_bicsi3): Likewise.
2129 (*iordi3_insn): Likewise.
2130 (*iordi_zesidi_di): Likewise.
2131 (*iordi_sesidi_di): Likewise.
2132 (*thumb1_iorsi3_insn): Likewise.
2133 (*xordi3_insn): Likewise.
2134 (*xordi_zesidi_di): Likewise.
2135 (*xordi_sesidi_di): Likewise.
2136 (*arm_xorsi3): Likewise.
2137 (*andsi_iorsi3_no): Likewise.
2138 (*smax_0): Likewise.
2139 (*smax_m1): Likewise.
2140 (*arm_smax_insn): Likewise.
2141 (*smin_0): Likewise.
2142 (*arm_smin_insn): Likewise.
2143 (*arm_umaxsi3): Likewise.
2144 (*arm_uminsi3): Likewise.
2145 (*minmax_arithsi): Likewise.
2146 (*minmax_arithsi_): Likewise.
2147 (*satsi_<SAT:code>): Likewise.
2148 (arm_ashldi3_1bit): Likewise.
2149 (arm_ashrdi3_1bit): Likewise.
2150 (arm_lshrdi3_1bit): Likewise.
2151 (*arm_negdi2): Likewise.
2152 (*thumb1_negdi2): Likewise.
2153 (*arm_negsi2): Likewise.
2154 (*thumb1_negsi2): Likewise.
2155 (*negdi_extendsid): Likewise.
2156 (*negdi_zero_extend): Likewise.
2157 (*arm_abssi2): Likewise.
2158 (*thumb1_abssi2): Likewise.
2159 (*arm_neg_abssi2): Likewise.
2160 (*thumb1_neg_abss): Likewise.
2161 (one_cmpldi2): Likewise.
2162 (extend<mode>di2): Likewise.
2163 (*compareqi_eq0): Likewise.
2164 (*arm_extendhisi2addsi): Likewise.
2165 (*arm_movdi): Likewise.
2166 (*thumb1_movdi_insn): Likewise.
2167 (*arm_movt): Likewise.
2168 (*thumb1_movsi_insn): Likewise.
2169 (pic_add_dot_plus_four): Likewise.
2170 (pic_add_dot_plus_eight): Likewise.
2171 (tls_load_dot_plus_eight): Likewise.
2172 (*thumb1_movhi_insn): Likewise.
2173 (*thumb1_movsf_insn): Likewise.
2174 (*movdf_soft_insn): Likewise.
2175 (*thumb_movdf_insn): Likewise.
2176 (cbranchsi4_insn): Likewise.
2177 (cbranchsi4_scratch): Likewise.
2178 (*negated_cbranchsi4): Likewise.
2179 (*tbit_cbranch): Likewise.
2180 (*tlobits_cbranch): Likewise.
2181 (*tstsi3_cbranch): Likewise.
2182 (*cbranchne_decr1): Likewise.
2183 (*addsi3_cbranch): Likewise.
2184 (*addsi3_cbranch_scratch): Likewise.
2185 (*arm_cmpdi_insn): Likewise.
2186 (*arm_cmpdi_unsig): Likewise.
2187 (*arm_cmpdi_zero): Likewise.
2188 (*thumb_cmpdi_zero): Likewise.
2189 (*deleted_compare): Likewise.
2190 (*mov_scc): Likewise.
2191 (*mov_negscc): Likewise.
2192 (*mov_notscc): Likewise.
2193 (*cstoresi_eq0_thumb1_insn): Likewise.
2194 (cstoresi_nltu_thumb1): Likewise.
2195 (cstoresi_ltu_thu): Likewise.
2196 (thumb1_addsi3_addgeu): Likewise.
2197 (*arm_jump): Likewise.
2198 (*thumb_jump): Likewise.
2199 (*check_arch2): Likewise.
2200 (arm_casesi_internal): Likewise.
2201 (thumb1_casesi_dispatch): Likewise.
2202 (*arm_indirect_jump): Likewise.
2203 (*thumb1_indirect_jump): Likewise.
2205 (*and_scc): Likewise.
2206 (*ior_scc): Likewise.
2207 (*compare_scc): Likewise.
2208 (*cond_move): Likewise.
2209 (*cond_arith): Likewise.
2210 (*cond_sub): Likewise.
2211 (*cmp_ite0): Likewise.
2212 (*cmp_ite1): Likewise.
2213 (*cmp_and): Likewise.
2214 (*cmp_ior): Likewise.
2215 (*ior_scc_scc): Likewise.
2216 (*ior_scc_scc_cmp): Likewise.
2217 (*and_scc_scc): Likewise.
2218 (*and_scc_scc_cmp): Likewise.
2219 (*and_scc_scc_nod): Likewise.
2220 (*negscc): Likewise.
2221 (movcond_addsi): Likewise.
2222 (movcond): Likewise.
2223 (*ifcompare_plus_move): Likewise.
2224 (*if_plus_move): Likewise.
2225 (*ifcompare_move_plus): Likewise.
2226 (*if_move_plus): Likewise.
2227 (*ifcompare_arith_arith): Likewise.
2228 (*if_arith_arith): Likewise.
2229 (*ifcompare_arith_move): Likewise.
2230 (*if_arith_move): Likewise.
2231 (*ifcompare_move_arith): Likewise.
2232 (*if_move_arith): Likewise.
2233 (*ifcompare_move_not): Likewise.
2234 (*if_move_not): Likewise.
2235 (*ifcompare_not_move): Likewise.
2236 (*if_not_move): Likewise.
2237 (*ifcompare_shift_move): Likewise.
2238 (*if_shift_move): Likewise.
2239 (*ifcompare_move_shift): Likewise.
2240 (*if_move_shift): Likewise.
2241 (*ifcompare_shift_shift): Likewise.
2242 (*ifcompare_not_arith): Likewise.
2243 (*ifcompare_arith_not): Likewise.
2244 (*if_arith_not): Likewise.
2245 (*ifcompare_neg_move): Likewise.
2246 (*if_neg_move): Likewise.
2247 (*ifcompare_move_neg): Likewise.
2248 (*if_move_neg): Likewise.
2249 (prologue_thumb1_interwork): Likewise.
2250 (*cond_move_not): Likewise.
2251 (*sign_extract_onebit): Likewise.
2252 (*not_signextract_onebit): Likewise.
2253 (stack_tie): Likewise.
2254 (align_4): Likewise.
2255 (align_8): Likewise.
2256 (consttable_end): Likewise.
2257 (consttable_1): Likewise.
2258 (consttable_2): Likewise.
2259 (consttable_4): Likewise.
2260 (consttable_8): Likewise.
2261 (consttable_16): Likewise.
2262 (*thumb1_tablejump): Likewise.
2263 (prefetch): Likewise.
2264 (force_register_use): Likewise.
2265 (thumb_eh_return): Likewise.
2266 (load_tp_hard): Likewise.
2267 (load_tp_soft): Likewise.
2268 (tlscall): Likewise.
2269 (*arm_movtas_ze): Likewise.
2270 (*arm_rev): Likewise.
2271 (*arm_revsh): Likewise.
2272 (*arm_rev16): Likewise.
2273 * config/arm/thumb2.md
2274 (*thumb2_smaxsi3): Likewise.
2275 (*thumb2_sminsi3): Likewise.
2276 (*thumb32_umaxsi3): Likewise.
2277 (*thumb2_uminsi3): Likewise.
2278 (*thumb2_negdi2): Likewise.
2279 (*thumb2_abssi2): Likewise.
2280 (*thumb2_neg_abss): Likewise.
2281 (*thumb2_movsi_insn): Likewise.
2282 (tls_load_dot_plus_four): Likewise.
2283 (*thumb2_movhi_insn): Likewise.
2284 (*thumb2_mov_scc): Likewise.
2285 (*thumb2_mov_negs): Likewise.
2286 (*thumb2_mov_negs): Likewise.
2287 (*thumb2_mov_nots): Likewise.
2288 (*thumb2_mov_nots): Likewise.
2289 (*thumb2_movsicc_): Likewise.
2290 (*thumb2_movsfcc_soft_insn): Likewise.
2291 (*thumb2_indirect_jump): Likewise.
2292 (*thumb2_and_scc): Likewise.
2293 (*thumb2_ior_scc): Likewise.
2294 (*thumb2_ior_scc_strict_it): Likewise.
2295 (*thumb2_cond_move): Likewise.
2296 (*thumb2_cond_arith): Likewise.
2297 (*thumb2_cond_ari): Likewise.
2298 (*thumb2_cond_sub): Likewise.
2299 (*thumb2_negscc): Likewise.
2300 (*thumb2_movcond): Likewise.
2301 (thumb2_casesi_internal): Likewise.
2302 (thumb2_casesi_internal_pic): Likewise.
2303 (*thumb2_alusi3_short): Likewise.
2304 (*thumb2_mov<mode>_shortim): Likewise.
2305 (*thumb2_addsi_short): Likewise.
2306 (*thumb2_subsi_short): Likewise.
2307 (thumb2_addsi3_compare0): Likewise.
2308 (*thumb2_cbz): Likewise.
2309 (*thumb2_cbnz): Likewise.
2310 (*thumb2_one_cmplsi2_short): Likewise.
2311 (*thumb2_negsi2_short): Likewise.
2312 (*orsi_notsi_si): Likewise.
2313 * config/arm/arm1020e.md: Update with new attributes.
2314 * config/arm/arm1026ejs.md: Update with new attributes.
2315 * config/arm/arm1136jfs.md: Update with new attributes.
2316 * config/arm/arm926ejs.md: Update with new attributes.
2317 * config/arm/cortex-a15.md: Update with new attributes.
2318 * config/arm/cortex-a5.md: Update with new attributes.
2319 * config/arm/cortex-a53.md: Update with new attributes.
2320 * config/arm/cortex-a7.md: Update with new attributes.
2321 * config/arm/cortex-a8.md: Update with new attributes.
2322 * config/arm/cortex-a9.md: Update with new attributes.
2323 * config/arm/cortex-m4.md: Update with new attributes.
2324 * config/arm/cortex-r4.md: Update with new attributes.
2325 * config/arm/fa526.md: Update with new attributes.
2326 * config/arm/fa606te.md: Update with new attributes.
2327 * config/arm/fa626te.md: Update with new attributes.
2328 * config/arm/fa726te.md: Update with new attributes.
2330 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2332 * config/aarch64/aarch64-simd.md
2333 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2334 <vwx> iterator to ensure correct register choice.
2335 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2336 (aarch64_sqdmull_n<mode>): Likewise.
2337 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2338 * config/aarch64/arm_neon.h
2339 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2340 (vml<as><q>_n_<su>16): Likewise.
2341 (vml<as>l_high_lane<q>_<su>16): Likewise.
2342 (vml<as>l_high_n_<su>16): Likewise.
2343 (vml<as>l_lane<q>_<su>16): Likewise.
2344 (vml<as>l_n_<su>16): Likewise.
2345 (vmul<q>_lane<q>_<su>16): Likewise.
2346 (vmul<q>_n_<su>16): Likewise.
2347 (vmull_lane<q>_<su>16): Likewise.
2348 (vmull_n_<su>16): Likewise.
2349 (vmull_high_lane<q>_<su>16): Likewise.
2350 (vmull_high_n_<su>16): Likewise.
2351 (vqrdmulh<q>_n_s16): Likewise.
2353 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2355 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2356 have the correct lane parameter.
2358 2013-09-06 Richard Biener <rguenther@suse.de>
2360 * cfganal.c (control_dependences::~control_dependences):
2361 Properly free all of the vector.
2363 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2366 * config/i386/i386.c (ix86_conditional_register_usage):
2367 Proper initialize extended SSE registers.
2369 2013-09-06 Jan Hubicka <jh@suse.cz>
2371 PR tree-optimization/58311
2372 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2374 2013-09-06 Jan Hubicka <jh@suse.cz>
2376 * Makefile.in (tree-sra.o): Update dependencies.
2377 * tree-sra.c: Include ipa-utils.h
2378 (scan_function): Use recursive_call_p.
2379 (has_caller_p): New function.
2380 (cgraph_for_node_and_aliases): Count also callers of aliases.
2382 2013-09-06 Jan Hubicka <jh@suse.cz>
2385 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2386 * tree-tailcall.c: Include ipa-utils.h.
2387 (find_tail_calls): Use it.
2388 * ipa-pure-const.c (check_call): Likewise.
2389 * ipa-utils.c (recursive_call_p): New function.
2390 * ipa-utils.h (recursive_call_p): Dclare.
2391 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2392 (symtab_semantically_equivalent_p): New function.
2393 * Makefile.in (tree-tailcall.o): Update dependencies.
2395 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2397 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2400 2013-09-06 Richard Biener <rguenther@suse.de>
2402 * lto-streamer.h (lto_global_var_decls): Remove.
2403 * Makefile.in (OBJS): Remove lto-symtab.o.
2404 (lto-symtab.o): Remove.
2405 (GTFILES): Remove lto-symtab.c
2406 * lto-symtab.c: Move to lto/
2408 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2410 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2411 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2412 (UNSPEC_FPINT_RINT): New constant definitions.
2413 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2414 definition with 2 attributes.
2415 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2416 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2419 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2421 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2422 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2423 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2424 attribute to "z196_alone".
2425 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2428 2013-09-06 Richard Biener <rguenther@suse.de>
2430 * basic-block.h (class control_dependences): New.
2431 * tree-ssa-dce.c (control_dependence_map): Remove.
2433 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2434 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2435 find_pdom, find_control_dependence, find_all_control_dependences):
2437 (mark_control_dependent_edges_necessary,
2438 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2439 tree_dce_done, perform_tree_ssa_dce): Adjust.
2440 * cfganal.c (set_control_dependence_map_bit,
2441 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2442 find_all_control_dependences): Move from tree-ssa-dce.c and
2443 implement as methods of control_dependences class.
2444 (control_dependences::control_dependences): New.
2445 (control_dependences::~control_dependences): Likewise.
2446 (control_dependences::get_edges_dependent_on): Likewise.
2447 (control_dependences::get_edge): Likewise.
2449 2013-09-04 Jan Hubicka <jh@suse.cz>
2451 * tree.c (types_same_for_odr): Drop overactive check.
2452 * ipa-devirt.c (hash_type_name): Likewise.
2454 2013-09-04 Jan Hubicka <jh@suse.cz>
2456 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2457 (analyze_functions): ... here.
2459 2013-09-04 Jan Hubicka <jh@suse.cz>
2462 * cgraphunit.c (analyze_functions): Clear AUX fields
2463 after processing; initialize assembler name has.
2465 2013-09-05 Jeff Law <law@redhat.com>
2467 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2468 from thread_around_empty_block. Record threading path into PATH.
2469 Recurse if threading through the initial block is successful.
2470 (thread_across_edge): Corresponding changes to slightly simplify.
2472 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2474 * config/aarch64/aarch64.md
2475 (type): Remove frecpe, frecps, frecpx.
2476 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2477 fix to be a TARGET_SIMD instruction.
2478 (aarch64_frecps): Remove.
2479 * config/aarch64/aarch64-simd.md
2480 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2481 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2483 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2484 Sofiane Naci <sofiane.naci@arm.com>
2486 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2487 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2488 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2489 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2490 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2491 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2492 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2493 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2494 into "clz, "rbit". Rename "shift" to "shift_imm".
2495 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2496 changes. Update for attribute changes all occurrences of arlo_* and
2498 * config/arm/arm-fixed.md: Update for attribute changes
2499 all occurrences of arlo_* types.
2500 * config/arm/thumb2.md: Update for attribute changes all occurrences
2502 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2503 (cortexa7_older_only): Likewise.
2504 (cortexa7_younger): Likewise.
2505 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2506 (1020alu_shift_op): Likewise.
2507 (1020alu_shift_reg_op): Likewise.
2508 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2509 (alu_shift_op): Likewise.
2510 (alu_shift_reg_op): Likewise.
2511 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2512 (11_alu_shift_op): Likewise.
2513 (11_alu_shift_reg_op): Likewise.
2514 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2515 (9_alu_shift_reg_op): Likewise.
2516 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2518 (cortex_a15_alu_shift): Likewise.
2519 (cortex_a15_alu_shift_reg): Likewise.
2520 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2522 (cortex_a5_alu_shift): Likewise.
2523 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2525 (cortex_a53_alu_shift): Likewise.
2526 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2528 (cortex_a7_alu_reg): Likewise.
2529 (cortex_a7_alu_shift): Likewise.
2530 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2532 (cortex_a8_alu_shift): Likewise.
2533 (cortex_a8_alu_shift_reg): Likewise.
2534 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2535 (cortex_a9_dp_shift): Likewise.
2536 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2538 * config/arm/cortex-r4.md
2539 (cortex_r4_alu): Update for attribute changes.
2540 (cortex_r4_mov): Likewise.
2541 (cortex_r4_alu_shift_reg): Likewise.
2542 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2543 (526_alu_shift_op): Likewise.
2544 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2545 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2546 (626te_alu_shift_op): Likewise.
2547 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2548 (726te_alu_shift_op): Likewise.
2549 (726te_alu_shift_reg_op): Likewise.
2550 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2551 (mp626_alu_shift_op): Likewise.
2552 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2553 (pj4_alu_conds): Likewise.
2554 (pj4_shift): Likewise.
2555 (pj4_shift_conds): Likewise.
2556 (pj4_alu_shift): Likewise.
2557 (pj4_alu_shift_conds): Likewise.
2558 * config/aarch64/aarch64.md: Update for attribute change
2559 all occurrences of arlo_* and shift* types.
2561 2013-09-05 Mike Stump <mikestump@comcast.net>
2563 * tree.h: Move documentation for tree_function_decl to tree-core.h
2564 with the declaration.
2566 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2569 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2570 looking for widest mode.
2572 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2574 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2576 2013-09-05 Richard Biener <rguenther@suse.de>
2578 PR tree-optimization/58137
2579 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2580 Do not create vectors of pointers.
2581 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2582 types for the components of the vector initializer.
2583 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2584 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2586 2013-09-05 Martin Jambor <mjambor@suse.cz>
2588 * ipa-prop.c (remove_described_reference): Accept missing references,
2589 return false if that hppens, otherwise return true.
2590 (cgraph_node_for_jfunc): New function.
2591 (try_decrement_rdesc_refcount): Likewise.
2592 (try_make_edge_direct_simple_call): Use them.
2593 (ipa_edge_removal_hook): Remove references from rdescs.
2594 (ipa_edge_duplication_hook): Clone rdescs and their references
2595 when the new edge has the same caller as the old one.
2596 * cgraph.c (cgraph_resolve_speculation): Remove speculative
2597 reference before removing any edges.
2599 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
2601 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
2603 * thumb2.md (thumb2_storewb_parisi): New pattern.
2605 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
2607 * config/aarch64/aarch64-option-extensions.def: Add
2608 AARCH64_OPT_EXTENSION of 'crc'.
2609 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
2610 (AARCH64_ISA_CRC): Ditto.
2611 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
2612 description of the CRC extension.
2614 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
2616 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
2617 * config/rs6000/linux.h: Ditto.
2618 * alpha/linux.h: Ditto.
2619 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
2620 no_c99_libc_has_function.
2621 * config/c6x/uclinux-elf.h: Ditto.
2622 * config/lm32/uclinux-elf.h: Ditto.
2623 * config/m68k/uclinux.h: Ditto.
2624 * config/moxie/uclinux.h: Ditto.
2625 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
2626 (crisv32-*-linux*, cris-*-linux*): Ditto.
2627 * config/bfin/bfin.c: Include "tm_p.h".
2629 2013-09-05 Richard Biener <rguenther@suse.de>
2631 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
2632 check for a definition without a basic-block.
2634 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2635 Sofiane Naci <sofiane.naci@arm.com>
2637 * config/aarch64/aarch64.md
2638 (*movti_aarch64): Rename r_2_f and f_2_r.
2639 (*movsf_aarch64): Likewise.
2640 (*movdf_aarch64): Likewise.
2641 (*movtf_aarch64): Likewise.
2642 (aarch64_movdi_<mode>low): Likewise.
2643 (aarch64_movdi_<mode>high): Likewise.
2644 (aarch64_mov<mode>high_di): Likewise.
2645 (aarch64_mov<mode>low_di): Likewise.
2646 (aarch64_movtilow_tilow): Likewise.
2647 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
2648 values to config/arm/types.md
2649 (attribute "conds"): Update for attribute change.
2650 (anddi3_insn): Likewise.
2651 (iordi3_insn): Likewise.
2652 (xordi3_insn): Likewise.
2653 (one_cmpldi2): Likewise.
2654 * config/arm/types.md (type): Add Neon types.
2655 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
2656 use "type" attribute.
2657 (movmisalign<mode>_neon_store): Likewise.
2658 (movmisalign<mode>_neon_load): Likewise.
2659 (vec_set<mode>_internal): Likewise.
2660 (vec_setv2di_internal): Likewise.
2661 (vec_extract<mode>): Likewise.
2662 (vec_extractv2di): Likewise.
2663 (add<mode>3_neon): Likewise.
2664 (adddi3_neon): Likewise.
2665 (sub<mode>3_neon): Likewise.
2666 (subdi3_neon): Likewise.
2667 (mul<mode>3_neon): Likewise.
2668 (mul<mode>3add<mode>_neon): Likewise.
2669 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
2670 (fma<VCVTF:mode>4)): Likewise.
2671 (fma<VCVTF:mode>4_intrinsic): Likewise.
2672 (fmsub<VCVTF:mode>4)): Likewise.
2673 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2674 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2675 (ior<mode>3): Likewise.
2676 (and<mode>3): Likewise.
2677 (anddi3_neon): Likewise.
2678 (orn<mode>3_neon): Likewise.
2679 (orndi3_neon): Likewise.
2680 (bic<mode>3_neon): Likewise.
2681 (bicdi3_neon): Likewise.
2682 (xor<mode>3): Likewise.
2683 (one_cmpl<mode>2): Likewise.
2684 (abs<mode>2): Likewise.
2685 (neg<mode>2): Likewise.
2686 (umin<mode>3_neon): Likewise.
2687 (umax<mode>3_neon): Likewise.
2688 (smin<mode>3_neon): Likewise.
2689 (smax<mode>3_neon): Likewise.
2690 (vashl<mode>3): Likewise.
2691 (vashr<mode>3_imm): Likewise.
2692 (vlshr<mode>3_imm): Likewise.
2693 (ashl<mode>3_signed): Likewise.
2694 (ashl<mode>3_unsigned): Likewise.
2695 (neon_load_count): Likewise.
2696 (ashldi3_neon_noclobber): Likewise.
2697 (signed_shift_di3_neon): Likewise.
2698 (unsigned_shift_di3_neon): Likewise.
2699 (ashrdi3_neon_imm_noclobber): Likewise.
2700 (lshrdi3_neon_imm_noclobber): Likewise.
2701 (widen_ssum<mode>3): Likewise.
2702 (widen_usum<mode>3): Likewise.
2703 (quad_halves_<code>v4si): Likewise.
2704 (quad_halves_<code>v4sf): Likewise.
2705 (quad_halves_<code>v8hi): Likewise.
2706 (quad_halves_<code>v16qi): Likewise.
2707 (reduc_splus_v2di): Likewise.
2708 (neon_vpadd_internal<mode>): Likewise.
2709 (neon_vpsmin<mode>): Likewise.
2710 (neon_vpsmax<mode>): Likewise.
2711 (neon_vpumin<mode>): Likewise.
2712 (neon_vpumax<mode>): Likewise.
2713 (ss_add<mode>_neon): Likewise.
2714 (us_add<mode>_neon): Likewise.
2715 (ss_sub<mode>_neon): Likewise.
2716 (us_sub<mode>_neon): Likewise.
2717 (neon_vadd<mode>_unspec): Likewise.
2718 (neon_vaddl<mode>): Likewise.
2719 (neon_vaddw<mode>): Likewise.
2720 (neon_vhadd<mode>): Likewise.
2721 (neon_vqadd<mode>): Likewise.
2722 (neon_vaddhn<mode>): Likewise.
2723 (neon_vmul<mode>): Likewise.
2724 (neon_vmla<mode>): Likewise.
2725 (neon_vmlal<mode>): Likewise.
2726 (neon_vmls<mode>): Likewise.
2727 (neon_vmlsl<mode>): Likewise.
2728 (neon_vqdmulh<mode>): Likewise.
2729 (neon_vqdmlal<mode>): Likewise.
2730 (neon_vqdmlsl<mode>): Likewise.
2731 (neon_vmull<mode>): Likewise.
2732 (neon_vqdmull<mode>): Likewise.
2733 (neon_vsub<mode>_unspec): Likewise.
2734 (neon_vsubl<mode>): Likewise.
2735 (neon_vsubw<mode>): Likewise.
2736 (neon_vqsub<mode>): Likewise.
2737 (neon_vhsub<mode>): Likewise.
2738 (neon_vsubhn<mode>): Likewise.
2739 (neon_vceq<mode>): Likewise.
2740 (neon_vcge<mode>): Likewise.
2741 (neon_vcgeu<mode>): Likewise.
2742 (neon_vcgt<mode>): Likewise.
2743 (neon_vcgtu<mode>): Likewise.
2744 (neon_vcle<mode>): Likewise.
2745 (neon_vclt<mode>): Likewise.
2746 (neon_vcage<mode>): Likewise.
2747 (neon_vcagt<mode>): Likewise.
2748 (neon_vtst<mode>): Likewise.
2749 (neon_vabd<mode>): Likewise.
2750 (neon_vabdl<mode>): Likewise.
2751 (neon_vaba<mode>): Likewise.
2752 (neon_vabal<mode>): Likewise.
2753 (neon_vmax<mode>): Likewise.
2754 (neon_vmin<mode>): Likewise.
2755 (neon_vpaddl<mode>): Likewise.
2756 (neon_vpadal<mode>): Likewise.
2757 (neon_vpmax<mode>): Likewise.
2758 (neon_vpmin<mode>): Likewise.
2759 (neon_vrecps<mode>): Likewise.
2760 (neon_vrsqrts<mode>): Likewise.
2761 (neon_vqabs<mode>): Likewise.
2762 (neon_vqneg<mode>): Likewise.
2763 (neon_vcls<mode>): Likewise.
2764 (clz<mode>2): Likewise.
2765 (popcount<mode>2): Likewise.
2766 (neon_vrecpe): Likewise.
2767 (neon_vrsqrte): Likewise.
2768 (neon_vget_lane<mode>_sext_internal): Likewise.
2769 (neon_vget_lane<mode>_zext_internal): Likewise.
2770 (neon_vdup_n<mode>): Likewise.
2771 (neon_vdup_nv2di): Likewise.
2772 (neon_vdpu_lane<mode>_internal): Likewise.
2773 (neon_vswp<mode>): Likewise.
2774 (float<mode><V_cvtto>2): Likewise.
2775 (floatuns<mode><V_cvtto>2): Likewise.
2776 (fix_trunc<mode><V_cvtto>)2): Likewise
2777 (fixuns_trunc<mode><V_cvtto)2): Likewise.
2778 (neon_vcvt<mode>): Likewise.
2779 (neon_vcvtv4sfv4hf): Likewise.
2780 (neon_vcvtv4hfv4sf): Likewise.
2781 (neon_vcvt_n<mode>): Likewise.
2782 (neon_vmovn<mode>): Likewise.
2783 (neon_vqmovn<mode>): Likewise.
2784 (neon_vqmovun<mode>): Likewise.
2785 (neon_vmovl<mode>): Likewise.
2786 (neon_vmul_lane<mode>): Likewise.
2787 (neon_vmull_lane<mode>): Likewise.
2788 (neon_vqdmull_lane<mode>): Likewise.
2789 (neon_vqdmulh_lane<mode>): Likewise.
2790 (neon_vmla_lane<mode>): Likewise.
2791 (neon_vmlal_lane<mode>): Likewise.
2792 (neon_vqdmlal_lane<mode>): Likewise.
2793 (neon_vmls_lane<mode>): Likewise.
2794 (neon_vmlsl_lane<mode>): Likewise.
2795 (neon_vqdmlsl_lane<mode>): Likewise.
2796 (neon_vext<mode>): Likewise.
2797 (neon_vrev64<mode>): Likewise.
2798 (neon_vrev32<mode>): Likewise.
2799 (neon_vrev16<mode>): Likewise.
2800 (neon_vbsl<mode>_internal): Likewise.
2801 (neon_vshl<mode>): Likewise.
2802 (neon_vqshl<mode>): Likewise.
2803 (neon_vshr_n<mode>): Likewise.
2804 (neon_vshrn_n<mode>): Likewise.
2805 (neon_vqshrn_n<mode>): Likewise.
2806 (neon_vqshrun_n<mode>): Likewise.
2807 (neon_vshl_n<mode>): Likewise.
2808 (neon_vqshl_n<mode>): Likewise.
2809 (neon_vqshlu_n<mode>): Likewise.
2810 (neon_vshll_n<mode>): Likewise.
2811 (neon_vsra_n<mode>): Likewise.
2812 (neon_vsri_n<mode>): Likewise.
2813 (neon_vsli_n<mode>): Likewise.
2814 (neon_vtbl1v8qi): Likewise.
2815 (neon_vtbl2v8qi): Likewise.
2816 (neon_vtbl3v8qi): Likewise.
2817 (neon_vtbl4v8qi): Likewise.
2818 (neon_vtbx1v8qi): Likewise.
2819 (neon_vtbx2v8qi): Likewise.
2820 (neon_vtbx3v8qi): Likewise.
2821 (neon_vtbx4v8qi): Likewise.
2822 (neon_vtrn<mode>_internal): Likewise.
2823 (neon_vzip<mode>_internal): Likewise.
2824 (neon_vuzp<mode>_internal): Likewise.
2825 (neon_vld1<mode>): Likewise.
2826 (neon_vld1_lane<mode>): Likewise.
2827 (neon_vld1_dup<mode>): Likewise.
2828 (neon_vld1_dupv2di): Likewise.
2829 (neon_vst1<mode>): Likewise.
2830 (neon_vst1_lane<mode>): Likewise.
2831 (neon_vld2<mode>): Likewise.
2832 (neon_vld2_lane<mode>): Likewise.
2833 (neon_vld2_dup<mode>): Likewise.
2834 (neon_vst2<mode>): Likewise.
2835 (neon_vst2_lane<mode>): Likewise.
2836 (neon_vld3<mode>): Likewise.
2837 (neon_vld3qa<mode>): Likewise.
2838 (neon_vld3qb<mode>): Likewise.
2839 (neon_vld3_lane<mode>): Likewise.
2840 (neon_vld3_dup<mode>): Likewise.
2841 (neon_vst3<mode>): Likewise.
2842 (neon_vst3qa<mode>): Likewise.
2843 (neon_vst3qb<mode>): Likewise.
2844 (neon_vst3_lane<mode>): Likewise.
2845 (neon_vld4<mode>): Likewise.
2846 (neon_vld4qa<mode>): Likewise.
2847 (neon_vld4qb<mode>): Likewise.
2848 (neon_vld4_lane<mode>): Likewise.
2849 (neon_vld4_dup<mode>): Likewise.
2850 (neon_vst4<mode>): Likewise.
2851 (neon_vst4qa<mode>): Likewise.
2852 (neon_vst4qb<mode>): Likewise.
2853 (neon_vst4_lane<mode>): Likewise.
2854 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2855 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2856 (neon_vec_<US>mult_lo_<mode>): Likewise.
2857 (neon_vec_<US>mult_hi_<mode>): Likewise.
2858 (neon_vec_<US>shiftl_<mode>): Likewise.
2859 (neon_unpack<US>_<mode>): Likewise.
2860 (neon_vec_<US>mult_<mode>): Likewise.
2861 (vec_pack_trunc_<mode>): Likewise.
2862 (neon_vec_pack_trunk_<mode>): Likewise.
2863 (neon_vabd<mode>_2): Likewise.
2864 (neon_vabd<mode>_3): Likewise.
2865 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2866 (thumb2_movsi_vfp): Likewise.
2867 (movdi_vfp): Likewise.
2868 (movdi_vfp_cortexa8): Likewise.
2869 (movhf_vfp_neon): Likewise.
2870 (movhf_vfp): Likewiwse.
2871 (movsf_vfp): Likewiwse.
2872 (thumb2_movsf_vfp): Likewiwse.
2873 (movdf_vfp): Likewise.
2874 (thumb2_movdf_vfp): Likewise.
2875 (movsfcc_vfp): Likewise.
2876 (thumb2_movsfcc_vfp): Likewise.
2877 (movdfcc_vfp): Likewise.
2878 (thumb2_movdfcc_vfp): Likewise.
2879 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
2880 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
2881 (v10_v2c): Likewise.
2882 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
2884 (cortex_a15_neon_int_2): Likewise.
2885 (cortex_a15_neon_int_3): Likewise.
2886 (cortex_a15_neon_int_4): Likewise.
2887 (cortex_a15_neon_int_5): Likewise.
2888 (cortex_a15_neon_vqneg_vqabs): Likewise.
2889 (cortex_a15_neon_vmov): Likewise.
2890 (cortex_a15_neon_vaba): Likewise.
2891 (cortex_a15_neon_vaba_qqq): Likewise.
2892 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2893 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2894 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
2895 scalar_64_32_long_scalar): Likewise.
2896 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2897 (cortex_a15_neon_mla_qqq_8_16): Likewise.
2898 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2899 lotype_qdd_64_32_long): Likewise.
2900 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2901 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2902 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
2903 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2904 (cortex_a15_neon_shift_1): Likewise.
2905 (cortex_a15_neon_shift_2): Likewise.
2906 (cortex_a15_neon_shift_3): Likewise.
2907 (cortex_a15_neon_vshl_ddd): Likewise.
2908 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2909 (cortex_a15_neon_vsra_vrsra): Likewise.
2910 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
2911 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
2912 (cortex_a15_neon_fp_vmul_ddd): Likewise.
2913 (cortex_a15_neon_fp_vmul_qqd): Likewise.
2914 (cortex_a15_neon_fp_vmla_ddd): Likewise.
2915 (cortex_a15_neon_fp_vmla_qqq): Likewise.
2916 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2917 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2918 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2919 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2920 (cortex_a15_neon_bp_simple): Likewise.
2921 (cortex_a15_neon_bp_2cycle): Likewise.
2922 (cortex_a15_neon_bp_3cycle): Likewise.
2923 (cortex_a15_neon_vld1_1_2_regs): Likewise.
2924 (cortex_a15_neon_vld1_3_4_regs): Likewise.
2925 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2926 (cortex_a15_neon_vld2_4_regs): Likewise.
2927 (cortex_a15_neon_vld3_vld4): Likewise.
2928 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2929 (cortex_a15_neon_vst1_3_4_regs): Likewise.
2930 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
2931 (cortex_a15_neon_vst3_vst4): Likewise.
2932 (cortex_a15_neon_vld1_vld2_lane): Likewise.
2933 (cortex_a15_neon_vld3_vld4_lane" 10
2934 (cortex_a15_neon_vst1_vst2_lane): Likewise.
2935 (cortex_a15_neon_vst3_vst4_lane): Likewise.
2936 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
2937 (cortex_a15_neon_ldm_2): Likewise.0
2938 (cortex_a15_neon_stm_2): Likewise.
2939 (cortex_a15_neon_mcr): Likewise.
2940 (cortex_a15_neon_mcr_2_mcrr): Likewise.
2941 (cortex_a15_neon_mrc): Likewise.
2942 (cortex_a15_neon_mrrc): Likewise.
2943 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2945 (cortex_a15_alu_shift): Likewise.
2946 (cortex_a15_alu_shift_reg): Likewise.
2947 (cortex_a15_mult32): Likewise.
2948 (cortex_a15_mult64): Likewise.
2949 (cortex_a15_block): Likewise.
2950 (cortex_a15_branch): Likewise.
2951 (cortex_a15_load1): Likewise.
2952 (cortex_a15_load3): Likewise.
2953 (cortex_a15_store1): Likewise.
2954 (cortex_a15_store3): Likewise.
2955 (cortex_a15_call): Likewise.
2956 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
2957 (cortex_a5_f2r): Likewise.
2958 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
2960 (cortex_a53_f2r): Likewise.
2961 * config/arm/cortex-a7.md
2962 (cortex_a7_branch): Update for attribute change.
2963 (cortex_a7_call): Likewise.
2964 (cortex_a7_alu_imm): Likewise.
2965 (cortex_a7_alu_reg): Likewise.
2966 (cortex_a7_alu_shift): Likewise.
2967 (cortex_a7_mul): Likewise.
2968 (cortex_a7_load1): Likewise.
2969 (cortex_a7_store1): Likewise.
2970 (cortex_a7_load2): Likewise.
2971 (cortex_a7_store2): Likewise.
2972 (cortex_a7_load3): Likewise.
2973 (cortex_a7_store3): Likewise.
2974 (cortex_a7_load4): Likewise.
2975 (cortex_a7_store4): Likewise.
2976 (cortex_a7_fpalu): Likewise.
2977 (cortex_a7_fconst): Likewise.
2978 (cortex_a7_fpmuls): Likewise.
2979 (cortex_a7_neon_mul): Likewise.
2980 (cortex_a7_fpmacs): Likewise.
2981 (cortex_a7_neon_mla: Likewise.
2982 (cortex_a7_fpmuld: Likewise.
2983 (cortex_a7_fpmacd: Likewise.
2984 (cortex_a7_fpfmad: Likewise.
2985 (cortex_a7_fdivs: Likewise.
2986 (cortex_a7_fdivd: Likewise.
2987 (cortex_a7_r2f: Likewise.
2988 (cortex_a7_f2r: Likewise.
2989 (cortex_a7_f_flags: Likewise.
2990 (cortex_a7_f_loads: Likewise.
2991 (cortex_a7_f_loadd: Likewise.
2992 (cortex_a7_f_stores: Likewise.
2993 (cortex_a7_f_stored: Likewise.
2994 (cortex_a7_neon): Likewise.
2995 * config/arm/cortex-a8-neon.md
2996 (cortex_a8_neon_mrc): Update for attribute change.
2997 (cortex_a8_neon_mrrc): Likewise.
2998 (cortex_a8_neon_int_1): Likewise.
2999 (cortex_a8_neon_int_2): Likewise.
3000 (cortex_a8_neon_int_3): Likewise.
3001 (cortex_a8_neon_int_4): Likewise.
3002 (cortex_a8_neon_int_5): Likewise.
3003 (cortex_a8_neon_vqneg_vqabs): Likewise.
3004 (cortex_a8_neon_vmov): Likewise.
3005 (cortex_a8_neon_vaba): Likewise.
3006 (cortex_a8_neon_vaba_qqq): Likewise.
3007 (cortex_a8_neon_vsma): Likewise.
3008 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3009 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3010 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3011 long_scalar): Likewise.
3012 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3013 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3014 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3015 long_scalar_qdd_64_32_long): Likewise.
3016 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3017 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3018 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3019 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3020 (cortex_a8_neon_shift_1): Likewise.
3021 (cortex_a8_neon_shift_2): Likewise.
3022 (cortex_a8_neon_shift_3): Likewise.
3023 (cortex_a8_neon_vshl_ddd): Likewise.
3024 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3025 (cortex_a8_neon_vsra_vrsra): Likewise.
3026 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3027 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3028 (cortex_a8_neon_fp_vsum): Likewise.
3029 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3030 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3031 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3032 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3033 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3034 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3035 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3036 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3037 (cortex_a8_neon_bp_simple): Likewise.
3038 (cortex_a8_neon_bp_2cycle): Likewise.
3039 (cortex_a8_neon_bp_3cycle): Likewise.
3040 (cortex_a8_neon_ldr): Likewise.
3041 (cortex_a8_neon_str): Likewise.
3042 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3043 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3044 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3045 (cortex_a8_neon_vld2_4_regs): Likewise.
3046 (cortex_a8_neon_vld3_vld4): Likewise.
3047 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3048 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3049 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3050 (cortex_a8_neon_vst3_vst4): Likewise.
3051 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3052 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3053 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3054 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3055 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3056 (cortex_a8_neon_mcr): Likewise.
3057 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3058 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3059 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3061 (ca9_neon_mrrc): Likewise.
3062 (cortex_a9_neon_int_1): Likewise.
3063 (cortex_a9_neon_int_2): Likewise.
3064 (cortex_a9_neon_int_3): Likewise.
3065 (cortex_a9_neon_int_4): Likewise.
3066 (cortex_a9_neon_int_5): Likewise.
3067 (cortex_a9_neon_vqneg_vqabs): Likewise.
3068 (cortex_a9_neon_vmov): Likewise.
3069 (cortex_a9_neon_vaba): Likewise.
3070 (cortex_a9_neon_vaba_qqq): Likewise.
3071 (cortex_a9_neon_vsma): Likewise.
3072 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3073 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3074 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3075 long_scalar): Likewise.
3076 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3077 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3078 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3079 long_scalar_qdd_64_32_long): Likewise.
3080 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3081 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3082 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3083 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3084 (cortex_a9_neon_shift_1): Likewise.
3085 (cortex_a9_neon_shift_2): Likewise.
3086 (cortex_a9_neon_shift_3): Likewise.
3087 (cortex_a9_neon_vshl_ddd): Likewise.
3088 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3089 (cortex_a9_neon_vsra_vrsra): Likewise.
3090 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3091 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3092 (cortex_a9_neon_fp_vsum): Likewise.
3093 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3094 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3095 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3096 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3097 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3098 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3099 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3100 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3101 (cortex_a9_neon_bp_simple): Likewise.
3102 (cortex_a9_neon_bp_2cycle): Likewise.
3103 (cortex_a9_neon_bp_3cycle): Likewise.
3104 (cortex_a9_neon_ldr): Likewise.
3105 (cortex_a9_neon_str): Likewise.
3106 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3107 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3108 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3109 (cortex_a9_neon_vld2_4_regs): Likewise.
3110 (cortex_a9_neon_vld3_vld4): Likewise.
3111 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3112 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3113 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3114 (cortex_a9_neon_vst3_vst4): Likewise.
3115 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3116 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3117 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3118 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3119 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3120 (cortex_a9_neon_mcr): Likewise.
3121 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3122 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3123 (cortex_a9_fps): Likewise.
3124 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3126 (cortex_m4_fmuls): Likewise.
3127 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3129 (cortex_r4_mrc): Likewise.
3130 * config/arm/iterators.md: Update comment referring to neon_type.
3131 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3132 (iwmmxt_movsi_insn): Likewise.
3133 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3135 (pj4_core_to_vfp): Likewise.
3136 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3138 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3139 (vfp_fstore): Likewise.
3140 * doc/md.texi: Change references to neon_type to refer to type.
3142 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3144 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3146 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3148 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3151 2013-09-04 Easwaran Raman <eraman@google.com>
3154 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3155 (build_and_add_sum): Use it.
3156 (appears_later_in_bb): Simplify code.
3158 2013-09-04 Teresa Johnson <tejohnson@google.com>
3160 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3162 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3164 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3166 2013-09-04 Jan Hubicka <jh@suse.cz>
3168 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3169 * ipa-devirt.c: Include diganostic.h
3170 (odr_type_d): Add types and types_set.
3171 (hash_type_name): Work for types with vtables during LTO.
3172 (odr_hasher::remove): Fix comment; destroy types_set.
3173 (add_type_duplicate): New function,
3174 (get_odr_type): Use it.
3175 (dump_type_inheritance_graph): Dump type duplicates.
3176 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3178 * tree.c (types_same_for_odr): Give exact answers on types with
3181 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3183 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3184 explaining their differences.
3186 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3188 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3190 2013-09-03 Jeff Law <law@redhat.com>
3192 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3193 when not threading through a joiner block. Pass joiner/no joiner
3194 state to register_jump_thread.
3195 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3196 state from argument rather than implying on path length.
3197 Dump the entire jump thread path into debugging dump.
3198 * tree-flow.h (register_jump_thread): Update prototype.
3200 2013-08-29 Xinliang David Li <davidxl@google.com>
3202 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3203 Remove a trivial gcc_assert.
3205 2013-08-29 Xinliang David Li <davidxl@google.com>
3207 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3208 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3209 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3210 Delay base decl alignment adjustment.
3211 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3212 * tree-vectorizer.h: New data structure.
3213 (set_dr_misalignment): New function.
3214 (dr_misalignment): Ditto.
3215 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3216 (vectorizable_load): Ditto.
3217 (ensure_base_align): New function.
3218 (vectorize_loops): Add dbg_cnt support.
3219 (execute_vect_slp): Ditto.
3220 * dbgcnt.def: New debug counter.
3221 * Makefile: New dependency.
3223 2013-09-03 Meador Inge <meadori@codesourcery.com>
3227 2013-08-30 Meador Inge <meadori@codesourcery.com>
3229 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3231 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3233 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3234 function descriptor.
3236 2013-09-03 Richard Biener <rguenther@suse.de>
3238 * tree-affine.c (add_elt_to_tree): Fix association issue,
3239 avoid useless converts and make sure to always return a
3240 properly typed result.
3242 2013-09-03 Richard Biener <rguenther@suse.de>
3245 * fold-const.c (negate_expr_p): Fix division case.
3246 (negate_expr): Likewise.
3248 2013-09-03 Richard Biener <rguenther@suse.de>
3251 * tree-streamer-out.c: Include tm.h.
3252 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3254 2013-09-03 Jan Hubicka <jh@suse.cz>
3256 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3258 2013-09-03 Alan Modra <amodra@gmail.com>
3260 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3261 multilibs to big-endian target and vice versa.
3262 * config/rs6000/t-linux64: Use := assignment on all vars.
3263 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3264 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3265 * config/rs6000/t-linux64le: New file.
3266 * config/rs6000/t-linux64bele: New file.
3267 * config/rs6000/t-linux64lebe: New file.
3269 2013-09-02 Jan Hubicka <jh@suse.cz>
3271 * ipa-inline-transform.c (inline_transform): Do not
3272 optimize_inline_calls when not optimizing.
3274 2013-09-02 Jan Hubicka <jh@suse.cz>
3276 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3277 duplicated nodes for assembler names.
3278 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3279 hard registers from assembler name hash.
3281 2013-09-02 Jan Hubicka <jh@suse.cz>
3283 * ipa-split.c (execute_split_functions): Split externally visible
3284 functions called once.
3286 2013-09-02 Martin Jambor <mjambor@suse.cz>
3289 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3290 linked list. When finding the correct duplicate, also consider also
3291 the caller in additon to its inlined_to node.
3293 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3295 * config/aarch64/aarch64-simd-builtins.def
3296 (dup_lane_scalar): Remove.
3297 * config/aarch64/aarch64-simd.md
3298 (aarch64_simd_dup): Add 'w->w' alternative.
3299 (aarch64_dup_lane<mode>): Allow for VALL.
3300 (aarch64_dup_lane_scalar<mode>): Remove.
3301 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3302 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3303 (aarch64_get_lane_unsigned<mode>): Likewise.
3304 (aarch64_get_lane<mode>): Likewise.
3305 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3306 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3307 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3308 (VCON): Change container of V2SF.
3309 (vswap_width_name): Likewise.
3310 * config/aarch64/arm_neon.h
3311 (__aarch64_vdup_lane_any): New.
3312 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3313 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3314 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3316 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3319 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3320 the source or the destination is a single hard register.
3322 2013-09-02 Richard Biener <rguenther@suse.de>
3325 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3328 2013-09-02 Richard Biener <rguenther@suse.de>
3330 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3331 arithmetic to sizetype.
3333 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3335 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3336 Find auto-increment use both before and after candidate.
3338 2013-09-02 Marek Polacek <polacek@redhat.com>
3340 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3342 2013-09-01 Jan Hubicka <jh@suse.cz>
3344 * Makefile.in: Add ipa-profile.o
3345 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3346 * cgraph.c (struct cgraph_propagate_frequency_data,
3347 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3348 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3349 * cgraph.h (cgraph_propagate_frequency): Remove.
3350 * ipa-inline-analysis.c: Include ipa-utils.h;
3351 drop duplicated cfgloop.h.
3352 (inline_update_callee_summaries): Update.
3353 * ipa-profile.c: New file.
3354 * ipa-utils.h (ipa_propagate_frequency): Declare.
3355 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3356 data-streamer.h, value-prof.h.
3357 (symtab_remove_unreachable_nodes): Update profile.
3358 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3359 account_time_size, cmp_counts, dump_histogram,
3360 ipa_profile_generate_summary, ipa_profile_write_summary,
3361 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3362 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3363 Move to ipa-profile.c.
3365 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3367 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3369 2013-09-01 Jan Hubicka <jh@suse.cz>
3371 * common.opt (fdevirtualize-speculatively): New function.
3372 * invoke.texi (fdevirtualize-speculatively): Document.
3373 * ipa-devirt.c: Include ipa-inline.h
3374 (likely_target_p): New function.
3375 (ipa_devirt): New function.
3376 (gate_ipa_devirt): New function.
3377 (pass_data_ipa_devirt): New static var.
3378 (pass_ipa_devirt): Likewise.
3379 (make_pass_ipa_devirt): New function.
3380 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3381 (common_handle_option): Disable devirtualization when
3382 value range profiling is available.
3383 * passes.def (pass_ipa_devirt): Add.
3384 * timever.def (TV_IPA_DEVIRT): New timevar.
3385 * tree-pass.h (make_pass_ipa_devirt):
3387 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3389 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3390 include sanitize(undefined).
3392 2013-08-31 Diego Novillo <dnovillo@google.com>
3394 * Makefile.in (TREE_CORE_H): Define.
3396 (GTFILES): Add tree-core.h.
3397 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3399 * tree-core.h: New file.
3400 Move all data structures, enum, typedefs, global
3401 declarations and constants from ...
3404 2013-08-31 Jan Hubicka <jh@suse.cz>
3406 * bulitins.c (expand_builtin): Do not early exit for gcov
3407 instrumented functions.
3409 2013-08-31 Marek Polacek <polacek@redhat.com>
3411 * ubsan.c: Include tm_p.h.
3413 2013-08-31 Jan Hubicka <jh@suse.cz>
3415 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3418 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3419 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3420 * tree-cfg.c (verify_gimple_label): ... here.
3421 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3422 (ipa_merge_profiles): New function.
3423 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3424 (lto_input_function_body): Likewise.
3425 * ipa-utils.h (ipa_merge_profiles): Declare.
3426 * lto-streamer.h (lto_input_function_body): Update prototype.
3427 (emit_label_in_global_context_p): Remove.
3428 * lto-symtab.c: Include ipa-utils.h
3429 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3431 2013-08-31 Jan Hubicka <jh@suse.cz>
3433 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3435 2013-08-31 Jan Hubicka <jh@suse.cz>
3437 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3439 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3441 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3442 "cmp" RTX before signed_comparison_operator check to account
3445 2013-08-30 Jan Hubicka <jh@suse.cz>
3447 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3448 (detect_type_change_1): Rename to ...
3449 (detect_type_change): ... this one; early return on non-polymorphic
3451 (detect_type_change_ssa): Add comp_type parameter; update
3452 use of detect_type_change.
3453 (compute_complex_assign_jump_func): Add param_type parameter;
3454 update use of detect_type_change_ssa.
3455 (compute_complex_ancestor_jump_func): Likewise.
3456 (ipa_get_callee_param_type): New function.
3457 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3458 update calls to the jump function computation functions.
3460 2013-08-30 Teresa Johnson <tejohnson@google.com>
3461 Steven Bosscher <steven@gcc.gnu.org>
3463 * cfgrtl.c (fixup_new_cold_bb): New routine.
3464 (commit_edge_insertions): Invoke fixup_partitions.
3465 (find_partition_fixes): New routine.
3466 (fixup_partitions): Ditto.
3467 (verify_hot_cold_block_grouping): Update comments.
3468 (rtl_verify_edges): Invoke find_partition_fixes.
3469 (rtl_verify_bb_pointers): Update comments.
3470 (rtl_verify_bb_layout): Ditto.
3471 * basic-block.h (probably_never_executed_edge_p): Declare.
3472 (fixup_partitions): Ditto.
3473 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3474 * bb-reorder.c (sanitize_hot_paths): New function.
3475 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3477 * predict.c (probably_never_executed_edge_p): New routine.
3478 * cfg.c (check_bb_profile): Add partition insanity warnings.
3480 2013-08-30 Meador Inge <meadori@codesourcery.com>
3482 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3484 2013-08-30 Marek Polacek <polacek@redhat.com>
3486 * Makefile.in (ubsan.o): Add.
3487 (c-family/c-ubsan.o): Add.
3488 (builtins.o): Add ubsan.h dependency.
3489 * ubsan.h: New file.
3490 * ubsan.c: New file.
3491 * common.opt: Add -fsanitize=undefined option.
3492 (flag_sanitize): Add variable.
3493 (fsanitize=): Add option. Add Driver.
3494 (fsanitize=thread): Remove option.
3495 (fsanitize=address): Likewise.
3496 (static-libubsan): New option.
3497 * doc/invoke.texi: Document the new flag and -static-libubsan.
3498 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3499 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3500 * builtin-attrs.def (ATTR_COLD): Define.
3501 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3502 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3503 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3504 * flag-types.h (sanitize_code): New enum.
3505 * opts.c (common_handle_option): Parse command line arguments
3506 of -fsanitize=. Add -fsanitize=unreachable option.
3507 * varasm.c (get_variable_section): Adjust.
3508 (assemble_noswitch_variable): Likewise.
3509 (assemble_variable): Likewise.
3510 (output_constant_def_contents): Likewise.
3511 (categorize_decl_for_section): Likewise.
3512 (place_block_symbol): Likewise.
3513 (output_object_block): Likewise.
3514 * builtins.def: Likewise.
3515 * toplev.c (compile_file): Likewise.
3516 (process_options): Likewise.
3517 * cppbuiltin.c: Likewise.
3518 * tsan.c (tsan_pass): Likewise.
3519 (tsan_gate): Likewise.
3520 (tsan_gate_O0): Likewise.
3521 * cfgexpand.c (partition_stack_vars): Likewise.
3522 (expand_stack_vars): Likewise.
3523 (defer_stack_allocation): Likewise.
3524 (expand_used_vars): Likewise.
3525 * cfgcleanup.c (old_insns_match_p): Likewise.
3526 * asan.c (asan_finish_file): Likewise.
3527 (asan_instrument): Likewise.
3528 (gate_asan): Likewise.
3529 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3530 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3531 (asan_global_struct): Use pointer_sized_int_node instead
3532 calling build_nonstandard_integer_type.
3533 (initialize_sanitizer_builtins): Likewise.
3534 (asan_finish_file): Likewise.
3535 * gcc.c: Document %{%:function(args):X}.
3536 (static_spec_functions): Add sanitize.
3537 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3538 store funcval != NULL there.
3539 (do_spec_1): Adjust handle_spec_function caller.
3540 (handle_braces): Allow %:function(args) as condition.
3541 (sanitize_spec_function): New function.
3542 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3543 (LIBUBSAN_SPEC): Likewise.
3544 (LIBUBSAN_EARLY_SPEC): Likewise.
3545 (SANITIZER_SPEC): Handle libubsan.
3546 (SANITIZER_EARLY_SPEC): Likewise.
3547 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3548 instead of fsanitize=address.
3549 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3550 instead of fsanitize=address*.
3551 * builtins.c: Include ubsan.h.
3552 (fold_builtin_0): Instrument __builtin_unreachable.
3553 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3554 instead of flag_asan.
3555 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3556 (pointer_sized_int_node): Define.
3557 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3559 2013-08-30 Mike Stump <mikestump@comcast.net>
3561 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3564 2013-08-29 Jan Hubicka <jh@suse.cz>
3566 * cgraph.c (cgraph_function_body_availability): Handle weakref
3568 * passes.def: Remove pass_fixup_cfg.
3569 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3570 track when we need to remove functions.
3571 (gate_ipa_inline): Execute inlining always; add comment why.
3572 (pass_data_ipa_inline): Remove TODO_remove_functions.
3573 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3574 do not produce summaries.
3575 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3576 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3577 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3578 constant pool and vtable.
3580 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3582 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3583 New arm_neon.h's internal macros to specify 64-bit constants.
3584 Avoid using stdint.h's macros.
3586 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3588 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3589 before applying REG_P.
3591 2013-08-30 Jakub Jelinek <jakub@redhat.com>
3593 PR tree-optimization/58277
3594 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
3595 after seeing too many stmts with vdef in between dombb and current
3596 bb, invalidate everything.
3598 2013-08-30 Richard Biener <rguenther@suse.de>
3600 * fold-const.c (fold_single_bit_test): Fix overflow test.
3602 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
3604 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
3605 and which can live in a register, always retrieve the value on entry.
3606 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
3607 passed by invisible reference specially.
3608 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
3609 (vt_add_function_parameter): Correctly deal with a parameter passed by
3610 invisible reference.
3612 2013-08-30 Jan Hubicka <jh@suse.cz>
3614 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
3616 2013-08-30 Richard Biener <rguenther@suse.de>
3618 PR tree-optimization/58228
3619 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
3620 allow invariant loads in nested loop vectorization.
3622 2013-08-30 Richard Biener <rguenther@suse.de>
3624 PR tree-optimization/58223
3625 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
3626 (has_anti_or_output_dependence): ... this and adjust to also
3627 look for output dependences.
3628 (mark_nodes_having_upstream_mem_writes): Adjust.
3629 (rdg_flag_uses): Likewise.
3631 2013-08-30 Richard Biener <rguenther@suse.de>
3633 PR tree-optimization/58010
3634 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
3635 assert that we have a loop-closed PHI.
3637 2013-08-29 Jan Hubicka <jh@suse.cz>
3639 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
3640 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
3641 * lto-section-in.c (lto_free_function_in_decl_state): New function.
3642 (lto_free_function_in_decl_state_for_node): New function.
3644 2013-08-29 Xinliang David Li <davidxl@google.com>
3646 * loop-unroll.c (report_unroll_peel): Minor message change.
3647 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
3648 Emit alignment peeling message with default -fopt-info.
3649 (vect_loop_versioning): Emit loop version info message.
3650 * tree-vectorizer.c (vectorize_loops): Minor message change.
3651 (execute_vect_slp): Ditto.
3653 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
3655 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
3656 of the clone from the DECL_NAME of the original function.
3658 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
3660 * passes.c (register_pass): Add overload.
3661 * tree-pass.h (register_pass): Forward declare it. Add comment.
3663 2013-08-29 Jan Hubicka <jh@suse.cz>
3665 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
3666 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
3667 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
3668 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3669 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
3670 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
3671 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
3672 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
3674 2013-08-29 Teresa Johnson <tejohnson@google.com>
3676 * dumpfile.c (dump_loc): Output column number.
3677 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
3678 * doc/invoke.texi: Document optall -fopt-info flag.
3679 * profile.c (read_profile_edge_counts): Use new dump framework.
3680 (compute_branch_probabilities): Ditto.
3681 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
3682 when pass not in any opt group.
3683 * pass_manager.h (pass_manager::get_pass_profile): New method.
3684 * value-prof.c (check_counter): Use new dump framework.
3685 (check_ic_target): Ditto.
3686 * coverage.c (get_coverage_counts): Ditto.
3687 (coverage_init): Setup new dump framework.
3689 2013-08-29 Richard Biener <rguenther@suse.de>
3691 PR tree-optimization/58246
3692 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
3693 handle the dominance check inside a basic-block.
3695 2013-08-29 Richard Biener <rguenther@suse.de>
3698 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
3699 of default defs that appear in abnormal PHI nodes.
3701 2013-08-29 Richard Biener <rguenther@suse.de>
3703 PR tree-optimization/57685
3704 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
3705 single-use operands to avoid exponential complexity.
3707 2013-08-28 Dehao Chen <dehao@google.com>
3709 * ipa-inline.c (edge_badness): Fix integer underflow.
3711 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
3713 * gtm-builtins.def (_ITM_free): Declare leaf.
3715 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3718 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
3719 (*tls_local_dynamic_base_64_largepic): Likewise.
3720 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
3721 Remove predicate from call operand.
3722 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
3723 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
3725 2013-08-28 Jeff Law <law@redhat.com>
3727 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
3728 checks for the number of predecessors and successors allowed.
3729 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
3730 which require copying a joiner block if there is a request which
3731 is a subpath that requires no joiner block copying.
3733 2013-08-28 Jan Hubicka <jh@suse.cz>
3735 * lto-streamer-out.c (DFS_write_tree_body): Drop
3736 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
3737 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
3738 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
3739 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
3740 Do not read DECL_ERROR_ISSUED.
3741 (unpack_ts_decl_with_vis_value_fields): Do not read
3743 (lto_input_ts_binfo_tree_pointers): Do not read
3744 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
3745 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
3746 write DECL_ERROR_ISSUED..
3747 (pack_ts_decl_with_vis_value_fields): Do not write
3749 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
3750 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
3751 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
3752 * tree.h (tree_decl_common): Update comment.
3753 (DECL_ERROR_ISSUED): Remove.
3755 2013-08-28 Jakub Jelinek <jakub@redhat.com>
3758 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
3760 2013-08-28 Jan Hubicka <jh@suse.cz>
3762 * builtins.def (free): Declare leaf.
3764 2013-08-27 David Malcolm <dmalcolm@redhat.com>
3767 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
3768 * configure: Regenerate.
3770 2013-08-27 Martin Jambor <mjambor@suse.cz>
3772 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
3773 (ipa_ancestor_jf_data): Likewise.
3774 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
3775 (ipa_get_jf_pass_through_type_preserved): New function.
3776 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
3777 (ipa_get_jf_ancestor_type_preserved): New function.
3778 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
3779 (ipa_get_jf_ancestor_result): Likewise.
3780 (propagate_vals_accross_pass_through): Use
3781 ipa_get_jf_pass_through_result to do all the value mappings.
3782 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
3783 type_preserved flag.
3784 (ipa_set_jf_cst_copy): New function.
3785 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
3786 (ipa_set_jf_arith_pass_through): Likewise.
3787 (ipa_set_ancestor_jf): Likewise.
3788 (compute_complex_assign_jump_func): Set type_preserved instead of
3790 (ipa_compute_jump_functions_for_edge): Likewise.
3791 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
3792 (update_jump_functions_after_inlining): Update type_preserved.
3793 Explicitely create jump functions when combining one with pass_through.
3794 (ipa_write_jump_function): Stream the type_preserved flags.
3795 (ipa_read_jump_function): Likewise.
3797 2013-08-27 Jakub Jelinek <jakub@redhat.com>
3798 Aldy Hernandez <aldyh@redhat.com>
3800 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
3801 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
3802 * function.h (struct function): Add has_force_vect_loops and
3804 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
3805 * gimple.c (gimple_build_omp_critical): Add KIND argument and
3807 * gimple.def: Update CLAUSES comments.
3808 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
3809 (gimple_build_omp_for): Add argument to prototype.
3810 (gimple_omp_for_kind): New.
3811 (gimple_omp_for_set_kind): New.
3812 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
3813 GOVD_DATA_SHARE_CLASS.
3814 (enum omp_region_type): Add ORT_SIMD.
3815 (gimple_add_tmp_var): Handle ORT_SIMD.
3816 (gimplify_var_or_parm_decl): Same.
3817 (is_gimple_stmt): Same.
3818 (omp_firstprivatize_variable): Same.
3819 (omp_add_variable): Only use splay_tree_insert if lookup failed.
3820 (omp_notice_variable): Handle ORT_SIMD.
3821 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
3822 (omp_check_private): Handle ORT_SIMD.
3823 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3825 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
3826 Handle OMP_CLAUSE_LASTPRIVATE.
3827 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3829 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
3830 (gimplify_expr): Handle OMP_SIMD.
3831 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
3832 (expand_GOMP_SIMD_VF): New.
3833 (expand_GOMP_SIMD_LAST_LANE): New.
3834 * internal-fn.def (GOMP_SIMD_LANE): New.
3835 (GOMP_SIMD_VF): New.
3836 (GOMP_SIMD_LAST_LANE): New.
3837 * omp-low.c: Include target.h.
3838 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
3840 (check_omp_nesting_restrictions): Same.
3842 (lower_rec_simd_input_clauses): New.
3843 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
3845 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
3846 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
3847 (expand_omp_build_assign): New.
3848 (expand_omp_for_init_counts): New.
3849 (expand_omp_for_init_vars): New.
3850 (extract_omp_for_update_vars): New.
3851 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
3852 and rewrite accordingly.
3853 (expand_omp_simd): New.
3854 (expand_omp_for): Use expand_omp_simd.
3855 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
3856 (lower_omp_for): Do not lower the body.
3857 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
3859 * tree-flow.h (find_omp_clause): Remove prototype.
3860 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
3861 forcing vectorization of the loop, or if flag_tree_vectorize.
3862 (gate_tree_if_conversion): Similarly.
3863 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
3864 gimple_build_omp_for.
3865 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
3866 * tree-parloops (create_parallel_loop): Pass kind argument to
3867 gimple_build_omp_for.
3868 * tree-pretty-print.c (dump_omp_clause): Add cases for
3869 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
3870 OMP_CLAUSE__SIMDUID_.
3871 (dump_generic_node): Handle OMP_SIMD.
3872 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
3873 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
3874 unroll OMP_SIMD loops here.
3875 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
3876 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
3878 (vect_analyze_data_refs): Handle simd loops.
3879 * tree-vect-loop.c (vectorizable_live_operation): Handle
3881 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
3882 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
3883 (vectorizable_load): Same.
3884 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
3885 (struct simduid_to_vf): New.
3886 (simduid_to_vf::hash): New.
3887 (simduid_to-vf::equal): New.
3888 (struct simd_array_to_simduid): New.
3889 (simd_array_to_simduid::hash): New.
3890 (simd_array_to_simduid::equal): New.
3891 (adjust_simduid_builtins): New.
3892 (struct note_simd_array_uses_struct): New.
3893 (note_simd_array_uses_cb): New.
3894 (note_simd_array_uses): New.
3895 (vectorize_loops): Handle simd hints and adjust simd builtins
3897 * tree-vectorizer.h (struct _stmt_vec_info): Add
3898 simd_lane_access_p field.
3899 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
3900 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
3901 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
3902 (omp_clause_code_name): Same.
3903 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
3904 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
3905 * tree.def (OMP_SIMD): New entry.
3906 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
3907 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
3908 (OMP_CLAUSE_DECL): Adjust range for new clauses.
3909 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
3910 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
3911 (OMP_CLAUSE_LINEAR_STEP): New.
3912 (OMP_CLAUSE_SAFELEN_EXPR): New.
3913 (OMP_CLAUSE__SIMDUID__DECL): New.
3914 (find_omp_clause): New prototype.
3916 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
3918 * config/i386/driver-i386.c (host_detect_local_cpu): Update
3919 Haswell processor detection.
3921 2013-08-27 Christian Widmer <shadow@umbrox.de>
3924 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
3925 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
3926 AVX2 capable processors.
3928 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
3930 * config/aarch64/arm_neon.h: Replace all inline asm implementations
3931 of vget_low_* with implementations in terms of other intrinsics.
3933 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
3936 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
3937 values to -1, 0 and 1.
3939 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
3941 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
3942 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
3943 (<optab><mode>3_insn): Remove.
3944 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
3945 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
3946 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
3947 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
3948 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
3949 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
3950 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
3951 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
3952 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
3953 (ror<mode>3_insn): Likewise.
3954 * config/aarch64/predicates.md (aarch64_simd_register): New.
3956 2013-08-27 Richard Biener <rguenther@suse.de>
3958 PR tree-optimization/57521
3959 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
3960 one edge is non-critical.
3961 (find_phi_replacement_condition): Make sure to use a non-critical
3962 edge. Cleanup and remove old bug workarounds.
3963 (bb_postdominates_preds): Remove.
3964 (if_convertible_loop_p_1): Do not compute post-dominators.
3965 (combine_blocks): Do not free post-dominators.
3966 (main_tree_if_conversion): Likewise.
3967 (pass_data_if_conversion): Add TODO_verify_ssa.
3969 2013-08-27 DJ Delorie <dj@redhat.com>
3971 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
3973 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
3975 * function.c (assign_parm_find_data_types): Set passed_mode and
3976 nominal_mode to the TYPE_MODE of nominal_type for the built
3977 pointer type in case of the struct-pass-by-reference.
3979 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3981 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
3982 (UINT16_TYPE): Change default to "unsigned int".
3984 * config/avr/avr.opt (mfract-convert-truncate): New option.
3985 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
3986 is set, round negative fractional integers according to n1169
3987 when converting to integer types.
3989 2013-08-26 Jan Hubicka <jh@suse.cz>
3991 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
3992 methods can not be called indirectly when their address is not taken.
3994 2013-08-26 Jan Hubicka <jh@suse.cz>
3996 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
3999 2013-08-26 Jan Hubicka <jh@suse.cz>
4001 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4004 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4006 * reload.c (find_valid_class): Allow classes that do not include
4007 FIRST_PSEUDO_REGISTER - 1.
4009 2013-08-26 Jan Hubicka <jh@suse.cz>
4011 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4012 fix edge count/frequency when speculation failed; fix type check
4013 for the direct call.
4015 2013-08-26 Jan Hubicka <jh@suse.cz>
4017 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4019 2013-08-26 Jan Hubicka <jh@suse.cz>
4021 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4022 to be changed by edge redirection.
4024 2013-08-26 Jan Hubicka <jh@suse.cz>
4026 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4027 formating; add sanity check.
4028 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4029 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4030 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4031 (resolve_noninline_speculation): Update callee keys, too.
4033 2013-08-26 Jan Hubicka <jh@suse.cz>
4035 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4036 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4038 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4040 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4043 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4045 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4046 (x86_64_elf_select_section): Ditto.
4047 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4049 (ix86_pass_by_reference): Ditto.
4050 (output_set_got): Ditto.
4051 (ix86_unary_operator_ok): Ditto.
4052 (ix86_expand_builtin): Ditto.
4054 2013-08-23 Jan Hubicka <jh@suse.cz>
4056 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4058 2013-08-23 Jan Hubicka <jh@suse.cz>
4060 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4061 (tree_decl_with_vis): Add FINAL field.
4063 2013-08-23 Jeff Law <law@redhat.com>
4065 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4066 the debugging dump when the expression is fully redundant.
4068 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4070 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4071 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4072 * pretty-print.c (pp_formatted_text_data): Likewise.
4073 (pp_write_text_to_stream): Likewise.
4074 (pp_write_text_as_dot_label_to_stream): Likewise.
4075 (pp_append_r): Likewise.
4076 (pp_format): Likewise.
4077 (pp_flush): Likewise.
4078 (pp_clear_output_area): Likewise.
4079 (pp_append_text): Likewise.
4080 (pp_formatted_text): Likewise.
4081 (pp_remaining_character_count_for_line): Likewise.
4082 (pp_newline): Likewise.
4083 (pp_character): Likewise.
4084 (output_buffer::~output_buffer): Define.
4085 (pretty_printer::~pretty_printer): Destruct output buffer.
4086 * pretty-print.h (output_buffer::~output_buffer): Declare.
4087 (pretty_printer::~pretty_printer): Declare virtual.
4089 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4092 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4093 HOST_WIDE_INT_M1U): New macros.
4094 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4095 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4096 unsigned -1 for lshift.
4097 * cse.c (cse_insn): Likewise.
4098 * double-int.c (rshift_double, lshift_double): Likewise.
4099 * builtins.c (fold_builtin_bitop): Likewise.
4100 * combine.c (force_to_mode): Likewise.
4101 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4102 * simplify-rtx.c (simplify_const_unary_operation,
4103 simplify_const_binary_operation): Likewise.
4104 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4105 check_va_list_escapes): Likewise.
4106 * rtlanal.c (nonzero_bits1): Likewise.
4107 * expmed.c (expand_smod_pow2): Likewise.
4108 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4110 2013-08-23 Jan Hubicka <jh@suse.cz>
4112 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4113 as having address taken.
4115 2013-08-23 Jan Hubicka <jh@suse.cz>
4117 * ipa-utils.h (method_class_type): Declare.
4118 * ipa-devirt.c (method_class_type): Export.
4120 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4121 do not walk base classes of anonymous types.
4123 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4125 PR rtl-optimization/58220
4127 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4128 handle SEQUENCE insns properly.
4130 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4132 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4134 (pp_newline_and_indent): Likewise.
4135 (pp_separate_with): Likewise.
4136 * pretty-print.c (pp_newline_and_flush): Define.
4137 (pp_newline_and_indent): Likewise.
4138 (pp_separate_with): Likewise.
4140 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4143 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4144 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4146 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4148 * config/i386/predicates.md (ext_sse_reg_operand): New.
4149 * config/i386/i386.md (*movti_internal): Use
4150 predicate to determine if EVEX is needed.
4151 (*movsi_internal): Ditto.
4152 (*movdf_internal): Ditto.
4153 (*movsf_internal): Ditto.
4154 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4156 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4158 PR tree-optimization/58209
4159 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4160 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4161 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4162 emit POINTER_PLUS_EXPR.
4163 (create_tailcall_accumulator): For pointer result type accumulate in
4166 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4168 * configure.ac: Add backslashes missing from the last change.
4169 * configure: Regenerate.
4171 2013-08-22 Jan Hubicka <jh@susue.cz>
4173 * ipa.c (function_and_variable_visibility): First remember function
4174 was global and then make it local.
4176 2013-08-22 Julian Brown <julian@codesourcery.com>
4178 * configure.ac: Add aarch64 to list of arches which use "nop" in
4180 * configure: Regenerate.
4182 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4184 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4185 gnu_libc_has_function.
4186 * config/s390/tpf.h: Likewise.
4188 2013-08-22 Jan Hubicka <jh@susue.cz>
4190 * timevar.c (validate_phases): Add cast.
4192 2013-08-22 Jan Hubicka <jh@susue.cz>
4194 * timevar.c (validate_phases): Use size_t for memory.
4195 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4197 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4199 * pretty-print.h (output_buffer::output_buffer): Declare.
4200 (pretty_printer::pretty_printer): Likewise.
4201 (pp_construct): Remove.
4202 * pretty-print.c (output_buffer::output_buffer): Define.
4203 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4204 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4205 (print_gimple_expr): Likewise.
4206 (print_gimple_seq): Likewise.
4207 (gimple_dump_bb): Likewise.
4208 * sched-vis.c (dump_value_slim): Likewise.
4209 (dump_insn_slim): Likewise.
4210 (dump_rtl_slim): Likewise.
4211 (str_pattern_slim): Likewise.
4212 * tree-mudflap.c (mf_varname_tree): Likewise.
4213 * graph.c (print_graph_cfg): Likewise.
4214 (start_graph_dump): Likewise.
4215 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4217 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4218 pretty printer initialization.
4219 * coretypes.h (diagnostic_context): Remove superflous type alias
4221 (pretty_printer): Likewise. Declare directly as a class.
4222 (pretty_print_info): Remove declaration as class.
4223 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4224 and pp_clear_output_area.
4225 (asan_add_global): Likewise.
4227 2013-08-22 Jan Hubicka <jh@suse.cz>
4229 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4230 * ipa-utils.h (update_type_inheritance_graph): Declare.
4231 (possible_polymorphic_call_target_p): Declare.
4232 (possible_polymorphic_call_target_p): New.
4233 * ipa-devirt.c: Update toplevel comments.
4234 (cached_polymorphic_call_targets): Move up.
4235 (odr_type_d): Move ID down.
4236 (polymorphic_type_binfo_p): Update comment.
4237 (odr_hasher::remove): Likewise;
4238 (get_odr_type): Set anonymous_namespace.
4239 (dump_odr_type): Dump it.
4240 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4241 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4242 (record_binfo): Add comment.
4243 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4245 (devirt_node_removal_hook): Do not iCE when cache is freed.
4246 (possible_polymorphic_call_target_p): New predicate.
4247 (update_type_inheritance_graph): New function.
4249 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4250 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4251 Sergey Lega <sergey.s.lega@intel.com>
4252 Anna Tikhonova <anna.tikhonova@intel.com>
4253 Ilya Tocar <ilya.tocar@intel.com>
4254 Andrey Turetskiy <andrey.turetskiy@intel.com>
4255 Ilya Verbin <ilya.verbin@intel.com>
4256 Kirill Yukhin <kirill.yukhin@intel.com>
4257 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4259 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4260 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4261 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4262 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4263 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4264 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4265 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4266 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4267 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4268 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4269 OPT_mavx512pf, OPT_mavx512er cases.
4270 * config/i386/constraints.md (v): New constraint.
4271 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4272 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4273 (bit_AVX512CD): New.
4274 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4275 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4276 * config/i386/i386-c.c (ix86_target_macros_internal):
4277 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4279 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4280 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4281 * config/i386/i386.c (regclass_map, dbx_register_map)
4282 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4283 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4284 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4286 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4287 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4288 -mavx512pf options. Fix formatting.
4289 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4290 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4291 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4292 -mavx512cd, -mavx512pf options.
4293 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4294 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4295 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4297 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4298 (ix86_set_reg_reg_cost): Ditto.
4299 (x86_order_regs_for_local_alloc): Ditto.
4300 (MAX_VECT_LEN): Extend to 64-byte.
4301 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4302 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4303 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4304 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4305 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4306 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4307 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4308 (SSE_REG_MODE_P): Support new modes.
4309 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4310 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4311 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4312 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4313 (REG_CLASS_CONTENTS): Add new registers.
4314 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4315 (EXT_REX_SSE_REGNO_P): New.
4316 (HI_REGISTER_NAMES): Add new registers.
4317 * config/i386/i386.md: Define constants for new registers.
4318 (mode): Add new 512-bit modes.
4319 (prefix): Support evex prefix.
4320 (isa): Support avx512f, noavx512f, fma_avx512f.
4321 (ssemodesuffix): Add new 512-bit modes.
4323 (*movxi_internal_avx512f): Ditto.
4324 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4326 (*movsi_internal): Likewise.
4327 (*movdf_internal): Likewise.
4328 (*movsf_internal): Likewise.
4329 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4330 (<code><mode>3): Likewise.
4331 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4333 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4334 with the new constraint "v".
4335 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4337 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4338 with the new constraint "v".
4339 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4340 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4341 (*<plusminus_insn><mode>3): Likewise.
4342 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4343 (*mul<mode>3): Likewise.
4344 (<sse>_vmmul<mode>3): Likewise.
4345 (<sse>_div<mode>3): Likewise.
4346 (<sse>_vmdiv<mode>3): Likewise.
4347 (<sse>_sqrt<mode>2): Likewise.
4348 (<sse>_vmsqrt<mode>2): Likewise.
4349 (*<code><mode>3_finite): Likewise.
4350 (*<code><mode>3) <smaxmin>: Likewise.
4351 (<sse>_vm<code><mode>3): Likewise.
4352 (*<code><mode>3) <any_logic>: Likewise.
4353 (*fma_fmadd_<mode>): Likewise.
4354 (*fma_fmsub_<mode>): Likewise.
4355 (*fma_fnmadd_<mode>): Likewise.
4356 (*fma_fnmsub_<mode>): Likewise.
4357 (*fma_fmaddsub_<mode>): Likewise.
4358 (*fma_fmsubadd_<mode>): Likewise.
4359 (*fmai_fmadd_<mode>): Likewise.
4360 (*fmai_fmsub_<mode>): Likewise.
4361 (*fmai_fnmadd_<mode>): Likewise.
4362 (*fmai_fnmsub_<mode>): Likewise.
4363 (sse_cvtsi2ss): Likewise.
4364 (sse_cvtsi2ssq): Likewise.
4365 (sse_cvtss2si): Likewise.
4366 (sse_cvtss2si_2): Likewise.
4367 (sse_cvtss2siq): Likewise.
4368 (sse_cvtss2siq_2): Likewise.
4369 (sse_cvttss2si): Likewise.
4370 (sse_cvtss2siq_2): Likewise.
4371 (float<sseintvecmodelower><mode>2): Likewise.
4372 (sse2_cvtsd2si_2): Likewise.
4373 (sse2_cvtsd2siq_2): Likewise.
4374 (*<plusminus_insn><mode>3): Likewise.
4375 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4376 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4377 (ashr<mode>3): Likewise.
4378 (<shift_insn><mode>3): Likewise.
4379 (avx2_<code><mode>3): Likewise.
4380 (*avx2_<code><mode>3): Likewise.
4381 (*andnot<mode>3): Likewise.
4382 (*<code><mode>3) <any_logic>: Likewise.
4383 (abs<mode>2): Likewise.
4384 (avx2_permvar<mode>): Likewise.
4385 (avx2_perm<mode>_1): Likewise.
4386 (*avx_vpermilp<mode>): Likewise.
4387 (avx_vpermilvar<mode>3): Likewise.
4388 (avx2_ashrv<mode>): Likewise.
4389 (avx2_<shift_insn>v<mode>): Likewise.
4390 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4392 * doc/rtl.texi: Document XImode.
4394 2013-08-21 Jeff Law <law@redhat.com>
4396 * tree-flow.h (register_jump_thread): Pass vector of edges
4397 instead of each important edge.
4398 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4399 thread path into a vector and pass that to register_jump_thread.
4400 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4401 passed in edge vector to the current 3-edge form.
4404 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4406 * dce.c (fini_dce): Call df_analyze again just in case
4407 delete_unmarked_insns removed anything.
4409 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4411 * reload.h (struct reg_equivs): Rename to ..
4412 (struct reg_equivs_s): .. this.
4414 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4416 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4418 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4420 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4422 2013-08-21 Jeff Law <law@redhat.com>
4424 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4425 simplify assignments too. If the RHS collapses to a singleton
4426 range, then return the value for the range.
4428 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4430 * config/i386/sse.md (V16): Rename to...
4432 (mov<mode>): Update iterator name.
4433 (*mov<mode>_internal): Ditto.
4434 (push<mode>1): Ditto.
4435 (movmisalign<mode>): Ditto.
4437 2013-08-20 Jan Hubicka <jh@suse.cz>
4440 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4441 entry for direct edges.
4442 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4444 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4446 Revert my last two changes, r201865 and r201864:
4449 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4451 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4452 instances can own GC refs.
4454 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4455 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4456 (gcc::context::gt_pch_nx): Likewise.
4457 (gcc::context::gt_pch_nx): Likewise.
4458 * ggc.h (gt_ggc_mx <T>): New.
4459 (gt_pch_nx_with_op <T>): New.
4460 (gt_pch_nx <T>): New.
4461 * passes.c (opt_pass::gt_ggc_mx): New.
4462 (opt_pass::gt_pch_nx): New.
4463 (opt_pass::gt_pch_nx_with_op): New.
4464 (pass_manager::gt_ggc_mx): New.
4465 (pass_manager::gt_pch_nx): New.
4466 (pass_manager::gt_pch_nx_with_op): New.
4467 (pass_manager::operator new): Use
4468 ggc_internal_cleared_alloc_stat rather than xcalloc.
4469 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4470 (pass_manager::gt_ggc_mx): New.
4471 (pass_manager::gt_pch_nx): New.
4472 (pass_manager::gt_pch_nx_with_op): New.
4473 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4474 (opt_pass::operator new): New.
4475 (opt_pass::gt_ggc_mx): New.
4476 (opt_pass::gt_pch_nx): New.
4477 (opt_pass::gt_pch_nx_with_op): New.
4480 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4482 * Makefile.in (GTFILES): Add context.h.
4483 * context.c (gcc::context::operator new): New.
4484 (gcc::context::gt_ggc_mx): New.
4485 (gcc::context::gt_pch_nx): New.
4486 (gcc::context::gt_pch_nx): New.
4487 * context.h (gcc::context): Add GTY((user)) marking.
4488 (gcc::context::operator new): New.
4489 (gcc::context::gt_ggc_mx): New.
4490 (gcc::context::gt_pch_nx): New.
4491 (gcc::context::gt_pch_nx): New.
4492 (g): Add GTY marking.
4493 (gt_ggc_mx (gcc::context *)): New.
4494 (gt_pch_nx (gcc::context *)): New.
4495 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4496 void *cookie)): New.
4497 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4499 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4501 * dce.c (fini_dce): Call df_analyze again just in case
4502 delete_unmarked_insns removed anything.
4504 2013-08-20 Teresa Johnson <tejohnson@google.com>
4506 PR rtl-optimizations/57451
4507 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4508 from crossing split section boundaries.
4510 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4512 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4513 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4514 with MULTLIB_DEFAULTS.
4516 2013-08-20 Nick Clifton <nickc@redhat.com>
4518 * target.def (narrow_volatile_bitfield): Note that the default
4519 value is false, not !TARGET_STRICT_ALIGN.
4520 * doc/tm.texi: Regenerate.
4522 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4524 Fix LIB_SPEC for systems without libpthread.
4526 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4527 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4529 * config/i386/linux-common.h: Likewise.
4530 * config/mips/linux-common.h: Likewise.
4532 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4534 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4537 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4539 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4540 instances can own GC refs.
4542 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4543 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4544 (gcc::context::gt_pch_nx): Likewise.
4545 (gcc::context::gt_pch_nx): Likewise.
4546 * ggc.h (gt_ggc_mx <T>): New.
4547 (gt_pch_nx_with_op <T>): New.
4548 (gt_pch_nx <T>): New.
4549 * passes.c (opt_pass::gt_ggc_mx): New.
4550 (opt_pass::gt_pch_nx): New.
4551 (opt_pass::gt_pch_nx_with_op): New.
4552 (pass_manager::gt_ggc_mx): New.
4553 (pass_manager::gt_pch_nx): New.
4554 (pass_manager::gt_pch_nx_with_op): New.
4555 (pass_manager::operator new): Use
4556 ggc_internal_cleared_alloc_stat rather than xcalloc.
4557 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4558 (pass_manager::gt_ggc_mx): New.
4559 (pass_manager::gt_pch_nx): New.
4560 (pass_manager::gt_pch_nx_with_op): New.
4561 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4562 (opt_pass::operator new): New.
4563 (opt_pass::gt_ggc_mx): New.
4564 (opt_pass::gt_pch_nx): New.
4565 (opt_pass::gt_pch_nx_with_op): New.
4567 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4569 * Makefile.in (GTFILES): Add context.h.
4570 * context.c (gcc::context::operator new): New.
4571 (gcc::context::gt_ggc_mx): New.
4572 (gcc::context::gt_pch_nx): New.
4573 (gcc::context::gt_pch_nx): New.
4574 * context.h (gcc::context): Add GTY((user)) marking.
4575 (gcc::context::operator new): New.
4576 (gcc::context::gt_ggc_mx): New.
4577 (gcc::context::gt_pch_nx): New.
4578 (gcc::context::gt_pch_nx): New.
4579 (g): Add GTY marking.
4580 (gt_ggc_mx (gcc::context *)): New.
4581 (gt_pch_nx (gcc::context *)): New.
4582 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4583 void *cookie)): New.
4584 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4586 2013-08-20 Alan Modra <amodra@gmail.com>
4589 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
4590 (rs6000_emit_epilogue): Likewise.
4592 2013-08-19 Dehao Chen <dehao@google.com>
4594 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
4596 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
4597 Jakub Jelinek <jakub@redhat.com>
4599 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
4600 (BUILT_IN_FABSD64): Likewise.
4601 (BUILT_IN_FABSD128): Likewise.
4602 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
4603 (fold_builtin_1): Likewise.
4604 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
4605 destination and source operands.
4606 (*abstd2_fpr): Likewise.
4607 (*nabstd2_fpr): Likewise.
4609 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
4611 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
4614 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
4616 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
4619 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
4621 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
4622 * builtins.c (default_libc_has_function): New.
4623 (gnu_libc_has_function): Ditto.
4624 (no_c99_libc_has_function): Ditto.
4625 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
4626 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
4627 (fold_builtin_sincos): Likewise.
4628 (fold_builtin_cexp): Likewise.
4629 * builtins.def (DEF_C94_BUILTIN): Likewise.
4630 (DEF_C99_BUILTIN): Likewise.
4631 (DEF_C99_C90RES_BUILTIN): Likewise.
4632 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
4633 definitions to using this define.
4634 * config/darwin-protos.h (darwin_libc_has_function): New.
4635 * config/darwin.c (darwin_libc_has_function): Ditto.
4636 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
4637 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
4638 * config/darwin.h: Ditto.
4639 * config/elfos.h: Ditto.
4640 * config/freebsd.h: Ditto.
4641 * config/i386/cygming.h: Ditto.
4642 * config/i386/djgpp.h: Ditto.
4643 * config/i386/i386-interix.h: Ditto.
4644 * config/microblaze/microblaze.h: Ditto.
4645 * config/mmix/mmix.h: Ditto.
4646 * config/gnu-user.h: Ditto.
4647 * config/ia64/hpux.h: Ditto.
4648 * config/pa/pa-hpux.h: Ditto.
4649 * config/pdp11/pdp11.h: Ditto.
4650 * config/picochip/picochip.h: Ditto.
4651 * config/linux.h: Ditto.
4652 * config/netbsd.h: Ditto.
4653 * config/openbsd.h: Ditto.
4654 * config/rs6000/aix43.h: Ditto.
4655 * config/rs6000/aix51.h: Ditto.
4656 * config/rs6000/aix52.h: Ditto.
4657 * config/rs6000/aix53.h: Ditto.
4658 * config/rs6000/aix61.h: Ditto.
4659 * config/rs6000/darwin.h: Ditto.
4660 * config/rs6000/linux.h: Ditto.
4661 * config/rs6000/linux64.h: Ditto.
4662 * config/s390/tpf.h: Ditto.
4663 * config/sol2-10.h: Ditto.
4664 * config/sol2.h: Ditto.
4665 * config/vms/vms.h: Ditto.
4666 * config/vxworks.h: Ditto.
4667 * config/linux-android.c (linux_android_libc_has_function):
4668 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
4669 * config/linux-protos.h (linux_android_libc_has_function):
4671 * config/i386/i386.c (ix86_libc_has_function): New.
4672 * config/i386/i386-protos.h
4673 (ix86_libc_has_function): New declaration.
4674 * config/i386/i386.md
4675 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
4676 ("isinf<mode>2): Likewise.
4677 * convert.c (convert_to_integer): Using new target hook
4678 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
4679 TARGET_C99_FUNCTIONS.
4680 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
4681 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
4682 * coretypes.h (function_class): New enum for different
4683 classes of functions.
4684 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
4685 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
4686 (TARGET_HAS_SINCOS): Likewise.
4687 (TARGET_LIBC_HAS_FUNCTION): New.
4688 * doc/tm.texi: Regenerated.
4689 * targhooks.h (default_libc_has_function): New declaration.
4690 (no_c99_libc_has_function): Ditto.
4691 (gnu_libc_has_function): Ditto.
4692 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
4693 and TARGET_HAS_SINCOS.
4695 2013-08-18 Jan Hubicka <jh@suse.cz>
4697 * Makeifle-in (ipa-devirt.o): New.
4698 (GTFILES): Add ipa-utils.h and ipa-devirt.c
4699 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
4700 (analyze_functions): Look into possible targets of polymorphic call.
4701 * dumpfile.c (dump_files): Add type-inheritance dump.
4702 * dumpfile.h (TDI_inheritance): New.
4703 * ipa-devirt.c: New file.
4704 * ipa-utils.h (odr_type_d): Forward declare.
4705 (odr_type): New type.
4706 (build_type_inheritance_graph): Declare.
4707 (possible_polymorphic_call_targets): Declare and introduce inline
4708 variant when only edge is pased.
4709 (dump_possible_polymorphic_call_targets): Likewise.
4710 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
4711 * tree.c (type_in_anonymous_namespace_p): Break out from ...
4712 (types_same_for_odr): ... here.
4713 * tree.h (type_in_anonymous_namespace_p): Declare.
4715 2013-08-18 Jakub Jelinek <jakub@redhat.com>
4717 PR tree-optimization/58006
4718 * tree-parloops.c (take_address_of): Don't ICE if get_name
4720 (eliminate_local_variables_stmt): Remove clobber stmts.
4722 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
4724 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
4725 error message is issued for an alias to undefined symbol.
4727 2013-08-18 Jan Hubicka <jh@suse.cz>
4729 * cgraph.c (cgraph_create_indirect_edge): Discover
4730 polymorphic calls and record basic info into indirect_info.
4731 * gimple-fold.c (gimple_fold_call): When doing BINFO based
4732 devirtualization, ignore objc function calls.
4733 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
4734 call with no parm index info.
4735 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
4736 * tree.c (virtual_method_call_p): New function.
4737 * tree.h (virtual_method_call_p): Declare.
4739 2013-08-16 Jan Hubicka <jh@suse.cz>
4742 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
4744 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
4746 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
4749 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4751 * gengtype.c (type_for_name): Add special-case support for
4752 locating types within the "gcc::" namespace.
4753 (open_base_files): Emit a "using namespace gcc" directive.
4755 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
4758 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
4759 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
4761 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
4762 array instead of each individual operand as a separate argument.
4763 (emit_fusion_gpr_load): Likewise.
4764 (expand_fusion_gpr_load): Add new function declaration.
4766 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
4767 signature to have the operands passed as an array, instead of as
4768 separate arguments. Allow ZERO_EXTEND to be in the memory
4769 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
4770 depend on the register live/dead flags when peepholes are run.
4771 (expand_fusion_gpr_load): New function to be called from the
4772 peephole2 pass, to change the register that addis sets to be the
4774 (emit_fusion_gpr_load): Change the calling signature to have the
4775 operands passed as an array, instead of as separate arguments.
4776 Allow ZERO_EXTEND to be in the memory address, and also
4777 SIGN_EXTEND if -mpower8-fusion-sign.
4779 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
4781 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
4782 adjust the register addis loads up in the peephole2 pass. Do not
4783 depend on the register live/dead state when the peephole pass is done.
4785 2013-08-16 David Malcolm <dmalcolm@redhat.com>
4787 * gengtype.c (create_user_defined_type): Ensure that the kind
4788 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
4789 declaration is seen before the GTY((user)) marking.
4791 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
4794 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
4796 2013-08-16 Jan Hubicka <jh@suse.cz>
4798 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
4799 arugment expected_type.
4800 (gimple_fold_call): Use it.
4801 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
4802 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
4803 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
4804 (try_make_edge_direct_virtual_call): Likewise.
4805 * tree.c (obj_type_ref_class): New.
4806 * tree.h (obj_type_ref_class): Use it.
4808 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
4810 * sched-vis.c (rtl_slim_pp_initialized): Remove.
4811 (rtl_slim_pp): Likewise.
4812 (init_rtl_slim_pretty_print): Likewise.
4813 (dump_value_slim): Don't call it. Use local pretty printer.
4814 (dump_insn_slim): Likewise.
4815 (dump_rtl_slim): Likewise.
4816 (str_pattern_slim): Likewise.
4817 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
4820 2013-08-16 Jakub Jelinek <jakub@redhat.com>
4822 PR tree-optimization/58164
4823 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
4824 walk gimple_goto_dest of GIMPLE_GOTO.
4826 PR tree-optimization/58165
4827 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
4828 bi_call must be the last stmt in a bb, don't split_block, instead
4829 use fallthru edge from it and give up if there is none.
4830 Release conds vector when returning early.
4832 2013-08-14 Xinliang David Li <davidxl@google.com>
4834 * config/i386/i386.c (ix86_option_override_internal):
4835 Remove unused variable and field.
4837 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4840 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
4841 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
4842 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
4843 and Linux, correct BLKmode alignment when 128-bit alignment is
4844 required and compatibility flag is not set.
4845 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
4846 for zero-size arguments when compatibility flag is not set.
4848 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4850 PR tree-optimization/58145
4851 * tree-sra.c (build_ref_for_offset): If prev_base has
4852 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
4854 2013-08-14 Xinliang David Li <davidxl@google.com>
4856 * config/i386/i386.c (ix86_option_override_internal):
4857 Fix uninitialized variable error.
4859 2013-08-14 Xinliang David Li <davidxl@google.com>
4861 * config/i386/i386.opt: Define two new options.
4862 * config/i386/x86-tune.def: Add arch selector field in macros.
4863 * config/i386/i386.h: Adjust macro definition.
4864 * config/i386/i386.c (ix86_option_override_internal):
4866 (parse_mtune_ctrl_str): New function.
4867 (set_ix86_tune_features): New function.
4868 (ix86_function_specific_restore): Call the new helper function.
4870 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
4872 PR rtl-optimization/57662
4873 * sel-sched.c (code_motion_process_successors): When the current insn
4874 is removed after the recursive traversal, break from the loop.
4875 Add comments and debug printouts.
4877 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4878 Alexandre Oliva <aoliva@redhat.com>
4881 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
4882 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
4883 there also UNSPEC_PLTOFF.
4885 2013-08-14 Marek Polacek <polacek@redhat.com>
4887 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
4888 NUM_CONDITIONS bit positions.
4890 2013-08-13 Cary Coutant <ccoutant@google.com>
4892 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
4893 (attr_checksum): Hash vector contents instead of pointer.
4894 (attr_checksum_ordered): Likewise.
4896 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
4898 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
4899 when Pmode != word_mode. Add length_address attribute.
4900 (sse3_monitor_<mode>): Merge from sse3_monitor and
4901 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
4902 Pmode != word_mode. Update insn length attribute.
4903 * config/i386/i386.c (ix86_option_override_internal): Update
4904 ix86_gen_monitor selection for merged sse3_monitor insn.
4906 2013-08-13 Julian Brown <julian@codesourcery.com>
4908 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
4909 perform invalid legitimization on greater-than-word-size modes for
4912 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
4914 * ira.c (setup_class_translate_array): Use aclass instead of cl
4915 for classes not fully covered by allocno classes.
4917 2013-08-13 Jakub Jelinek <jakub@redhat.com>
4919 PR tree-optimization/57661
4920 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
4921 * tree-inline.c (tree_function_versioning): Initialize it.
4922 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
4923 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
4924 that is not being copied.
4927 * asan.c (instrument_strlen_call): Fix typo in comment.
4928 Use char * type even for the lhs of POINTER_PLUS_EXPR.
4930 2013-08-13 Steve Ellcey <sellcey@mips.com>
4932 * config/mips/mips.md (prefetch): Use lw instead of ld on
4933 loongson in 32bit mode.
4935 2013-08-13 Nick Clifton <nickc@redhat.com>
4937 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
4939 2013-08-13 Jan Hubicka <jh@suse.cz>
4941 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
4942 introduced edge; fix typo in sanity check.
4943 (cgraph_resolve_speculation): Export; improve diagnostic.
4944 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
4945 speculation at type mismatch.
4946 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
4947 (cgraph_resolve_speculation): Declare.
4948 (symtab_can_be_discarded): New function.
4949 * value-prof.c (gimple_ic_transform): Remove actual transform code.
4950 * ipa-inline-transform.c (speculation_removed): New global var.
4951 (clone_inlined_nodes): See if speculation can be removed.
4952 (inline_call): If speculations was removed, we growths may not match.
4953 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
4954 (speculation_useful_p): New function.
4955 (resolve_noninline_speculation): New function.
4956 (inline_small_functions): Resolve useless speculations.
4957 * ipa-inline.h (speculation_useful_p): Declare
4958 * ipa.c (can_replace_by_local_alias): Simplify.
4959 (ipa_profile): Produce speculative calls in non-lto, too;
4960 add simple cost model; produce local aliases.
4962 2013-08-13 David Malcolm <dmalcolm@redhat.com>
4964 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
4967 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
4969 * config/i386/i386.c (ix86_function_versions): Use error + inform.
4971 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
4973 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
4974 iterator instead of X87MODEF.
4976 2013-08-12 Perez Read <netfirewall@gmail.com>
4979 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
4980 operand 0 for intel asm alternative.
4981 (*movabs<mode>_2): Ditto for operand 1.
4983 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
4985 * config/aarch64/arm_none.h
4986 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
4988 2013-08-12 Nick Clifton <nickc@redhat.com>
4990 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
4992 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4994 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
4995 expand for QI/HImode operand to produce more effictive code for
4996 unsigned char(short) --> float(double) conversion.
4998 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5000 * doc/invoke.texi: Mention that -ftls-model does not force the final
5003 2013-08-12 Marek Polacek <polacek@redhat.com>
5004 Marc Glisse <marc.glisse@inria.fr>
5006 PR tree-optimization/57980
5007 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5008 when creating -1 constant.
5010 2013-08-10 Jan Hubicka <jh@suse.cz>
5012 Workaround binutils PR14342.
5013 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5014 (gimple_init_edge_profiler): Likewise.
5015 (gimple_gen_ic_func_profiler): Likewise.
5017 2013-08-09 Jan Hubicka <jh@suse.cz>
5019 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5021 2013-08-09 Xinliang David Li <davidxl@google.com>
5023 * config/i386/stringop.def: New file.
5024 * config/i386/stringop.opt: New file.
5025 * config/i386/i386-opts.h: Include stringopt.def.
5026 * config/i386/i386.opt: Include stringopt.opt.
5027 * config/i386/i386.c (ix86_option_override_internal):
5028 Override default size based stringop inline strategies with options.
5029 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5032 2013-08-09 Jan Hubicka <jh@suse.cz>
5034 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5036 2013-08-09 Jan Hubicka <jh@suse.cz>
5038 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5040 (dump_cgraph_node): Dump profile-id.
5041 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5042 and common_target_probability.
5043 * lto-cgraph.c (lto_output_edge): Stream common targets.
5044 (lto_output_node): Stream profile ids.
5045 (input_node): Stream profile ids.
5046 (input_edge): Stream common targets.
5047 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5048 * ipa.c: Include value-prof.h
5049 (ipa_profile_generate_summary): Turn indirect call statement histograms
5050 into common targets.
5051 (ipa_profile): Turn common targets into speculative edges.
5053 2013-08-09 Jan Hubicka <jh@suse.cz>
5055 * cgraph.h (cgraph_node): Add profile_id.
5056 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5057 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5058 (del_node_map): Update.
5059 (find_func_by_funcdef_no): Replace by ...
5060 (find_func_by_profile_id): ... this one.
5061 (gimple_ic_transform): Do not remove useful histograms when
5062 speculation is not done; dump info when indirect call removal
5064 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5065 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5066 (__gcov_indirect_call_profiler_v2): .. this one.
5067 * profile.h (init_node_map): Update.
5068 * coverage.c (coverage_compute_profile_id): New function.
5069 * coverage.h (coverage_compute_profile_id): Declare.
5070 * tree-profile.c (init_ic_make_global_vars): Make
5071 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5072 (gimple_init_edge_profiler): Update prototype of
5073 __gcov_indirect_call_profiler.
5074 (gimple_gen_ic_func_profiler): Simplify.
5075 (tree_profiling): Use init_node_map
5077 2013-08-09 Jan Hubicka <jh@suse.cz>
5079 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5080 non-speculative refs.
5081 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5082 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5083 (cgraph_set_call_stmt): Likewise.
5084 (cgraph_create_edge_1): Fix release checking compilatoin;
5086 (cgraph_free_edge): Free indirect info.
5087 (cgraph_turn_edge_to_speculative): New function.
5088 (cgraph_speculative_call_info): New function.
5089 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5090 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5091 (dump_cgraph_node): Dump speculation.
5092 (verify_edge_count_and_frequency): Accept speculative edges.
5093 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5094 (verify_cgraph_node): Handle speculation.
5095 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5096 (cgraph_set_call_stmt): Update prototype.
5097 (cgraph_make_edge_direct): Update prototype.
5098 (cgraph_speculative_call_info): Declare.
5099 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5100 to change; update call of ipa_find_references.
5101 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5102 lto_stmt_uid and speculative flags.
5103 (ipa_dump_references): Dump speculation.
5104 (ipa_clone_references): Clone speculative flag.
5105 (ipa_clone_referring): Likewise.
5106 (ipa_clone_ref): New function.
5107 (ipa_find_reference): Look into lto_stmt_uids
5108 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5109 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5110 (ipa_find_reference): Update declaration.
5111 (ipa_clone_ref): Declare.
5112 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5113 stream speculative flag.
5114 (lto_output_ref): Stream statements uids and speculation.
5115 (input_ref): Likewise.
5116 (input_edge): Stream speuclation.
5117 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5118 (cgraph_set_call_stmt_including_clones): Handle speculation.
5119 * ipa-inline.c (heap_edge_removal_hook): New function.
5120 (inline_small_functions): Register it.
5121 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5122 also initialize refs.
5123 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5125 (try_make_edge_direct_simple_call): Likewise.
5126 (try_make_edge_direct_simple_call): Likewise.
5127 (update_indirect_edges_after_inlining): Likewise.
5128 (remove_described_reference): Look proper lto_stmt_uid.
5129 (propagate_controlled_uses): Likewise.
5130 (propagate_controlled_uses): Liekwise.
5131 * tree-inline.c (copy_bb): Copy speculative edges.
5132 (redirect_all_calls): New function.
5133 (copy_cfg_body): Do redirection after loop info is updated.
5134 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5136 2013-08-09 Jan Hubicka <jh@suse.cz>
5138 * lto-streamer-out.c (output_function): Renumber PHIs.
5139 * lto-streamer-in.c (input_function): Likewise.
5141 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5143 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5144 (get_lane_unsigned): Likewise.
5145 (dup_lane_scalar): Likewise.
5146 (get_lane): enable for VALL.
5147 * config/aarch64/aarch64-simd.md
5148 (aarch64_dup_lane_scalar<mode>): Remove.
5149 (aarch64_get_lane_signed<mode>): Likewise.
5150 (aarch64_get_lane_unsigned<mode>): Likewise.
5151 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5152 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5153 (aarch64_get_lane<mode>): Enable for all vector modes.
5154 (aarch64_get_lanedi): Remove misleading constraints.
5155 * config/aarch64/arm_neon.h
5156 (__aarch64_vget_lane_any): Define.
5157 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5158 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5159 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5160 * config/aarch64/iterators.md (VDQQH): New.
5164 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5166 * configure.ac: Add GAS check for LEON instructions on SPARC.
5167 * configure: Regenerate.
5168 * config.in: Likewise.
5169 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5171 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5172 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5174 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5175 (AS_LEON_FLAG): New macro.
5176 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5177 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5178 Deal with LEON and LEON3 for the memory model.
5179 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5180 (atomic_compare_and_swap<mode>_1): Likewise.
5181 (*atomic_compare_and_swap<mode>_1): Likewise.
5183 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5185 * config/arm/neon.md (vcond): Fix floating-point vector
5186 comparisons against 0.
5188 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5190 * lra-constraints.c (emit_spill_move): Remove assert.
5191 (process_alt_operands): Add more debugging
5192 output. Increase reject for spilling into memory. Decrease
5193 reject for reloading scratch.
5194 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5196 2013-08-08 Steve Ellcey <sellcey@mips.com>
5198 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5199 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5200 micromips incompatible. Add nan2008.
5201 (MULTILIB_DIRNAMES): Add nan2008.
5202 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5203 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5204 and micromips incompatible. Add nan2008.
5205 (MULTILIB_DIRNAMES): Add nan2008.
5206 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5208 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5210 PR rtl-optimization/58079
5211 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5212 simplify_comparison has widened a comparison with an integer.
5214 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5216 * config/arm/neon.md (movmisalign<mode>): Disable when we
5217 don't allow unaligned accesses.
5218 (*movmisalign<mode>_neon_store): Likewise.
5219 (*movmisalign<mode>_neon_load): Likewise.
5220 (*movmisalign<mode>_neon_store): Likewise.
5221 (*movmisalign<mode>_neon_load): Likewise.
5223 2013-08-08 Jan Hubicka <jh@suse.cz>
5225 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5226 (make_pass_rebuild_cgraph_edges): Also clear references.
5227 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5228 * ipa-inline-transform.c (inline_transform): Remove all references
5230 * cgraphunit.c (expand_function): Remove all references after
5232 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5233 (ipa_find_reference): Rewrite to iterator.
5234 (remove_stmt_references): Likewise.
5235 (ipa_clear_stmts_in_references): New function.
5236 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5237 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5239 * ipa-split.c (split_function): Remove references in split function.
5241 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5244 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5245 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5247 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5250 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5251 suggested mode for the assignment isn't compatible with the
5254 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5257 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5259 2013-08-07 Xinliang David Li <davidxl@google.com>
5261 * config/i386/i386.opt: New option -mtune-ctrl=.
5262 * config/i386/x86-tune.def: New file.
5263 * config/i386/i386.h: include x86-tune.def.
5264 * config/i386/i386.c (ix86_option_override_internal):
5265 Parsing -mtune-ctrl= option and set tune features.
5267 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5270 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5271 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5274 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5276 * rtl.h (update_alignments): Declare.
5277 * final.c (grow_label_align): New function extracted from...
5278 (shorten_branches): ...here. Call it.
5279 (update_alignments): New function.
5280 * reorg.c (sibling_labels): New variable.
5281 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5282 the new label along with it onto the sibling_labels vector.
5283 (fill_simple_delay_slots): Adjust call to get_label_before.
5284 (fill_slots_from_thread): Likewise.
5285 (relax_delay_slots): Likewise.
5286 (make_return_insns): Likewise.
5287 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5289 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5291 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5292 document its semantics.
5293 (diagnostic_report_diagnostic): Adjust accordingly.
5295 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5297 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5298 (sparc_option_override): ...and port to new C++ pass API.
5299 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5301 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5303 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5305 2013-08-06 Caroline Tice <cmtice@google.com>
5307 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5308 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5309 * tree-pass.h: Add pass_vtable_verify.
5310 * varasm.c (assemble_variable): Add code to properly set the comdat
5311 section and name for the .vtable_map_vars section.
5312 (assemble_vtyv_preinit_initializer): New function.
5313 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5315 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5316 * vtable-verify.c: New file.
5317 * vtable-verify.h: New file.
5318 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5319 initialiation levels.
5320 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5321 * passes.def: Insert pass_vtable_verify.
5322 * aclocal.m4: Reorder includes.
5323 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5324 -fvtv-counts options.
5325 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5326 as appropriate, if -fvtable-verify=... is used.
5327 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5328 -fvtable-verify=... is used.
5329 * Makefile.in (OBJS): Add vtable-verify.o to list.
5330 (vtable-verify.o): Add new build rule.
5331 (GTFILES): Add vtable-verify.c to list.
5332 * common.opt (fvtable-verify=): New flag.
5333 (vtv_priority): Values for fvtable-verify= flag.
5334 (fvtv-counts): New flag.
5335 (fvtv-debug): New flag.
5336 * tree.h (save_vtable_map_decl): New extern function decl.
5338 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5340 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5341 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5342 (pass_data_rl78_devirt): ...new pass_data instance and...
5343 (make_pass_rl78_devirt): ...new function.
5344 (rl78_asm_file_start): Port pass registration to new C++ API.
5346 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5348 * coretypes.h (rtl_opt_pass): Add.
5349 (gcc::context): Add.
5350 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5351 (epiphany_init): Port to new C++ pass API.
5352 (epiphany_optimize_mode_switching): Likewise.
5353 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5354 (pass_manager::get_pass_mode_switching): New.
5355 (pass_manager::get_pass_peephole2): New.
5356 * mode-switching.c (pass_mode_switching): Add clone method.
5357 * recog.c (pass_peephole2): Add clone method.
5358 (pass_split_all_insns): Add clone method.
5360 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5362 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5363 (mips_option_override): ...here, porting to new C++ API for passes.
5365 2013-08-06 Jan Hubicka <jh@suse.cz>
5367 * cgraph.c (cgraph_get_body): New function based on lto.c
5369 * cgraph.h (cgraph_get_body): Declare.
5370 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5372 * cgraphunit.c (expand_function): Get body prior expanding.
5373 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5374 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5376 * passes.c (do_per_function_toporder): Get body.
5377 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5378 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5380 2013-08-06 Martin Jambor <mjambor@suse.cz>
5383 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5384 is not re-finalized. Rename second parameter to no_collect.
5386 2013-08-06 Martin Jambor <mjambor@suse.cz>
5389 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5390 MEM_REF has proper alignment information.
5392 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5395 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5397 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5398 rtx (*) (rtx, ...) with insn_gen_fn.
5399 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5400 insn_gen_fn::stored_funcptr. Add initializer braces.
5402 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5404 Rewrite how instances of passes are cloned to remove assumptions
5405 about their sizes (thus allowing pass subclasses to have
5406 additional data fields, albeit non-GC-managed ones at this point).
5408 * passes.c (make_pass_instance): Now that passes have clone
5409 methods, rewrite this function to eliminate XNEW and memcpy
5410 calls that used hardcoded sizes. Since this function no longer
5411 creates pass instances, rename it to...
5412 (add_pass_instance): ...this. Document the old way that passes were
5413 numbered and flagged, and rework this function to continue using it.
5414 (next_pass_1): Add an initial_pass argument for use by
5416 (position_pass): When adding multiple instances of a pass, use
5417 the pass's clone method, rather than relying on the XNEW/memcpy
5418 within the former make_pass_instance (now add_pass_instance).
5419 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5420 the initial instance of the current pass within the pass manager.
5422 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5424 This is the automated part of the conversion of passes from C
5425 structs to C++ classes.
5427 Patch autogenerated by refactor_passes.py from
5428 https://github.com/davidmalcolm/gcc-refactoring-scripts
5429 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5431 * asan.c (pass_asan): Convert from a global struct to a subclass of
5432 gimple_opt_pass along with...
5433 (pass_data_asan): ...new pass_data instance and...
5434 (make_pass_asan): ...new function.
5435 (pass_asan_O0): Convert from a global struct to a subclass of
5436 gimple_opt_pass along with...
5437 (pass_data_asan_O0): ...new pass_data instance and...
5438 (make_pass_asan_O0): ...new function.
5439 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5440 subclass of rtl_opt_pass along with...
5441 (pass_data_inc_dec): ...new pass_data instance and...
5442 (make_pass_inc_dec): ...new function.
5443 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5444 a subclass of rtl_opt_pass along with...
5445 (pass_data_reorder_blocks): ...new pass_data instance and...
5446 (make_pass_reorder_blocks): ...new function.
5447 (pass_duplicate_computed_gotos): Convert from a global struct to a
5448 subclass of rtl_opt_pass along with...
5449 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5450 (make_pass_duplicate_computed_gotos): ...new function.
5451 (pass_partition_blocks): Convert from a global struct to a subclass of
5452 rtl_opt_pass along with...
5453 (pass_data_partition_blocks): ...new pass_data instance and...
5454 (make_pass_partition_blocks): ...new function.
5455 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5456 struct to a subclass of rtl_opt_pass along with...
5457 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5459 (make_pass_branch_target_load_optimize1): ...new function.
5460 (pass_branch_target_load_optimize2): Convert from a global struct to a
5461 subclass of rtl_opt_pass along with...
5462 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5464 (make_pass_branch_target_load_optimize2): ...new function.
5465 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5466 of rtl_opt_pass along with...
5467 (pass_data_jump): ...new pass_data instance and...
5468 (make_pass_jump): ...new function.
5469 (pass_jump2): Convert from a global struct to a subclass of
5470 rtl_opt_pass along with...
5471 (pass_data_jump2): ...new pass_data instance and...
5472 (make_pass_jump2): ...new function.
5473 * cfgexpand.c (pass_expand): Convert from a global struct to a
5474 subclass of rtl_opt_pass along with...
5475 (pass_data_expand): ...new pass_data instance and...
5476 (make_pass_expand): ...new function.
5477 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5478 of rtl_opt_pass along with...
5479 (pass_data_free_cfg): ...new pass_data instance and...
5480 (make_pass_free_cfg): ...new function.
5481 (pass_into_cfg_layout_mode): Convert from a global struct to a
5482 subclass of rtl_opt_pass along with...
5483 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5484 (make_pass_into_cfg_layout_mode): ...new function.
5485 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5486 subclass of rtl_opt_pass along with...
5487 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5488 (make_pass_outof_cfg_layout_mode): ...new function.
5489 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5490 struct to a subclass of gimple_opt_pass along with...
5491 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5492 (make_pass_build_cgraph_edges): ...new function.
5493 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5494 subclass of gimple_opt_pass along with...
5495 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5496 (make_pass_rebuild_cgraph_edges): ...new function.
5497 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5498 subclass of gimple_opt_pass along with...
5499 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5501 (make_pass_remove_cgraph_callee_edges): ...new function.
5502 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5503 struct to a subclass of rtl_opt_pass along with...
5504 (pass_data_stack_adjustments): ...new pass_data instance and...
5505 (make_pass_stack_adjustments): ...new function.
5506 * combine.c (pass_combine): Convert from a global struct to a subclass
5507 of rtl_opt_pass along with...
5508 (pass_data_combine): ...new pass_data instance and...
5509 (make_pass_combine): ...new function.
5510 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5511 global struct to a subclass of rtl_opt_pass along with...
5512 (pass_data_compare_elim_after_reload): ...new pass_data instance
5514 (make_pass_compare_elim_after_reload): ...new function.
5515 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5516 of rtl_opt_pass along with...
5517 (pass_data_rtl_cprop): ...new pass_data instance and...
5518 (make_pass_rtl_cprop): ...new function.
5519 * cse.c (pass_cse): Convert from a global struct to a subclass of
5520 rtl_opt_pass along with...
5521 (pass_data_cse): ...new pass_data instance and...
5522 (make_pass_cse): ...new function.
5523 (pass_cse2): Convert from a global struct to a subclass of
5524 rtl_opt_pass along with...
5525 (pass_data_cse2): ...new pass_data instance and...
5526 (make_pass_cse2): ...new function.
5527 (pass_cse_after_global_opts): Convert from a global struct to a
5528 subclass of rtl_opt_pass along with...
5529 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5530 (make_pass_cse_after_global_opts): ...new function.
5531 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5532 of rtl_opt_pass along with...
5533 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5534 (make_pass_ud_rtl_dce): ...new function.
5535 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5536 rtl_opt_pass along with...
5537 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5538 (make_pass_fast_rtl_dce): ...new function.
5539 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5540 a subclass of rtl_opt_pass along with...
5541 (pass_data_df_initialize_opt): ...new pass_data instance and...
5542 (make_pass_df_initialize_opt): ...new function.
5543 (pass_df_initialize_no_opt): Convert from a global struct to a
5544 subclass of rtl_opt_pass along with...
5545 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5546 (make_pass_df_initialize_no_opt): ...new function.
5547 (pass_df_finish): Convert from a global struct to a subclass of
5548 rtl_opt_pass along with...
5549 (pass_data_df_finish): ...new pass_data instance and...
5550 (make_pass_df_finish): ...new function.
5551 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5552 rtl_opt_pass along with...
5553 (pass_data_rtl_dse1): ...new pass_data instance and...
5554 (make_pass_rtl_dse1): ...new function.
5555 (pass_rtl_dse2): Convert from a global struct to a subclass of
5556 rtl_opt_pass along with...
5557 (pass_data_rtl_dse2): ...new pass_data instance and...
5558 (make_pass_rtl_dse2): ...new function.
5559 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5560 subclass of rtl_opt_pass along with...
5561 (pass_data_dwarf2_frame): ...new pass_data instance and...
5562 (make_pass_dwarf2_frame): ...new function.
5563 * except.c (pass_set_nothrow_function_flags): Convert from a global
5564 struct to a subclass of rtl_opt_pass along with...
5565 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5567 (make_pass_set_nothrow_function_flags): ...new function.
5568 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5569 subclass of rtl_opt_pass along with...
5570 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5572 (make_pass_convert_to_eh_region_ranges): ...new function.
5573 * final.c (pass_compute_alignments): Convert from a global struct to a
5574 subclass of rtl_opt_pass along with...
5575 (pass_data_compute_alignments): ...new pass_data instance and...
5576 (make_pass_compute_alignments): ...new function.
5577 (pass_final): Convert from a global struct to a subclass of
5578 rtl_opt_pass along with...
5579 (pass_data_final): ...new pass_data instance and...
5580 (make_pass_final): ...new function.
5581 (pass_shorten_branches): Convert from a global struct to a subclass of
5582 rtl_opt_pass along with...
5583 (pass_data_shorten_branches): ...new pass_data instance and...
5584 (make_pass_shorten_branches): ...new function.
5585 (pass_clean_state): Convert from a global struct to a subclass of
5586 rtl_opt_pass along with...
5587 (pass_data_clean_state): ...new pass_data instance and...
5588 (make_pass_clean_state): ...new function.
5589 * function.c (pass_instantiate_virtual_regs): Convert from a global
5590 struct to a subclass of rtl_opt_pass along with...
5591 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
5592 (make_pass_instantiate_virtual_regs): ...new function.
5593 (pass_leaf_regs): Convert from a global struct to a subclass of
5594 rtl_opt_pass along with...
5595 (pass_data_leaf_regs): ...new pass_data instance and...
5596 (make_pass_leaf_regs): ...new function.
5597 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
5598 subclass of rtl_opt_pass along with...
5599 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
5601 (make_pass_thread_prologue_and_epilogue): ...new function.
5602 (pass_match_asm_constraints): Convert from a global struct to a
5603 subclass of rtl_opt_pass along with...
5604 (pass_data_match_asm_constraints): ...new pass_data instance and...
5605 (make_pass_match_asm_constraints): ...new function.
5606 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
5607 subclass of rtl_opt_pass along with...
5608 (pass_data_rtl_fwprop): ...new pass_data instance and...
5609 (make_pass_rtl_fwprop): ...new function.
5610 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
5611 rtl_opt_pass along with...
5612 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
5613 (make_pass_rtl_fwprop_addr): ...new function.
5614 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
5615 rtl_opt_pass along with...
5616 (pass_data_rtl_pre): ...new pass_data instance and...
5617 (make_pass_rtl_pre): ...new function.
5618 (pass_rtl_hoist): Convert from a global struct to a subclass of
5619 rtl_opt_pass along with...
5620 (pass_data_rtl_hoist): ...new pass_data instance and...
5621 (make_pass_rtl_hoist): ...new function.
5622 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
5623 subclass of gimple_opt_pass along with...
5624 (pass_data_lower_cf): ...new pass_data instance and...
5625 (make_pass_lower_cf): ...new function.
5626 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
5627 from a global struct to a subclass of gimple_opt_pass along with...
5628 (pass_data_strength_reduction): ...new pass_data instance and...
5629 (make_pass_strength_reduction): ...new function.
5630 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
5631 of rtl_opt_pass along with...
5632 (pass_data_rtl_ifcvt): ...new pass_data instance and...
5633 (make_pass_rtl_ifcvt): ...new function.
5634 (pass_if_after_combine): Convert from a global struct to a subclass of
5635 rtl_opt_pass along with...
5636 (pass_data_if_after_combine): ...new pass_data instance and...
5637 (make_pass_if_after_combine): ...new function.
5638 (pass_if_after_reload): Convert from a global struct to a subclass of
5639 rtl_opt_pass along with...
5640 (pass_data_if_after_reload): ...new pass_data instance and...
5641 (make_pass_if_after_reload): ...new function.
5642 * init-regs.c (pass_initialize_regs): Convert from a global struct to
5643 a subclass of rtl_opt_pass along with...
5644 (pass_data_initialize_regs): ...new pass_data instance and...
5645 (make_pass_initialize_regs): ...new function.
5646 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
5647 of ipa_opt_pass_d along with...
5648 (pass_data_ipa_cp): ...new pass_data instance and...
5649 (make_pass_ipa_cp): ...new function.
5650 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
5651 global struct to a subclass of gimple_opt_pass along with...
5652 (pass_data_inline_parameters): ...new pass_data instance and...
5653 (make_pass_inline_parameters): ...new function.
5654 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
5655 subclass of gimple_opt_pass along with...
5656 (pass_data_early_inline): ...new pass_data instance and...
5657 (make_pass_early_inline): ...new function.
5658 (pass_ipa_inline): Convert from a global struct to a subclass of
5659 ipa_opt_pass_d along with...
5660 (pass_data_ipa_inline): ...new pass_data instance and...
5661 (make_pass_ipa_inline): ...new function.
5662 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
5663 struct to a subclass of gimple_opt_pass along with...
5664 (pass_data_local_pure_const): ...new pass_data instance and...
5665 (make_pass_local_pure_const): ...new function.
5666 (pass_ipa_pure_const): Convert from a global struct to a subclass of
5667 ipa_opt_pass_d along with...
5668 (pass_data_ipa_pure_const): ...new pass_data instance and...
5669 (make_pass_ipa_pure_const): ...new function.
5670 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
5671 to a subclass of ipa_opt_pass_d along with...
5672 (pass_data_ipa_reference): ...new pass_data instance and...
5673 (make_pass_ipa_reference): ...new function.
5674 * ipa-split.c (pass_split_functions): Convert from a global struct to
5675 a subclass of gimple_opt_pass along with...
5676 (pass_data_split_functions): ...new pass_data instance and...
5677 (make_pass_split_functions): ...new function.
5678 (pass_feedback_split_functions): Convert from a global struct to a
5679 subclass of gimple_opt_pass along with...
5680 (pass_data_feedback_split_functions): ...new pass_data instance and...
5681 (make_pass_feedback_split_functions): ...new function.
5682 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
5683 global struct to a subclass of simple_ipa_opt_pass along with...
5684 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
5686 (make_pass_ipa_function_and_variable_visibility): ...new function.
5687 (pass_ipa_free_inline_summary): Convert from a global struct to a
5688 subclass of simple_ipa_opt_pass along with...
5689 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
5690 (make_pass_ipa_free_inline_summary): ...new function.
5691 (pass_ipa_whole_program_visibility): Convert from a global struct to a
5692 subclass of ipa_opt_pass_d along with...
5693 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
5695 (make_pass_ipa_whole_program_visibility): ...new function.
5696 (pass_ipa_profile): Convert from a global struct to a subclass of
5697 ipa_opt_pass_d along with...
5698 (pass_data_ipa_profile): ...new pass_data instance and...
5699 (make_pass_ipa_profile): ...new function.
5700 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
5701 ipa_opt_pass_d along with...
5702 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
5703 (make_pass_ipa_cdtor_merge): ...new function.
5704 * ira.c (pass_ira): Convert from a global struct to a subclass of
5705 rtl_opt_pass along with...
5706 (pass_data_ira): ...new pass_data instance and...
5707 (make_pass_ira): ...new function.
5708 (pass_reload): Convert from a global struct to a subclass of
5709 rtl_opt_pass along with...
5710 (pass_data_reload): ...new pass_data instance and...
5711 (make_pass_reload): ...new function.
5712 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
5713 subclass of rtl_opt_pass along with...
5714 (pass_data_cleanup_barriers): ...new pass_data instance and...
5715 (make_pass_cleanup_barriers): ...new function.
5716 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
5717 of rtl_opt_pass along with...
5718 (pass_data_loop2): ...new pass_data instance and...
5719 (make_pass_loop2): ...new function.
5720 (pass_rtl_loop_init): Convert from a global struct to a subclass of
5721 rtl_opt_pass along with...
5722 (pass_data_rtl_loop_init): ...new pass_data instance and...
5723 (make_pass_rtl_loop_init): ...new function.
5724 (pass_rtl_loop_done): Convert from a global struct to a subclass of
5725 rtl_opt_pass along with...
5726 (pass_data_rtl_loop_done): ...new pass_data instance and...
5727 (make_pass_rtl_loop_done): ...new function.
5728 (pass_rtl_move_loop_invariants): Convert from a global struct to a
5729 subclass of rtl_opt_pass along with...
5730 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
5731 (make_pass_rtl_move_loop_invariants): ...new function.
5732 (pass_rtl_unswitch): Convert from a global struct to a subclass of
5733 rtl_opt_pass along with...
5734 (pass_data_rtl_unswitch): ...new pass_data instance and...
5735 (make_pass_rtl_unswitch): ...new function.
5736 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
5737 subclass of rtl_opt_pass along with...
5738 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
5740 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5741 (pass_rtl_doloop): Convert from a global struct to a subclass of
5742 rtl_opt_pass along with...
5743 (pass_data_rtl_doloop): ...new pass_data instance and...
5744 (make_pass_rtl_doloop): ...new function.
5745 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
5746 a subclass of rtl_opt_pass along with...
5747 (pass_data_lower_subreg): ...new pass_data instance and...
5748 (make_pass_lower_subreg): ...new function.
5749 (pass_lower_subreg2): Convert from a global struct to a subclass of
5750 rtl_opt_pass along with...
5751 (pass_data_lower_subreg2): ...new pass_data instance and...
5752 (make_pass_lower_subreg2): ...new function.
5753 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
5754 struct to a subclass of ipa_opt_pass_d along with...
5755 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
5756 (make_pass_ipa_lto_gimple_out): ...new function.
5757 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
5758 of ipa_opt_pass_d along with...
5759 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
5760 (make_pass_ipa_lto_finish_out): ...new function.
5761 * mode-switching.c (pass_mode_switching): Convert from a global struct
5762 to a subclass of rtl_opt_pass along with...
5763 (pass_data_mode_switching): ...new pass_data instance and...
5764 (make_pass_mode_switching): ...new function.
5765 * modulo-sched.c (pass_sms): Convert from a global struct to a
5766 subclass of rtl_opt_pass along with...
5767 (pass_data_sms): ...new pass_data instance and...
5768 (make_pass_sms): ...new function.
5769 * omp-low.c (pass_expand_omp): Convert from a global struct to a
5770 subclass of gimple_opt_pass along with...
5771 (pass_data_expand_omp): ...new pass_data instance and...
5772 (make_pass_expand_omp): ...new function.
5773 (pass_lower_omp): Convert from a global struct to a subclass of
5774 gimple_opt_pass along with...
5775 (pass_data_lower_omp): ...new pass_data instance and...
5776 (make_pass_lower_omp): ...new function.
5777 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
5778 of gimple_opt_pass along with...
5779 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
5780 (make_pass_diagnose_omp_blocks): ...new function.
5781 * passes.c (pass_early_local_passes): Convert from a global struct to
5782 a subclass of simple_ipa_opt_pass along with...
5783 (pass_data_early_local_passes): ...new pass_data instance and...
5784 (make_pass_early_local_passes): ...new function.
5785 (pass_all_early_optimizations): Convert from a global struct to a
5786 subclass of gimple_opt_pass along with...
5787 (pass_data_all_early_optimizations): ...new pass_data instance and...
5788 (make_pass_all_early_optimizations): ...new function.
5789 (pass_all_optimizations): Convert from a global struct to a subclass
5790 of gimple_opt_pass along with...
5791 (pass_data_all_optimizations): ...new pass_data instance and...
5792 (make_pass_all_optimizations): ...new function.
5793 (pass_all_optimizations_g): Convert from a global struct to a subclass
5794 of gimple_opt_pass along with...
5795 (pass_data_all_optimizations_g): ...new pass_data instance and...
5796 (make_pass_all_optimizations_g): ...new function.
5797 (pass_rest_of_compilation): Convert from a global struct to a subclass
5798 of rtl_opt_pass along with...
5799 (pass_data_rest_of_compilation): ...new pass_data instance and...
5800 (make_pass_rest_of_compilation): ...new function.
5801 (pass_postreload): Convert from a global struct to a subclass of
5802 rtl_opt_pass along with...
5803 (pass_data_postreload): ...new pass_data instance and...
5804 (make_pass_postreload): ...new function.
5805 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
5806 subclass of rtl_opt_pass along with...
5807 (pass_data_gcse2): ...new pass_data instance and...
5808 (make_pass_gcse2): ...new function.
5809 * postreload.c (pass_postreload_cse): Convert from a global struct to
5810 a subclass of rtl_opt_pass along with...
5811 (pass_data_postreload_cse): ...new pass_data instance and...
5812 (make_pass_postreload_cse): ...new function.
5813 * predict.c (pass_profile): Convert from a global struct to a subclass
5814 of gimple_opt_pass along with...
5815 (pass_data_profile): ...new pass_data instance and...
5816 (make_pass_profile): ...new function.
5817 (pass_strip_predict_hints): Convert from a global struct to a subclass
5818 of gimple_opt_pass along with...
5819 (pass_data_strip_predict_hints): ...new pass_data instance and...
5820 (make_pass_strip_predict_hints): ...new function.
5821 * recog.c (pass_peephole2): Convert from a global struct to a subclass
5822 of rtl_opt_pass along with...
5823 (pass_data_peephole2): ...new pass_data instance and...
5824 (make_pass_peephole2): ...new function.
5825 (pass_split_all_insns): Convert from a global struct to a subclass of
5826 rtl_opt_pass along with...
5827 (pass_data_split_all_insns): ...new pass_data instance and...
5828 (make_pass_split_all_insns): ...new function.
5829 (pass_split_after_reload): Convert from a global struct to a subclass
5830 of rtl_opt_pass along with...
5831 (pass_data_split_after_reload): ...new pass_data instance and...
5832 (make_pass_split_after_reload): ...new function.
5833 (pass_split_before_regstack): Convert from a global struct to a
5834 subclass of rtl_opt_pass along with...
5835 (pass_data_split_before_regstack): ...new pass_data instance and...
5836 (make_pass_split_before_regstack): ...new function.
5837 (pass_split_before_sched2): Convert from a global struct to a subclass
5838 of rtl_opt_pass along with...
5839 (pass_data_split_before_sched2): ...new pass_data instance and...
5840 (make_pass_split_before_sched2): ...new function.
5841 (pass_split_for_shorten_branches): Convert from a global struct to a
5842 subclass of rtl_opt_pass along with...
5843 (pass_data_split_for_shorten_branches): ...new pass_data instance
5845 (make_pass_split_for_shorten_branches): ...new function.
5846 * ree.c (pass_ree): Convert from a global struct to a subclass of
5847 rtl_opt_pass along with...
5848 (pass_data_ree): ...new pass_data instance and...
5849 (make_pass_ree): ...new function.
5850 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
5851 subclass of rtl_opt_pass along with...
5852 (pass_data_stack_regs): ...new pass_data instance and...
5853 (make_pass_stack_regs): ...new function.
5854 (pass_stack_regs_run): Convert from a global struct to a subclass of
5855 rtl_opt_pass along with...
5856 (pass_data_stack_regs_run): ...new pass_data instance and...
5857 (make_pass_stack_regs_run): ...new function.
5858 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
5859 subclass of rtl_opt_pass along with...
5860 (pass_data_cprop_hardreg): ...new pass_data instance and...
5861 (make_pass_cprop_hardreg): ...new function.
5862 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
5863 subclass of rtl_opt_pass along with...
5864 (pass_data_reginfo_init): ...new pass_data instance and...
5865 (make_pass_reginfo_init): ...new function.
5866 * regmove.c (pass_regmove): Convert from a global struct to a subclass
5867 of rtl_opt_pass along with...
5868 (pass_data_regmove): ...new pass_data instance and...
5869 (make_pass_regmove): ...new function.
5870 * regrename.c (pass_regrename): Convert from a global struct to a
5871 subclass of rtl_opt_pass along with...
5872 (pass_data_regrename): ...new pass_data instance and...
5873 (make_pass_regrename): ...new function.
5874 * reorg.c (pass_delay_slots): Convert from a global struct to a
5875 subclass of rtl_opt_pass along with...
5876 (pass_data_delay_slots): ...new pass_data instance and...
5877 (make_pass_delay_slots): ...new function.
5878 (pass_machine_reorg): Convert from a global struct to a subclass of
5879 rtl_opt_pass along with...
5880 (pass_data_machine_reorg): ...new pass_data instance and...
5881 (make_pass_machine_reorg): ...new function.
5882 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
5883 of rtl_opt_pass along with...
5884 (pass_data_sched): ...new pass_data instance and...
5885 (make_pass_sched): ...new function.
5886 (pass_sched2): Convert from a global struct to a subclass of
5887 rtl_opt_pass along with...
5888 (pass_data_sched2): ...new pass_data instance and...
5889 (make_pass_sched2): ...new function.
5890 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
5891 to a subclass of rtl_opt_pass along with...
5892 (pass_data_stack_ptr_mod): ...new pass_data instance and...
5893 (make_pass_stack_ptr_mod): ...new function.
5894 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
5895 to a subclass of rtl_opt_pass along with...
5896 (pass_data_rtl_store_motion): ...new pass_data instance and...
5897 (make_pass_rtl_store_motion): ...new function.
5898 * tracer.c (pass_tracer): Convert from a global struct to a subclass
5899 of gimple_opt_pass along with...
5900 (pass_data_tracer): ...new pass_data instance and...
5901 (make_pass_tracer): ...new function.
5902 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
5903 to a subclass of gimple_opt_pass along with...
5904 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
5905 (make_pass_diagnose_tm_blocks): ...new function.
5906 (pass_lower_tm): Convert from a global struct to a subclass of
5907 gimple_opt_pass along with...
5908 (pass_data_lower_tm): ...new pass_data instance and...
5909 (make_pass_lower_tm): ...new function.
5910 (pass_tm_init): Convert from a global struct to a subclass of
5911 gimple_opt_pass along with...
5912 (pass_data_tm_init): ...new pass_data instance and...
5913 (make_pass_tm_init): ...new function.
5914 (pass_tm_mark): Convert from a global struct to a subclass of
5915 gimple_opt_pass along with...
5916 (pass_data_tm_mark): ...new pass_data instance and...
5917 (make_pass_tm_mark): ...new function.
5918 (pass_tm_edges): Convert from a global struct to a subclass of
5919 gimple_opt_pass along with...
5920 (pass_data_tm_edges): ...new pass_data instance and...
5921 (make_pass_tm_edges): ...new function.
5922 (pass_tm_memopt): Convert from a global struct to a subclass of
5923 gimple_opt_pass along with...
5924 (pass_data_tm_memopt): ...new pass_data instance and...
5925 (make_pass_tm_memopt): ...new function.
5926 (pass_ipa_tm): Convert from a global struct to a subclass of
5927 simple_ipa_opt_pass along with...
5928 (pass_data_ipa_tm): ...new pass_data instance and...
5929 (make_pass_ipa_tm): ...new function.
5930 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
5931 subclass of gimple_opt_pass along with...
5932 (pass_data_call_cdce): ...new pass_data instance and...
5933 (make_pass_call_cdce): ...new function.
5934 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
5935 subclass of gimple_opt_pass along with...
5936 (pass_data_build_cfg): ...new pass_data instance and...
5937 (make_pass_build_cfg): ...new function.
5938 (pass_split_crit_edges): Convert from a global struct to a subclass of
5939 gimple_opt_pass along with...
5940 (pass_data_split_crit_edges): ...new pass_data instance and...
5941 (make_pass_split_crit_edges): ...new function.
5942 (pass_warn_function_return): Convert from a global struct to a
5943 subclass of gimple_opt_pass along with...
5944 (pass_data_warn_function_return): ...new pass_data instance and...
5945 (make_pass_warn_function_return): ...new function.
5946 (pass_warn_function_noreturn): Convert from a global struct to a
5947 subclass of gimple_opt_pass along with...
5948 (pass_data_warn_function_noreturn): ...new pass_data instance and...
5949 (make_pass_warn_function_noreturn): ...new function.
5950 (pass_warn_unused_result): Convert from a global struct to a subclass
5951 of gimple_opt_pass along with...
5952 (pass_data_warn_unused_result): ...new pass_data instance and...
5953 (make_pass_warn_unused_result): ...new function.
5954 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
5955 a subclass of gimple_opt_pass along with...
5956 (pass_data_merge_phi): ...new pass_data instance and...
5957 (make_pass_merge_phi): ...new function.
5958 * tree-complex.c (pass_lower_complex): Convert from a global struct to
5959 a subclass of gimple_opt_pass along with...
5960 (pass_data_lower_complex): ...new pass_data instance and...
5961 (make_pass_lower_complex): ...new function.
5962 (pass_lower_complex_O0): Convert from a global struct to a subclass of
5963 gimple_opt_pass along with...
5964 (pass_data_lower_complex_O0): ...new pass_data instance and...
5965 (make_pass_lower_complex_O0): ...new function.
5966 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
5967 subclass of gimple_opt_pass along with...
5968 (pass_data_lower_eh): ...new pass_data instance and...
5969 (make_pass_lower_eh): ...new function.
5970 (pass_refactor_eh): Convert from a global struct to a subclass of
5971 gimple_opt_pass along with...
5972 (pass_data_refactor_eh): ...new pass_data instance and...
5973 (make_pass_refactor_eh): ...new function.
5974 (pass_lower_resx): Convert from a global struct to a subclass of
5975 gimple_opt_pass along with...
5976 (pass_data_lower_resx): ...new pass_data instance and...
5977 (make_pass_lower_resx): ...new function.
5978 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
5979 of gimple_opt_pass along with...
5980 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
5981 (make_pass_lower_eh_dispatch): ...new function.
5982 (pass_cleanup_eh): Convert from a global struct to a subclass of
5983 gimple_opt_pass along with...
5984 (pass_data_cleanup_eh): ...new pass_data instance and...
5985 (make_pass_cleanup_eh): ...new function.
5986 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
5987 to a subclass of simple_ipa_opt_pass along with...
5988 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
5989 (make_pass_ipa_lower_emutls): ...new function.
5990 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
5991 a subclass of gimple_opt_pass along with...
5992 (pass_data_if_conversion): ...new pass_data instance and...
5993 (make_pass_if_conversion): ...new function.
5994 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
5995 subclass of gimple_opt_pass along with...
5996 (pass_data_build_ssa): ...new pass_data instance and...
5997 (make_pass_build_ssa): ...new function.
5998 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
5999 global struct to a subclass of gimple_opt_pass along with...
6000 (pass_data_loop_distribution): ...new pass_data instance and...
6001 (make_pass_loop_distribution): ...new function.
6002 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6003 subclass of gimple_opt_pass along with...
6004 (pass_data_mudflap_1): ...new pass_data instance and...
6005 (make_pass_mudflap_1): ...new function.
6006 (pass_mudflap_2): Convert from a global struct to a subclass of
6007 gimple_opt_pass along with...
6008 (pass_data_mudflap_2): ...new pass_data instance and...
6009 (make_pass_mudflap_2): ...new function.
6010 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6011 subclass of gimple_opt_pass along with...
6012 (pass_data_mudflap_1): ...new pass_data instance and...
6013 (make_pass_mudflap_1): ...new function.
6014 (pass_mudflap_2): Convert from a global struct to a subclass of
6015 gimple_opt_pass along with...
6016 (pass_data_mudflap_2): ...new pass_data instance and...
6017 (make_pass_mudflap_2): ...new function.
6018 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6019 gimple_opt_pass along with...
6020 (pass_data_nrv): ...new pass_data instance and...
6021 (make_pass_nrv): ...new function.
6022 (pass_return_slot): Convert from a global struct to a subclass of
6023 gimple_opt_pass along with...
6024 (pass_data_return_slot): ...new pass_data instance and...
6025 (make_pass_return_slot): ...new function.
6026 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6027 to a subclass of gimple_opt_pass along with...
6028 (pass_data_object_sizes): ...new pass_data instance and...
6029 (make_pass_object_sizes): ...new function.
6030 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6031 global struct to a subclass of gimple_opt_pass along with...
6032 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6034 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6035 (pass_fixup_cfg): Convert from a global struct to a subclass of
6036 gimple_opt_pass along with...
6037 (pass_data_fixup_cfg): ...new pass_data instance and...
6038 (make_pass_fixup_cfg): ...new function.
6039 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6040 (make_pass_mudflap_1): ...new function.
6041 (pass_mudflap_2): Replace declaration with that of...
6042 (make_pass_mudflap_2): ...new function.
6043 (pass_asan): Replace declaration with that of...
6044 (make_pass_asan): ...new function.
6045 (pass_asan_O0): Replace declaration with that of...
6046 (make_pass_asan_O0): ...new function.
6047 (pass_tsan): Replace declaration with that of...
6048 (make_pass_tsan): ...new function.
6049 (pass_tsan_O0): Replace declaration with that of...
6050 (make_pass_tsan_O0): ...new function.
6051 (pass_lower_cf): Replace declaration with that of...
6052 (make_pass_lower_cf): ...new function.
6053 (pass_refactor_eh): Replace declaration with that of...
6054 (make_pass_refactor_eh): ...new function.
6055 (pass_lower_eh): Replace declaration with that of...
6056 (make_pass_lower_eh): ...new function.
6057 (pass_lower_eh_dispatch): Replace declaration with that of...
6058 (make_pass_lower_eh_dispatch): ...new function.
6059 (pass_lower_resx): Replace declaration with that of...
6060 (make_pass_lower_resx): ...new function.
6061 (pass_build_cfg): Replace declaration with that of...
6062 (make_pass_build_cfg): ...new function.
6063 (pass_early_tree_profile): Replace declaration with that of...
6064 (make_pass_early_tree_profile): ...new function.
6065 (pass_cleanup_eh): Replace declaration with that of...
6066 (make_pass_cleanup_eh): ...new function.
6067 (pass_sra): Replace declaration with that of...
6068 (make_pass_sra): ...new function.
6069 (pass_sra_early): Replace declaration with that of...
6070 (make_pass_sra_early): ...new function.
6071 (pass_early_ipa_sra): Replace declaration with that of...
6072 (make_pass_early_ipa_sra): ...new function.
6073 (pass_tail_recursion): Replace declaration with that of...
6074 (make_pass_tail_recursion): ...new function.
6075 (pass_tail_calls): Replace declaration with that of...
6076 (make_pass_tail_calls): ...new function.
6077 (pass_tree_loop): Replace declaration with that of...
6078 (make_pass_tree_loop): ...new function.
6079 (pass_tree_loop_init): Replace declaration with that of...
6080 (make_pass_tree_loop_init): ...new function.
6081 (pass_lim): Replace declaration with that of...
6082 (make_pass_lim): ...new function.
6083 (pass_tree_unswitch): Replace declaration with that of...
6084 (make_pass_tree_unswitch): ...new function.
6085 (pass_predcom): Replace declaration with that of...
6086 (make_pass_predcom): ...new function.
6087 (pass_iv_canon): Replace declaration with that of...
6088 (make_pass_iv_canon): ...new function.
6089 (pass_scev_cprop): Replace declaration with that of...
6090 (make_pass_scev_cprop): ...new function.
6091 (pass_empty_loop): Replace declaration with that of...
6092 (make_pass_empty_loop): ...new function.
6093 (pass_record_bounds): Replace declaration with that of...
6094 (make_pass_record_bounds): ...new function.
6095 (pass_graphite): Replace declaration with that of...
6096 (make_pass_graphite): ...new function.
6097 (pass_graphite_transforms): Replace declaration with that of...
6098 (make_pass_graphite_transforms): ...new function.
6099 (pass_if_conversion): Replace declaration with that of...
6100 (make_pass_if_conversion): ...new function.
6101 (pass_loop_distribution): Replace declaration with that of...
6102 (make_pass_loop_distribution): ...new function.
6103 (pass_vectorize): Replace declaration with that of...
6104 (make_pass_vectorize): ...new function.
6105 (pass_slp_vectorize): Replace declaration with that of...
6106 (make_pass_slp_vectorize): ...new function.
6107 (pass_complete_unroll): Replace declaration with that of...
6108 (make_pass_complete_unroll): ...new function.
6109 (pass_complete_unrolli): Replace declaration with that of...
6110 (make_pass_complete_unrolli): ...new function.
6111 (pass_parallelize_loops): Replace declaration with that of...
6112 (make_pass_parallelize_loops): ...new function.
6113 (pass_loop_prefetch): Replace declaration with that of...
6114 (make_pass_loop_prefetch): ...new function.
6115 (pass_iv_optimize): Replace declaration with that of...
6116 (make_pass_iv_optimize): ...new function.
6117 (pass_tree_loop_done): Replace declaration with that of...
6118 (make_pass_tree_loop_done): ...new function.
6119 (pass_ch): Replace declaration with that of...
6120 (make_pass_ch): ...new function.
6121 (pass_ccp): Replace declaration with that of...
6122 (make_pass_ccp): ...new function.
6123 (pass_phi_only_cprop): Replace declaration with that of...
6124 (make_pass_phi_only_cprop): ...new function.
6125 (pass_build_ssa): Replace declaration with that of...
6126 (make_pass_build_ssa): ...new function.
6127 (pass_build_alias): Replace declaration with that of...
6128 (make_pass_build_alias): ...new function.
6129 (pass_build_ealias): Replace declaration with that of...
6130 (make_pass_build_ealias): ...new function.
6131 (pass_dominator): Replace declaration with that of...
6132 (make_pass_dominator): ...new function.
6133 (pass_dce): Replace declaration with that of...
6134 (make_pass_dce): ...new function.
6135 (pass_dce_loop): Replace declaration with that of...
6136 (make_pass_dce_loop): ...new function.
6137 (pass_cd_dce): Replace declaration with that of...
6138 (make_pass_cd_dce): ...new function.
6139 (pass_call_cdce): Replace declaration with that of...
6140 (make_pass_call_cdce): ...new function.
6141 (pass_merge_phi): Replace declaration with that of...
6142 (make_pass_merge_phi): ...new function.
6143 (pass_split_crit_edges): Replace declaration with that of...
6144 (make_pass_split_crit_edges): ...new function.
6145 (pass_pre): Replace declaration with that of...
6146 (make_pass_pre): ...new function.
6147 (pass_profile): Replace declaration with that of...
6148 (make_pass_profile): ...new function.
6149 (pass_strip_predict_hints): Replace declaration with that of...
6150 (make_pass_strip_predict_hints): ...new function.
6151 (pass_lower_complex_O0): Replace declaration with that of...
6152 (make_pass_lower_complex_O0): ...new function.
6153 (pass_lower_complex): Replace declaration with that of...
6154 (make_pass_lower_complex): ...new function.
6155 (pass_lower_vector): Replace declaration with that of...
6156 (make_pass_lower_vector): ...new function.
6157 (pass_lower_vector_ssa): Replace declaration with that of...
6158 (make_pass_lower_vector_ssa): ...new function.
6159 (pass_lower_omp): Replace declaration with that of...
6160 (make_pass_lower_omp): ...new function.
6161 (pass_diagnose_omp_blocks): Replace declaration with that of...
6162 (make_pass_diagnose_omp_blocks): ...new function.
6163 (pass_expand_omp): Replace declaration with that of...
6164 (make_pass_expand_omp): ...new function.
6165 (pass_expand_omp_ssa): Replace declaration with that of...
6166 (make_pass_expand_omp_ssa): ...new function.
6167 (pass_object_sizes): Replace declaration with that of...
6168 (make_pass_object_sizes): ...new function.
6169 (pass_strlen): Replace declaration with that of...
6170 (make_pass_strlen): ...new function.
6171 (pass_fold_builtins): Replace declaration with that of...
6172 (make_pass_fold_builtins): ...new function.
6173 (pass_stdarg): Replace declaration with that of...
6174 (make_pass_stdarg): ...new function.
6175 (pass_early_warn_uninitialized): Replace declaration with that of...
6176 (make_pass_early_warn_uninitialized): ...new function.
6177 (pass_late_warn_uninitialized): Replace declaration with that of...
6178 (make_pass_late_warn_uninitialized): ...new function.
6179 (pass_cse_reciprocals): Replace declaration with that of...
6180 (make_pass_cse_reciprocals): ...new function.
6181 (pass_cse_sincos): Replace declaration with that of...
6182 (make_pass_cse_sincos): ...new function.
6183 (pass_optimize_bswap): Replace declaration with that of...
6184 (make_pass_optimize_bswap): ...new function.
6185 (pass_optimize_widening_mul): Replace declaration with that of...
6186 (make_pass_optimize_widening_mul): ...new function.
6187 (pass_warn_function_return): Replace declaration with that of...
6188 (make_pass_warn_function_return): ...new function.
6189 (pass_warn_function_noreturn): Replace declaration with that of...
6190 (make_pass_warn_function_noreturn): ...new function.
6191 (pass_cselim): Replace declaration with that of...
6192 (make_pass_cselim): ...new function.
6193 (pass_phiopt): Replace declaration with that of...
6194 (make_pass_phiopt): ...new function.
6195 (pass_forwprop): Replace declaration with that of...
6196 (make_pass_forwprop): ...new function.
6197 (pass_phiprop): Replace declaration with that of...
6198 (make_pass_phiprop): ...new function.
6199 (pass_tree_ifcombine): Replace declaration with that of...
6200 (make_pass_tree_ifcombine): ...new function.
6201 (pass_dse): Replace declaration with that of...
6202 (make_pass_dse): ...new function.
6203 (pass_nrv): Replace declaration with that of...
6204 (make_pass_nrv): ...new function.
6205 (pass_rename_ssa_copies): Replace declaration with that of...
6206 (make_pass_rename_ssa_copies): ...new function.
6207 (pass_sink_code): Replace declaration with that of...
6208 (make_pass_sink_code): ...new function.
6209 (pass_fre): Replace declaration with that of...
6210 (make_pass_fre): ...new function.
6211 (pass_check_data_deps): Replace declaration with that of...
6212 (make_pass_check_data_deps): ...new function.
6213 (pass_copy_prop): Replace declaration with that of...
6214 (make_pass_copy_prop): ...new function.
6215 (pass_vrp): Replace declaration with that of...
6216 (make_pass_vrp): ...new function.
6217 (pass_uncprop): Replace declaration with that of...
6218 (make_pass_uncprop): ...new function.
6219 (pass_return_slot): Replace declaration with that of...
6220 (make_pass_return_slot): ...new function.
6221 (pass_reassoc): Replace declaration with that of...
6222 (make_pass_reassoc): ...new function.
6223 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6224 (make_pass_rebuild_cgraph_edges): ...new function.
6225 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6226 (make_pass_remove_cgraph_callee_edges): ...new function.
6227 (pass_build_cgraph_edges): Replace declaration with that of...
6228 (make_pass_build_cgraph_edges): ...new function.
6229 (pass_local_pure_const): Replace declaration with that of...
6230 (make_pass_local_pure_const): ...new function.
6231 (pass_tracer): Replace declaration with that of...
6232 (make_pass_tracer): ...new function.
6233 (pass_warn_unused_result): Replace declaration with that of...
6234 (make_pass_warn_unused_result): ...new function.
6235 (pass_diagnose_tm_blocks): Replace declaration with that of...
6236 (make_pass_diagnose_tm_blocks): ...new function.
6237 (pass_lower_tm): Replace declaration with that of...
6238 (make_pass_lower_tm): ...new function.
6239 (pass_tm_init): Replace declaration with that of...
6240 (make_pass_tm_init): ...new function.
6241 (pass_tm_mark): Replace declaration with that of...
6242 (make_pass_tm_mark): ...new function.
6243 (pass_tm_memopt): Replace declaration with that of...
6244 (make_pass_tm_memopt): ...new function.
6245 (pass_tm_edges): Replace declaration with that of...
6246 (make_pass_tm_edges): ...new function.
6247 (pass_split_functions): Replace declaration with that of...
6248 (make_pass_split_functions): ...new function.
6249 (pass_feedback_split_functions): Replace declaration with that of...
6250 (make_pass_feedback_split_functions): ...new function.
6251 (pass_strength_reduction): Replace declaration with that of...
6252 (make_pass_strength_reduction): ...new function.
6253 (pass_ipa_lower_emutls): Replace declaration with that of...
6254 (make_pass_ipa_lower_emutls): ...new function.
6255 (pass_ipa_function_and_variable_visibility): Replace declaration with
6257 (make_pass_ipa_function_and_variable_visibility): ...new function.
6258 (pass_ipa_tree_profile): Replace declaration with that of...
6259 (make_pass_ipa_tree_profile): ...new function.
6260 (pass_early_local_passes): Replace declaration with that of...
6261 (make_pass_early_local_passes): ...new function.
6262 (pass_ipa_whole_program_visibility): Replace declaration with that
6264 (make_pass_ipa_whole_program_visibility): ...new function.
6265 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6266 (make_pass_ipa_lto_gimple_out): ...new function.
6267 (pass_ipa_increase_alignment): Replace declaration with that of...
6268 (make_pass_ipa_increase_alignment): ...new function.
6269 (pass_ipa_inline): Replace declaration with that of...
6270 (make_pass_ipa_inline): ...new function.
6271 (pass_ipa_free_lang_data): Replace declaration with that of...
6272 (make_pass_ipa_free_lang_data): ...new function.
6273 (pass_ipa_free_inline_summary): Replace declaration with that of...
6274 (make_pass_ipa_free_inline_summary): ...new function.
6275 (pass_ipa_cp): Replace declaration with that of...
6276 (make_pass_ipa_cp): ...new function.
6277 (pass_ipa_reference): Replace declaration with that of...
6278 (make_pass_ipa_reference): ...new function.
6279 (pass_ipa_pure_const): Replace declaration with that of...
6280 (make_pass_ipa_pure_const): ...new function.
6281 (pass_ipa_pta): Replace declaration with that of...
6282 (make_pass_ipa_pta): ...new function.
6283 (pass_ipa_lto_finish_out): Replace declaration with that of...
6284 (make_pass_ipa_lto_finish_out): ...new function.
6285 (pass_ipa_tm): Replace declaration with that of...
6286 (make_pass_ipa_tm): ...new function.
6287 (pass_ipa_profile): Replace declaration with that of...
6288 (make_pass_ipa_profile): ...new function.
6289 (pass_ipa_cdtor_merge): Replace declaration with that of...
6290 (make_pass_ipa_cdtor_merge): ...new function.
6291 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6293 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6294 (pass_init_datastructures): Replace declaration with that of...
6295 (make_pass_init_datastructures): ...new function.
6296 (pass_fixup_cfg): Replace declaration with that of...
6297 (make_pass_fixup_cfg): ...new function.
6298 (pass_expand): Replace declaration with that of...
6299 (make_pass_expand): ...new function.
6300 (pass_instantiate_virtual_regs): Replace declaration with that of...
6301 (make_pass_instantiate_virtual_regs): ...new function.
6302 (pass_rtl_fwprop): Replace declaration with that of...
6303 (make_pass_rtl_fwprop): ...new function.
6304 (pass_rtl_fwprop_addr): Replace declaration with that of...
6305 (make_pass_rtl_fwprop_addr): ...new function.
6306 (pass_jump): Replace declaration with that of...
6307 (make_pass_jump): ...new function.
6308 (pass_jump2): Replace declaration with that of...
6309 (make_pass_jump2): ...new function.
6310 (pass_lower_subreg): Replace declaration with that of...
6311 (make_pass_lower_subreg): ...new function.
6312 (pass_cse): Replace declaration with that of...
6313 (make_pass_cse): ...new function.
6314 (pass_fast_rtl_dce): Replace declaration with that of...
6315 (make_pass_fast_rtl_dce): ...new function.
6316 (pass_ud_rtl_dce): Replace declaration with that of...
6317 (make_pass_ud_rtl_dce): ...new function.
6318 (pass_rtl_dce): Replace declaration with that of...
6319 (make_pass_rtl_dce): ...new function.
6320 (pass_rtl_dse1): Replace declaration with that of...
6321 (make_pass_rtl_dse1): ...new function.
6322 (pass_rtl_dse2): Replace declaration with that of...
6323 (make_pass_rtl_dse2): ...new function.
6324 (pass_rtl_dse3): Replace declaration with that of...
6325 (make_pass_rtl_dse3): ...new function.
6326 (pass_rtl_cprop): Replace declaration with that of...
6327 (make_pass_rtl_cprop): ...new function.
6328 (pass_rtl_pre): Replace declaration with that of...
6329 (make_pass_rtl_pre): ...new function.
6330 (pass_rtl_hoist): Replace declaration with that of...
6331 (make_pass_rtl_hoist): ...new function.
6332 (pass_rtl_store_motion): Replace declaration with that of...
6333 (make_pass_rtl_store_motion): ...new function.
6334 (pass_cse_after_global_opts): Replace declaration with that of...
6335 (make_pass_cse_after_global_opts): ...new function.
6336 (pass_rtl_ifcvt): Replace declaration with that of...
6337 (make_pass_rtl_ifcvt): ...new function.
6338 (pass_into_cfg_layout_mode): Replace declaration with that of...
6339 (make_pass_into_cfg_layout_mode): ...new function.
6340 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6341 (make_pass_outof_cfg_layout_mode): ...new function.
6342 (pass_loop2): Replace declaration with that of...
6343 (make_pass_loop2): ...new function.
6344 (pass_rtl_loop_init): Replace declaration with that of...
6345 (make_pass_rtl_loop_init): ...new function.
6346 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6347 (make_pass_rtl_move_loop_invariants): ...new function.
6348 (pass_rtl_unswitch): Replace declaration with that of...
6349 (make_pass_rtl_unswitch): ...new function.
6350 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6351 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6352 (pass_rtl_doloop): Replace declaration with that of...
6353 (make_pass_rtl_doloop): ...new function.
6354 (pass_rtl_loop_done): Replace declaration with that of...
6355 (make_pass_rtl_loop_done): ...new function.
6356 (pass_web): Replace declaration with that of...
6357 (make_pass_web): ...new function.
6358 (pass_cse2): Replace declaration with that of...
6359 (make_pass_cse2): ...new function.
6360 (pass_df_initialize_opt): Replace declaration with that of...
6361 (make_pass_df_initialize_opt): ...new function.
6362 (pass_df_initialize_no_opt): Replace declaration with that of...
6363 (make_pass_df_initialize_no_opt): ...new function.
6364 (pass_reginfo_init): Replace declaration with that of...
6365 (make_pass_reginfo_init): ...new function.
6366 (pass_inc_dec): Replace declaration with that of...
6367 (make_pass_inc_dec): ...new function.
6368 (pass_stack_ptr_mod): Replace declaration with that of...
6369 (make_pass_stack_ptr_mod): ...new function.
6370 (pass_initialize_regs): Replace declaration with that of...
6371 (make_pass_initialize_regs): ...new function.
6372 (pass_combine): Replace declaration with that of...
6373 (make_pass_combine): ...new function.
6374 (pass_if_after_combine): Replace declaration with that of...
6375 (make_pass_if_after_combine): ...new function.
6376 (pass_ree): Replace declaration with that of...
6377 (make_pass_ree): ...new function.
6378 (pass_partition_blocks): Replace declaration with that of...
6379 (make_pass_partition_blocks): ...new function.
6380 (pass_match_asm_constraints): Replace declaration with that of...
6381 (make_pass_match_asm_constraints): ...new function.
6382 (pass_regmove): Replace declaration with that of...
6383 (make_pass_regmove): ...new function.
6384 (pass_split_all_insns): Replace declaration with that of...
6385 (make_pass_split_all_insns): ...new function.
6386 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6387 (make_pass_fast_rtl_byte_dce): ...new function.
6388 (pass_lower_subreg2): Replace declaration with that of...
6389 (make_pass_lower_subreg2): ...new function.
6390 (pass_mode_switching): Replace declaration with that of...
6391 (make_pass_mode_switching): ...new function.
6392 (pass_sms): Replace declaration with that of...
6393 (make_pass_sms): ...new function.
6394 (pass_sched): Replace declaration with that of...
6395 (make_pass_sched): ...new function.
6396 (pass_ira): Replace declaration with that of...
6397 (make_pass_ira): ...new function.
6398 (pass_reload): Replace declaration with that of...
6399 (make_pass_reload): ...new function.
6400 (pass_clean_state): Replace declaration with that of...
6401 (make_pass_clean_state): ...new function.
6402 (pass_branch_prob): Replace declaration with that of...
6403 (make_pass_branch_prob): ...new function.
6404 (pass_value_profile_transformations): Replace declaration with that
6406 (make_pass_value_profile_transformations): ...new function.
6407 (pass_postreload_cse): Replace declaration with that of...
6408 (make_pass_postreload_cse): ...new function.
6409 (pass_gcse2): Replace declaration with that of...
6410 (make_pass_gcse2): ...new function.
6411 (pass_split_after_reload): Replace declaration with that of...
6412 (make_pass_split_after_reload): ...new function.
6413 (pass_branch_target_load_optimize1): Replace declaration with that
6415 (make_pass_branch_target_load_optimize1): ...new function.
6416 (pass_thread_prologue_and_epilogue): Replace declaration with that
6418 (make_pass_thread_prologue_and_epilogue): ...new function.
6419 (pass_stack_adjustments): Replace declaration with that of...
6420 (make_pass_stack_adjustments): ...new function.
6421 (pass_peephole2): Replace declaration with that of...
6422 (make_pass_peephole2): ...new function.
6423 (pass_if_after_reload): Replace declaration with that of...
6424 (make_pass_if_after_reload): ...new function.
6425 (pass_regrename): Replace declaration with that of...
6426 (make_pass_regrename): ...new function.
6427 (pass_cprop_hardreg): Replace declaration with that of...
6428 (make_pass_cprop_hardreg): ...new function.
6429 (pass_reorder_blocks): Replace declaration with that of...
6430 (make_pass_reorder_blocks): ...new function.
6431 (pass_branch_target_load_optimize2): Replace declaration with that
6433 (make_pass_branch_target_load_optimize2): ...new function.
6434 (pass_leaf_regs): Replace declaration with that of...
6435 (make_pass_leaf_regs): ...new function.
6436 (pass_split_before_sched2): Replace declaration with that of...
6437 (make_pass_split_before_sched2): ...new function.
6438 (pass_compare_elim_after_reload): Replace declaration with that of...
6439 (make_pass_compare_elim_after_reload): ...new function.
6440 (pass_sched2): Replace declaration with that of...
6441 (make_pass_sched2): ...new function.
6442 (pass_stack_regs): Replace declaration with that of...
6443 (make_pass_stack_regs): ...new function.
6444 (pass_stack_regs_run): Replace declaration with that of...
6445 (make_pass_stack_regs_run): ...new function.
6446 (pass_df_finish): Replace declaration with that of...
6447 (make_pass_df_finish): ...new function.
6448 (pass_compute_alignments): Replace declaration with that of...
6449 (make_pass_compute_alignments): ...new function.
6450 (pass_duplicate_computed_gotos): Replace declaration with that of...
6451 (make_pass_duplicate_computed_gotos): ...new function.
6452 (pass_variable_tracking): Replace declaration with that of...
6453 (make_pass_variable_tracking): ...new function.
6454 (pass_free_cfg): Replace declaration with that of...
6455 (make_pass_free_cfg): ...new function.
6456 (pass_machine_reorg): Replace declaration with that of...
6457 (make_pass_machine_reorg): ...new function.
6458 (pass_cleanup_barriers): Replace declaration with that of...
6459 (make_pass_cleanup_barriers): ...new function.
6460 (pass_delay_slots): Replace declaration with that of...
6461 (make_pass_delay_slots): ...new function.
6462 (pass_split_for_shorten_branches): Replace declaration with that of...
6463 (make_pass_split_for_shorten_branches): ...new function.
6464 (pass_split_before_regstack): Replace declaration with that of...
6465 (make_pass_split_before_regstack): ...new function.
6466 (pass_convert_to_eh_region_ranges): Replace declaration with that
6468 (make_pass_convert_to_eh_region_ranges): ...new function.
6469 (pass_shorten_branches): Replace declaration with that of...
6470 (make_pass_shorten_branches): ...new function.
6471 (pass_set_nothrow_function_flags): Replace declaration with that of...
6472 (make_pass_set_nothrow_function_flags): ...new function.
6473 (pass_dwarf2_frame): Replace declaration with that of...
6474 (make_pass_dwarf2_frame): ...new function.
6475 (pass_final): Replace declaration with that of...
6476 (make_pass_final): ...new function.
6477 (pass_rtl_seqabstr): Replace declaration with that of...
6478 (make_pass_rtl_seqabstr): ...new function.
6479 (pass_release_ssa_names): Replace declaration with that of...
6480 (make_pass_release_ssa_names): ...new function.
6481 (pass_early_inline): Replace declaration with that of...
6482 (make_pass_early_inline): ...new function.
6483 (pass_inline_parameters): Replace declaration with that of...
6484 (make_pass_inline_parameters): ...new function.
6485 (pass_update_address_taken): Replace declaration with that of...
6486 (make_pass_update_address_taken): ...new function.
6487 (pass_convert_switch): Replace declaration with that of...
6488 (make_pass_convert_switch): ...new function.
6489 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6490 to a subclass of simple_ipa_opt_pass along with...
6491 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6492 (make_pass_ipa_tree_profile): ...new function.
6493 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6494 subclass of gimple_opt_pass along with...
6495 (pass_data_sra_early): ...new pass_data instance and...
6496 (make_pass_sra_early): ...new function.
6497 (pass_sra): Convert from a global struct to a subclass of
6498 gimple_opt_pass along with...
6499 (pass_data_sra): ...new pass_data instance and...
6500 (make_pass_sra): ...new function.
6501 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6502 gimple_opt_pass along with...
6503 (pass_data_early_ipa_sra): ...new pass_data instance and...
6504 (make_pass_early_ipa_sra): ...new function.
6505 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6506 subclass of gimple_opt_pass along with...
6507 (pass_data_ccp): ...new pass_data instance and...
6508 (make_pass_ccp): ...new function.
6509 (pass_fold_builtins): Convert from a global struct to a subclass of
6510 gimple_opt_pass along with...
6511 (pass_data_fold_builtins): ...new pass_data instance and...
6512 (make_pass_fold_builtins): ...new function.
6513 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6514 subclass of gimple_opt_pass along with...
6515 (pass_data_copy_prop): ...new pass_data instance and...
6516 (make_pass_copy_prop): ...new function.
6517 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6518 global struct to a subclass of gimple_opt_pass along with...
6519 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6520 (make_pass_rename_ssa_copies): ...new function.
6521 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6522 subclass of gimple_opt_pass along with...
6523 (pass_data_dce): ...new pass_data instance and...
6524 (make_pass_dce): ...new function.
6525 (pass_dce_loop): Convert from a global struct to a subclass of
6526 gimple_opt_pass along with...
6527 (pass_data_dce_loop): ...new pass_data instance and...
6528 (make_pass_dce_loop): ...new function.
6529 (pass_cd_dce): Convert from a global struct to a subclass of
6530 gimple_opt_pass along with...
6531 (pass_data_cd_dce): ...new pass_data instance and...
6532 (make_pass_cd_dce): ...new function.
6533 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6534 subclass of gimple_opt_pass along with...
6535 (pass_data_dominator): ...new pass_data instance and...
6536 (make_pass_dominator): ...new function.
6537 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6538 gimple_opt_pass along with...
6539 (pass_data_phi_only_cprop): ...new pass_data instance and...
6540 (make_pass_phi_only_cprop): ...new function.
6541 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6542 subclass of gimple_opt_pass along with...
6543 (pass_data_dse): ...new pass_data instance and...
6544 (make_pass_dse): ...new function.
6545 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6546 a subclass of gimple_opt_pass along with...
6547 (pass_data_forwprop): ...new pass_data instance and...
6548 (make_pass_forwprop): ...new function.
6549 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6550 struct to a subclass of gimple_opt_pass along with...
6551 (pass_data_tree_ifcombine): ...new pass_data instance and...
6552 (make_pass_tree_ifcombine): ...new function.
6553 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6554 subclass of gimple_opt_pass along with...
6555 (pass_data_ch): ...new pass_data instance and...
6556 (make_pass_ch): ...new function.
6557 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6558 subclass of gimple_opt_pass along with...
6559 (pass_data_tree_loop): ...new pass_data instance and...
6560 (make_pass_tree_loop): ...new function.
6561 (pass_tree_loop_init): Convert from a global struct to a subclass of
6562 gimple_opt_pass along with...
6563 (pass_data_tree_loop_init): ...new pass_data instance and...
6564 (make_pass_tree_loop_init): ...new function.
6565 (pass_lim): Convert from a global struct to a subclass of
6566 gimple_opt_pass along with...
6567 (pass_data_lim): ...new pass_data instance and...
6568 (make_pass_lim): ...new function.
6569 (pass_tree_unswitch): Convert from a global struct to a subclass of
6570 gimple_opt_pass along with...
6571 (pass_data_tree_unswitch): ...new pass_data instance and...
6572 (make_pass_tree_unswitch): ...new function.
6573 (pass_predcom): Convert from a global struct to a subclass of
6574 gimple_opt_pass along with...
6575 (pass_data_predcom): ...new pass_data instance and...
6576 (make_pass_predcom): ...new function.
6577 (pass_vectorize): Convert from a global struct to a subclass of
6578 gimple_opt_pass along with...
6579 (pass_data_vectorize): ...new pass_data instance and...
6580 (make_pass_vectorize): ...new function.
6581 (pass_graphite): Convert from a global struct to a subclass of
6582 gimple_opt_pass along with...
6583 (pass_data_graphite): ...new pass_data instance and...
6584 (make_pass_graphite): ...new function.
6585 (pass_graphite_transforms): Convert from a global struct to a subclass
6586 of gimple_opt_pass along with...
6587 (pass_data_graphite_transforms): ...new pass_data instance and...
6588 (make_pass_graphite_transforms): ...new function.
6589 (pass_check_data_deps): Convert from a global struct to a subclass of
6590 gimple_opt_pass along with...
6591 (pass_data_check_data_deps): ...new pass_data instance and...
6592 (make_pass_check_data_deps): ...new function.
6593 (pass_iv_canon): Convert from a global struct to a subclass of
6594 gimple_opt_pass along with...
6595 (pass_data_iv_canon): ...new pass_data instance and...
6596 (make_pass_iv_canon): ...new function.
6597 (pass_scev_cprop): Convert from a global struct to a subclass of
6598 gimple_opt_pass along with...
6599 (pass_data_scev_cprop): ...new pass_data instance and...
6600 (make_pass_scev_cprop): ...new function.
6601 (pass_record_bounds): Convert from a global struct to a subclass of
6602 gimple_opt_pass along with...
6603 (pass_data_record_bounds): ...new pass_data instance and...
6604 (make_pass_record_bounds): ...new function.
6605 (pass_complete_unroll): Convert from a global struct to a subclass of
6606 gimple_opt_pass along with...
6607 (pass_data_complete_unroll): ...new pass_data instance and...
6608 (make_pass_complete_unroll): ...new function.
6609 (pass_complete_unrolli): Convert from a global struct to a subclass of
6610 gimple_opt_pass along with...
6611 (pass_data_complete_unrolli): ...new pass_data instance and...
6612 (make_pass_complete_unrolli): ...new function.
6613 (pass_parallelize_loops): Convert from a global struct to a subclass
6614 of gimple_opt_pass along with...
6615 (pass_data_parallelize_loops): ...new pass_data instance and...
6616 (make_pass_parallelize_loops): ...new function.
6617 (pass_loop_prefetch): Convert from a global struct to a subclass of
6618 gimple_opt_pass along with...
6619 (pass_data_loop_prefetch): ...new pass_data instance and...
6620 (make_pass_loop_prefetch): ...new function.
6621 (pass_iv_optimize): Convert from a global struct to a subclass of
6622 gimple_opt_pass along with...
6623 (pass_data_iv_optimize): ...new pass_data instance and...
6624 (make_pass_iv_optimize): ...new function.
6625 (pass_tree_loop_done): Convert from a global struct to a subclass of
6626 gimple_opt_pass along with...
6627 (pass_data_tree_loop_done): ...new pass_data instance and...
6628 (make_pass_tree_loop_done): ...new function.
6629 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
6630 struct to a subclass of gimple_opt_pass along with...
6631 (pass_data_cse_reciprocals): ...new pass_data instance and...
6632 (make_pass_cse_reciprocals): ...new function.
6633 (pass_cse_sincos): Convert from a global struct to a subclass of
6634 gimple_opt_pass along with...
6635 (pass_data_cse_sincos): ...new pass_data instance and...
6636 (make_pass_cse_sincos): ...new function.
6637 (pass_optimize_bswap): Convert from a global struct to a subclass of
6638 gimple_opt_pass along with...
6639 (pass_data_optimize_bswap): ...new pass_data instance and...
6640 (make_pass_optimize_bswap): ...new function.
6641 (pass_optimize_widening_mul): Convert from a global struct to a
6642 subclass of gimple_opt_pass along with...
6643 (pass_data_optimize_widening_mul): ...new pass_data instance and...
6644 (make_pass_optimize_widening_mul): ...new function.
6645 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
6646 subclass of gimple_opt_pass along with...
6647 (pass_data_phiopt): ...new pass_data instance and...
6648 (make_pass_phiopt): ...new function.
6649 (pass_cselim): Convert from a global struct to a subclass of
6650 gimple_opt_pass along with...
6651 (pass_data_cselim): ...new pass_data instance and...
6652 (make_pass_cselim): ...new function.
6653 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
6654 subclass of gimple_opt_pass along with...
6655 (pass_data_phiprop): ...new pass_data instance and...
6656 (make_pass_phiprop): ...new function.
6657 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
6658 subclass of gimple_opt_pass along with...
6659 (pass_data_pre): ...new pass_data instance and...
6660 (make_pass_pre): ...new function.
6661 (pass_fre): Convert from a global struct to a subclass of
6662 gimple_opt_pass along with...
6663 (pass_data_fre): ...new pass_data instance and...
6664 (make_pass_fre): ...new function.
6665 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
6666 subclass of gimple_opt_pass along with...
6667 (pass_data_reassoc): ...new pass_data instance and...
6668 (make_pass_reassoc): ...new function.
6669 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
6670 subclass of gimple_opt_pass along with...
6671 (pass_data_sink_code): ...new pass_data instance and...
6672 (make_pass_sink_code): ...new function.
6673 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
6674 subclass of gimple_opt_pass along with...
6675 (pass_data_strlen): ...new pass_data instance and...
6676 (make_pass_strlen): ...new function.
6677 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
6678 struct to a subclass of gimple_opt_pass along with...
6679 (pass_data_build_alias): ...new pass_data instance and...
6680 (make_pass_build_alias): ...new function.
6681 (pass_build_ealias): Convert from a global struct to a subclass of
6682 gimple_opt_pass along with...
6683 (pass_data_build_ealias): ...new pass_data instance and...
6684 (make_pass_build_ealias): ...new function.
6685 (pass_ipa_pta): Convert from a global struct to a subclass of
6686 simple_ipa_opt_pass along with...
6687 (pass_data_ipa_pta): ...new pass_data instance and...
6688 (make_pass_ipa_pta): ...new function.
6689 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
6690 subclass of gimple_opt_pass along with...
6691 (pass_data_uncprop): ...new pass_data instance and...
6692 (make_pass_uncprop): ...new function.
6693 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
6694 global struct to a subclass of gimple_opt_pass along with...
6695 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
6696 (make_pass_late_warn_uninitialized): ...new function.
6697 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
6698 to a subclass of gimple_opt_pass along with...
6699 (pass_data_init_datastructures): ...new pass_data instance and...
6700 (make_pass_init_datastructures): ...new function.
6701 (pass_early_warn_uninitialized): Convert from a global struct to a
6702 subclass of gimple_opt_pass along with...
6703 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
6704 (make_pass_early_warn_uninitialized): ...new function.
6705 (pass_update_address_taken): Convert from a global struct to a
6706 subclass of gimple_opt_pass along with...
6707 (pass_data_update_address_taken): ...new pass_data instance and...
6708 (make_pass_update_address_taken): ...new function.
6709 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
6710 struct to a subclass of gimple_opt_pass along with...
6711 (pass_data_release_ssa_names): ...new pass_data instance and...
6712 (make_pass_release_ssa_names): ...new function.
6713 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
6714 subclass of gimple_opt_pass along with...
6715 (pass_data_stdarg): ...new pass_data instance and...
6716 (make_pass_stdarg): ...new function.
6717 * tree-switch-conversion.c (pass_convert_switch): Convert from a
6718 global struct to a subclass of gimple_opt_pass along with...
6719 (pass_data_convert_switch): ...new pass_data instance and...
6720 (make_pass_convert_switch): ...new function.
6721 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
6722 to a subclass of gimple_opt_pass along with...
6723 (pass_data_tail_recursion): ...new pass_data instance and...
6724 (make_pass_tail_recursion): ...new function.
6725 (pass_tail_calls): Convert from a global struct to a subclass of
6726 gimple_opt_pass along with...
6727 (pass_data_tail_calls): ...new pass_data instance and...
6728 (make_pass_tail_calls): ...new function.
6729 * tree-vect-generic.c (pass_lower_vector): Convert from a global
6730 struct to a subclass of gimple_opt_pass along with...
6731 (pass_data_lower_vector): ...new pass_data instance and...
6732 (make_pass_lower_vector): ...new function.
6733 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
6734 gimple_opt_pass along with...
6735 (pass_data_lower_vector_ssa): ...new pass_data instance and...
6736 (make_pass_lower_vector_ssa): ...new function.
6737 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
6738 to a subclass of gimple_opt_pass along with...
6739 (pass_data_slp_vectorize): ...new pass_data instance and...
6740 (make_pass_slp_vectorize): ...new function.
6741 (pass_ipa_increase_alignment): Convert from a global struct to a
6742 subclass of simple_ipa_opt_pass along with...
6743 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
6744 (make_pass_ipa_increase_alignment): ...new function.
6745 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
6746 gimple_opt_pass along with...
6747 (pass_data_vrp): ...new pass_data instance and...
6748 (make_pass_vrp): ...new function.
6749 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
6750 subclass of simple_ipa_opt_pass along with...
6751 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
6752 (make_pass_ipa_free_lang_data): ...new function.
6753 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
6754 gimple_opt_pass along with...
6755 (pass_data_tsan): ...new pass_data instance and...
6756 (make_pass_tsan): ...new function.
6757 (pass_tsan_O0): Convert from a global struct to a subclass of
6758 gimple_opt_pass along with...
6759 (pass_data_tsan_O0): ...new pass_data instance and...
6760 (make_pass_tsan_O0): ...new function.
6761 * var-tracking.c (pass_variable_tracking): Convert from a global
6762 struct to a subclass of rtl_opt_pass along with...
6763 (pass_data_variable_tracking): ...new pass_data instance and...
6764 (make_pass_variable_tracking): ...new function.
6765 * web.c (pass_web): Convert from a global struct to a subclass of
6766 rtl_opt_pass along with...
6767 (pass_data_web): ...new pass_data instance and...
6768 (make_pass_web): ...new function.
6769 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
6770 declaration with that of...
6771 (make_pass_mode_switch_use): ...new function.
6772 (pass_resolve_sw_modes): Replace declaration with that of...
6773 (make_pass_resolve_sw_modes): ...new function.
6774 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
6775 from a global struct to a subclass of rtl_opt_pass along with...
6776 (pass_data_mode_switch_use): ...new pass_data instance and...
6777 (make_pass_mode_switch_use): ...new function.
6778 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
6779 from a global struct to a subclass of rtl_opt_pass along with...
6780 (pass_data_resolve_sw_modes): ...new pass_data instance and...
6781 (make_pass_resolve_sw_modes): ...new function.
6782 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
6783 struct to a subclass of rtl_opt_pass along with...
6784 (pass_data_insert_vzeroupper): ...new pass_data instance and...
6785 (make_pass_insert_vzeroupper): ...new function.
6786 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
6787 global struct to a subclass of rtl_opt_pass along with...
6788 (pass_data_work_around_errata): ...new pass_data instance and...
6789 (make_pass_work_around_errata): ...new function.
6790 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
6791 struct to a subclass of rtl_opt_pass along with...
6792 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
6793 (make_pass_mips_machine_reorg2): ...new function.
6795 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6797 * passes.c (pass_manager::operator new): New.
6799 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6801 Handwritten part of conversion of passes to C++ classes.
6803 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
6804 (toplev.o): Add dep on PASS_MANAGER_H.
6805 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
6806 of early local pases to reflect this moving from a global to a
6807 member of gcc::pass_manager.
6808 (cgraph_add_new_function): Likewise.
6809 * lto-cgraph.c (lto_output_node): Update for conversion of
6810 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
6811 * passes.c (opt_pass::clone): New.
6812 (opt_pass::gate): New.
6813 (opt_pass::execute): New.
6814 (opt_pass::opt_pass): New.
6815 (pass_manager::execute_early_local_passes): New.
6816 (pass_manager::execute_pass_mode_switching): new.
6817 (finish_optimization_passes): Convert to...
6818 (pass_manager::finish_optimization_passes): ...this.
6819 (finish_optimization_passes): Update for conversion of passes to
6821 (register_dump_files_1): Use has_gate since we cannot portably
6822 check a vtable entry against NULL.
6823 (dump_one_pass): Likewise.
6824 (ipa_write_summaries_2): Likewise.
6825 (ipa_write_optimization_summaries_1): Likewise.
6826 (ipa_read_summaries_1): Likewise.
6827 (ipa_read_optimization_summaries_1): Likewise.
6828 (execute_ipa_stmt_fixups): Likewise.
6829 (pass_manager::pass_manager): Rewrite pass-creation, invoking
6830 pass-creation functions rather than wiring up globals, and
6831 storing the results in fields of pass_manager generated using
6833 (pass_manager::dump_profile_report): Update for conversion of
6834 passes to C++ classes.
6835 (pass_manager::execute_ipa_summary_passes): Likewise.
6836 (execute_one_ipa_transform_pass): Likewise.
6837 (execute_one_pass): Use has_gate and has_execute since we cannot
6838 portably check a vtable entry against NULL.
6839 * pass_manager.h (pass_manager::finish_optimization_passes): New.
6840 (pass_manager): Use pass-instances.def to add fields for the
6841 various pass instances.
6842 * toplev.c (finalize): Update for move of
6843 finish_optimization_passes to a method of gcc::pass_manager.
6844 * toplev.h (finish_optimization_passes): Move to method of class
6846 * tree-pass.h (struct pass_data): New.
6847 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
6848 (opt_pass::gate): Convert to virtual function.
6849 (opt_pass::~opt_pass): New.
6850 (opt_pass::clone): New.
6851 (opt_pass::execute): Convert to virtual function.
6852 (opt_pass::opt_pass): New.
6853 (opt_pass::ctxt_): new.
6854 (gimple_opt_pass): Convert to subclass of opt_pass.
6855 (gimple_opt_pass::gimple_opt_pass): New.
6856 (rtl_opt_pass): Convert to subclass of opt_pass.
6857 (rtl_opt_pass::rtl_opt_pass): New.
6858 (ipa_opt_pass_d): Convert to subclass of opt_pass.
6859 (ipa_opt_pass_d::ipa_opt_pass_d): New.
6860 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
6861 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
6862 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
6863 invocation of pass_mode_switching to reflect this moving from a
6864 global to a member of gcc::pass_manager.
6865 (ix86_option_override): Rework how pass_insert_vzeroupper is
6866 added to the pass_manager to reflect autogenerated changes.
6867 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
6869 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
6871 PR rtl-optimization/57708
6872 * recog.c (peep2_find_free_register): Validate all regs in a
6875 2013-08-05 Jan Hubicka <jh@suse.cz>
6878 * cgraph.c (verify_cgraph_node): Accept local flags from other
6880 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
6881 (function_and_variable_visibility): Likewise.
6882 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
6884 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6886 * graph.c (init_graph_slim_pretty_print): Remove.
6887 (print_graph_cfg): Do not call it. Use local pretty printer.
6888 (start_graph_dump): Likewise.
6890 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6892 * gimple-pretty-print.c (buffer): Remove.
6893 (initialized): Likewise.
6894 (maybe_init_pretty_print): Likewise.
6895 (print_gimple_stmt): Do not call it. Use non-static local
6896 pretty_printer variable.
6897 (print_gimple_expr): Likewise.
6898 (print_gimple_seq): Likewise.
6899 (gimple_dump_bb): Likewise.
6901 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6903 * asan.c (asan_pp): Remove.
6904 (asan_pp_initialized): Likewise.
6905 (asan_pp_initialize): Likewise.
6906 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
6907 (asan_emit_stack_protection): Tidy. Use local pretty printer.
6908 (asan_add_global): Likewise.
6910 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
6912 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
6913 * diagnostic.h (diagnostic_flush_buffer): Adjust.
6914 * pretty-print.c (pp_formatted_text_data): Likewise.
6915 (pp_indent): Rename from pp_base_indent.
6916 (pp_format): Rename from pp_base_format.
6917 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
6918 (pp_format_verbatim): Rename from pp_base_format_verbatim.
6919 (pp_flush): Rename from pp_base_flush.
6920 (pp_set_line_maximum_length): Rename from
6921 pp_base_set_line_maximum_length.
6922 (pp_clear_output_area): Rename from pp_base_clear_output_area.
6923 (pp_set_prefix): Rename from pp_base_set_prefix.
6924 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
6925 (pp_emit_prefix): Rename from pp_base_emit_prefix.
6926 (pp_append_text): Rename from pp_base_append_text.
6927 (pp_formatted_text): Rename from pp_base_formatted_text.
6928 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
6929 (pp_remaining_character_count_for_line): Rename from
6930 pp_base_remaining_character_count_for_line.
6931 (pp_newline): Rename from pp_base_newline.
6932 (pp_character): Rename from pp_base_character.
6933 (pp_string): Rename from pp_base_string.
6934 (pp_maybe_space): Rename from pp_base_maybe_space.
6935 * asan.c (asan_pp_string): Adjust.
6936 (asan_emit_stack_protection): Likewise.
6937 (asan_add_global): Likewise.
6938 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
6939 * tree-mudflap.c (mf_varname_tree): Likewise.
6940 * tree-pretty-print.c (pp_tree_identifier): Rename from
6941 pp_base_tree_identifier.
6942 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
6943 Declare as function.
6945 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6947 * pretty-print.h (pp_bar_bar): New.
6948 (pp_ampersand_ampersand): Likewise.
6949 (pp_less_equal): Likewise.
6950 (pp_greater_equal): Likewise.
6951 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
6952 printer functions instead of pp_string or operators and punctuators.
6953 (dump_gimple_call): Likewise.
6954 (dump_gimple_omp_for): Likewise.
6955 (dump_gimple_transaction): Likewise.
6956 (dump_gimple_phi): Likewise.
6957 (pp_gimple_stmt_1): Likewise.
6958 * sched-vis.c (print_insn): Likewise.
6959 * tree-mudflap.c (mf_varname_tree): Likewise.
6960 * tree-pretty-print.c (dump_block_node): Likewise.
6961 (dump_generic_node): Likewise.
6963 2013-08-02 Jan Hubicka <jh@suse.cz>
6965 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
6967 * lto-streamer-out.c (tree_is_indexable): Results decls and
6968 parm decls are not indexable.
6969 (DFS_write_tree_body): Do not follow args and results.
6970 (hash_tree): Likewise.
6971 (output_functions): Rearrange so struct function is needed
6972 only when real body is output; be able to also ouptut abstract
6973 functions; output DECL_ARGUMENTS and DECL_RESULT.
6974 (lto_output): When not in WPA, ale store abstract functions.
6975 (write_symbol): Do not care about RESULT_DECL.
6976 (output_symbol_p): Handle correctly sbtract decls.
6977 * lto-streamer-in.c (input_function): Rearrange so struct
6978 function can be NULL at entry; allow streaming of
6979 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
6980 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
6981 sanity check during LTO.
6982 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
6983 RESULT_DECl and DECL_ARGUMENTS.
6984 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
6987 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6989 * pretty-print.h (pp_underscore): New.
6991 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
6992 printer functions instead of pp_character.
6993 (dump_binary_rhs): Likewise.
6994 (dump_ternary_rhs): Likewise.
6995 (dump_gimple_call_args): Likewise.
6996 (pp_points_to_solution): Likewise.
6997 (dump_gimple_call): Likewise.
6998 (dump_gimple_switch): Likewise.
6999 (dump_gimple_cond): Likewise.
7000 (dump_gimple_bind): Likewise.
7001 (dump_gimple_try): Likewise.
7002 (dump_gimple_omp_for): Likewise.
7003 (dump_gimple_omp_continue): Likewise.
7004 (dump_gimple_omp_single): Likewise.
7005 (dump_gimple_omp_sections): Likewise.
7006 (dump_gimple_omp_block): Likewise.
7007 (dump_gimple_omp_critical): Likewise.
7008 (dump_gimple_transaction): Likewise.
7009 (dump_gimple_asm): Likewise.
7010 (dump_gimple_phi): Likewise.
7011 (dump_gimple_omp_parallel): Likewise.
7012 (dump_gimple_omp_task): Likewise.
7013 (dump_gimple_omp_atomic_load): Likewise.
7014 (dump_gimple_omp_atomic_store): Likewise.
7015 (dump_gimple_mem_ops): Likewise.
7016 (pp_gimple_stmt_1): Likewise.
7017 (pp_cfg_jump): Likewise.
7018 (dump_implicit_edges): Likewise.
7019 (gimple_dump_bb_for_graph): Likewise.
7020 * graph.c (draw_cfg_node): Likewise.
7021 * langhooks.c (lhd_print_error_function): Likewise.
7022 * sched-vis.c (print_exp): Likewise.
7023 (print_value): Likewise.
7024 (print_pattern): Likewise.
7025 (print_insn): Likewise.
7026 (rtl_dump_bb_for_graph): Likewise.
7027 * tree-pretty-print.c (dump_function_declaration): Likewise.
7028 (dump_array_domain): Likewise.
7029 (dump_omp_clause): Likewise.
7030 (dump_location): Likewise.
7031 (dump_generic_node): Likewise.
7032 (print_struct_decl): Likewise.
7033 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7035 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7037 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7038 candidate table when replacing a candidate statement.
7039 (replace_rhs_if_not_dup): Likewise.
7040 (replace_one_candidate): Likewise.
7042 2013-08-02 Jan Hubicka <jh@suse.cz>
7043 Martin Liska <marxin.liska@gmail.com>
7045 * cgraphunit.c (add_new_function): Fix logic when adding from
7047 (assemble_thunk): Rename to ...
7048 (expand_thunk); .. this one; export; get it working with
7049 general functions; make produced gimple valid.
7050 * cgraph.h (expand_thunk): Declare.
7052 2013-08-02 Jan Hubicka <jh@suse.cz>
7054 * ipa-cp.c (gather_context_independent_values): Use
7055 ipa_get_param_move_cost.
7056 (get_replacement_map): Remove PARAM; move parameter folding
7058 (create_specialized_node): Update.
7059 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7060 assert that we have gimple body; update move_cost.
7061 (count_formal_params): Assert that we have gimple body.
7062 (ipa_dump_param): New function.
7063 (ipa_alloc_node_params): Break out from ...
7064 (ipa_initialize_node_params): ... here.
7065 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7066 (ipa_write_node_info): Stream move costs.
7067 (ipa_read_node_info): Read move costs.
7068 (ipa_update_after_lto_read): Do not recompute node params.
7069 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7070 (ipa_get_param): Check we are not in WPA.
7071 (ipa_get_param_move_cost): New.
7072 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7073 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7074 parm numbers to be present.
7076 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7078 PR rtl-optimization/58048
7079 * lra-constraints.c (process_alt_operands): Don't check asm
7080 operand on register.
7082 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7084 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7085 the implied StoreLoad barrier for atomic operations if before.
7087 2013-08-02 Jan Hubicka <jh@suse.cz>
7088 Martin Liska <marxin.liska@gmail.com>
7090 * cgraph.c (cgraph_function_body_availability): Do not check
7092 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7093 symtab_node_availability): Declare.
7094 * ipa.c (can_replace_by_local_alias): New.
7095 (function_and_variable_visibility): Use it.
7096 * symtab.c (symtab_for_node_and_aliases,
7097 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7099 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7101 PR rtl-optimization/57963
7102 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7103 (lra_constraints): Use them.
7105 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7107 * config/arm/types.md (define_attr "type"): Add "load_acq"
7109 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7111 (cortex_a53_store1): Likewise.
7113 2013-08-01 Jan Hubicka <jh@suse.cz>
7115 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7116 partitions are not needed.
7118 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7120 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7121 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7122 MAYBE_NON_Q_CLASS_P where appropriate.
7124 2013-08-01 Jan Hubicka <jh@suse.cz>
7126 * cgraph.h (release_function_body): Declare.
7127 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7128 of unused delcarations.
7130 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7132 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7133 RTL form when subtracting a constant.
7135 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7137 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7138 Generate canonical plus rtx with negated immediate instead of minus
7140 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7142 2013-08-01 Jan Hubicka <jh@suse.cz>
7144 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7145 (cgraph_release_function_body): Likewise.
7146 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7147 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7148 to used_as_abstract_origin.
7149 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7150 symbols used as abstract origins.
7151 * cgraphunit.c (analyze_functions): Update.
7152 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7153 used_as_abstract_origin.
7154 * tree-inline.c (tree_function_versioning): Update
7155 used_as_abstract_origin; be ready for DECL_RESULT and
7156 DECL_ARGUMENTS to be NULL.
7158 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7159 for abstract functions.
7160 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7163 2013-08-01 Jan Hubicka <jh@suse.cz>
7165 * profile.c (compute_value_histograms): Fix thinko.
7167 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7169 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7170 aarch-common-protos.h to extra_headers.
7171 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7172 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7173 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7175 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7177 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7178 Include "../arm/types.md". Define "type" attribute for all patterns.
7179 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7182 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7184 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7185 to support power8 load fusion.
7186 (fusion_gpr_mem_load): Likewise.
7188 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7190 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7191 declarations for power8 load fusion.
7192 (emit_fusion_gpr_load): Likewise.
7194 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7195 tuning for power8, turn on fusion mode by default. Turn on sign
7196 extending fusion mode if normal fusion mode is on, and we are at
7198 (fusion_gpr_load_p): New function, return true if we can fuse an
7199 addis instruction with a dependent load to a GPR.
7200 (emit_fusion_gpr_load): Emit the instructions for power8 load
7203 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7204 (VSX load fusion peepholes): New peepholes to fuse together an
7205 addi instruction with a VSX load instruction.
7207 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7208 peepholes to fuse an addis instruction with a load to a GPR base
7209 register. If we are supporting sign extending fusions, convert
7210 sign extending loads to zero extending loads and add an explicit
7213 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7215 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7216 aarch-common-protos.h to extra_headers.
7217 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7218 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7219 (arm_early_store_addr_dep): Likewise.
7220 (arm_no_early_alu_shift_dep): Likewise.
7221 (arm_no_early_alu_shift_value_dep): Likewise.
7222 (arm_no_early_mul_dep): Likewise.
7223 (arm_no_early_store_addr_dep): Likewise.
7224 (arm_mac_accumulator_is_mul_result): Likewise.
7225 (arm_mac_accumulator_is_result): Likewise.
7226 * config/arm/aarch-common.c: ... here. New file.
7227 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7229 (arm_early_store_addr_dep): Likewise.
7230 (arm_no_early_alu_shift_dep): Likewise.
7231 (arm_no_early_alu_shift_value_dep): Likewise.
7232 (arm_no_early_mul_dep): Likewise.
7233 (arm_no_early_store_addr_dep): Likewise.
7234 (arm_mac_accumulator_is_mul_result): Likewise.
7235 (arm_mac_accumulator_is_result): Likewise.
7236 * config/arm/aarch-common-protos.h: ... here. New file.
7237 * config/arm/t-arm (aarch-common.o): Define.
7239 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7241 * config/arm/arm.md: Include new file "types.md".
7242 (define_attr "type"): Move from here to ...
7243 (define_attr "mul32"): Likewise.
7244 (define_attr "mul64"): Likewise.
7245 * config/arm/types.md: ... here. New file.
7247 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7249 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7250 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7252 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7254 * gen-pass-instances.awk: Fix offset of substr().
7256 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7258 * Makefile.in (pass-instances.def): New.
7259 (passes.o): Replace dependency on passes.def with one on
7262 * gen-pass-instances.awk: New.
7264 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7265 rather than passes.def, updating local definition of NEXT_PASS
7266 macro to add an extra NUM parameter (currently unused).
7268 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7270 * Makefile.in (PASS_MANAGER_H): New.
7271 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7272 (passes.o): Likewise.
7273 (statistics.o): Likewise.
7274 (cgraphunit.o): Likewise.
7275 (context.o): Depend on PASS_MANAGER_H.
7277 * pass_manager.h: New.
7279 * cgraphunit.c (cgraph_add_new_function): Update for moves
7280 of globals to fields of pass_manager.
7281 (analyze_function): Likewise.
7282 (expand_function): Likewise.
7283 (ipa_passes): Likewise.
7284 (compile): Likewise.
7286 * context.c (context::context): New.
7287 * context.h (context::context): New.
7288 (context::get_passes): New.
7289 (context::passes_): New.
7291 * lto-cgraph.c (input_node): Update for moves of globals to
7292 fields of pass_manager.
7294 * passes.c (all_passes): Remove, in favor of a field of the
7295 same name within the new class pass_manager.
7296 (all_small_ipa_passes): Likewise.
7297 (all_lowering_passes): Likewise.
7298 (all_regular_ipa_passes): Likewise.
7299 (all_late_ipa_passes): Likewise.
7300 (all_lto_gen_passes): Likewise.
7301 (passes_by_id): Likewise.
7302 (passes_by_id_size): Likewise.
7303 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7304 the new class pass_manager.
7305 (set_pass_for_id): Convert to...
7306 (pass_manager::set_pass_for_id): ...method.
7307 (get_pass_for_id): Convert to...
7308 (pass_manager::get_pass_for_id): ...method.
7309 (register_one_dump_file): Move body of implementation into...
7310 (pass_manager::register_one_dump_file): ...here.
7311 (register_dump_files_1): Convert to...
7312 (pass_manager::register_dump_files_1): ...method.
7313 (register_dump_files): Convert to...
7314 (pass_manager::register_dump_files): ...method.
7315 (create_pass_tab): Update for moves of globals to fields of
7317 (dump_passes): Move body of implementation into...
7318 (pass_manager::dump_passes): ...here.
7319 (register_pass): Move body of implementation into...
7320 (pass_manager::register_pass): ...here.
7321 (init_optimization_passes): Convert into...
7322 (pass_manager::pass_manager): ...constructor for new
7323 pass_manager class, and initialize the pass_lists array.
7324 (check_profile_consistency): Update for moves of globals to
7325 fields of pass_manager.
7326 (dump_profile_report): Move body of implementation into...
7327 (pass_manager::dump_profile_report): ...here.
7328 (ipa_write_summaries_1): Update for moves of pass lists from
7329 being globals to fields of pass_manager.
7330 (ipa_write_optimization_summaries): Likewise.
7331 (ipa_read_summaries): Likewise.
7332 (ipa_read_optimization_summaries): Likewise.
7333 (execute_all_ipa_stmt_fixups): Likewise.
7335 * statistics.c (statistics_fini): Update for moves of globals to
7336 fields of pass_manager.
7338 * toplev.c (general_init): Replace call to
7339 init_optimization_passes with construction of the pass_manager
7342 * tree-pass.h (all_passes): Remove, in favor of a field of the
7343 same name within the new class pass_manager.
7344 (all_small_ipa_passes): Likewise.
7345 (all_lowering_passes): Likewise.
7346 (all_regular_ipa_passes): Likewise.
7347 (all_lto_gen_passes): Likewise.
7348 (all_late_ipa_passes): Likewise.
7349 (passes_by_id): Likewise.
7350 (passes_by_id_size): Likewise.
7351 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7352 the new class pass_manager.
7353 (get_pass_for_id): Remove.
7355 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7357 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7360 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7362 * arm.md (mulhi3): New expand pattern.
7364 2013-07-30 Jan Hubicka <jh@suse.cz>
7365 Martin Liska <marxin.liska@gmail.com>
7367 * profile.c (compute_value_histograms): Do not ICE when
7368 there is mismatch only on some counters.
7370 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7372 PR rtl-optimization/57637
7373 * function.c (move_insn_for_shrink_wrap): Also check the
7374 GEN set of the LIVE problem for the liveness analysis
7375 if it exists, otherwise give up.
7377 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7379 PR tree-optimization/57993
7380 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7381 replaced statement in the candidate table.
7382 (phi_add_costs): Return infinite cost when the hidden basis does
7383 not dominate all phis on which the candidate is dependent.
7384 (replace_one_candidate): Record replaced statement in the
7387 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7389 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7390 (ashlv2si3): New expander.
7391 (*ashlv2si3_i): New define_insn_and_split.
7392 * predicates.md (float_operation): Allow patterns with three
7395 PR rtl-optimization/58021
7396 * mode-switching.c (create_pre_exit): Always split off preceding
7397 insns if we are not at the basic block head.
7399 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7401 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7402 (UCLIBC_DYNAMIC_LINKER): New macro.
7403 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7405 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7406 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7408 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7409 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7410 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7411 for SF and DF modes. Use ieee_quad_format for TF mode.
7412 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7413 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7414 (mips_option_override): Handle `-mnan=legacy'.
7415 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7416 `-mabs=2008' and `-mnan=2008'.
7417 (OPTION_DEFAULT_SPECS): Add "nan" default.
7418 (ASM_SPEC): Handle `-mnan='.
7419 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7420 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7421 comment accordingly.
7422 (neg<mode>2): Likewise.
7423 * config/mips/mips.opt (mabs, mnan): New options.
7424 * doc/install.texi (Configuration): Document `--with-nan=' option.
7425 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7427 (MIPS Options): Document them.
7428 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7429 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7430 * configure: Regenerate.
7431 * config.in: Regenerate.
7433 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7435 * config/i386/i386.md (float post-reload splitters): Do not check
7436 for subregs of SSE registers.
7438 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7439 H.J. Lu <hongjiu.lu@intel.com>
7443 * config/i386/i386.md (post-reload splitter
7444 to avoid partial SSE reg dependency stalls): New pattern.
7446 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7448 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7450 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7452 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7453 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7455 2013-07-26 Cary Coutant <ccoutant@google.com>
7457 * dwarf2out.c (die_checksum_ordered): Don't include template
7458 instantiations in signature.
7459 (is_template_parameter): New function.
7460 (is_template_instantiation): New function.
7461 (generate_skeleton_bottom_up): Don't include template instantiations
7463 (generate_skeleton): Likewise.
7464 (break_out_comdat_types): Move recursive call to break out nested
7466 (prune_unused_types_mark_generic_parms_dies): Call
7467 is_template_parameter.
7469 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7471 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7472 uses vector registers.
7473 * config/aarch64/iterators.md: Add attributes rtn and vas.
7475 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7476 Richard Earnshaw <richard.earnshaw@arm.com>
7478 * combine.c (simplify_comparison): Re-canonicalize operands
7480 * config/arm/arm.md (movcond_addsi): New splitter.
7482 2013-07-25 Sterling Augustine <saugustine@google.com>
7484 * dwarf2out.c (size_of_pubnames): Move code to...
7485 (include_pubname_in_output): ...here. New.
7486 (want_pubnames): Rearrange.
7487 (output_pubnames): Call include_pubname_in_output. Move assertion.
7489 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7491 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7493 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7496 * doc/extend.texi: Remove obsolete builtins. Fix
7497 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7499 2013-07-25 Jan Hubicka <jh@suse.cz>
7501 * cgraph.c (release_function_body): Break out from ...
7502 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7504 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7505 old_tree in the map.
7506 (create_specialized_node): Update.
7507 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7509 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7510 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7511 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7512 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7515 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7517 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7518 addsi3_carryin_alt2_<optab>): Correct output template.
7520 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7522 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7523 Adjust for arm_restrict_it.
7524 Remove trailing whitespace.
7526 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7528 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7529 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7531 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7533 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7535 PR rtl-optimization/57960
7536 * lra-constraints.c (process_alt_operands): Use the right mode
7537 when checking strict_low.
7539 2013-07-25 Jan Hubicka <jh@suse.cz>
7541 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7542 * cgraph.c (cgraph_remove_node): Do not release function body
7543 when in cgraph streaming.
7544 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7545 in other partitions are not considered reachable; fix handling of
7548 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7550 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7552 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7557 * config/arm/arm.md ("*sibcall_insn): Replace use of
7558 Ss with US. Adjust output for v5 and v4t.
7559 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7561 * config/arm/constraints.md ("Ss"): Rename to US.
7563 2013-07-25 Terry Guo <terry.guo@arm.com>
7565 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7566 shift_add/shift_sub0/shift_sub1 RTXs.
7568 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7569 Anton Blanchard <anton@au1.ibm.com>
7571 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7572 (altivec_vpks<VI_char>ss): Likewise.
7573 (altivec_vpks<VI_char>us): Likewise.
7574 (altivec_vpku<VI_char>us): Likewise.
7575 (altivec_vpku<VI_char>um): Likewise.
7577 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7579 Introduce context class.
7581 * Makefile.in (CONTEXT_H): New.
7582 (OBJS): Add context.o.
7583 (toplev.o): Add CONTEXT_H to dependencies.
7586 * toplev.c (general_init): Create the singleton gcc::context instance.
7592 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
7594 PR rtl-optimization/57968
7595 * mode-switching.c (create_pre_exit): Allow instructions that
7596 don't set a return register to need a non-exit mode.
7598 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7599 Anton Blanchard <anton@au1.ibm.com>
7601 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
7602 operands to vperm for little endian.
7603 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
7604 of lvsl to create the control mask for a vperm for little endian.
7606 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7607 Anton Blanchard <anton@au1.ibm.com>
7609 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
7610 two operands for little-endian.
7612 2013-07-23 Steve Ellcey <sellcey@mips.com>
7614 * config/mips/mips.c (mips_case_values_threshold): New.
7615 (TARGET_CASE_VALUES_THRESHOLD): Define.
7617 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7618 Anton Blanchard <anton@au1.ibm.com>
7620 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
7621 selection of field for vector splat in little endian mode.
7623 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7625 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
7626 expanders to rs6000.md.
7627 (ior<mode>3): Likewise.
7628 (and<mode>3): Likewise.
7629 (one_cmpl<mode>2): Likewise.
7630 (nor<mode>3): Likewise.
7631 (andc<mode>3): Likewise.
7632 (eqv<mode>3): Likewise.
7633 (nand<mode>3): Likewise.
7634 (orc<mode>3): Likewise.
7636 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
7639 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
7640 to split multi-word logical operations.
7641 (rs6000_split_logical_di): Likewise.
7642 (rs6000_split_logical): Likewise.
7644 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
7645 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
7646 and allow TImode operations in 32-bit.
7647 (vsx_and<mode>3_64bit): Likewise.
7648 (vsx_ior<mode>3_32bit): Likewise.
7649 (vsx_ior<mode>3_64bit): Likewise.
7650 (vsx_xor<mode>3_32bit): Likewise.
7651 (vsx_xor<mode>3_64bit): Likewise.
7652 (vsx_one_cmpl<mode>2_32bit): Likewise.
7653 (vsx_one_cmpl<mode>2_64bit): Likewise.
7654 (vsx_nor<mode>3_32bit): Likewise.
7655 (vsx_nor<mode>3_64bit): Likewise.
7656 (vsx_andc<mode>3_32bit): Likewise.
7657 (vsx_andc<mode>3_64bit): Likewise.
7658 (vsx_eqv<mode>3_32bit): Likewise.
7659 (vsx_eqv<mode>3_64bit): Likewise.
7660 (vsx_nand<mode>3_32bit): Likewise.
7661 (vsx_nand<mode>3_64bit): Likewise.
7662 (vsx_orc<mode>3_32bit): Likewise.
7663 (vsx_orc<mode>3_64bit): Likewise.
7665 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
7666 logical types in GPRs.
7668 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
7669 logical insns to rs6000.md, and allow TImode operations in
7671 (altivec_ior<mode>3): Likewise.
7672 (altivec_xor<mode>3): Likewise.
7673 (altivec_one_cmpl<mode>2): Likewise.
7674 (altivec_nor<mode>3): Likewise.
7675 (altivec_andc<mode>3): Likewise.
7677 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
7678 attributes for moving the 128-bit logical operations into
7680 (BOOL_REGS_OUTPUT): Likewise.
7681 (BOOL_REGS_OP1): Likewise.
7682 (BOOL_REGS_OP2): Likewise.
7683 (BOOL_REGS_UNARY): Likewise.
7684 (BOOL_REGS_AND_CR0): Likewise.
7685 (one_cmpl<mode>2): Add support for DI logical operations on
7686 32-bit, splitting the operations to 32-bit.
7690 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
7691 changes to combine the 32/64-bit code, allow logical operations on
7692 TI mode in 32-bit, and to use similar match_operator patterns like
7693 scalar mode uses. Combine the Altivec and VSX code for logical
7694 operations, and move it here.
7695 (ior<mode>3, 128-bit types): Likewise.
7696 (xor<mode>3, 128-bit types): Likewise.
7697 (one_cmpl<mode>3, 128-bit types): Likewise.
7698 (nor<mode>3, 128-bit types): Likewise.
7699 (andc<mode>3, 128-bit types): Likewise.
7700 (eqv<mode>3, 128-bit types): Likewise.
7701 (nand<mode>3, 128-bit types): Likewise.
7702 (orc<mode>3, 128-bit types): Likewise.
7703 (and<mode>3_internal): Likewise.
7704 (bool<mode>3_internal): Likewise.
7705 (boolc<mode>3_internal1): Likewise.
7706 (boolc<mode>3_internal2): Likewise.
7707 (boolcc<mode>3_internal1): Likewise.
7708 (boolcc<mode>3_internal2): Likewise.
7709 (eqv<mode>3_internal1): Likewise.
7710 (eqv<mode>3_internal2): Likewise.
7711 (one_cmpl1<mode>3_internal): Likewise.
7713 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7715 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7716 Rename flag_stack_usage to flag_stack_usage_info.
7718 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
7720 * config/microblaze/sync.md: New file.
7721 * config/microblaze/microblaze.md: Include sync.md
7722 * config/microblaze/microblaze.c: Add print_operand 'y'.
7723 * config/microblaze/constraints.md: Add memory_contraint
7724 'Q' which is a single register.
7726 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
7728 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
7730 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7732 * reload.c (find_reloads): Exit loop once we find this operand
7733 cannot be reloaded somehow for this alternative.
7735 * reload.c (find_reloads): Exit loop once we find a hard register.
7737 * rtlanal.c (computed_jump_p): Exit loop once we find label
7740 * i386.c (ix86_pad_returns): Exit loop after setting replace.
7742 * cfgloopmanip.c (remove_path): Exit loop after setting
7745 * gensupport.c (subst_dup): Avoid loop if code is not
7746 MATCH_DUP nor MATCH_OP_DUP.
7748 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
7750 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
7752 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7754 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
7755 true for SP_REGNUM if mode == ptr_mode.
7756 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
7757 with value R0_REGNUM + 31.
7759 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7761 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
7762 pad pointer-typed argument downward.
7764 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7766 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
7767 and __ILP32__ when the ILP32 model is in use.
7769 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7771 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
7772 (aarch64_load_symref_appropriately): In the case of
7773 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
7774 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
7775 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
7776 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
7777 if 'mode' doesn't equal to 'ptr_mode'.
7778 (aarch64_output_mi_thunk): Add an assertion on the alignment of
7779 'vcall_offset'; change to call aarch64_emit_move differently depending
7780 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
7781 to calculate the upper bound of 'vcall_offset'.
7782 (aarch64_cannot_force_const_mem): Change to also return true if
7784 (aarch64_legitimize_reload_address): In the case of large
7785 displacements, add new local variable 'xmode' and an assertion
7786 based on it; change to use 'xmode' to generate the new rtx and
7788 (aarch64_asm_trampoline_template): Change to generate the template
7789 differently depending on TARGET_ILP32 or not; change to use
7790 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
7791 (aarch64_trampoline_size): Removed.
7792 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
7793 and replace immediate literals with it. Change to use 'ptr_mode'
7794 instead of 'DImode' and call convert_memory_address if the mode
7795 of 'fnaddr' doesn't equal to 'ptr_mode'.
7796 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
7798 (aarch64_elf_asm_destructor): Likewise.
7799 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
7800 on TARGET_ILP32 instead of aarch64_trampoline_size.
7801 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
7802 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
7803 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
7804 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
7805 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
7806 (storewb_pair<GPI:mode>_<P:mode>): ... this.
7807 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
7808 depending on the value of 'mode'.
7809 (add_losym_<mode>): New.
7810 (ldr_got_small_<mode>): New, based on ldr_got_small.
7811 (ldr_got_small): Remove.
7812 (ldr_got_small_sidi): New.
7813 * config/aarch64/iterators.md (P): New.
7814 (PTR): Change to 'ptr_mode' in the condition.
7816 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7818 * config.gcc (aarch64*-*-*): Support --with-abi.
7819 (aarch64*-*-elf): Support --with-multilib-list.
7820 (aarch64*-*-linux*): Likewise.
7821 (supported_defaults): Add abi to aarch64*-*-*.
7822 * configure.ac: Mention AArch64 for --with-multilib-list.
7823 * configure: Re-generated.
7824 * config/aarch64/biarchilp32.h: New file.
7825 * config/aarch64/biarchlp64.h: New file.
7826 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
7828 (MULTILIB_DEFAULTS): Ditto.
7829 (DRIVER_SELF_SPECS): Ditto.
7830 (ASM_SPEC): Update to also substitute -mabi.
7831 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
7832 file whose name depends on -mabi= and -mbig-endian.
7833 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
7835 (POINTER_SIZE): New define.
7836 (POINTERS_EXTEND_UNSIGNED): Ditto.
7837 (enum aarch64_abi_type): New enumeration tag.
7838 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
7839 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
7840 (TARGET_ILP32): New define.
7841 * config/aarch64/aarch64.opt (mabi): New.
7843 (ilp32, lp64): New values for -mabi.
7844 * config/aarch64/t-aarch64 (comma): New define.
7845 (MULTILIB_OPTIONS): Ditto.
7846 (MULTILIB_DIRNAMES): Ditto.
7847 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
7848 * doc/invoke.texi: Document -mabi for AArch64.
7850 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
7852 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
7854 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7855 Anton Blanchard <anton@au1.ibm.com>
7857 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
7858 endianness when selecting field to splat.
7860 2013-07-22 Eric Christopher <echristo@gmail.com>
7862 * dwarf2out.c (die_odr_checksum): New function to use
7863 CHECKSUM_ macros and ULEB128 for DIE tag.
7864 (generate_type_signature): Use.
7866 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
7868 * config.gcc (sparc*-*-*): Accept leon3 processor.
7869 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
7870 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
7871 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
7872 * config/sparc/sparc.opt (enum processor_type): Add leon3.
7873 (mfix-ut699): Adjust comment.
7874 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
7875 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
7876 (CPP_CPU_SPEC): Likewise.
7877 (ASM_CPU_SPEC): Likewise.
7878 * config/sparc/sparc.c (leon3_cost): New constant.
7879 (sparc_option_override): Add leon3 support.
7880 (mem_ref): New function.
7881 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
7882 (sparc_do_work_around_errata): Look into the instruction in the delay
7883 slot and adjust accordingly. Add fix for the data cache nullify issues
7884 of the UT699. Change insertion position for the NOP.
7885 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
7886 (leon3_load): New reservation.
7887 (leon_store): Bump latency to 2.
7888 (grfpu): New automaton.
7889 (grfpu_alu): New unit.
7890 (grfpu_ds): Likewise.
7891 (leon_fp_alu): Adjust.
7892 (leon_fp_mult): Delete.
7893 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
7894 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
7895 * config/sparc/sparc.md (cpu): Add leon3.
7896 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
7898 (atomic_test_and_set): Likewise.
7901 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
7903 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
7904 default for R5900 targets.
7905 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
7906 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
7907 * config/mips/mips.c (mips_option_override): Report an error for
7908 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
7909 for -march=r5900 -mhard-float.
7911 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7913 * df-problems.c (can_move_insns_across): Exit loop once we
7914 find a non-fixed, non-global register.
7916 * ipa-pure-const.c (propagate_nothrow): Exit loop after
7919 * omega.c (omega_eliminate_red): Break after setting red_found.
7920 (omega_problem_has_red_equations): Similarly after setting found.
7921 (omega_query_variable): Similarly after setting coupled.
7923 2013-07-22 Marek Polacek <polacek@redhat.com>
7925 * gimplify.c: Don't include gimple.h twice.
7927 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7929 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
7930 instead of TARGET_THUMB1.
7931 (Pz): New constraint.
7932 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
7934 (compare_negsi_si): Likewise.
7935 (compare_addsi2_op0): Likewise.
7936 (compare_addsi2_op1): Likewise.
7937 (addsi3_carryin_<optab>): Likewise.
7938 (addsi3_carryin_alt2_<optab>): Likewise.
7939 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
7940 for arm_restrict_it.
7941 (subsi3_carryin): Likewise.
7942 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
7943 (minmax_arithsi): Disable for arm_restrict_it.
7944 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
7945 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
7946 (satsi_<SAT:code>_shift): Likewise.
7947 (arm_shiftsi3): Add alternative for 16-bit encoding.
7948 (arm32_movhf): Disable for arm_restrict_it.
7949 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
7950 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
7952 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
7954 * config/arm/arm.md (attribute "insn"): Delete.
7955 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
7956 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
7957 (not_shiftsi): Update for attribute change.
7958 (not_shiftsi_compare0): Likewise.
7959 (not_shiftsi_compare0_scratch): Likewise.
7960 (arm_one_cmplsi2): Likewise.
7961 (thumb1_one_cmplsi2): Likewise.
7962 (notsi_compare0): Likewise.
7963 (notsi_compare0_scratch): Likewise.
7964 (thumb1_movdi_insn): Likewise.
7965 (arm_movsi_insn): Likewise.
7966 (movhi_insn_arch4): Likewise.
7967 (movhi_bytes): Likewise.
7968 (arm_movqi_insn): Likewise.
7969 (thumb1_movqi_insn): Likewise.
7970 (arm32_movhf): Likewise.
7971 (thumb1_movhf): Likewise.
7972 (arm_movsf_soft_insn): Likewise.
7973 (thumb1_movsf_insn): Likewise.
7974 (thumb_movdf_insn): Likewise.
7975 (movsicc_insn): Likewise.
7976 (movsfcc_soft_insn): Likewise.
7977 (and_scc): Likewise.
7978 (cond_move): Likewise.
7979 (if_move_not): Likewise.
7980 (if_not_move): Likewise.
7981 (if_shift_move): Likewise.
7982 (if_move_shift): Likewise.
7983 (if_shift_shift): Likewise.
7984 (if_not_arith): Likewise.
7985 (if_arith_not): Likewise.
7986 (cond_move_not): Likewise.
7987 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
7988 (neon_mov<mode>): Likewise.
7989 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
7990 (thumb2_movsi_vfp): Likewise.
7991 (movsf_vfp): Likewise.
7992 (thumb2_movsf_vfp): Likewise.
7993 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7995 (cortexa7_older_only): Likewise.
7996 (cortexa7_younger): Likewise.
7997 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
7998 (1020alu_shift_op): Likewise.
7999 (1020alu_shift_reg_op): Likewise.
8000 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8001 (alu_shift_op): Likewise.
8002 (alu_shift_reg_op): Likewise.
8003 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8004 (11_alu_shift_op): Likewise.
8005 (11_alu_shift_reg_op): Likewise.
8006 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8007 (9_alu_shift_reg_op): Likewise.
8008 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8010 (cortex_a15_alu_shift): Likewise.
8011 (cortex_a15_alu_shift_reg): Likewise.
8012 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8013 (cortex_a5_alu_shift): Likewise.
8014 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8016 (cortex_a53_alu_shift): Likewise.
8017 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8019 (cortex_a7_alu_reg): Likewise.
8020 (cortex_a7_alu_shift): Likewise.
8021 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8022 (cortex_a8_alu_shift): Likewise.
8023 (cortex_a8_alu_shift_reg): Likewise.
8024 (cortex_a8_mov): Likewise.
8025 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8026 (cortex_a9_dp_shift): Likewise.
8027 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8028 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8029 (cortex_r4_mov): Likewise.
8030 (cortex_r4_alu_shift): Likewise.
8031 (cortex_r4_alu_shift_reg): Likewise.
8032 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8033 (526_alu_shift_op): Likewise.
8034 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8035 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8036 (626te_alu_shift_op): Likewise.
8037 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8038 (726te_alu_op): Likewise.
8039 (726te_alu_shift_op): Likewise.
8040 (726te_alu_shift_reg_op): Likewise.
8041 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8042 (mp626_alu_shift_op): Likewise.
8043 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8044 (pj4_alu_e1_conds): Likewise.
8045 (pj4_alu): Likewise.
8046 (pj4_alu_conds): Likewise.
8047 (pj4_shift): Likewise.
8048 (pj4_shift_conds): Likewise.
8049 (pj4_alu_shift): Likewise.
8050 (pj4_alu_shift_conds): Likewise.
8052 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8054 * config/arm/predicates.md (shiftable_operator_strict_it):
8056 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8057 Disable cond_exec version for arm_restrict_it.
8058 (thumb2_smaxsi3): Convert to generate cond_exec.
8059 (thumb2_sminsi3): Likewise.
8060 (thumb32_umaxsi3): Likewise.
8061 (thumb2_uminsi3): Likewise.
8062 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8063 (thumb2_neg_abssi2): Likewise.
8064 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8065 (thumb2_movsicc_insn): Adjust alternatives.
8066 (thumb2_mov_negscc): Disable for arm_restrict_it.
8067 (thumb2_mov_negscc_strict_it): New pattern.
8068 (thumb2_mov_notscc_strict_it): New pattern.
8069 (thumb2_mov_notscc): Disable for arm_restrict_it.
8070 (thumb2_ior_scc): Likewise.
8071 (thumb2_ior_scc_strict_it): New pattern.
8072 (thumb2_cond_move): Adjust for arm_restrict_it.
8073 (thumb2_cond_arith): Disable for arm_restrict_it.
8074 (thumb2_cond_arith_strict_it): New pattern.
8075 (thumb2_cond_sub): Adjust for arm_restrict_it.
8076 (thumb2_movcond): Likewise.
8077 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8078 (thumb2_zero_extendhisi2_v6): Likewise.
8079 (thumb2_zero_extendqisi2_v6): Likewise.
8080 (orsi_notsi_si): Likewise.
8081 (orsi_not_shiftsi_si): Likewise.
8083 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8085 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8086 instruction sequence is 1 byte shorter.
8088 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8090 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8091 it is not needed after split.
8093 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8096 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8097 second label for nonlocal goto receivers. Don't output pic base labels
8098 unless we're producing PIC; mark that action unreachable().
8099 (ix86_save_reg): If the function contains a nonlocal label, save the
8101 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8102 * config/darwin.c (emitted_pic_label_num): New GTY.
8103 (update_pic_label_number_if_needed): New.
8104 (machopic_output_function_base_name): Adjust for nonlocal receiver
8106 (machopic_should_output_picbase_label): New.
8107 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8108 (nonlocal_goto_receiver): New insn and split.
8110 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8112 * config/aarch64/aarch64-builtins.c
8113 (aarch64_fold_builtin): Fold abs in all modes.
8114 * config/aarch64/aarch64-simd-builtins.def
8115 (abs): Enable for all modes.
8116 * config/aarch64/arm_neon.h
8117 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8118 (vabs_f64): Add missing intrinsic.
8120 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8122 * config/aarch64/arm_neon.h (vabs_s64): New function
8124 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8127 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8128 * config/avr/avr.md (adjust_len): Add `round'.
8129 * config/avr/avr-protos.h (avr_out_round): New prototype.
8130 (avr_out_plus): Add `out_label' argument.
8131 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8132 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8133 Handle the case where `insn' is just a pattern.
8134 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8135 (avr_out_round): New function.
8136 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8138 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8140 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8141 Add check for flag_stack_usage to handle -fstack-usage support
8143 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8145 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8146 interaction for new Power8 flags and VSX.
8148 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8151 * tree-inline.c (expand_call_inline): Emit errors during
8152 early_inlining only if optimization is not turned on.
8154 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8158 * passes.c (init_optimization_passes): Move the construction of
8159 the pass hierarchy into a new passes.def file.
8161 * Makefile.in (passes.o): Add dependency on passes.def.
8163 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8165 * passes.c (init_optimization_passes): Introduce macros for
8166 constructing the tree of passes (INSERT_PASSES_AFTER,
8167 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8168 TERMINATE_PASS_LIST).
8170 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8171 Wei Mi <wmi@google.com>
8173 PR rtl-optimization/57878
8174 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8176 (reload_pseudo_compare_func): Check nregs first for reload
8179 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8181 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8183 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8185 * read-rtl.c (validate_const_int): Once an invalid character is
8186 seen, quit the loop.
8188 * gengtype.c (write_roots): Similarly once we find the "deletable"
8189 or "if_marked" option.
8191 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8193 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8194 "xtab" and "sat". Move value "clz" from here to ...
8195 (attriubte "type"): ... here.
8196 (satsi_<SAT:code>): Delete "insn" attribute.
8197 (satsi_<SAT:code>_shift): Likewise.
8198 (arm_zero_extendqisi2addsi): Likewise.
8199 (arm_extendqisi2addsi): Likewise.
8200 (clzsi2): Update for attribute changes.
8201 (rbitsi2): Likewise.
8202 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8204 (arm_usatsihi): Likewise.
8205 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8207 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8209 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8210 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8211 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8212 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8213 in alphabetical order.
8214 (attribute "core_cycles"): Update for attribute changes.
8215 (arm_addsi3): Likewise.
8216 (addsi3_compare0): Likewise.
8217 (addsi3_compare0_scratch): Likewise.
8218 (addsi3_compare_op1): Likewise.
8219 (addsi3_compare_op2): Likewise.
8220 (compare_addsi2_op0): Likewise.
8221 (compare_addsi2_op1): Likewise.
8222 (addsi3_carryin_shift_<optab>): Likewise.
8223 (subsi3_carryin_shift): Likewise.
8224 (rsbsi3_carryin_shift): Likewise.
8225 (arm_subsi3_insn): Likewise.
8226 (subsi3_compare0): Likewise.
8227 (subsi3_compare): Likewise.
8228 (arm_andsi3_insn): Likewise.
8229 (thumb1_andsi3_insn): Likewise.
8230 (andsi3_compare0): Likewise.
8231 (andsi3_compare0_scratch): Likewise.
8232 (zeroextractsi_compare0_scratch
8233 (andsi_not_shiftsi_si): Likewise.
8234 (iorsi3_insn): Likewise.
8235 (iorsi3_compare0): Likewise.
8236 (iorsi3_compare0_scratch): Likewise.
8237 (arm_xorsi3): Likewise.
8238 (thumb1_xorsi3_insn): Likewise.
8239 (xorsi3_compare0): Likewise.
8240 (xorsi3_compare0_scratch): Likewise.
8241 (satsi_<SAT:code>_shift): Likewise.
8243 (arm_shiftsi3): Likewise.
8244 (shiftsi3_compare0): Likewise.
8245 (not_shiftsi): Likewise.
8246 (not_shiftsi_compare0): Likewise.
8247 (not_shiftsi_compare0_scratch): Likewise.
8248 (arm_one_cmplsi2): Likewise.
8249 (thumb_one_complsi2): Likewise.
8250 (notsi_compare0): Likewise.
8251 (notsi_compare0_scratch): Likewise.
8252 (thumb1_zero_extendhisi2): Likewise.
8253 (arm_zero_extendhisi2): Likewise.
8254 (arm_zero_extendhisi2_v6): Likewise.
8255 (arm_zero_extendhisi2addsi): Likewise.
8256 (thumb1_zero_extendqisi2): Likewise.
8257 (thumb1_zero_extendqisi2_v6): Likewise.
8258 (arm_zero_extendqisi2): Likewise.
8259 (arm_zero_extendqisi2_v6): Likewise.
8260 (arm_zero_extendqisi2addsi): Likewise.
8261 (thumb1_extendhisi2): Likewise.
8262 (arm_extendhisi2): Likewise.
8263 (arm_extendhisi2_v6): Likewise.
8264 (arm_extendqisi): Likewise.
8265 (arm_extendqisi_v6): Likewise.
8266 (arm_extendqisi2addsi): Likewise.
8267 (thumb1_extendqisi2): Likewise.
8268 (thumb1_movdi_insn): Likewise.
8269 (arm_movsi_insn): Likewise.
8270 (movsi_compare0): Likewise.
8271 (movhi_insn_arch4): Likewise.
8272 (movhi_bytes): Likewise.
8273 (arm_movqi_insn): Likewise.
8274 (thumb1_movqi_insn): Likewise.
8275 (arm32_movhf): Likewise.
8276 (thumb1_movhf): Likewise.
8277 (arm_movsf_soft_insn): Likewise.
8278 (thumb1_movsf_insn): Likewise.
8279 (movdf_soft_insn): Likewise.
8280 (thumb_movdf_insn): Likewise.
8281 (arm_cmpsi_insn): Likewise.
8282 (cmpsi_shiftsi): Likewise.
8283 (cmpsi_shiftsi_swp): Likewise.
8284 (arm_cmpsi_negshiftsi_si): Likewise.
8285 (movsicc_insn): Likewise.
8286 (movsfcc_soft_insn): Likewise.
8287 (arith_shiftsi): Likewise.
8288 (arith_shiftsi_compare0
8289 (arith_shiftsi_compare0_scratch
8290 (sub_shiftsi): Likewise.
8291 (sub_shiftsi_compare0
8292 (sub_shiftsi_compare0_scratch
8293 (and_scc): Likewise.
8294 (cond_move): Likewise.
8295 (if_plus_move): Likewise.
8296 (if_move_plus): Likewise.
8297 (if_move_not): Likewise.
8298 (if_not_move): Likewise.
8299 (if_shift_move): Likewise.
8300 (if_move_shift): Likewise.
8301 (if_shift_shift): Likewise.
8302 (if_not_arith): Likewise.
8303 (if_arith_not): Likewise.
8304 (cond_move_not): Likewise.
8305 (thumb1_ashlsi3): Set type attribute.
8306 (thumb1_ashrsi3): Likewise.
8307 (thumb1_lshrsi3): Likewise.
8308 (thumb1_rotrsi3): Likewise.
8309 (shiftsi3_compare0_scratch): Likewise.
8310 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8311 (neon_mov<mode>): Likewise.
8312 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8314 (thumb2_movsi_insn): Likewise.
8315 (thumb2_cmpsi_neg_shiftsi): Likewise.
8316 (thumb2_extendqisi_v6): Likewise.
8317 (thumb2_zero_extendhisi2_v6): Likewise.
8318 (thumb2_zero_extendqisi2_v6): Likewise.
8319 (thumb2_shiftsi3_short): Likewise.
8320 (thumb2_addsi3_compare0_scratch): Likewise.
8321 (orsi_not_shiftsi_si): Likewise.
8322 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8323 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8325 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8326 (1020alu_shift_op): Likewise.
8327 (1020alu_shift_reg_op): Likewise.
8328 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8329 (alu_shift_op): Likewise.
8330 (alu_shift_reg_op): Likewise.
8331 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8332 (11_alu_shift_op): Likewise.
8333 (11_alu_shift_reg_op): Likewise.
8334 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8335 (9_alu_shift_reg_op): Likewise.
8336 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8338 (cortex_a15_alu_shift): Likewise.
8339 (cortex_a15_alu_shift_reg): Likewise.
8340 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8342 (cortex_a5_alu_shift): Likewise.
8343 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8345 (cortex_a53_alu_shift): Likewise.
8346 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8348 (cortex_a7_alu_reg): Likewise.
8349 (cortex_a7_alu_shift): Likewise.
8350 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8352 (cortex_a8_alu_shift): Likewise.
8353 (cortex_a8_alu_shift_reg): Likewise.
8354 (cortex_a8_mov): Likewise.
8355 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8356 (cortex_a9_dp_shift): Likewise.
8357 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8359 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8361 (cortex_r4_mov): Likewise.
8362 (cortex_r4_alu_shift): Likewise.
8363 (cortex_r4_alu_shift_reg): Likewise.
8364 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8365 (526_alu_shift_op): Likewise.
8366 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8367 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8368 (626te_alu_shift_op): Likewise.
8369 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8370 (726te_alu_op): Likewise.
8371 (726te_alu_shift_op): Likewise.
8372 (726te_alu_shift_reg_op): Likewise.
8373 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8374 (mp626_alu_shift_op): Likewise.
8375 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8376 (pj4_alu_e1_conds): Likewise.
8377 (pj4_alu): Likewise.
8378 (pj4_alu_conds): Likewise.
8379 (pj4_shift): Likewise.
8380 (pj4_shift_conds): Likewise.
8381 (pj4_alu_shift): Likewise.
8382 (pj4_alu_shift_conds): Likewise.
8383 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8385 (cortexa7_older_only): Likewise.
8386 (cortexa7_younger): Likewise.
8388 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8390 * ipa-pure-const.c (generate_summary): Rename to...
8391 (pure_const_generate_summary): ... this.
8393 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8395 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8397 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8400 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8403 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8405 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8406 enabled without -march=zEC12.
8407 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8410 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8412 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8413 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8414 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8415 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8416 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8417 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8418 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8420 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8421 and ISA_HAS_NMADD3_NMSUB3.
8422 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8423 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8424 (nmsub4<mode>, nmsub3<mode>): Likewise.
8425 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8427 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8429 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8430 TARGET_MIPS5400 checking.
8432 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8433 Peter Bergner <bergner@vnet.ibm.com>
8435 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8436 registers in the comment.
8437 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8438 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8439 rather than FIRST_PSEUDO_REGISTERS.
8441 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8443 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8444 enable extra ISA flags with TARGET_HTM.
8446 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8448 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8451 2013-07-15 Cong Hou <congh@google.com>
8453 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8454 in compare function for sorting.
8456 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8458 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8459 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8460 * config/rs6000/rs6000.opt: Add -mhtm option.
8461 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8462 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8463 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8464 __HTM__ if the HTM instructions are available.
8465 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8466 htm_spr_reg_operand): New define_predicates.
8467 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8468 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8470 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8471 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8472 HTM builtin functions.
8473 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8474 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8475 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8476 (rs6000_builtin_mask_calculate): Likewise.
8477 (rs6000_option_override_internal): Likewise.
8478 (bdesc_htm): Add new HTM builtin support.
8479 (htm_spr_num): New function.
8480 (htm_spr_regno): Likewise.
8481 (rs6000_htm_spr_icode): Likewise.
8482 (htm_expand_builtin): Likewise.
8483 (htm_init_builtins): Likewise.
8484 (rs6000_expand_builtin): Add support for HTM builtin functions.
8485 (rs6000_init_builtins): Likewise.
8486 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8488 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8489 (TARGET_HTM, MASK_HTM): Define macros.
8490 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8491 (FIXED_REGISTERS): Likewise.
8492 (CALL_USED_REGISTERS): Likewise.
8493 (CALL_REALLY_USED_REGISTERS): Likewise.
8494 (REG_ALLOC_ORDER): Likewise.
8495 (enum reg_class): Likewise.
8496 (REG_CLASS_NAMES): Likewise.
8497 (REG_CLASS_CONTENTS): Likewise.
8498 (REGISTER_NAMES): Likewise.
8499 (ADDITIONAL_REGISTER_NAMES): Likewise.
8500 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8501 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8502 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8503 * config/rs6000/htm.md: New file.
8504 * config/rs6000/htmintrin.h: New file.
8505 * config/rs6000/htmxlintrin.h: New file.
8507 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8509 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8510 Define SYMBOL_TINY_GOT, update comment.
8511 * config/aarch64/aarch64.c
8512 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8513 (aarch64_expand_mov_immediate): Likewise.
8514 (aarch64_print_operand): Likewise.
8515 (aarch64_classify_symbol): Likewise.
8516 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8517 (ldr_got_tiny): Define.
8519 2013-07-13 Tobias Grosser <tobias@grosser.es>
8521 PR tree-optimization/54094
8522 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8523 scheduling dimension for the parallelism check from the polyhedral
8524 information in the AST.
8525 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8528 2013-07-13 Jason Merrill <jason@redhat.com>
8530 * print-tree.c (debug_vec_tree): Use debug_raw.
8531 (debug_raw (vec<tree, va_gc> &)): New.
8532 (debug_raw (vec<tree, va_gc> *)): New.
8533 * tree.h: Declare them.
8535 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8537 * ifcvt.c (ifcvt_after_combine): New static variable.
8538 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8540 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8541 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8542 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8544 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8546 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8548 2013-07-12 Michael Matz <matz@suse.de>
8551 * convert.c (convert_to_real): Reject non-float inner types.
8553 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8555 * config/aarch64/aarch64-protos.h
8556 (aarch64_simd_immediate_valid_for_move): Remove.
8557 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8558 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8560 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8562 2013-07-11 Steve Ellcey <sellcey@mips.com>
8564 * config/mips/mips.c (mips_conditional_register_usage): Do not
8565 use t[0-7] registers in MIPS16 mode when optimizing for size.
8567 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8569 * config/i386/i386.c (dispatch_function_versions): Fix array
8570 indexing of function_version_info to match actual_versions.
8572 2013-07-11 Teresa Johnson <tejohnson@google.com>
8574 * vec.h (struct va_gc): Move release out-of-line.
8575 (va_gc::release): Call ggc_free on released vec.
8577 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8579 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8580 Require GOT register as additional operand in UNSPEC.
8581 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8582 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8583 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8584 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8585 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8586 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8587 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8589 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
8592 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
8593 name seen by assembler/linker rather if available.
8595 2013-07-11 Andreas Schwab <schwab@suse.de>
8597 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
8599 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
8601 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
8603 2013-07-10 Joseph Myers <joseph@codesourcery.com>
8605 * doc/tm.texi.in: Move hook documentation to ....
8606 * target.def: ... here.
8608 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
8610 * doc/tm.texi: Regenerate.
8612 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
8615 * doc/invoke.texi: Document -Wconditionally-supported.
8617 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8620 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
8623 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8626 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
8627 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
8628 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
8629 Remove duplicate devices.
8630 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
8631 * config/avr/t-multilib: Regenerate.
8632 * config/avr/avr-tables.opt: Regenerate.
8633 * doc/avr-mmcu.texi: Regenerate.
8635 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8638 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
8640 2013-07-10 Graham Stott <graham.stott@btinternet.com>
8642 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
8643 the cost of MULT when optimizing for size.
8645 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8647 * config/cr16/cr16-protos.h: Don't include target.h.
8649 2013-07-09 Joseph Myers <joseph@codesourcery.com>
8651 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
8652 adjust register size for TDmode and TFmode for VSX registers.
8654 2013-07-08 Kai Tietz <ktietz@redhat.com>
8657 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
8658 hook_bool_const_tree_true.
8660 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8662 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
8663 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
8664 * config/s390/s390.md: Define FPR*_REGNUM constants.
8665 Fix FPR2_REGNUM constant (18 -> 17).
8666 ("*trunc<BFP:mode><DFP_ALL:mode>2")
8667 ("*trunc<DFP_ALL:mode><BFP:mode>2")
8668 ("trunc<BFP:mode><DFP_ALL:mode>2")
8669 ("trunc<DFP_ALL:mode><BFP:mode>2")
8670 ("*extend<BFP:mode><DFP_ALL:mode>2")
8671 ("*extend<DFP_ALL:mode><BFP:mode>2")
8672 ("extend<BFP:mode><DFP_ALL:mode>2")
8673 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
8676 2013-07-08 Graham Stott <graham.stott@btinternet.com>
8678 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
8680 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8682 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
8683 and cfun_fpr_bit_p to cfun_fpr_save_p.
8684 (s390_frame_area, s390_register_info, s390_frame_info)
8685 (s390_emit_prologue, s390_emit_epilogue)
8686 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
8688 * config/s390/s390.h: Define *_REGNUM macros for floating point
8691 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
8693 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
8695 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
8697 PR rtl-optimization/57786
8698 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
8699 and break out of the loop when it is set to false.
8701 2013-07-08 Jakub Jelinek <jakub@redhat.com>
8704 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
8705 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
8706 (const_int 63)) 0)).
8707 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
8708 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
8709 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
8711 PR rtl-optimization/57829
8712 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
8713 mask bits outside of mode are just sign-extension from mode to HWI.
8715 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8717 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
8718 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
8719 adjust_address instead of change_address to keep info about alignment.
8720 (emit_strmov): Remove.
8721 (emit_memmov): New function.
8722 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
8723 (expand_movmem_epilogue): Likewise and return updated rtx for
8725 (expand_constant_movmem_prologue): Likewise and return updated rtx for
8726 destination and source.
8727 (decide_alignment): Refactor, handle vector_loop.
8728 (ix86_expand_movmem): Likewise.
8729 (ix86_expand_setmem): Likewise.
8730 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
8732 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
8734 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
8735 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
8737 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8739 * config/i386/sse.md (sse_movlhps): Change alternative 3
8740 of operand 2 to "m".
8742 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
8745 * config/i386/sse.md (iptr): New mode attribute.
8746 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
8747 (<sse>_vm<plusminus_insn><mode>3): Ditto.
8748 (<sse>_vmmul<mode>3): Ditto.
8749 (<sse>_vmdiv<mode>3): Ditto.
8750 (sse_vmrcpv4sf2): Ditto.
8751 (<sse>_vmsqrt<mode>2): Ditto.
8752 (sse_vmrsqrtv4sf2): Ditto.
8753 (<sse>_vm<code><mode>3): Ditto.
8754 (avx_vmcmp<mode>3): Ditto.
8755 (<sse>_vmmaskcmp<mode>3): Ditto.
8756 (<sse>_comi): Ditto.
8757 (<sse>_ucomi): Ditto.
8758 (*xop_vmfrcz_<mode>): Ditto.
8759 (*fmai_fmadd_<mode>): Ditto.
8760 (*fmai_fmsub_<mode>): Ditto.
8761 (*fmai_fnmadd_<mode>): Ditto.
8762 (*fmai_fnmsub_<mode>): Ditto.
8763 (*fma4i_vmfmadd_<mode>): Ditto.
8764 (*fma4i_vmfmsub_<mode>): Ditto.
8765 (*fma4i_vmfnmadd_<mode>): Ditto.
8766 (*fma4i_vmfnmsub_<mode>): Ditto.
8767 (*xop_vmfrcz_<mode>): Ditto.
8768 (sse_cvtps2pi): Ditto.
8769 (sse_cvttps2pi): Ditto.
8770 (sse_cvtss2si): Ditto.
8771 (sse_cvtss2si_2): Ditto.
8772 (sse_cvtss2siq_2): Ditto.
8773 (sse_cvttss2si): Ditto.
8774 (sse_cvttss2siq): Ditto.
8775 (sse_cvtsd2si): Ditto.
8776 (sse_cvtsd2si_2): Ditto.
8777 (sse_cvtsd2siq_2): Ditto.
8778 (sse_cvttsd2si): Ditto.
8779 (sse_cvttsd2siq): Ditto.
8780 (sse_cvtsd2ss): Ditto.
8781 (sse_cvtss2sd): Ditto.
8782 (avx2_pbroadcast<mode>): Ditto.
8783 (avx2_pbroadcast<mode>_1): Ditto.
8784 (*avx_vperm_broadcast_v4sf): Ditto.
8786 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
8787 (sse_movlhps): Ditto.
8788 (sse_storehps): Ditto.
8789 (sse_loadhps): Ditto.
8790 (sse_storelps): Ditto.
8791 (sse_loadlps): Ditto.
8792 (*vec_concatv4sf): Ditto.
8793 (*vec_interleave_highv2df): Ditto.
8794 (*vec_interleave_lowv2df): Ditto.
8795 (*vec_extractv2df_1_sse): Ditto.
8796 (*vec_extractv2df_0_sse): Ditto.
8797 (sse2_storelpd): Ditto.
8798 (sse2_loadlpd): Ditto.
8799 (sse2_movsd): Ditto.
8800 (*vec_concatv4si): Ditto.
8801 (vec_concatv2di): Ditto.
8803 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
8804 for Intel asm dialect.
8805 (mmx_punpcklwd): Ditto.
8806 (mmx_punpckldq): Ditto.
8808 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
8809 for intel assembler dialect.
8811 2013-07-06 Jakub Jelinek <jakub@redhat.com>
8814 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
8815 for BUILT_IN_C{LZ,LRSB}*.
8816 * tree.h (CASE_INT_FN): Add FN##IMAX case.
8817 * tree-vrp.c (extract_range_basic): Handle
8818 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
8819 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
8820 fall thru to code calling set_value*.
8821 * builtins.c (expand_builtin): Remove *IMAX cases.
8822 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
8823 if width is bigger than 2*HWI.
8825 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
8827 PR rtl-optimization/55342
8828 * lra-int.h (lra_subreg_reload_pseudos): New.
8829 * lra.c: Add undoing optional reloads to the block diagram.
8830 (lra_subreg_reload_pseudos): New.
8831 (lra_optional_reload_pseudos): Change comments.
8832 (lra): Init and clear lra_subreg_reload_pseudos. Clear
8833 lra_optional_reload_pseudos after undo transformations.
8834 * lra-assigns.c (pseudo_prefix_title): New.
8835 (lra_setup_reg_renumber): Use it.
8836 (spill_for): Ditto. Check subreg reload pseudos too.
8837 (assign_by_spills): Consider subreg reload pseudos too.
8838 * lra-constraints.c (simplify_operand_subreg): Use
8839 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
8840 (curr_insn_transform): Recognize and do optional reloads.
8841 (undo_optional_reloads): New.
8842 (lra_undo_inheritance): Call undo_optional_reloads.
8844 2013-07-05 Thomas Quinot <quinot@adacore.com>
8846 * tree-complex.c (expand_complex_operations_1): Fix typo.
8848 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
8850 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
8851 (tune_params): New member 'const vec_costs'.
8852 * config/aarch64/aarch64.c (generic_vector_cost): New.
8853 (generic_tunings): New member 'generic_vector_cost'.
8854 (aarch64_builtin_vectorization_cost): New.
8855 (aarch64_add_stmt_cost): New.
8856 (TARGET_VECTORIZE_ADD_STMT_COST): New.
8857 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
8859 2013-07-03 Jakub Jelinek <jakub@redhat.com>
8862 * config/i386/predicates.md (vsib_address_operand): Disallow
8863 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
8865 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
8868 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
8869 expand_builtin_setjmp_receiver.
8870 (expand_label): Adjust, call expand_builtin_setjmp_receiver
8871 with NULL for the label parameter.
8872 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
8873 the frame-pointer. Adjust comments.
8874 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
8875 only if LABEL is non-NULL.
8877 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
8879 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
8880 (ARM_ABI_AAPCS64): Ditto.
8882 (ARM_DEFAULT_ABI): Ditto.
8884 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
8886 * config/aarch64/aarch64-builtins.c
8887 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
8888 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
8890 * config/aarch64/aarch64-simd.md
8891 (aarch64_ld1<VALL:mode>): New.
8892 (aarch64_st1<VALL:mode>): Likewise.
8893 * config/aarch64/arm_neon.h
8894 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
8896 2013-07-02 Sriraman Tallam <tmsriram@google.com>
8898 * config/i386/i386.c (gate_insert_vzeroupper): Check if
8900 (ix86_option_override_internal):Turn on all -mavx target flags by
8901 default as they are dependent on AVX anyway.
8903 2013-07-02 Cary Coutant <ccoutant@google.com>
8905 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
8907 (loc_checksum_ordered): Likewise.
8908 (hash_loc_operands): Remove inline keyword.
8910 2013-07-02 Jakub Jelinek <jakub@redhat.com>
8912 PR tree-optimization/57741
8913 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
8914 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
8915 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
8916 Allow REAL_CST step_exprs if flag_associative_math.
8917 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
8919 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8921 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
8923 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8925 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
8927 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8929 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
8931 (iorsi3_insn): Likewise.
8932 (arm_xorsi3): Likewise.
8934 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
8936 * arm.md (attribute "wtype"): Delete. Move attribute values from here
8938 (attribute "type"): ... here, and prefix with "wmmx_".
8939 (attribute "core_cycles"): Update for attribute changes.
8940 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
8941 (tbcstv4hi): Likewise.
8942 (tbcstv2si): Likewise.
8943 (iwmmxt_iordi3): Likewise.
8944 (iwmmxt_xordi3): Likewise.
8945 (iwmmxt_anddi3): Likewise.
8946 (iwmmxt_nanddi3): Likewise.
8947 (iwmmxt_arm_movdi): Likewise.
8948 (iwmmxt_movsi_insn): Likewise.
8949 (mov<mode>_internal): Likewise.
8950 (and<mode>3_iwmmxt): Likewise.
8951 (ior<mode>3_iwmmxt): Likewise.
8952 (xor<mode>3_iwmmxt): Likewise.
8953 (add<mode>3_iwmmxt): Likewise.
8954 (ssaddv8qi3): Likewise.
8955 (ssaddv4hi3): Likewise.
8956 (ssaddv2si3): Likewise.
8957 (usaddv8qi3): Likewise.
8958 (usaddv4hi3): Likewise.
8959 (usaddv2si3): Likewise.
8960 (sub<mode>3_iwmmxt): Likewise.
8961 (sssubv8qi3): Likewise.
8962 (sssubv4hi3): Likewise.
8963 (sssubv2si3): Likewise.
8964 (ussubv8qi3): Likewise.
8965 (ussubv4hi3): Likewise.
8966 (ussubv2si3): Likewise.
8967 (mulv4hi3_iwmmxt): Likewise.
8968 (smulv4hi3_highpart): Likewise.
8969 (umulv4hi3_highpart): Likewise.
8970 (iwmmxt_wmacs): Likewise.
8971 (iwmmxt_wmacsz): Likewise.
8972 (iwmmxt_wmacu): Likewise.
8973 (iwmmxt_wmacuz): Likewise.
8974 (iwmmxt_clrdi): Likewise.
8975 (iwmmxt_clrv8qi): Likewise.
8976 (iwmmxt_clr4hi): Likewise.
8977 (iwmmxt_clr2si): Likewise.
8978 (iwmmxt_uavgrndv8qi3): Likewise.
8979 (iwmmxt_uavgrndv4hi3): Likewise.
8980 (iwmmxt_uavgv8qi3): Likewise.
8981 (iwmmxt_uavgv4hi3): Likewise.
8982 (iwmmxt_tinsrb): Likewise.
8983 (iwmmxt_tinsrh): Likewise.
8984 (iwmmxt_tinsrw): Likewise.
8985 (iwmmxt_textrmub): Likewise.
8986 (iwmmxt_textrmsb): Likewise.
8987 (iwmmxt_textrmuh): Likewise.
8988 (iwmmxt_textrmsh): Likewise.
8989 (iwmmxt_textrmw): Likewise.
8990 (iwmxxt_wshufh): Likewise.
8991 (eqv8qi3): Likewise.
8992 (eqv4hi3): Likewise.
8993 (eqv2si3): Likewise.
8994 (gtuv8qi3): Likewise.
8995 (gtuv4hi3): Likewise.
8996 (gtuv2si3): Likewise.
8997 (gtv8qi3): Likewise.
8998 (gtv4hi3): Likewise.
8999 (gtv2si3): Likewise.
9000 (smax<mode>3_iwmmxt): Likewise.
9001 (umax<mode>3_iwmmxt): Likewise.
9002 (smin<mode>3_iwmmxt): Likewise.
9003 (umin<mode>3_iwmmxt): Likewise.
9004 (iwmmxt_wpackhss): Likewise.
9005 (iwmmxt_wpackwss): Likewise.
9006 (iwmmxt_wpackdss): Likewise.
9007 (iwmmxt_wpackhus): Likewise.
9008 (iwmmxt_wpackwus): Likewise.
9009 (iwmmxt_wpackdus): Likewise.
9010 (iwmmxt_wunpckihb): Likewise.
9011 (iwmmxt_wunpckihh): Likewise.
9012 (iwmmxt_wunpckihw): Likewise.
9013 (iwmmxt_wunpckilb): Likewise.
9014 (iwmmxt_wunpckilh): Likewise.
9015 (iwmmxt_wunpckilw): Likewise.
9016 (iwmmxt_wunpckehub): Likewise.
9017 (iwmmxt_wunpckehuh): Likewise.
9018 (iwmmxt_wunpckehuw): Likewise.
9019 (iwmmxt_wunpckehsb): Likewise.
9020 (iwmmxt_wunpckehsh): Likewise.
9021 (iwmmxt_wunpckehsw): Likewise.
9022 (iwmmxt_wunpckelub): Likewise.
9023 (iwmmxt_wunpckeluh): Likewise.
9024 (iwmmxt_wunpckeluw): Likewise.
9025 (iwmmxt_wunpckelsb): Likewise.
9026 (iwmmxt_wunpckelsh): Likewise.
9027 (iwmmxt_wunpckelsw): Likewise.
9028 (ror<mode>3): Likewise.
9029 (ashr<mode>3_iwmmxt): Likewise.
9030 (lshr<mode>3_iwmmxt): Likewise.
9031 (ashl<mode>3_iwmmxt): Likewise.
9032 (ror<mode>3_di): Likewise.
9033 (ashr<mode>3_di): Likewise.
9034 (lshr<mode>3_di): Likewise.
9035 (ashl<mode>3_di): Likewise.
9036 (iwmmxt_wmadds): Likewise.
9037 (iwmmxt_wmaddu): Likewise.
9038 (iwmmxt_tmia): Likewise.
9039 (iwmmxt_tmiaph): Likewise.
9040 (iwmmxt_tmiabb): Likewise.
9041 (iwmmxt_tmiatb): Likewise.
9042 (iwmmxt_tmiabt): Likewise.
9043 (iwmmxt_tmiatt): Likewise.
9044 (iwmmxt_tmovmskb): Likewise.
9045 (iwmmxt_tmovmskh): Likewise.
9046 (iwmmxt_tmovmskw): Likewise.
9047 (iwmmxt_waccb): Likewise.
9048 (iwmmxt_wacch): Likewise.
9049 (iwmmxt_waccw): Likewise.
9050 (iwmmxt_waligni): Likewise.
9051 (iwmmxt_walignr): Likewise.
9052 (iwmmxt_walignr0): Likewise.
9053 (iwmmxt_walignr1): Likewise.
9054 (iwmmxt_walignr2): Likewise.
9055 (iwmmxt_walignr3): Likewise.
9056 (iwmmxt_wsadb): Likewise.
9057 (iwmmxt_wsadh): Likewise.
9058 (iwmmxt_wsadbz): Likewise.
9059 (iwmmxt_wsadhz): Likewise.
9060 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9061 (iwmmxt_wabsdiffb): Likewise.
9062 (iwmmxt_wabsdiffh): Likewise.
9063 (iwmmxt_wabsdiffw): Likewise.
9064 (iwmmxt_waddsubhx): Likewise
9065 (iwmmxt_wsubaddhx): Likewise.
9066 (addc<mode>3): Likewise.
9067 (iwmmxt_avg4): Likewise.
9068 (iwmmxt_avg4r): Likewise.
9069 (iwmmxt_wmaddsx): Likewise.
9070 (iwmmxt_wmaddux): Likewise.
9071 (iwmmxt_wmaddsn): Likewise.
9072 (iwmmxt_wmaddun): Likewise.
9073 (iwmmxt_wmulwsm): Likewise.
9074 (iwmmxt_wmulwum): Likewise.
9075 (iwmmxt_wmulsmr): Likewise.
9076 (iwmmxt_wmulumr): Likewise.
9077 (iwmmxt_wmulwsmr): Likewise.
9078 (iwmmxt_wmulwumr): Likewise.
9079 (iwmmxt_wmulwl): Likewise.
9080 (iwmmxt_wqmulm): Likewise.
9081 (iwmmxt_wqmulwm): Likewise.
9082 (iwmmxt_wqmulmr): Likewise.
9083 (iwmmxt_wqmulwmr): Likewise.
9084 (iwmmxt_waddbhusm): Likewise.
9085 (iwmmxt_waddbhusl): Likewise.
9086 (iwmmxt_wqmiabb): Likewise.
9087 (iwmmxt_wqmiabt): Likewise.
9088 (iwmmxt_wqmiatb): Likewise.
9089 (iwmmxt_wqmiatt): Likewise.
9090 (iwmmxt_wqmiabbn): Likewise.
9091 (iwmmxt_wqmiabtn): Likewise.
9092 (iwmmxt_wqmiatbn): Likewise.
9093 (iwmmxt_wqmiattn): Likewise.
9094 (iwmmxt_wmiabb): Likewise.
9095 (iwmmxt_wmiabt): Likewise.
9096 (iwmmxt_wmiatb): Likewise.
9097 (iwmmxt_wmiatt): Likewise.
9098 (iwmmxt_wmiabbn): Likewise.
9099 (iwmmxt_wmiabtn): Likewise.
9100 (iwmmxt_wmiatbn): Likewise.
9101 (iwmmxt_wmiattn): Likewise.
9102 (iwmmxt_wmiawbb): Likewise.
9103 (iwmmxt_wmiawbt): Likewise.
9104 (iwmmxt_wmiawtb): Likewise.
9105 (iwmmxt_wmiawtt): Likewise.
9106 (iwmmxt_wmiawbbn): Likewise.
9107 (iwmmxt_wmiawbtn): Likewise.
9108 (iwmmxt_wmiawtbn): Likewise.
9109 (iwmmxt_wmiawttn): Likewise.
9110 (iwmmxt_wmerge): Likewise.
9111 (iwmmxt_tandc<mode>3): Likewise.
9112 (iwmmxt_torc<mode>3): Likewise.
9113 (iwmmxt_torvsc<mode>3): Likewise.
9114 (iwmmxt_textrc<mode>3): Likewise.
9115 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9116 (wmmxt_pack): Likewise.
9117 (wmmxt_mult_c1): Likewise.
9118 (wmmxt_mult_c2): Likewise.
9119 (wmmxt_alu_c1): Likewise.
9120 (wmmxt_alu_c2): Likewise.
9121 (wmmxt_alu_c3): Likewise.
9122 (wmmxt_transfer_c1): Likewise.
9123 (wmmxt_transfer_c2): Likewise.
9124 (wmmxt_transfer_c3): Likewise.
9125 (marvell_f_iwmmxt_wstr): Likewise.
9126 (marvell_f_iwmmxt_wldr): Likewise.
9128 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9130 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9132 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9135 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9136 * lra-constraints.c (need_for_split_p): Check call used hard regs
9137 living through calls.
9139 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9140 call used regs for call insn.
9142 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9145 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9146 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9148 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9150 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9152 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9154 * lra-constraints.c (need_for_split_p): Check call used hard regs
9155 living through calls.
9157 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9160 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9161 to tie with any other modes. Eliminate Altivec vector mode tests,
9162 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9163 code, to return 0 if testing MODE2 for a condition, if we've
9164 already tested MODE1 for the same condition.
9166 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9168 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9171 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9173 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9174 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9176 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9178 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9180 (aarch64_symbolic_constant_p): Remove.
9181 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9182 static. Fix line length and white space.
9183 (aarch64_symbolic_constant_p): Remove.
9184 * config/aarch64/predicates.md (aarch64_valid_symref):
9185 Use aarch64_classify_symbol_expression.
9187 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9189 * config/arm/constraints.md (Ts): New constraint.
9190 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9192 (compare_scc): Use "Ts" constraint for operand 0.
9193 (ior_scc_scc): Likewise.
9194 (and_scc_scc): Likewise.
9195 (and_scc_scc_nodom): Likewise.
9196 (ior_scc_scc_cmp): Likewise for operand 7.
9197 (and_scc_scc_cmp): Likewise.
9198 * config/arm/thumb2.md (thumb2_movsi_insn):
9199 Add alternatives for 16-bit encodings.
9200 (thumb2_movhi_insn): Likewise.
9201 (thumb2_movsicc_insn): Likewise.
9202 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9203 (thumb2_negscc): Use "Ts" constraint.
9204 Move mvn instruction outside cond_exec block.
9205 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9206 for 16-bit encodings.
9208 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9210 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9212 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9213 (mulsi3subsi): Likewise.
9214 (mulsidi3adddi): Likewise.
9215 (mulsidi3_v6): Likewise.
9216 (umulsidi3_v6): Likewise.
9217 (umulsidi3adddi_v6): Likewise.
9218 (smulsi3_highpart_v6): Likewise.
9219 (umulsi3_highpart_v6): Likewise.
9220 (mulhisi3tb): Likewise.
9221 (mulhisi3bt): Likewise.
9222 (mulhisi3tt): Likewise.
9223 (maddhisi4): Likewise.
9224 (maddhisi4tb): Likewise.
9225 (maddhisi4tt): Likewise.
9226 (maddhidi4): Likewise.
9227 (maddhidi4tb): Likewise.
9228 (maddhidi4tt): Likewise.
9229 (zeroextractsi_compare0_scratch): Likewise.
9230 (insv_zero): Likewise.
9231 (insv_t2): Likewise.
9232 (anddi_notzesidi_di): Likewise.
9233 (anddi_notsesidi_di): Likewise.
9234 (andsi_notsi_si): Likewise.
9235 (iordi_zesidi_di): Likewise.
9236 (xordi_zesidi_di): Likewise.
9237 (andsi_iorsi3_notsi): Likewise.
9239 (smax_m1): Likewise.
9241 (not_shiftsi): Likewise.
9242 (unaligned_loadsi): Likewise.
9243 (unaligned_loadhis): Likewise.
9244 (unaligned_loadhiu): Likewise.
9245 (unaligned_storesi): Likewise.
9246 (unaligned_storehi): Likewise.
9247 (extv_reg): Likewise.
9248 (extzv_t2): Likewise.
9250 (udivsi3): Likewise.
9251 (arm_zero_extendhisi2addsi): Likewise.
9252 (arm_zero_extendqisi2addsi): Likewise.
9253 (compareqi_eq0): Likewise.
9254 (arm_extendhisi2_v6): Likewise.
9255 (arm_extendqisi2addsi): Likewise.
9256 (arm_movt): Likewise.
9257 (thumb2_ldrd): Likewise.
9258 (thumb2_ldrd_base): Likewise.
9259 (thumb2_ldrd_base_neg): Likewise.
9260 (thumb2_strd): Likewise.
9261 (thumb2_strd_base): Likewise.
9262 (thumb2_strd_base_neg): Likewise.
9263 (arm_negsi2): Add alternative for 16-bit encoding.
9264 (arm_one_cmplsi2): Likewise.
9266 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9268 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9269 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9270 (movdfcc): Likewise.
9271 * config/arm/vfp.md (*thumb2_movsf_vfp):
9272 Disable predication for arm_restrict_it.
9273 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9274 (*thumb2_movdfcc_vfp): Likewise.
9275 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9276 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9277 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9278 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9279 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9280 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9281 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9282 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9283 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9284 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9285 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9286 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9287 Disable predication for arm_restrict_it.
9289 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9291 * config/i386/bmiintrin.h (_bextr_u32): New.
9292 (_bextr_u64): Ditto.
9294 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9296 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9297 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9298 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9299 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9300 * config/mips/n32-elf.h: ...this new file.
9302 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9305 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9306 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9308 2013-06-27 Catherine Moore <clm@codesourcery.com>
9310 * config/mips/mips-tables.opt: Regenerate.
9311 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9312 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9313 * doc/invoke.texi: Add -m14kc.
9315 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9318 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9319 constraints of operand 1 and 2.
9322 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9323 to match RTL canonicalization. Swap predicates and
9324 constraints of operand 1 and 2.
9326 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9328 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9329 Process OP_INOUT regs for splitting too.
9331 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9333 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9334 decl before the loop, initialize to NULL.
9335 (vectorizable_load): Initialize ptr_incr to NULL.
9337 2013-06-27 Martin Jambor <mjambor@suse.cz>
9340 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9341 * ipa-ref.c (ipa_maybe_record_reference): New function.
9342 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9343 * ipa-cp.c (create_specialized_node): Record potential references from
9345 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9347 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9349 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9350 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9351 'mode' to force_reg.
9352 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9353 (aarch64_expand_mov_immediate): Likewise.
9355 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9357 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9358 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9360 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9362 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9363 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9364 (struct machine_function): Add tbegin_p.
9365 (s390_canonicalize_comparison): Fold CC mode compares to
9366 conditional jump if possible.
9367 (s390_emit_jump): Return the emitted jump.
9368 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9369 Handle CCRAWmode compares.
9370 (s390_option_override): Default to -mhtm if available.
9371 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9372 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9373 FPRs instead of df_regs_ever_live_p.
9374 (s390_optimize_nonescaping_tx): New function.
9375 (s390_init_frame_layout): Extend clobbered_regs array to cover
9377 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9378 (s390_expand_tbegin): New function.
9379 (enum s390_builtin): New enum definition.
9380 (code_for_builtin): New array definition.
9381 (s390_init_builtins): New function.
9382 (s390_expand_builtin): New function.
9383 (TARGET_INIT_BUILTINS): Define.
9384 (TARGET_EXPAND_BUILTIN): Define.
9385 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9386 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9387 (s390_alc_comparison): Likewise.
9388 * config/s390/s390-modes.def: Add CCRAWmode.
9389 * config/s390/s390.h (processor_flags): Add PF_TX.
9390 (TARGET_CPU_HTM): Define macro.
9391 (TARGET_HTM): Define macro.
9392 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9393 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9394 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9395 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9396 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9397 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9398 ("*cc_to_int"): Move up.
9399 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9400 constants other than 0.
9401 ("*ccraw_to_int"): New insn and splitter definition.
9402 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9403 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9404 ("tx_assist"): New expander.
9405 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9406 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9407 * config/s390/s390.opt: Add -mhtm option.
9408 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9409 * config/s390/htmxlintrin.h: New file.
9410 * config/s390/htmintrin.h: New file.
9411 * config/s390/s390intrin.h: New file.
9412 * doc/extend.texi: Document htm builtins.
9413 * config.gcc: Add the new header files to extra_headers.
9415 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9417 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9418 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9420 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9421 Pat Haugen <pthaugen@us.ibm.com>
9422 Peter Bergner <bergner@vnet.ibm.com>
9424 * config/rs6000/power8.md: New.
9425 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9426 setting for power8 entry.
9427 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9428 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9429 test for Power4/Power5 only.
9430 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9432 (force_new_group): Adjust comment.
9433 * config/rs6000/rs6000.md: Include power8.md.
9435 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9437 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9438 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9440 (tune_params): Update comment.
9441 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9442 (arm_max_conditional_execute): New function.
9443 (thumb2_final_prescan_insn): Use max_insn_skipped and
9444 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9446 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9448 PR tree-optimization/57705
9449 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9450 SSA_NAME step, provided that it is not defined inside the loop.
9451 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9452 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9454 2013-06-25 Martin Jambor <mjambor@suse.cz>
9457 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9458 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9460 (ipa_note_param_call): Initialize member_ptr flag.
9461 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9462 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9463 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9464 (ipa_read_indirect_edge_info): Likewise.
9466 2013-06-25 Richard Biener <rguenther@suse.de>
9469 * passes.c (init_optimization_passes): Move pass_fold_builtins
9470 and pass_dce earlier with -Og.
9472 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9474 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9475 <BIT_FIELD_REF>: Remove trailing TAB.
9476 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9479 2013-06-24 Martin Jambor <mjambor@suse.cz>
9481 PR tree-optimization/57358
9482 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9483 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9484 (ipa_analyze_params_uses): Generate pessimistic info when true.
9486 2013-06-24 Martin Jambor <mjambor@suse.cz>
9488 PR tree-optimization/57539
9489 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9490 global.inlined_to of the new node to it. All callers changed.
9491 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9492 inlining_into, pass it to cgraph_clone_node.
9493 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9494 ipa_free_edge_args_substructures.
9495 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9496 rdesc linked list. Do not assert rdesc edges have inlined caller.
9497 Assert we have found an rdesc in the rdesc list.
9499 2013-06-24 Richard Biener <rguenther@suse.de>
9501 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9502 (pointer_set_lookup): Declare.
9503 (class pointer_map): New template class implementing a
9504 generic pointer to T map.
9505 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9506 pointer_map<T>::contains, pointer_map<T>::insert,
9507 pointer_map<T>::traverse): New functions.
9508 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9509 (pointer_set_lookup): New function.
9510 (pointer_set_contains): Use pointer_set_lookup.
9511 (pointer_set_insert): Likewise.
9512 (insert_aux): Remove.
9513 (struct pointer_map_t): Embed a pointer_set_t.
9514 (pointer_map_create): Adjust.
9515 (pointer_map_destroy): Likewise.
9516 (pointer_map_contains): Likewise.
9517 (pointer_map_insert): Likewise.
9518 (pointer_map_traverse): Likewise.
9519 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9520 pointer_map<unsigned> instead of a pointer_map_t.
9521 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9522 (streamer_tree_cache_lookup): Likewise.
9523 (streamer_tree_cache_create): Likewise.
9524 (streamer_tree_cache_delete): Likewise.
9525 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9526 pointer_map<unsigned> instead of a pointer_map_t.
9527 (lto_init_tree_ref_encoder): Adjust.
9528 (lto_destroy_tree_ref_encoder): Likewise.
9529 * lto-section-out.c (lto_output_decl_index): Likewise.
9530 (lto_record_function_out_decl_state): Likewise.
9531 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9533 2013-06-24 Richard Biener <rguenther@suse.de>
9535 PR tree-optimization/57488
9536 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9538 2013-06-24 Alan Modra <amodra@gmail.com>
9540 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9541 (gen_easy_altivec_constant): Likewise.
9542 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9543 easy_vector_constant_msb): Likewise.
9545 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9548 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9549 add missing return true.
9551 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9554 * config/sh/predicates.md (general_extend_operand): Invoke
9555 general_movsrc_operand for memory operands.
9556 (general_movsrc_operand): Allow reg+reg addressing, do not use
9557 general_operand for memory operands.
9559 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9561 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9562 when current target options does not apply.
9563 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9564 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9565 * config/i386/bmiintrin.h: Pass appropriate target
9566 attributes to header.
9567 * config/i386/mmintrin.h: Ditto.
9568 * config/i386/nmmintrin.h: Ditto.
9569 * config/i386/avx2intrin.h: Ditto.
9570 * config/i386/fxsrintrin.h: Ditto.
9571 * config/i386/tbmintrin.h: Ditto.
9572 * config/i386/xsaveintrin.h: Ditto.
9573 * config/i386/f16cintrin.h: Ditto.
9574 * config/i386/xtestintrin.h: Ditto.
9575 * config/i386/xsaveoptintrin.h: Ditto.
9576 * config/i386/bmi2intrin.h: Ditto.
9577 * config/i386/lzcntintrin.h: Ditto.
9578 * config/i386/smmintrin.h: Ditto.
9579 * config/i386/wmmintrin.h: Ditto.
9580 * config/i386/x86intrin.h: Remove all header include guards.
9581 * config/i386/prfchwintrin.h: Ditto.
9582 * config/i386/pmmintrin.h: Ditto.
9583 * config/i386/tmmintrin.h: Ditto.
9584 * config/i386/xmmintrin.h: Ditto.
9585 * config/i386/popcntintrin.h: Ditto.
9586 * config/i386/rdseedintrin.h: Ditto.
9587 * config/i386/ammintrin.h: Ditto.
9588 * config/i386/emmintrin.h: Ditto.
9589 * config/i386/immintrin.h: Remove all header include guards.
9590 * config/i386/fma4intrin.h: Ditto.
9591 * config/i386/lwpintrin.h: Ditto.
9592 * config/i386/xopintrin.h: Ditto.
9593 * config/i386/ia32intrin.h: Ditto.
9594 * config/i386/avxintrin.h: Ditto.
9595 * config/i386/rtmintrin.h: Ditto.
9596 * config/i386/fmaintrin.h: Ditto.
9597 * config/i386/mm3dnow.h: Ditto.
9599 2013-06-22 Sriraman Tallam <tmsriram@google.com>
9601 * common/config/i386/i386-common.c: Handle LZCNT.
9603 2013-06-22 Andi Kleen <ak@linux.intel.com>
9605 * doc/extend.texi: Use __atomic_store_n instead of
9606 __atomic_store in HLE example.
9608 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
9610 * config/sh/sh.c: Remove <cstdlib> workaround.
9612 2013-06-21 Andi Kleen <ak@linux.intel.com>
9614 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
9616 2013-06-21 Andi Kleen <ak@linux.intel.com>
9618 * doc/extend.texi: Document that __atomic_clear and
9619 __atomic_test_and_set should only be used with bool.
9621 2013-06-20 Jan Hubicka <jh@suse.cz>
9623 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
9625 * tree.c (decls_same_for_odr): New function.
9626 (same_for_edr): New function.
9627 (types_same_for_odr): New function.
9628 (get_binfo_at_offset): Use it.
9629 * tree.h (types_same_for_odr): Declare.
9631 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
9632 Jason Merrill <jason@redhat.com>
9634 * system.h: Include <cstdlib> as well as <stdlib.h>.
9636 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
9639 * config/i386/i386.c (construct_container): Report error if
9640 long double is used with disabled x87 float returns.
9642 2013-06-20 Jan Hubicka <jh@suse.cz>
9644 * lto-cgraph.c (input_symtab): Do not set cgraph state.
9646 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
9648 PR rtl-optimization/57425
9649 PR rtl-optimization/57569
9650 * alias.c (write_dependence_p): Remove parameters mem_mode and
9651 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
9652 Changed all callers.
9653 (canon_anti_dependence): Get comments and semantics in sync.
9654 Add parameter mem_canonicalized. Changed all callers.
9655 * rtl.h (canon_anti_dependence): Update prototype.
9657 2013-06-20 Richard Biener <rguenther@suse.de>
9659 * data-streamer-in.c (streamer_read_uhwi): Optimize single
9660 byte case, inline streamer_read_uchar and defer section
9663 2013-06-20 Richard Biener <rguenther@suse.de>
9665 PR tree-optimization/57584
9666 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
9667 SSA names into the expanded expression that take part in
9668 abnormal coalescing.
9670 2013-06-19 Sharad Singhai <singhai@google.com>
9672 * gcov.c (print_usage): Handle new option.
9673 (process_args): Ditto.
9674 (get_gcov_intermediate_filename): New function.
9675 (output_intermediate_file): New function.
9676 (output_gcov_file): New function
9677 (generate_results): Handle new option.
9678 (release_function): Relase demangled name.
9679 (read_graph_file): Handle demangled name.
9680 (output_lines): Ditto.
9681 * doc/gcov.texi: Document gcov intermediate format.
9683 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
9686 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
9687 (lra_emit_add): Use the functions. Add comment about Y as an
9690 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
9693 * collect2.c (collect_atexit): New.
9694 (collect_exit): Delete.
9695 (main): Register collect_atexit with atexit.
9696 (collect_wait): Change collect_exit to exit.
9698 * collect2.h (collect_exit): Delete.
9699 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
9701 2013-06-19 Wei Mi <wmi@google.com>
9703 PR rtl-optimization/57518
9704 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
9705 if regno is used in paradoxical subreg.
9706 (update_equiv_regs): Check pdx_subregs[regno] before
9707 set a reg to be equivalent with a mem.
9709 2013-06-19 Matthias Klose <doko@ubuntu.com>
9712 * file-find.h (find_a_file): Add a mode parameter.
9713 * file-find.c (find_a_file): Likewise.
9714 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
9715 with X_OK for the executables.
9716 * collect2.c (main): Call find_a_file with X_OK.
9718 2013-06-19 Steve Ellcey <sellcey@mips.com>
9721 * config/mips/mips.md (casesi_internal_mips16_<mode>):
9722 Use NEXT_INSN instead of next_real_insn.
9724 2013-06-19 Jan Hubicka <jh@suse.cz>
9726 * cgraph.h (const_value_known_p): Replace by ...
9727 (ctor_for_folding): .. this one.
9728 * cgraphunit.c (process_function_and_variable_attributes): Use it.
9729 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
9730 * expr.c (expand_expr_real_1): Likewise.
9731 (string_constant): Likewise.
9732 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
9733 * ipa.c (process_references): Likewise.
9734 (symtab_remove_unreachable_nodes): Likewise.
9735 * ipa-inline-analysis.c (param_change_prob): Likewise.
9736 * gimple-fold.c (canonicalize_constructor_val): Likewise.
9737 (get_base_constructor): Likwise.
9738 * varpool.c (varpool_remove_node): Likewise.
9739 (varpool_remove_initializer): LIkewise.
9740 (dump_varpool_node): LIkwise.
9741 (const_value_known_p): Rewrite to ...
9742 (ctor_for_folding): ... this one.
9744 2013-06-19 Jakub Jelinek <jakub@redhat.com>
9747 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
9748 PERSONALITY in $PATH derived prefixes.
9750 2013-06-19 Jeff Law <law@redhat.com>
9752 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
9755 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
9756 (simplify_bitwise_binary): Use it to simpify certain binary ops on
9759 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
9761 * config/arm/vfp.md: Move VFP instruction classification documentation
9763 * config/arm/arm.md: ... here. Update instruction classification
9766 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
9768 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
9769 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
9770 pattern. Use more efficient sequences on ARMv5 and Thumb2.
9772 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
9775 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9776 with NEXT_INSN. Use tablejump_p to check for jump table data
9779 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
9782 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
9783 that now in C++ the value is correct per the C++ standards.
9785 2013-06-19 Richard Biener <rguenther@suse.de>
9787 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
9790 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9793 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9796 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9797 with next_active_insn.
9799 2013-06-18 Sriraman Tallam <tmsriram@google.com>
9801 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
9802 functions are inlined during failures to flag an error.
9803 * tree-inline.c (expand_call_inline): Allow the error to be flagged
9804 in early inline pass.
9806 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
9808 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
9811 2013-06-18 Julian Brown <julian@codesourcery.com>
9813 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
9814 Permit virtual register pre-reload if !strict.
9815 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
9817 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
9819 * config/arm/neon.md (movmisalign<mode>): Use
9820 neon_perm_struct_or_reg_operand instead of
9821 neon_struct_or_register_operand.
9822 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
9823 neon_permissive_struct_operand instead of neon_struct_operand.
9824 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
9825 neon_vector_mem_operand.
9826 * config/arm/predicates.md (neon_struct_operand): Adjust call to
9827 neon_vector_mem_operand.
9828 (neon_permissive_struct_operand): New.
9829 (neon_struct_or_register_operand): Rename to...
9830 (neon_perm_struct_or_reg_operand): This. Adjust call to
9831 neon_vector_mem_operand.
9833 2013-06-18 Richard Biener <rguenther@suse.de>
9835 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
9836 * lto-streamer.h: Include pointer-set.h.
9837 (struct lto_decl_slot): Remove.
9838 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
9839 Remove next_index entry.
9840 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9841 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9842 (lto_init_tree_ref_encoder): Adjust.
9843 (lto_destroy_tree_ref_encoder): Likewise.
9844 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9845 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9846 (lto_output_decl_index): Adjust.
9847 (lto_new_out_decl_state): Likewise.
9848 (lto_record_function_out_decl_state): Likewise.
9849 * lto-streamer-out.c (copy_function): Likewise.
9851 2013-06-18 Richard Biener <rguenther@suse.de>
9853 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
9854 * cgraphunit.c: Include cfgloop.h.
9855 (init_lowered_empty_function): Initialize the loop tree.
9856 (assemble_thunk): Insert new BBs into loops.
9858 2013-06-18 Richard Biener <rguenther@suse.de>
9860 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
9861 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
9862 the map from cache entry to cache index optional.
9863 (streamer_tree_cache_replace_tree): Adjust accordingly.
9864 (streamer_tree_cache_append): Likewise.
9865 (streamer_tree_cache_delete): Likewise.
9866 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
9867 streamer cache map from cache entry to cache index.
9868 * lto-streamer-out.c (create_output_block): Adjust.
9870 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
9872 * config/arm/arm.md (attribute "insn"): Move multiplication and
9873 division attributes to...
9874 (attribute "type"): ... here. Remove mult.
9875 (attribute "mul32"): New attribute.
9876 (attribute "mul64"): Add umaal.
9877 (*arm_mulsi3): Update attributes.
9878 (*arm_mulsi3_v6): Likewise.
9879 (*thumb_mulsi3): Likewise.
9880 (*thumb_mulsi3_v6): Likewise.
9881 (*mulsi3_compare0): Likewise.
9882 (*mulsi3_compare0_v6): Likewise.
9883 (*mulsi_compare0_scratch): Likewise.
9884 (*mulsi_compare0_scratch_v6): Likewise.
9885 (*mulsi3addsi): Likewise.
9886 (*mulsi3addsi_v6): Likewise.
9887 (*mulsi3addsi_compare0): Likewise.
9888 (*mulsi3addsi_compare0_v6): Likewise.
9889 (*mulsi3addsi_compare0_scratch): Likewise.
9890 (*mulsi3addsi_compare0_scratch_v6): Likewise.
9891 (*mulsi3subsi): Likewise.
9892 (*mulsidi3adddi): Likewise.
9893 (*mulsi3addsi_v6): Likewise.
9894 (*mulsidi3adddi_v6): Likewise.
9895 (*mulsidi3_nov6): Likewise.
9896 (*mulsidi3_v6): Likewise.
9897 (*umulsidi3_nov6): Likewise.
9898 (*umulsidi3_v6): Likewise.
9899 (*umulsidi3adddi): Likewise.
9900 (*umulsidi3adddi_v6): Likewise.
9901 (*smulsi3_highpart_nov6): Likewise.
9902 (*smulsi3_highpart_v6): Likewise.
9903 (*umulsi3_highpart_nov6): Likewise.
9904 (*umulsi3_highpart_v6): Likewise.
9905 (mulhisi3): Likewise.
9906 (*mulhisi3tb): Likewise.
9907 (*mulhisi3bt): Likewise.
9908 (*mulhisi3tt): Likewise.
9909 (maddhisi4): Likewise.
9910 (*maddhisi4tb): Likewise.
9911 (*maddhisi4tt): Likewise.
9912 (maddhidi4): Likewise.
9913 (*maddhidi4tb): Likewise.
9914 (*maddhidi4tt): Likewise.
9916 (udivsi3): Likewise.
9917 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
9918 (thumb2_mulsi_short_compare0): Likewise.
9919 (thumb2_mulsi_short_compare0_scratch): Likewise.
9920 * config/arm/arm1020e.md (1020mult1): Update attribute change.
9921 (1020mult2): Likewise.
9922 (1020mult3): Likewise.
9923 (1020mult4): Likewise.
9924 (1020mult5): Likewise.
9925 (1020mult6): Likewise.
9926 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
9928 (cortex_a15_mult64): Likewise.
9929 (cortex_a15_sdiv): Likewise.
9930 (cortex_a15_udiv): Likewise.
9931 * config/arm/arm1026ejs.md (mult1): Update attribute change.
9937 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
9938 (pj4_ir_div): Likewise.
9939 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
9940 (11_mult2): Likewise.
9941 (11_mult3): Likewise.
9942 (11_mult4): Likewise.
9943 (11_mult5): Likewise.
9944 (11_mult6): Likewise.
9945 (11_mult7): Likewise.
9946 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
9947 (cortex_a8_mla): Likewise.
9948 (cortex_a8_mull): Likewise.
9949 (cortex_a8_smulwy): Likewise.
9950 (cortex_a8_smlald): Likewise.
9951 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
9952 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
9953 (cortex_r4_mul_3): Likewise.
9954 (cortex_r4_mla_4): Likewise.
9955 (cortex_r4_mla_3): Likewise.
9956 (cortex_r4_smlald): Likewise.
9957 (cortex_r4_mull): Likewise.
9958 (cortex_r4_sdiv): Likewise.
9959 (cortex_r4_udiv): Likewise.
9960 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
9961 (cortex_a7_idiv): Likewise.
9962 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
9963 (9_mult2): Likewise.
9964 (9_mult3): Likewise.
9965 (9_mult4): Likewise.
9966 (9_mult5): Likewise.
9967 (9_mult6): Likewise.
9968 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
9969 (cortex_a53_sdiv): Likewise.
9970 (cortex_a53_udiv): Likewise.
9971 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
9972 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
9973 (mp626_mult2): Likewise.
9974 (mp626_mult3): Likewise.
9975 (mp626_mult4): Likewise.
9976 * config/arm/fa526.md (526_mult1): Update attribute change.
9977 (526_mult2): Likewise.
9978 * config/arm/arm-generic.md (mult): Update attribute change.
9979 (mult_ldsched_strongarm): Likewise.
9980 (mult_ldsched): Likewise.
9981 (multi_cycle): Likewise.
9982 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
9983 * config/arm/fa606te.md (606te_mult1): Update attribute change.
9984 (606te_mult2): Likewise.
9985 (606te_mult3): Likewise.
9986 (606te_mult4): Likewise.
9987 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
9988 (cortex_a9_mac16): Likewise.
9989 (cortex_a9_multiply): Likewise.
9990 (cortex_a9_mac): Likewise.
9991 (cortex_a9_multiply_long): Likewise.
9992 * config/arm/fa626te.md (626te_mult1): Update attribute change.
9993 (626te_mult2): Likewise.
9994 (626te_mult3): Likewise.
9995 (626te_mult4): Likewise.
9997 2013-06-18 Richard Biener <rguenther@suse.de>
10000 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10002 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10005 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10006 with next_active_insn.
10008 2013-06-18 Alan Modra <amodra@gmail.com>
10010 * config/rs6000/rs6000.h (enum data_align): New.
10011 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10012 (DATA_ABI_ALIGNMENT): Define.
10013 (CONSTANT_ALIGNMENT): Correct comment.
10014 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10015 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10017 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10019 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10020 ATTRIBUTE_UNUSED marking.
10022 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10024 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10025 alternative and update.
10026 (aarch64_dup_lanedi): Delete.
10027 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10028 * config/aarch64/aarch64-simd-builtins.def: Update.
10030 2013-06-17 Richard Biener <rguenther@suse.de>
10032 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10033 (lto_input_scc): Declare.
10034 (lto_input_tree_1): Likewise.
10035 (struct lto_stats_d): Add num_tree_bodies_output and
10036 num_pickle_refs_output.
10037 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10038 (lto_read_tree_1): Split out from ...
10039 (lto_read_tree): ... this.
10040 (lto_input_scc): New function.
10041 (lto_input_tree_1): Split out from ...
10042 (lto_input_tree): ... this. Handle LTO_tree_scc.
10043 (lto_data_in_create): Create the streamer cache without hashes.
10044 * lto-streamer-out.c (create_output_block): Create the streamer
10045 cache with hashes when not doing WPA.
10046 (lto_write_tree_1): Split out from ...
10047 (lto_write_tree): ... this.
10048 (get_symbol_initial_value): New function.
10049 (lto_output_tree_1): Split out from ...
10050 (lto_output_tree): ... this. Write trees as series of SCCs
10051 using a DFS walk via DFS_write_tree.
10052 (struct sccs, struct scc_entry): New types.
10053 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10054 (DFS_write_tree_body): New function.
10055 (DFS_write_tree): Likewise.
10056 (hash_tree): Likewise.
10057 (scc_entry_compare): Likewise.
10058 (hash_scc): Likewise.
10059 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10060 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10061 TREE_CHAIN as regular reference.
10062 (streamer_read_integer_cst): Remove.
10063 (streamer_get_pickled_tree): Adjust.
10064 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10065 of DECL_EXTERNALs in BLOCK_VARS for now.
10066 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10068 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10069 Add hash value argument and record that if hashes are recorded
10071 (streamer_tree_cache_insert_1): Adjust.
10072 (streamer_tree_cache_insert): Likewise.
10073 (streamer_tree_cache_insert_at): Rename to ...
10074 (streamer_tree_cache_replace_tree): ... this and adjust.
10075 (streamer_tree_cache_append): Adjust.
10076 (record_common_node): Likewise.
10077 (streamer_tree_cache_create): Add argument whether to
10078 record hash values together with trees.
10079 (streamer_tree_cache_delete): Adjust.
10080 * tree-streamer.h (struct streamer_tree_cache_d): Add
10082 (streamer_read_integer_cst): Remove.
10083 (streamer_tree_cache_insert): Adjust.
10084 (streamer_tree_cache_append): Likewise.
10085 (streamer_tree_cache_insert_at): Rename to ...
10086 (streamer_tree_cache_replace_tree): ... this and adjust.
10087 (streamer_tree_cache_create): Add argument whether to record hashes.
10088 (streamer_tree_cache_get): Rename to ...
10089 (streamer_tree_cache_get_tree): ... this.
10090 (streamer_tree_cache_get_hash): New function.
10091 * tree.c (cache_integer_cst): New function.
10092 * tree.h (cache_integer_cst): Declare.
10093 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10094 * lto-symtab.c (lto_varpool_replace_node): Only release
10095 DECL_INITIAL of non-prevailing decls.
10096 * varpool.c (varpool_remove_initializer): Do not release
10097 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10099 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10101 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10102 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10103 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10104 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10105 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10106 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10107 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10108 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10109 instead of TARGET_64BIT.
10110 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10111 Require ISA_HAS_<D>DIV.
10113 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10115 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10116 (mips*-*-linux*): Move default with_llsc setting to where other
10118 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10120 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10121 Likewise. Remove default with_tune setting. Move default float
10122 setting to its own block. Handle with_llsc in the same block as above.
10124 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10126 PR rtl-optimization/57425
10127 PR rtl-optimization/57569
10128 * alias.c (write_dependence_p): Add new parameters mem_mode,
10129 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10130 Changed all callers.
10131 (canon_anti_dependence): New function.
10132 * cse.c (check_dependence): Use canon_anti_dependence.
10133 * cselib.c (cselib_invalidate_mem): Likewise.
10134 * rtl.h (canon_anti_dependence): Declare.
10136 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10138 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10139 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10140 ".set mips3" for 64-bit targets.
10142 2013-06-15 Dehao Chen <dehao@google.com>
10144 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10145 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10146 (gimple_check_call_args): Likewise.
10147 * value-prof.c (check_ic_target): Likewise.
10148 * ipa-inline.c (early_inliner): Likewise.
10149 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10150 * cgraph.c (cgraph_create_edge_1): Likewise.
10151 (cgraph_make_edge_direct): Likewise.
10153 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10156 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10157 rs6000_output_move_128bit to handle emitting quad memory
10158 operations. Set attribute length to 8 bytes.
10160 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10162 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10164 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10165 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10166 (aarch64_<su>mlsl<mode>): Likewise.
10168 2013-06-14 Mike Stump <mikestump@comcast.net>
10170 * Makefile.in (TARGET_H): Add insn-codes.h.
10172 2013-06-14 Alan Modra <amodra@gmail.com>
10174 PR middle-end/57134
10175 PR middle-end/57586
10176 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10177 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10178 bitfield expansion when EXPAND_MEMORY.
10179 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10181 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10183 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10184 test for clearing quad memory on 32-bit later.
10186 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10188 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10189 (fold_negate_expr): Likewise.
10190 (fold_real_zero_addition_p): Handle vectors.
10191 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10193 2013-06-14 Alan Modra <amodra@gmail.com>
10195 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10197 2013-06-13 Jan Hubicka <jh@suse.cz>
10199 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10200 Local comdats are not externally visible.
10201 * symtab.c (dump_symtab_base): Dump externally visible.
10202 (verify_symtab_base): Verify back links in the symtab hash.
10204 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10206 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10207 CONVERT_EXPR as equal nodes.
10209 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10211 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10213 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10215 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10216 Generalize to complex and vector.
10217 * tree.c (build_all_ones_cst): New function.
10218 * tree.h (build_all_ones_cst): Declare it.
10220 2013-06-13 Alan Modra <amodra@gmail.com>
10222 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10223 * config/rs6000/rs6000.md (signbittf2): New insn.
10224 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10225 (abstf2_internal, cmptf_internal2): Likewise.
10226 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10228 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10229 Pat Haugen <pthaugen@us.ibm.com>
10230 Peter Bergner <bergner@vnet.ibm.com>
10232 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10233 power8 byte, half-word, and quad-word atomic instructions.
10234 (emit_store_conditional): Likewise.
10235 (rs6000_expand_atomic_compare_and_swap): Likewise.
10236 (rs6000_expand_atomic_op): Likewise.
10238 * config/rs6000/sync.md (larx): Add new modes for power8.
10240 (AINT): New mode iterator to include TImode as well as normal
10241 integer modes on power8.
10242 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10243 that VSX registers are not considered. Use AINT mode iterator
10244 instead of INT1 to allow inclusion of quad word atomic operations
10246 (load_locked<mode>): Likewise.
10247 (store_conditional<mode>): Likewise.
10248 (atomic_compare_and_swap<mode>): Likewise.
10249 (atomic_exchange<mode>): Likewise.
10250 (atomic_nand<mode>): Likewise.
10251 (atomic_fetch_<fetchop_name><mode>): Likewise.
10252 (atomic_nand_fetch<mode>): Likewise.
10253 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10255 (ATOMIC): On power8, add QImode, HImode modes.
10256 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10257 modes that promote to SImode.
10258 (load_lockedti): Convert TImode arguments to PTImode, so that we
10259 get a guaranteed even/odd register pair.
10260 (load_lockedpti): Likewise.
10261 (store_conditionalti): Likewise.
10262 (store_conditionalpti): Likewise.
10264 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10265 atomic load/store instructions.
10268 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10270 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10272 (insn_count): New attribute, with most cases extracted from...
10273 (length): ...here. Redefine most cases in terms of insn_count.
10274 (single_insn): Delete.
10275 (can_delay): Use insn_count to check for single instructions.
10276 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10277 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10278 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10279 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10280 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10281 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10282 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10283 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10284 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10285 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10286 rather than "length".
10287 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10288 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10289 Use "insn_count" rather than "length".
10290 * config/mips/mips-dsp.md
10291 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10292 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10295 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10297 PR tree-optimization/57361
10298 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10300 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10302 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10304 (aarch64_simd_combine<mode>): New instruction expansion.
10305 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10306 function prototype.
10307 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10308 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10310 2013-06-12 Jan Hubicka <jh@suse.cz>
10312 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10313 decl has when in streaming stage.
10314 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10315 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10317 2013-06-12 Roland Stigge <stigge@antcom.de>
10320 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10322 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10324 PR tree-optimization/57537
10325 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10326 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10328 2013-06-12 Richard Biener <rguenther@suse.de>
10330 * data-streamer.h (streamer_write_char_stream): CSE
10331 obs->current_pointer.
10332 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10333 streamer_write_char_stream manually and optimize the resulting loop.
10334 (streamer_write_hwi_stream): Likewise.
10336 2013-06-12 Jan Hubicka <jh@suse.cz>
10338 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10339 * cgraph.h (varpool_create_empty_node): Declare.
10340 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10342 * symtab.c (symtab_unregister_node): Be lax about missin entries
10344 (symtab_get_node): Update comment.
10345 * varpool.c (varpool_create_empty_node): Break out from ...
10346 (varpool_node_for_decl): ... here.
10347 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10349 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10351 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10352 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10353 part. Use straight-line flow at the end.
10354 <COMPONENT_REF>: Remove superfluous else.
10355 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10357 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10360 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10361 target hook even for !TREE_PUBLIC decls. If no resolution info
10362 is available, return false for common and external decls.
10364 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10366 * config/rl78/constraints.md (U): New constraint.
10367 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10370 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10373 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10374 to allow returning address to AT_PLATFORM name.
10376 2013-06-11 Jan Hubicka <jh@suse.cz>
10378 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10379 * cgraph.h (symtab_node_base): Add weakref flag.
10380 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10381 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10382 (output_weakrefs): Use weakref flag.
10383 * fold-const.c (simple_operand_p): Handle WEAK.
10384 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10385 * ipa.c (varpool_externally_visible_p): Drop weakref.
10386 (function_and_variable_visibility): Update comment; fix weakref
10387 sanity checks; do not clear DECL_WEAK on them.
10388 * lto-cgraph.c (lto_output_node): update.
10389 (lto_output_varpool_node): Update.
10390 (input_overwrite_node): Update.
10391 (input_node): Update.
10392 (input_varpool_node): Update.
10393 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10394 (lto_symtab_merge_symbols): Add sanity check.
10395 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10396 * passes.c (rest_of_decl_compilation): Set static flag, too.
10397 * symtab.c (dump_symtab_base): Dump weakref.
10398 (verify_symtab_base): Sanity check weakrefs.
10399 (symtab_make_decl_local): Remove duplicated code.
10400 (symtab_alias_ultimate_target): Simplify.
10401 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10403 2013-06-11 Tom de Vries <tom@codesourcery.com>
10405 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10406 sequence_vect == NULL.
10408 2013-06-11 DJ Delorie <dj@redhat.com>
10410 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10411 (rl78_unwind_word_mode): New.
10413 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10415 * final.c (debug_prefix_maps): Make static.
10417 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10419 * function.c (initial_trampoline): Remove stray copy.
10421 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10423 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10425 2013-06-11 Martin Jambor <mjambor@suse.cz>
10427 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10428 within bounds at the beginning of the function.
10430 2013-06-11 Alan Modra <amodra@gmail.com>
10432 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10434 (get_named_section): Don't NULL !DECL_P decl.
10436 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10438 * doc/invoke.texi (core-avx2): Document.
10440 (atom): Updated with MOVBE.
10442 2013-06-11 Richard Biener <rguenther@suse.de>
10444 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10446 2013-06-11 Anton Blanchard <anton@samba.org>
10448 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10449 correct shift value in little-endian mode.
10451 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10454 * varasm.c (get_variable_align): Move #endif to the right place.
10456 2013-06-10 Cary Coutant <ccoutant@google.com>
10458 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10459 for hash so that hash table traversal order is deterministic.
10461 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10462 Pat Haugen <pthaugen@us.ibm.com>
10463 Peter Bergner <bergner@vnet.ibm.com>
10465 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10466 of vectors in GPRS if they are direct moves or quad word load or
10469 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10471 (direct_move_p): Likewise.
10472 (quad_load_store_p): Likewise.
10474 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10475 classes into bins based on the physical register type.
10476 (reg_class_to_reg_type): Likewise.
10477 (IS_STD_REG_TYPE): Likewise.
10478 (IS_FP_VECT_REG_TYPE): Likewise.
10479 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10480 use direct move instructions.
10481 (reload_gpr_vsx): Likewise.
10482 (reload_vsx_gpr): Likewise.
10483 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10484 information that is a simplification of register classes. Also
10485 precalculate direct move reload helpers.
10486 (direct_move_p): New function to return true if the operation can
10487 be done as a direct move instruciton.
10488 (quad_load_store_p): New function to return true if the operation
10489 is a quad memory operation.
10490 (rs6000_legitimize_address): If quad memory, only allow register
10491 indirect for TImode addresses.
10492 (rs6000_legitimate_address_p): Likewise.
10493 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10494 (rs6000_reload_register_type): Likewise.
10495 (register_to_reg_type): Return register type.
10496 (rs6000_secondary_reload_simple_move): New helper function for
10497 secondary reload and secondary memory needed to identify anything
10498 that is a simple move, and does not need reloading.
10499 (rs6000_secondary_reload_direct_move): New helper function for
10500 secondary reload to identify cases that can be done with several
10501 instructions via the direct move instructions.
10502 (rs6000_secondary_reload_move): New helper function for secondary
10503 reload to identify moves between register types that can be done.
10504 (rs6000_secondary_reload): Add support for quad memory operations
10505 and for direct move.
10506 (rs6000_secondary_memory_needed): Likewise.
10507 (rs6000_debug_secondary_memory_needed): Change argument names.
10508 (rs6000_output_move_128bit): New function to return the move to
10509 use for 128-bit moves, including knowing about the various
10510 limitations of quad memory operations.
10512 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10513 memory operations. call rs6000_output_move_128bit for the actual
10514 instruciton(s) to generate.
10515 (vsx_movti_64bit): Likewise.
10517 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10518 (UNSPEC_P8V_MTVSRWZ): Likewise.
10519 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10520 (UNSPEC_P8V_MTVSRD): Likewise.
10521 (UNSPEC_P8V_XXPERMDI): Likewise.
10522 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10523 (UNSPEC_FUSION_GPR): Likewise.
10524 (FMOVE128_GPR): New iterator for direct move.
10525 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10526 (f32_sv): Likewise.
10527 (f32_dm): Likewise.
10528 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10529 loads and direct move instructions.
10530 (zero_extendsidi2_lfiwzx): Likewise.
10531 (extendsidi2_lfiwax): Likewise.
10532 (extendsidi2_nocell): Likewise.
10533 (floatsi<mode>2_lfiwax): Likewise.
10534 (lfiwax): Likewise.
10535 (floatunssi<mode>2_lfiwzx): Likewise.
10536 (lfiwzx): Likewise.
10537 (fix_trunc<mode>_stfiwx): Likewise.
10538 (fixuns_trunc<mode>_stfiwx): Likewise.
10539 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10540 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10541 (parity<mode>2_cmpb): Set length/type attr.
10542 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10543 for 'mr.' to fast_compare.
10544 (bpermd_<mode>): Change type attr to popcnt.
10545 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10546 (p8_mtvsrwz_1): Likewise.
10547 (p8_mtvsrwz_2): Likewise.
10548 (reload_fpr_from_gpr<mode>): Likewise.
10549 (p8_mtvsrd_1): Likewise.
10550 (p8_mtvsrd_2): Likewise.
10551 (p8_xxpermdi_<mode>): Likewise.
10552 (reload_vsx_from_gpr<mode>): Likewise.
10553 (reload_vsx_from_gprsf): Likewise.
10554 (p8_mfvsrd_3_<mode>): LIkewise.
10555 (reload_gpr_from_vsx<mode>): Likewise.
10556 (reload_gpr_from_vsxsf): Likewise.
10557 (p8_mfvsrd_4_disf): Likewise.
10558 (multi-word GPR splits): Do not split direct moves or quad memory
10561 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10563 * tree-into-ssa.c (interesting_blocks): Make static.
10565 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10568 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10569 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10570 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10571 (get_variable_align): New function.
10572 (get_variable_section, emit_bss, emit_common,
10573 assemble_variable_contents, place_block_symbol): Use
10574 get_variable_align instead of DECL_ALIGN.
10575 (assemble_noswitch_variable): Add align argument, use it
10576 instead of DECL_ALIGN.
10577 (assemble_variable): Adjust caller. Use get_variable_align
10578 instead of DECL_ALIGN.
10579 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10581 (DATA_ABI_ALIGNMENT): Define.
10582 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10583 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10584 opt is false, only return the psABI mandated alignment increase.
10585 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10586 (DATA_ABI_ALIGNMENT): ... this.
10587 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10588 (DATA_ABI_ALIGNMENT): ... this.
10589 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
10590 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
10591 (DATA_ABI_ALIGNMENT): ... this.
10592 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
10593 * doc/tm.texi: Regenerated.
10595 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
10597 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
10598 cmp_code to construct REG_EQUAL note.
10600 2013-06-09 Jakub Jelinek <jakub@redhat.com>
10603 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
10604 that operands[2] doesn't overlap with operands[0].
10606 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
10607 Jan Hubicka <jh@suse.cz>
10609 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
10610 hack to mark symbols as used.
10612 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
10614 PR rtl-optimization/57559
10615 * lra-constraints.c (process_alt_operands): Don't discourage
10616 memory with known offset for offsetable memory constraint.
10617 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
10619 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
10621 * varasm.c (struct oc_local_state): Reorder fields.
10622 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
10623 and adjust accordingly.
10624 (output_constructor): Reorder initialization code and adjust call to
10625 output_constructor_bitfield.
10627 2013-06-07 Jan Hubicka <jh@suse.cz>
10629 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
10631 2013-06-07 David Malcolm <dmalcolm@redhat.com>
10633 * tree-object-size.c (unknown): Make const.
10635 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10637 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
10638 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
10639 for last alternative in the cpu_facility attribute.
10641 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10644 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
10645 (xordi3): Change operand 2 constraint to arm_xordi_operand.
10646 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
10647 * config/arm/constraints.md (Dg): New constraint.
10648 * config/arm/neon.md (xordi3_neon): Remove.
10649 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
10650 * config/arm/predicates.md (arm_xordi_operand): New predicate.
10652 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10654 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
10655 Clean up alternatives.
10657 2013-06-07 Alan Modra <amodra@gmail.com>
10659 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
10662 2013-06-07 Alan Modra <amodra@gmail.com>
10664 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
10666 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10668 * config/arm/constraints.md (Df): New constraint.
10669 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
10670 Correct length attribute for last two alternatives.
10672 2013-06-07 Alan Modra <amodra@gmail.com>
10674 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10675 override user -mfp-in-toc.
10676 (offsettable_ok_by_alignment): Consider just the current access
10677 rather than the whole object, unless BLKmode. Handle
10678 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
10679 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
10680 for -mcmodel=medium.
10681 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
10682 override user -mfp-in-toc or -msum-in-toc. Default to
10683 -mno-fp-in-toc for -mcmodel=medium.
10685 2013-06-06 DJ Delorie <dj@redhat.com>
10687 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
10688 TARGET_VALID_POINTER_MODE.
10690 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
10691 Pat Haugen <pthaugen@us.ibm.com>
10692 Peter Bergner <bergner@vnet.ibm.com>
10694 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
10695 Document new power8 builtins.
10697 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
10698 condition code register, to allow 128-bit logical operations to be
10699 done in the VSX or GPR registers.
10700 (nor<mode>3): Use the canonical form for nor.
10701 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
10702 vclz*, and vpopcnt* vector instructions.
10703 (nand<mode>3): Likewise.
10704 (orc<mode>3): Likewise.
10705 (clz<mode>2): LIkewise.
10706 (popcount<mode>2): Likewise.
10708 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
10709 that only the GPRs are recognized.
10711 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10712 support for new power8 builtins.
10714 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
10716 (xscvdpspn): Likewise.
10722 (vpopcnt): Likewise.
10723 (vpopcntb): Likewise.
10724 (vpopcnth): Likewise.
10725 (vpopcntw): Likewise.
10726 (vpopcntd): Likewise.
10728 (vmrgew): Likewise.
10729 (vmrgow): Likewise.
10731 (eqv_v16qi3): Likewise.
10732 (eqv_v8hi3): Likewise.
10733 (eqv_v4si3): Likewise.
10734 (eqv_v2di3): Likewise.
10735 (eqv_v4sf3): Likewise.
10736 (eqv_v2df3): Likewise.
10738 (nand_v16qi3): Likewise.
10739 (nand_v8hi3): Likewise.
10740 (nand_v4si3): Likewise.
10741 (nand_v2di3): Likewise.
10742 (nand_v4sf3): Likewise.
10743 (nand_v2df3): Likewise.
10745 (orc_v16qi3): Likewise.
10746 (orc_v8hi3): Likewise.
10747 (orc_v4si3): Likewise.
10748 (orc_v2di3): Likewise.
10749 (orc_v4sf3): Likewise.
10750 (orc_v2df3): Likewise.
10752 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
10753 allow power8 quad mode in 64-bit.
10754 (rs6000_builtin_vectorized_function): Add support to vectorize
10755 ISA 2.07 count leading zeros, population count builtins.
10756 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
10757 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
10758 (builtin_function_type): Add vgbbd builtin function which takes an
10760 (altivec_expand_vec_perm_const): Add support for new power8 merge
10763 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
10764 that does not include TImdoe for use with 32-bit.
10765 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
10767 (UNSPEC_VSX_CVDPSPN): Likewise.
10768 (vsx_xscvdpspn): Likewise.
10769 (vsx_xscvspdpn): Likewise.
10770 (vsx_xscvdpspn_scalar): Likewise.
10771 (vsx_xscvspdpn_directmove): Likewise.
10772 (vsx_and<mode>3): Split logical operations into 32-bit and
10773 64-bit. Add support to do logical operations on TImode as well as
10774 VSX vector types. Allow logical operations to be done in either
10775 VSX registers or in general purpose registers in 64-bit mode. Add
10776 splitters if GPRs were used. For AND, add clobber of CCmode to
10777 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
10779 (vsx_and<mode>3_32bit): Likewise.
10780 (vsx_and<mode>3_64bit): Likewise.
10781 (vsx_ior<mode>3): Likewise.
10782 (vsx_ior<mode>3_32bit): Likewise.
10783 (vsx_ior<mode>3_64bit): Likewise.
10784 (vsx_xor<mode>3): Likewise.
10785 (vsx_xor<mode>3_32bit): Likewise.
10786 (vsx_xor<mode>3_64bit): Likewise.
10787 (vsx_one_cmpl<mode>2): Likewise.
10788 (vsx_one_cmpl<mode>2_32bit): Likewise.
10789 (vsx_one_cmpl<mode>2_64bit): Likewise.
10790 (vsx_nor<mode>3): Likewise.
10791 (vsx_nor<mode>3_32bit): Likewise.
10792 (vsx_nor<mode>3_64bit): Likewise.
10793 (vsx_andc<mode>3): Likewise.
10794 (vsx_andc<mode>3_32bit): Likewise.
10795 (vsx_andc<mode>3_64bit): Likewise.
10796 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
10797 and xxlorc instructions.
10798 (vsx_eqv<mode>3_64bit): Likewise.
10799 (vsx_nand<mode>3_32bit): Likewise.
10800 (vsx_nand<mode>3_64bit): Likewise.
10801 (vsx_orc<mode>3_32bit): Likewise.
10802 (vsx_orc<mode>3_64bit): Likewise.
10804 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
10806 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
10808 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
10809 (p8_vmrgow): Likewise.
10810 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
10811 GPRs to be split under VSX.
10812 (p8v_clz<mode>2): Add power8 count leading zero support.
10813 (p8v_popcount<mode>2): Add power8 population count support.
10814 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
10817 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
10820 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
10822 (vec_nand): Likewise.
10823 (vec_vclz): Likewise.
10824 (vec_vclzb): Likewise.
10825 (vec_vclzd): Likewise.
10826 (vec_vclzh): Likewise.
10827 (vec_vclzw): Likewise.
10828 (vec_vgbbd): Likewise.
10829 (vec_vmrgew): Likewise.
10830 (vec_vmrgow): Likewise.
10831 (vec_vpopcnt): Likewise.
10832 (vec_vpopcntb): Likewise.
10833 (vec_vpopcntd): Likewise.
10834 (vec_vpopcnth): Likewise.
10835 (vec_vpopcntw): Likewise.
10837 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10839 PR rtl-optimization/57468
10840 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
10843 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10845 PR rtl-optimization/57459
10846 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
10847 type when setting live regs.
10849 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10851 * config/s390/s390.opt (mlra): New option.
10852 * config/s390/s390.c (s390_decompose_address): Check displacement
10853 for all registers for LRA.
10854 (s390_secondary_reload): Don't used secondary reloads for LRA.
10855 (s390_lra_p): New function.
10856 (TARGET_LRA_P): Define.
10857 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
10858 of attribute cpu_facility to zarch for the last alternative.
10859 (*cmpmem_short): Ditto.
10861 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
10863 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
10864 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
10865 (arm_expand_prologue): Likewise.
10867 2013-06-06 Teresa Johnson <tejohnson@google.com>
10870 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
10871 as this is now done by redirect_edge_and_branch_force.
10872 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
10873 barriers, and fix interaction with splitting.
10874 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
10875 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
10876 reflect changes made in the routine.
10877 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
10878 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
10879 since this is called in cfglayout mode, and replace partition fixup
10880 with assert as that is now done by force_nonfallthru_and_redirect.
10881 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
10882 already be marked with region crossing note.
10883 (insert_section_boundary_note): Make non-static, gate on flag
10884 has_bb_partition, rewrite to also check for multiple partitions.
10885 (rest_of_handle_reorder_blocks): Remove call to
10886 insert_section_boundary_note, now done later during free_cfg.
10887 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
10888 * bb-reorder.h (insert_section_boundary_note): Declare.
10889 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
10890 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
10891 invoke insert_section_boundary_note.
10892 (try_redirect_by_replacing_jump): Remove unnecessary
10893 check for region crossing note.
10894 (fixup_partition_crossing): New function.
10895 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
10896 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
10897 in non-cfglayout mode.
10898 (force_nonfallthru_and_redirect): Fixup partition boundaries,
10899 remove old code that tried to do this. Emit barrier correctly
10900 when we are in cfglayout mode.
10901 (last_bb_in_partition): New function.
10902 (rtl_split_edge): Correctly fixup partition boundaries.
10903 (commit_one_edge_insertion): Remove old code that tried to
10904 fixup region crossing edge since this is now handled in
10905 split_block, and set up insertion point correctly since
10906 block may now end in a jump.
10907 (verify_hot_cold_block_grouping): Guard against checking when not in
10908 linearized RTL mode.
10909 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
10911 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
10912 rtl_verify_flow_info, so not called in cfglayout mode.
10913 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
10914 (fixup_reorder_chain): Remove old code that attempted to fixup region
10915 crossing note as this is now handled in force_nonfallthru_and_redirect.
10916 (duplicate_insn_chain): Don't duplicate switch section notes.
10917 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
10919 * basic-block.h (emit_barrier_after_bb): Declare.
10921 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10923 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
10924 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
10925 arm_usatsihi): Adjust alternatives for arm_restrict_it.
10927 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10929 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
10931 * config/arm/ldmstm.md: Regenerate.
10933 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10935 * config/arm/sync.md (atomic_loaddi_1):
10936 Disable predication for arm_restrict_it.
10937 (arm_load_exclusive<mode>): Likewise.
10938 (arm_load_exclusivesi): Likewise.
10939 (arm_load_exclusivedi): Likewise.
10940 (arm_load_acquire_exclusive<mode>): Likewise.
10941 (arm_load_acquire_exclusivesi): Likewise.
10942 (arm_load_acquire_exclusivedi): Likewise.
10943 (arm_store_exclusive<mode>): Likewise.
10944 (arm_store_exclusive<mode>): Likewise.
10945 (arm_store_release_exclusivedi): Likewise.
10946 (arm_store_release_exclusive<mode>): Likewise.
10948 2013-06-06 Richard Biener <rguenther@suse.de>
10950 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
10952 (lto_tag_is_tree_code_p): Adjust.
10953 (lto_tag_is_gimple_code_p): Likewise.
10954 (lto_gimple_code_to_tag): Likewise.
10955 (lto_tag_to_gimple_code): Likewise.
10956 (lto_tree_code_to_tag): Likewise.
10957 (lto_tag_to_tree_code): Likewise.
10958 * data-streamer.h (streamer_write_hwi_in_range): Use
10959 uhwi streaming to stream the normalized range.
10960 (streamer_read_hwi_in_range): Likewise.
10962 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10964 * config/arm/arm.md (enabled_for_depr_it): New attribute.
10965 (predicable_short_it): Likewise.
10966 (predicated): Likewise.
10967 (enabled): Handle above.
10968 (define_cond_exec): Set predicated attribute to yes.
10970 2013-06-05 Mike Stump <mikestump@comcast.net>
10972 * gdbinit.in (__FUNCTION__): Add.
10974 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
10976 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
10977 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
10979 2013-06-05 Jan Hubicka <jh@suse.cz>
10981 * varasm.c (mark_decl_referenced): Revert the removal until targets
10984 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
10986 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
10987 instead of mark_decl_referenced.
10989 2013-06-05 Jan Hubicka <jh@suse.cz>
10991 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
10992 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
10993 and symtab_used_from_object_file_p.
10994 (cgraph_make_node_local_1): Clear forced_by_abi.
10995 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
10996 * cgraph.h (symtab_node_base): Add forced_by_abi.
10997 (decide_is_variable_needed): Remove.
10998 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
10999 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11000 (decide_is_symbol_needed): ... this one; handle symbols in general;
11001 always analyze virtuals; honnor forced_by_abi.
11002 (cgraph_finalize_function): Update.
11003 (varpool_finalize_decl): Update.
11004 (symbol_defined_and_needed): Remove.
11005 (analyze_functions): Update.
11006 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11007 output_refs, input_overwrite_node): Handle forced_by_abi.
11008 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11009 (address_taken_from_non_vtable_p): ... this one.
11010 (comdat_can_be_unshared_p_1): New function.
11011 (cgraph_comdat_can_be_unshared_p): Rename to ...
11012 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11013 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11014 (function_and_variable_visibility): Clear forced_by_abi as needed.
11015 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11016 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11017 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11018 * varpool.c (decide_is_variable_needed): Remove.
11020 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11022 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11023 (arm_option_override): Override arm_restrict_it where appropriate.
11024 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11025 * config/arm/arm.opt (mrestrict-it): New command-line option.
11026 * doc/invoke.texi: Document -mrestrict-it.
11028 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11030 * tsan.c (tsan_atomic_table): Make const.
11032 2013-06-05 Richard Biener <rguenther@suse.de>
11034 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11035 index associated with the tree we are supposed to replace.
11036 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11037 TREE_ASM_WRITTEN as zero for everything but SSA names.
11039 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11041 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11043 2013-06-04 Jan Hubicka <jh@suse.cz>
11045 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11046 (input_node, input_varpool_node): Handle correctly external same
11048 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11049 nodes at ltrans stage.
11051 2013-06-04 Jan Hubicka <jh@suse.cz>
11053 * ipa-inline.c (update_caller_keys): Fix availability test.
11054 (update_callee_keys): Likewise.
11055 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11056 to follow ELF standard.
11058 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11060 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11061 (mips64r5900el-*-elf*): New configurations.
11062 * config/mips/mips-cpus.def (r5900): New processor.
11063 * config/mips/mips-tables.opt: Regenerate.
11064 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11065 (mips_issue_rate): Handle PROCESSOR_R5900.
11066 (mips_reorg_process_insns): Force reorder mode for the R5900.
11067 * config/mips/mips.h (TARGET_MIPS5900): Define.
11068 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11070 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11072 * config/mips/mips.md (processor): Add r5900.
11073 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11075 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11077 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11078 into function to generate MOVI instruction.
11079 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11080 (aarch64_preferred_simd_mode): Turn into wrapper.
11081 (aarch64_output_scalar_simd_mov_immediate): New function.
11082 * config/aarch64/aarch64-protos.h: Add prototype for above.
11084 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11086 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11087 element_char member.
11088 (sizetochar): Return signed char.
11089 (aarch64_simd_valid_immediate): Remove elchar and other
11090 unnecessary variables.
11091 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11092 Calculate element_char as required.
11093 * config/aarch64/aarch64-protos.h: Update and move prototype
11094 for aarch64_output_simd_mov_immediate.
11095 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11098 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11100 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11101 information completed by aarch64_simd_valid_immediate.
11102 (aarch64_legitimate_constant_p): Update arguments.
11103 (aarch64_simd_valid_immediate): Work with struct rather than many
11105 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11106 (aarch64_simd_make_constant): Update arguments.
11107 (aarch64_output_simd_mov_immediate): Work with struct rather than
11108 many pointers. Output immediate directly rather than as operand.
11109 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11111 * config/aarch64/constraints.md (Dn): Update arguments.
11113 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11115 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11117 (aarch64_simd_immediate_valid_for_move): Remove.
11118 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11119 (aarch64_simd_make_constant): Update call.
11120 (aarch64_output_simd_mov_immediate): Update call.
11121 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11123 * config/aarch64/constraints.md (Dn): Update call.
11125 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11127 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11128 return type to bool for prototype.
11129 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11130 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11131 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11133 2013-06-04 Catherine Moore <clm@codesourcery.com>
11135 * config/mips/mips.opt (meva): New.
11136 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11137 (ASM_SPEC): Handle -meva.
11138 * doc/invoke.texi (meva): Document.
11140 2013-06-04 Alan Modra <amodra@gmail.com>
11142 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11145 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11147 * rtl.def: Add extra fourth optional field to define_cond_exec.
11148 * gensupport.c (process_one_cond_exec): Process attributes from
11150 * doc/md.texi: Document fourth field in define_cond_exec.
11152 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11154 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11155 out the processing order as in store_bit_field_1.
11157 2013-06-04 Jan Hubicka <jh@suse.cz>
11159 PR middle-end/57500
11160 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11161 non-VAR_DECL node if it does not exist yet.
11163 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11165 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11166 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11167 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11168 target_cpu_default setting.
11170 2013-06-03 Teresa Johnson <tejohnson@google.com>
11172 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11173 default to -fopt-info=optimized instead of all.
11174 * doc/invoke.texi: Ditto.
11175 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11176 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11177 (execute_vect_slp): Emit BB vectorization success under
11178 MSG_OPTIMIZED_LOCATIONS.
11179 * tree-vect-slp.c (vect_slp_transform_bb): Change
11180 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11181 * tree-vect-loop.c (vect_transform_loop): Ditto.
11183 2013-06-03 Jason Merrill <jason@redhat.com>
11186 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11187 Use TARGET_EXPR for C++.
11189 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11191 PR rtl-optimization/57268
11192 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11193 if DEBUG_INSN_P (insn).
11196 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11198 PR rtl-optimization/57268
11199 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11200 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11202 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11204 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11205 (ix86_avoid_lea_for_addr): Likewise.
11206 (exact_dependency_1): Likewise.
11207 (ix86_adjust_cost): Likewise.
11208 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11210 (do_reorder_for_imul): Fix typo, formatting and
11211 !reload_completed check removed.
11212 (ix86_sched_reorder): Fix typo and formatting.
11213 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11216 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11218 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11220 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11222 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11223 <STRING_CST>: Likewise.
11224 <VECTOR_CST>: Likewise.
11226 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11227 Mikael Morin <mikael@gcc.gnu.org>
11229 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11230 * config.in: Regenerated.
11231 * configure: Regenerated.
11233 2013-06-01 Jan Hubicka <jh@suse.cz>
11235 PR middle-end/57366
11236 * cgraphunit.c (compile): When weakref is not supported,
11237 set up transparent aliases before final output pass.
11238 * varasm.c (assemble_alias): Do not try to do it here.
11240 2013-06-01 Jan Hubicka <jh@suse.cz>
11242 PR middle-end/57467
11243 * passes.c (for_per_function): Skip unanalyzed functions.
11245 2013-06-01 Jan Hubicka <jh@suse.cz>
11247 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11248 (lto_symtab_merge_symbols_1): ... this one.
11249 (lto_symtab_merge_cgraph_nodes): Rename to ...
11250 (lto_symtab_merge_symbols): ... this one; simplify.
11251 * cgraph.c (same_body_aliases_done): Rename to ...
11252 (cpp_implicit_aliases_done): ... this one.
11253 (cgraph_create_function_alias): Update.
11254 (cgraph_same_body_alias): Update.
11255 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11256 (verify_edge_corresponds_to_fndecl): Simplify.
11257 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11258 (cgraph_node): Remove same_body_alias.
11259 (varpool_node): Remove alias_of and extra_name_alias.
11260 (same_body_aliases_done): Rename to ..
11261 (cpp_implicit_aliases_done): ... this one.
11262 (symtab_alias_ultimate_target): Add default parameter.
11263 (symtab_resolve_alias): New function.
11264 (fixup_same_cpp_alias_visibility): Declare.
11265 (cgraph_function_node): Add default parameter.
11266 (cgraph_node_asm_name): Likewise.
11267 (cgraph_function_or_thunk_node): Add default parameter; do
11268 not ICE when it is NULL.
11269 (varpool_variable_node): Likewise.
11270 * tree-emutls.c (create_emultls_var): Update.
11271 (ipa_lower_emutls): Update.
11272 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11273 (cgraph_reset_node): Reset alias info.
11274 (cgraph_finalize_function): Update.
11275 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11276 (analyze_function): Simplify.
11277 (cgraph_process_same_body_aliases): Simplify.
11278 (analyze_functions): Fixup same body aliases.
11279 (handle_alias_pairs): Simplify.
11280 (assemble_thunk): Update.
11281 (assemble_thunks_and_aliases): Update.
11282 (output_weakrefs): Rewrite.
11283 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11284 (lto_output_varpool_node): Likewise.
11285 (compute_ltrans_boundary): Remve assert.
11286 (get_alias_symbol): New functoin.
11287 (input_node): Rewrite alias handling.
11288 (input_varpool_node): Likewise.
11289 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11290 * ipa.c (process_references): Handle weakrefs correctly.
11291 (symtab_remove_unreachable_nodes): Likewise.
11292 * trans-mem.c (get_cg_data): Update.
11293 (ipa_tm_create_version_alias): Update.
11294 (ipa_tm_execute): Update.
11295 * symtab.c (dump_symtab_base): Dump aliases.
11296 (verify_symtab_base): Verify aliases.
11297 (symtab_node_availability): New function.
11298 (symtab_alias_ultimate_target): Simplify.
11299 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11300 handle all the fixup cases.
11301 (symtab_resolve_alias): New function.
11302 * passes.c (ipa_write_summaries): Handle weakrefs.
11303 * varpool.c (varpool_analyze_node): Simplify.
11304 (assemble_aliases): Update.
11305 (varpool_create_variable_alias): Simplify.
11306 (varpool_extra_name_alias): Simplify.
11307 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11308 (lto_symtab_merge_symbols): ... this one.
11310 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11313 PR rtl-optimization/57268
11314 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11315 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11317 2013-06-01 Tobias Burnus <burnus@net-b.de>
11319 Partially reverted:
11320 2013-05-31 Tobias Burnus <burnus@net-b.de>
11322 PR middle-end/57073
11323 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11326 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11328 PR rtl-optimization/57268
11329 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11330 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11332 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11334 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11335 unordered comparison operators when -fno-trapping-math is in effect
11337 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11338 and implement unordered comparison operators properly on the e500.
11340 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11342 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11343 for constant scalar integers.
11344 (simplify_relational_operation_1): Likewise.
11346 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11348 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11349 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11352 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11353 Igor Zamyatin <igor.zamyatin@intel.com>
11355 Silvermont (SLM) architecture performance tuning.
11356 * config/i386/i386.h (enum ix86_tune_indices): Add
11357 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11358 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11360 * config/i386/i386.c (initial_ix86_tune_features)
11361 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11362 (ix86_lea_outperforms): Handle Silvermont tuning.
11363 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11365 (ix86_use_lea_for_mov): Likewise.
11366 (ix86_avoid_lea_for_addr): Likewise.
11367 (ix86_lea_for_add_ok): Likewise.
11368 (exact_dependency_1): New function.
11369 (exact_store_load_dependency): Likewise.
11370 (ix86_adjust_cost): Handle Silvermont tuning.
11371 (do_reoder_for_imul): Likewise.
11372 (swap_top_of_ready_list): New function.
11373 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11375 * config/i386/i386.md (peepholes that split memory operand in fp
11378 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11380 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11381 Remove un-necessary braces.
11383 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11385 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11386 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11388 2013-05-31 Tobias Burnus <burnus@net-b.de>
11390 PR middle-end/57073
11391 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11393 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11396 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11397 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11398 * config/arm/neon.md (iordi3_neon): Remove.
11399 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11400 * config/arm/predicates.md (imm_for_neon_logic_operand):
11401 Move to earlier in the file.
11402 (neon_logic_op2): Likewise.
11403 (arm_iordi_operand_neon): New predicate.
11405 2013-05-31 Richard Biener <rguenther@suse.de>
11407 PR tree-optimization/57478
11408 PR tree-optimization/57453
11409 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11412 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11414 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11415 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11417 2013-05-30 Tobias Burnus <burnus@net-b.de>
11418 Thomas Koenig <tkoenig@gcc.gnu.org>
11420 PR middle-end/57073
11421 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11422 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11424 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11426 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11428 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11430 * target.def (register_usage_leveling_p): New hook.
11431 * targhooks.c (default_register_usage_leveling_p): New.
11432 * targhooks.h (default_register_usage_leveling_p): New prototype.
11433 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11434 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11435 * doc/tm.texi: Update.
11436 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11438 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11440 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11441 (*insv_reg<mode>): New define_insn.
11443 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11445 PR rtl-optimization/57439
11446 * postreload.c (move2add_valid_value_p): Check that we have
11447 a zero subreg_regno_offset when accessing the register in
11448 the requested mode.
11450 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11451 Igor Zamyatin <igor.zamyatin@intel.com>
11453 Silvermont (SLM) architecture pipeline model, tuning and
11455 * config.gcc: Add slm config options and target.
11457 * config/i386/slm.md: New.
11459 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11461 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11463 (ix86_target_macros_internal): Likewise.
11465 * config/i386/i386.c (slm_cost): New cost.
11466 (m_SLM): New macro flag.
11467 (initial_ix86_tune_features): Set m_SLM.
11468 (x86_accumulate_outgoing_args): Likewise.
11469 (x86_arch_always_fancy_math_387): Likewise.
11470 (processor_target_table): Add slm cost.
11471 (cpu_names): Add slm cpu name.
11472 (x86_option_override_internal): Set SLM ISA.
11473 (ix86_issue_rate): New case PROCESSOR_SLM.
11474 (ia32_multipass_dfa_lookahead): Likewise.
11475 (fold_builtin_cpu): Add slm.
11477 * config/i386/i386.h (TARGET_SLM): New target macro.
11478 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11479 (processor_type): Add PROCESSOR_SLM.
11481 * config/i386/i386.md (cpu): Add new value "slm".
11482 (slm.md): Include slm.md.
11484 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11485 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11487 * config/arm/arm-protos.h: Add and update function protos.
11488 * config/arm/arm.c (use_simple_return_p): New added.
11489 (thumb2_expand_return): Check simple_return flag.
11490 * config/arm/arm.md: Add simple_return and conditional simple_return.
11491 * config/arm/iterators.md: Add iterator for return and simple_return.
11493 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11495 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11496 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11497 (arm_emit_vfp_multi_reg_pop): Likewise.
11498 (thumb2_emit_ldrd_pop): Likewise.
11499 (arm_expand_epilogue): Add misc REG_CFA notes.
11500 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11502 2013-05-29 Lawrence Crowl <crowl@google.com>
11504 * config/arm/t-arm: Update for below.
11506 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11507 Change type to hash_table. Update dependent calls and types.
11509 * config/i386/t-cygming: Update for below.
11511 * config/i386/t-interix: Update for below.
11513 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11514 Change type to hash_table. Update dependent calls and types.
11515 (i386_find_on_wrapper_list::wrappers): Likewise.
11517 * config/ia64/t-ia64: Update for below.
11519 * config/ia64/ia64.c (bundle_state_table):
11520 Change type to hash_table. Update dependent calls and types.
11522 * config/mips/mips.c (mips_reorg_process_insns::htab):
11523 Change type to hash_table. Update dependent calls and types.
11525 * config/sol2.c (solaris_comdat_htab):
11526 Change type to hash_table. Update dependent calls and types.
11528 * config/t-sol2: Update for above.
11530 2013-05-29 Teresa Johnson <tejohnson@google.com>
11532 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11533 functions are not yet marked as defined.
11535 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11536 Pat Haugen <pthaugen@us.ibm.com>
11537 Peter Bergner <bergner@vnet.ibm.com>
11539 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11543 (vrotl<mode>3): Likewise.
11544 (vashl<mode>3): Likewise.
11545 (vlshr<mode>3): Likewise.
11546 (vashr<mode>3): Likewise.
11548 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11549 support for power8 V2DI builtins.
11551 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11552 power8 V2DI builtins.
11553 (vupkhsw): Likewise.
11554 (vupklsw): Likewise.
11555 (vaddudm): Likewise.
11556 (vminsd): Likewise.
11557 (vmaxsd): Likewise.
11558 (vminud): Likewise.
11559 (vmaxud): Likewise.
11560 (vpkudum): Likewise.
11561 (vpksdss): Likewise.
11562 (vpkudus): Likewise.
11563 (vpksdus): Likewise.
11568 (vsubudm): Likewise.
11569 (vcmpequd): Likewise.
11570 (vcmpgtsd): Likewise.
11571 (vcmpgtud): Likewise.
11572 (vcmpequd_p): Likewise.
11573 (vcmpgtsd_p): Likewise.
11574 (vcmpgtud_p): Likewise.
11575 (vupkhsw): Likewise.
11576 (vupklsw): Likewise.
11577 (vaddudm): Likewise.
11578 (vmaxsd): Likewise.
11579 (vmaxud): Likewise.
11580 (vminsd): Likewise.
11581 (vminud): Likewise.
11582 (vpksdss): Likewise.
11583 (vpksdus): Likewise.
11584 (vpkudum): Likewise.
11585 (vpkudus): Likewise.
11590 (vsubudm): Likewise.
11592 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
11593 support for power8 V2DI instructions.
11595 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
11596 power8 V2DI instructions. Combine pack and unpack insns to use an
11597 iterator for each mode. Check whether a particular mode supports
11598 Altivec instructions instead of just checking TARGET_ALTIVEC.
11599 (UNSPEC_VPKUWUM): Likewise.
11600 (UNSPEC_VPKSHSS): Likewise.
11601 (UNSPEC_VPKSWSS): Likewise.
11602 (UNSPEC_VPKUHUS): Likewise.
11603 (UNSPEC_VPKSHUS): Likewise.
11604 (UNSPEC_VPKUWUS): Likewise.
11605 (UNSPEC_VPKSWUS): Likewise.
11606 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
11607 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
11608 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
11609 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
11610 (UNSPEC_VUPKHSB): Likewise.
11611 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
11612 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
11613 (UNSPEC_VUPKHSH): Likewise.
11614 (UNSPEC_VUPKLSB): Likewise.
11615 (UNSPEC_VUPKLSH): Likewise.
11617 (VI_char): Likewise.
11618 (VI_scalar): Likewise.
11619 (VI_unit): Likewise.
11621 (VP_small): Likewise.
11622 (VP_small_lc): Likewise.
11623 (VU_char): Likewise.
11624 (add<mode>3): Likewise.
11625 (altivec_vaddcuw): Likewise.
11626 (altivec_vaddu<VI_char>s): Likewise.
11627 (altivec_vadds<VI_char>s): Likewise.
11628 (sub<mode>3): Likewise.
11629 (altivec_vsubcuw): Likewise.
11630 (altivec_vsubu<VI_char>s): Likewise.
11631 (altivec_vsubs<VI_char>s): Likewise.
11632 (altivec_vavgs<VI_char>): Likewise.
11633 (altivec_vcmpbfp): Likewise.
11634 (altivec_eq<mode>): Likewise.
11635 (altivec_gt<mode>): Likewise.
11636 (altivec_gtu<mode>): Likewise.
11637 (umax<mode>3): Likewise.
11638 (smax<mode>3): Likewise.
11639 (umin<mode>3): Likewise.
11640 (smin<mode>3): Likewise.
11641 (altivec_vpkuhum): Likewise.
11642 (altivec_vpkuwum): Likewise.
11643 (altivec_vpkshss): Likewise.
11644 (altivec_vpkswss): Likewise.
11645 (altivec_vpkuhus): Likewise.
11646 (altivec_vpkshus): Likewise.
11647 (altivec_vpkuwus): Likewise.
11648 (altivec_vpkswus): Likewise.
11649 (altivec_vpks<VI_char>ss): Likewise.
11650 (altivec_vpks<VI_char>us): Likewise.
11651 (altivec_vpku<VI_char>us): Likewise.
11652 (altivec_vpku<VI_char>um): Likewise.
11653 (altivec_vrl<VI_char>): Likewise.
11654 (altivec_vsl<VI_char>): Likewise.
11655 (altivec_vsr<VI_char>): Likewise.
11656 (altivec_vsra<VI_char>): Likewise.
11657 (altivec_vsldoi_<mode>): Likewise.
11658 (altivec_vupkhsb): Likewise.
11659 (altivec_vupkhs<VU_char>): Likewise.
11660 (altivec_vupkls<VU_char>): Likewise.
11661 (altivec_vupkhsh): Likewise.
11662 (altivec_vupklsb): Likewise.
11663 (altivec_vupklsh): Likewise.
11664 (altivec_vcmpequ<VI_char>_p): Likewise.
11665 (altivec_vcmpgts<VI_char>_p): Likewise.
11666 (altivec_vcmpgtu<VI_char>_p): Likewise.
11667 (abs<mode>2): Likewise.
11668 (vec_unpacks_hi_v16qi): Likewise.
11669 (vec_unpacks_hi_v8hi): Likewise.
11670 (vec_unpacks_lo_v16qi): Likewise.
11671 (vec_unpacks_hi_<VP_small_lc>): Likewise.
11672 (vec_unpacks_lo_v8hi): Likewise.
11673 (vec_unpacks_lo_<VP_small_lc>): Likewise.
11674 (vec_pack_trunc_v8h): Likewise.
11675 (vec_pack_trunc_v4si): Likewise.
11676 (vec_pack_trunc_<mode>): Likewise.
11678 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
11680 (vec_vmaxsd): Likewise.
11681 (vec_vmaxud): Likewise.
11682 (vec_vminsd): Likewise.
11683 (vec_vminud): Likewise.
11684 (vec_vpksdss): Likewise.
11685 (vec_vpksdus): Likewise.
11686 (vec_vpkudum): Likewise.
11687 (vec_vpkudus): Likewise.
11688 (vec_vrld): Likewise.
11689 (vec_vsld): Likewise.
11690 (vec_vsrad): Likewise.
11691 (vec_vsrd): Likewise.
11692 (vec_vsubudm): Likewise.
11693 (vec_vupkhsw): Likewise.
11694 (vec_vupklsw): Likewise.
11696 2013-05-29 Jan Hubicka <jh@suse.cz>
11698 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
11699 flags; reorder rest of fields in more consistent way.
11700 (varpool_node): Remove analyzed, finalized and alias.
11701 (cgraph_ndoe): Likewise.
11702 (symtab_alias_ultimate_target): New function.
11703 (cgraph_function_node): Move offline.
11704 (cgraph_reset_node): Declare.
11705 (cgraph_comdat_can_be_unshared_p): Remove.
11706 (varpool_remove_initializer): Declare.
11707 (varpool_first_defined_variable, varpool_next_defined_variable
11708 cgraph_first_defined_function, cgraph_next_defined_function): Update.
11709 (cgraph_function_with_gimple_body_p): Update.
11710 (varpool_all_refs_explicit_p): Update.
11711 (symtab_alias_target): New function.
11712 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
11713 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
11714 (cgraph_function_or_thunk_node): Simplify using
11715 symtab_alias_ultimate_target.
11716 (varpool_variable_node): Likewise.
11717 * cgraph.c (cgraph_create_function_alias): Update.
11718 (cgraph_add_thunk): Update.
11719 (cgraph_remove_node): Update.
11720 (dump_cgraph_node): Do not dump removed flags.
11721 (cgraph_function_body_availability): Update.
11722 (cgraph_propagate_frequency): Update.
11723 (verify_cgraph_node): Check sanity of local flag.
11724 (cgraph_function_node): Move here from cgraph.h; revamp for
11725 cgraph_function_or_thunk_node.
11726 * lto-symtab.c (lto_varpool_replace_node): Update.
11727 (lto_symtab_resolve_can_prevail_p): Update.
11728 (lto_symtab_merge_cgraph_nodes): Update.
11729 * ipa-cp.c (determine_versionability, initialize_node_lattices,
11730 propagate_constants_accross_call, devirtualization_time_bonus,
11731 ipcp_propagate_stage): Update.
11732 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
11733 * ipa-inline-transform.c (clone_inlined_nodes,
11734 preserve_function_body_p): Update.
11735 * ipa-reference.c (propagate): Update.
11736 (write_node_summary_p): Update.
11737 * toplev.c (wrapup_global_declaration_2): Update.
11738 * cgraphunit.c (cgraph_analyze_function): Rename to ...
11739 (analyze_function) ... this one.
11740 (cgraph_process_new_functions): Update.
11741 (cgraph_reset_node): Export.
11742 (cgraph_finalize_function): Update.
11743 (cgraph_add_new_function): Update.
11744 (process_function_and_variable_attributes): Update.
11745 (varpool_finalize_decl): Update.
11746 (symbol_finalized): Remove.
11747 (symbol_finalized_and_needed): Rename to ...
11748 (symbol_defined_and_needed): ... update.
11749 (cgraph_analyze_functions): Update.
11750 (handle_alias_pairs): Update.
11751 (mark_functions_to_output): Update.
11752 (assemble_thunk): Update.
11753 (output_in_order): Update.
11754 (output_weakrefs): Update.
11755 (finalize_compilation_unit): Update.
11756 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
11757 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
11758 input_node, input_varpool_node): Update.
11759 * dbxout.c (dbxout_expand_expr): Update.
11760 * cgraphclones.c (cgraph_clone_node): Update.
11761 (cgraph_copy_node_for_versioning): Update.
11762 (cgraph_materialize_clone): Update.
11763 (cgraph_materialize_all_clones): Update.
11764 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
11765 propagate_pure_const, propagate_nothrow): Update.
11766 * lto-streamer-out.c (lto_output, write_symbol): Update.
11767 * ipa-utils.c (ipa_reverse_postorder): Update.
11768 * ipa-inline.c (can_inline_edge_p): Update.
11769 (update_caller_keys, ipa_inline): Update.
11770 * dwarf2out.c (reference_to_unused,
11771 premark_types_used_by_global_vars_helper): Update.
11772 * tree-eh.c (tree_could_trap_p): Update.
11773 * ipa-split.c (consider_split, execute_split_functions): Update.
11774 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
11775 has_addr_references_p): Update; move ahead in file for better
11777 (process_references): Simplify.
11778 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
11779 bodies are removed.
11780 (cgraph_comdat_can_be_unshared_p): Make static.
11781 (cgraph_externally_visible_p): Update.
11782 (varpool_externally_visible_p): Update.
11783 (function_and_variable_visibility): Update.
11784 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
11785 ipa_tm_mark_force_output_node): Update.
11786 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
11787 estimate_edge_devirt_benefit, inline_generate_summary,
11788 inline_write_summary): Update.
11789 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
11790 * ipa-prop.c (ipa_compute_jump_functions): Update.
11791 (ipa_print_node_params, ipa_prop_read_section,
11792 ipa_update_after_lto_read, read_replacements_section): Update.
11793 * varasm.c (mark_decl_referenced): Update.
11794 (assemble_alias, dump_tm_clone_pairs): Update.
11795 * tree-inline.c (copy_bb): Update.
11796 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
11798 * symtab.c (dump_symtab_base): Print new flags.
11799 (verify_symtab_base): Verify new flags.
11800 (symtab_alias_ultimate_target): New function.
11801 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
11802 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
11804 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
11806 * i386.c (ix86_get_function_versions_dispatcher,
11807 ix86_generate_version_dispatcher_body): Update.
11808 (fold_builtin_cpu): Use varpool_add_new_variable.
11809 * varpool.c (varpool_remove_initializer): Break out from ...
11810 (varpool_remove_node): ... this one.
11811 (dump_varpool_node, varpool_node_for_asm,
11812 cgraph_variable_initializer_availability, varpool_analyze_node,
11813 varpool_assemble_decl, varpool_remove_unreferenced_decls,
11814 varpool_finalize_named_section_flags, varpool_create_variable_alias):
11817 2013-05-29 Jan Hubicka <jh@suse.cz>
11819 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
11821 2013-05-29 Easwaran Raman <eraman@google.com>
11823 PR tree-optimization/57442
11824 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
11825 when control exits the main loop.
11827 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
11829 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
11831 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
11832 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
11833 * rx/t-rx: Add rx100 under multi library matches option for nofpu
11836 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11838 PR tree-optimization/57441
11839 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
11840 Don't limit size of incr_vec to number of candidates.
11842 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
11844 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
11845 and mips16 directories.
11846 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
11847 (MULTILIB_DIRNAMES): Ditto.
11848 (MULTILIB_EXCEPTIONS): Add new exceptions.
11849 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
11850 (MULTILIB_DIRNAMES): Ditto.
11851 (MULTILIB_EXCEPTIONS): Add new exceptions.
11853 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11854 Marcus Shawcroft <marcus.shawcroft@arm.com>
11856 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
11857 SYMBOL_TINY_ABSOLUTE.
11858 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
11859 SYMBOL_TINY_ABSOLUTE.
11860 (aarch64_expand_mov_immediate): Likewise.
11861 (aarch64_classify_symbol): Likewise.
11862 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
11863 Permit SYMBOL_TINY_ABSOLUTE.
11864 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
11866 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11867 Marcus Shawcroft <marcus.shawcroft@arm.com>
11869 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
11870 Refactor if/switch. Replace gcc_assert with if.
11872 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11874 * config/i386/i386.c (initial_ix86_tune_features): Enable
11875 FP Reassociation for AMD bdver1 and bdver2.
11877 2013-05-29 Martin Jambor <mjambor@suse.cz>
11879 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
11880 and IMAGPART_EXPR do not occur within other handled_components.
11882 2013-05-29 Richard Biener <rguenther@suse.de>
11884 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
11885 access on whether the use is in the BB we currently try to
11887 (vect_bb_vectorization_profitable_p): Pass the BB we currently
11888 vectorize to vect_bb_slp_scalar_cost.
11890 2013-05-29 Richard Biener <rguenther@suse.de>
11892 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
11893 computing scalar cost offsetted by stmts that are kept live
11895 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
11896 for computation of scalar cost.
11898 2013-05-28 Steve Ellcey <sellcey@mips.com>
11900 * config/mips/mips-cpus.def (mips32r2): Change processor type.
11902 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11904 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
11905 array notation built-in reduction functions.
11906 * doc/passes.texi (Passes): Added documentation about changes done
11908 * doc/invoke.texi (C Dialect Options): Added documentation about
11909 the -fcilkplus flag.
11910 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
11911 (BUILTINS_DEF): Depend on cilkplus.def.
11912 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
11913 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
11914 * cilkplus.def: New file.
11916 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
11918 PR rtl-optimization/57439
11919 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
11921 2013-05-28 Easwaran Raman <eraman@google.com>
11923 PR tree-optimization/57337
11924 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
11925 (find_insert_point): Correctly identify the insertion point
11926 when two statements with the same UID is compared.
11928 2013-05-28 Richard Biener <rguenther@suse.de>
11930 PR tree-optimization/56787
11931 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
11932 from the list of data references.
11933 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
11935 (vect_analyze_loop_operations): Likewise.
11936 (vect_transform_loop): Remove clobbers.
11938 2013-05-28 Martin Jambor <mjambor@suse.cz>
11940 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
11941 and REALPART_EXPRs have scalar type.
11943 2013-05-28 Richard Biener <rguenther@suse.de>
11945 PR tree-optimization/57411
11946 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
11948 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
11949 virtual operand propagation.
11951 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11953 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
11954 destination register for bmasksi_vis.
11955 (vector_init_bshuffle): Likewise.
11956 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
11958 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11960 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
11961 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
11962 mode if the instruction isn't available in the original mode.
11963 * config/sparc/sparc.opt (mfix-ut699): New option.
11964 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
11965 (divdf3): Turn into expander.
11966 (divdf3_nofix): New insn.
11967 (divdf3_fix): Likewise.
11968 (divsf3): Disable if -mfix-ut699.
11969 (sqrtdf2): Turn into expander.
11970 (sqrtdf2_nofix): New insn.
11971 (sqrtdf2_fix): Likewise.
11972 (sqrtsf2): Disable if -mfix-ut699.
11974 2013-05-27 Richard Biener <rguenther@suse.de>
11976 PR middle-end/57412
11977 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
11978 block for the new loop.
11980 2013-05-27 Richard Biener <rguenther@suse.de>
11982 PR tree-optimization/57343
11983 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
11984 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
11985 (number_of_iterations_cond): Do not build the folded tree.
11987 2013-05-27 Richard Biener <rguenther@suse.de>
11990 PR middle-end/57381
11991 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11992 OEP_CONSTANT_ADDRESS_OF retained.
11994 PR tree-optimization/57417
11995 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
11996 for unchanged base.
11997 (set_ssa_val_to): Compare addresses using
11998 get_addr_base_and_unit_offset.
12000 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12002 PR rtl-optimization/56833
12003 * postreload.c (move2add_record_mode): New function.
12004 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12005 (move2add_use_add2_insn): Use move2add_record_sym_value.
12006 (move2add_use_add3_insn): Likewise.
12007 (reload_cse_move2add): Use move2add_valid_value_p and
12008 move2add_record_mode. Invalidate call-clobbered and REG_INC
12009 affected regs by setting reg_mode to VOIDmode.
12010 (move2add_note_store): Don't pretend the inside of a SUBREG is
12011 the actual destination. Invalidate single/leading registers by
12012 setting reg_mode to VOIDmode.
12013 Use move2add_record_sym_value, move2add_valid_value_p and
12014 move2add_record_mode.
12016 2013-05-27 Richard Biener <rguenther@suse.de>
12018 PR tree-optimization/57396
12019 * tree-affine.c (double_int_constant_multiple_p): Properly
12020 return false for val == 0 and div != 0.
12022 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12024 * config/mips/mips.h: Use #elif in preprocessor conditions.
12026 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12029 * config/mips/constraints.md (kl): New constraint.
12030 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12031 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12032 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12033 from LO for MIPS16.
12034 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12036 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12039 * config/mips/mips.c (mips_can_inline_p): New function.
12040 (TARGET_CAN_INLINE_P): Define.
12042 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12044 * sched-int.h (ds_t, dw_t): Make unsigned int.
12045 Fix documentation that describes how all the ds_t bits are used.
12046 Reserve the last bit for delayed-branch scheduling.
12047 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12048 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12049 (gen_dep_weak_1): Remove prototype.
12050 * sched-deps.c (get_dep_weak_1): Make static.
12051 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12052 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12053 * doc/tm.texi: Regenerate.
12054 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12056 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12059 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12061 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12062 Sandra Loosemore <sandra@codesourcery.com>
12064 * config.gcc (powerpc-*): Allow native for with-cpu.
12066 2013-05-24 Jeff Law <law@redhat.com>
12068 PR tree-optimization/57124
12069 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12070 conversion feeding a condition if the range has an overflow
12071 if -fstrict-overflow. Add warnings for when we do make the
12074 2013-05-24 Dehao Chen <dehao@google.com>
12076 * tree-cfg.c (locus_discrim_map): Fix the typo.
12077 (locus_discrim_hasher): Likewise.
12078 (locus_discrim_hasher::hash): Likewise.
12079 (locus_discrim_hasher::equal): Likewise.
12081 2013-05-24 Martin Jambor <mjambor@suse.cz>
12083 PR tree-optimization/57294
12084 * cgraph.h (ipa_record_stmt_references): Declare.
12085 * cgraphbuild.c (ipa_record_stmt_references): New function.
12086 (build_cgraph_edges): Use ipa_record_stmt_references.
12087 (rebuild_cgraph_edges): Likewise.
12088 (cgraph_rebuild_references): Likewise.
12089 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12090 associated with the old statement and build references from the
12091 newly built statements.
12092 * ipa-ref.c (ipa_remove_stmt_references): New function.
12093 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12095 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12097 * lra-constraints.c (emit_spill_move): Use smaller mode for
12099 (check_and_process_move): Consider mem-reg moves for secondary
12101 (curr_insn_transform): Don't lose insns emitted before for
12102 secondary memory moves.
12103 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12104 reg set up in the current insn.
12106 2013-05-24 Dehao Chen <dehao@google.com>
12108 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12110 (locus_descrim_hasher::equal): Likewise.
12111 (build_gimple_cfg): New discriminator assignment algorithm.
12112 (make_edges): Likewise.
12113 (next_discriminator_for_locus): Likewise.
12114 (same_line_p): Likewise.
12115 (assign_discriminators): Likewise.
12116 (make_cond_expr_edges): Likewise.
12117 (make_gimple_switch_edges): Likewise.
12118 (make_goto_expr_edges): Likewise.
12119 (make_gimple_asm_edges): Likewise.
12121 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12123 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12124 X format specifier to only display bottom 16 bits.
12125 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12126 immediate to match for operand 2, since it will be masked.
12128 2013-05-24 Richard Biener <rguenther@suse.de>
12130 PR tree-optimization/57287
12131 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12132 all SSA names that occur in abnormal PHIs.
12134 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12137 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12138 that index is not negative.
12140 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12142 PR rtl-optimization/55177
12143 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12144 (simplify_byte_swapping_operation): New.
12145 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12146 (simplify_relational_operation_1): Deal with BSWAP.
12148 2013-05-23 Richard Henderson <rth@redhat.com>
12151 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12152 (ix86_reorg): Call it.
12154 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12157 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12158 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12159 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12161 2013-05-23 Christian Bruel <christian.bruel@st.com>
12164 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12166 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12167 Marcus Shawcroft <marcus.shawcroft@arm.com>
12169 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12170 * config/aarch64/constraints.md (Usa): Remove.
12171 * doc/md.texi (AArch64 Usa): Remove.
12173 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12174 Marcus Shawcroft <marcus.shawcroft@arm.com>
12176 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12177 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12178 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12179 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12181 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12183 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12184 instruction (AdvSIMD).
12185 * config/aarch64/aarch64-builtins.c
12186 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12187 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12189 2013-05-23 Martin Jambor <mjambor@suse.cz>
12191 PR middle-end/57347
12192 * tree.h (contains_bitfld_component_ref_p): Declare.
12193 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12194 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12196 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12197 not access a bit-field. Assert all final offsets are byte-aligned.
12199 2013-05-23 Richard Biener <rguenther@suse.de>
12201 PR tree-optimization/57380
12202 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12203 least one invariant or re-used load.
12204 * passes.c (init_optimization_passes): Move pass_phiprop before
12207 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12209 * config/aarch64/aarch64-simd.md
12210 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12212 2013-05-23 Richard Biener <rguenther@suse.de>
12214 PR middle-end/57381
12215 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12216 OEP_CONSTANT_ADDRESS_OF retained.
12218 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12220 PR middle-end/57344
12221 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12222 don't lower unit. Handle unit not being always BITS_PER_WORD.
12224 2013-05-23 Richard Biener <rguenther@suse.de>
12226 PR rtl-optimization/57341
12227 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12228 instead of true_dependence.
12230 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12232 * bb-reorder.c (branch_threshold): Make const.
12233 (exec_threshold): Ditto.
12235 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12236 Pat Haugen <pthaugen@us.ibm.com>
12237 Peter Bergner <bergner@vnet.ibm.com>
12239 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12240 documentation for the power8 crypto builtins.
12242 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12244 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12245 macros for defining power8 builtin functions.
12246 (BU_P8V_AV_2): Likewise.
12247 (BU_P8V_AV_P): Likewise.
12248 (BU_P8V_VSX_1): Likewise.
12249 (BU_P8V_OVERLOAD_1): Likewise.
12250 (BU_P8V_OVERLOAD_2): Likewise.
12251 (BU_CRYPTO_1): Likewise.
12252 (BU_CRYPTO_2): Likewise.
12253 (BU_CRYPTO_3): Likewise.
12254 (BU_CRYPTO_OVERLOAD_1): Likewise.
12255 (BU_CRYPTO_OVERLOAD_2): Likewise.
12256 (XSCVSPDP): Fix typo, point to the correct instruction.
12257 (VCIPHER): Add power8 crypto builtins.
12258 (VCIPHERLAST): Likewise.
12259 (VNCIPHER): Likewise.
12260 (VNCIPHERLAST): Likewise.
12261 (VPMSUMB): Likewise.
12262 (VPMSUMH): Likewise.
12263 (VPMSUMW): Likewise.
12264 (VPERMXOR_V2DI): Likewise.
12265 (VPERMXOR_V4SI: Likewise.
12266 (VPERMXOR_V8HI: Likewise.
12267 (VPERMXOR_V16QI: Likewise.
12268 (VSHASIGMAW): Likewise.
12269 (VSHASIGMAD): Likewise.
12270 (VPMSUM): Likewise.
12271 (VPERMXOR): Likewise.
12272 (VSHASIGMA): Likewise.
12274 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12275 __CRYPTO__ if the crypto instructions are available.
12276 (altivec_overloaded_builtins): Add support for overloaded power8
12279 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12280 support for power8 crypto builtins.
12281 (builtin_function_type): Likewise.
12282 (altivec_init_builtins): Add support for builtins that take vector
12283 long long (V2DI) arguments.
12285 * config/rs6000/crypto.md: New file, define power8 crypto
12288 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12289 Pat Haugen <pthaugen@us.ibm.com>
12290 Peter Bergner <bergner@vnet.ibm.com>
12292 * doc/invoke.texi (Option Summary): Add power8 options.
12293 (RS/6000 and PowerPC Options): Likewise.
12295 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12296 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12297 wm, wn, wr documentation.
12299 * config/rs6000/constraints.md (wm): New constraint for VSX
12300 registers if direct move instructions are enabled.
12301 (wn): New constraint for no registers.
12302 (wq): New constraint for quad word even GPR registers.
12303 (wr): New constraint if 64-bit instructions are enabled.
12304 (wv): New constraint if power8 vector instructions are enabled.
12305 (wQ): New constraint for quad word memory locations.
12307 * config/rs6000/predicates.md (const_0_to_15_operand): New
12308 constraint for 0..15 for crypto instructions.
12309 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12310 as GPR and floating point registers.
12311 (int_reg_operand): New predicate to match only GPR registers.
12312 (base_reg_operand): New predicate to match base registers.
12313 (quad_int_reg_operand): New predicate to match even GPR registers
12314 for quad memory operations.
12315 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12316 operations in both GPR and VSX registers.
12317 (quad_memory_operand): New predicate for quad memory operations.
12318 (reg_or_indexed_operand): New predicate for direct move support.
12320 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12321 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12322 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12323 (POWERPC_MASKS): Add power8 options.
12324 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12327 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12328 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12330 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12331 (-mpower8-fusion): New power8 options.
12332 (-mpower8-fusion-sign): Likewise.
12333 (-mpower8-vector): Likewise.
12334 (-mcrypto): Likewise.
12335 (-mdirect-move): Likewise.
12336 (-mquad-memory): Likewise.
12338 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12339 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12341 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12342 (rs6000_debug_vector_unit): Add p8_vector.
12343 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12344 definitions. Also print fusion state.
12345 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12346 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12347 (rs6000_option_override_internal): Add support for power8.
12348 (rs6000_common_init_builtins): Add debugging for skipped builtins
12349 if -mdebug=builtin.
12350 (rs6000_adjust_cost): Add power8 support.
12351 (rs6000_issue_rate): Likewise.
12352 (insn_must_be_first_in_group): Likewise.
12353 (insn_must_be_last_in_group): Likewise.
12354 (force_new_group): Likewise.
12355 (rs6000_register_move_cost): Likewise.
12356 (rs6000_opt_masks): Likewise.
12358 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12359 power8 capable assembler, default to power7 options.
12360 (TARGET_DIRECT_MOVE): Likewise.
12361 (TARGET_CRYPTO): Likewise.
12362 (TARGET_P8_VECTOR): Likewise.
12363 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12364 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12365 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12366 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12367 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12368 (TARGET_XSCVDPSPN): Likewise.
12369 (TARGET_XSCVSPDPN): Likewsie.
12370 (TARGET_SYNC_HI_QI): Likewise.
12371 (TARGET_SYNC_TI): Likewise.
12372 (MASK_CRYPTO): Likewise.
12373 (MASK_DIRECT_MOVE): Likewise.
12374 (MASK_P8_FUSION): Likewise.
12375 (MASK_P8_VECTOR): Likewise.
12376 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12377 temporary used by some of the direct move instructions to get two FP
12378 temporary registers does not force creation of a stack frame.
12379 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12380 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12381 that any VSX registers are tieable, even if they are also an
12382 Altivec vector mode.
12383 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12384 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12385 (RS6000_BTM_CRYPTO): Likewise.
12386 (RS6000_BTM_COMMON): Likewise.
12388 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12389 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12390 (enum rs6000_vector): Add power8 vector support.
12392 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12396 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12397 (any_sibcall_could_use_r3): this and handle indirect calls.
12398 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12400 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12402 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12404 2013-05-22 Richard Biener <rguenther@suse.de>
12406 PR middle-end/57349
12407 * profile.c (branch_prob): Do not split blocks that are
12408 abnormally receiving from ECF_RETURNS_TWICE functions.
12410 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12412 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12413 address mode. Move pointer mode initialization to the same place.
12415 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12417 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12418 while it has any effect.
12420 2013-05-21 Easwaran Raman <eraman@google.com>
12422 PR tree-optimization/57322
12423 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12424 UID of the statement added to the BB to be 1.
12426 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12428 PR tree-optimization/57331
12429 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12430 of conversion from pointer type to integral type with integer.
12432 2013-05-21 Martin Jambor <mjambor@suse.cz>
12435 * ipa-prop.c (ipa_read_node_info): Process param_used and
12436 controlled_uses in the same order as when writing.
12438 2013-05-21 Magnus Granberg <baldrick@free.fr>
12441 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12443 2013-05-21 Richard Biener <rguenther@suse.de>
12445 PR tree-optimization/57318
12446 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12447 estimate stmts with side-effects as likely eliminated.
12449 2013-05-21 Richard Biener <rguenther@suse.de>
12451 PR tree-optimization/57330
12452 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12453 preserve the call stmts fntype.
12455 2013-05-21 Richard Biener <rguenther@suse.de>
12457 PR tree-optimization/57303
12458 * tree-ssa-sink.c (statement_sink_location): Improve killing
12459 stmt detection and properly handle self-assignments.
12461 2013-05-21 Christian Bruel <christian.bruel@st.com>
12463 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12464 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12465 registers. Set register size out of the PARALLEL loop.
12467 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12470 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12471 (*fmasf4, *fmasf4_media): New insns.
12473 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12475 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12476 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12477 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12478 (mips_idiv_insns): Update the comments to say that the returned
12479 instruction counts are in units of BASE_INSN_LENGTH.
12480 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12481 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12482 using 2 rather than 4 as the length of indirect MIPS16 and
12483 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12484 length of a NOP. Don't divide MIPS16 lengths by 2.
12485 (mips16_split_long_branches): Assume a branch is long if the
12486 length is greater than 4 rather than 8.
12487 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12488 rather than multiplying them by 2. Multiply instruction counts
12489 by BASE_INSN_LENGTH rather than 4.
12490 (*jump_mips16, tls_get_tp_mips16_<mode>)
12491 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12493 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12495 * config/mips/mips.md (extended_mips16): Remove branch case.
12496 (length): Remove duplicated extended_mips16 test.
12498 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12500 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12502 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12504 * recog.h (Recog_data): Rename to...
12505 (recog_data_d): ...this.
12506 (recog_data): Update accordingly.
12507 * recog.c (recog_data): Likewise.
12508 * reload.c (save_recog_data): Likewise.
12509 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12510 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12512 2013-05-17 Julian Brown <julian@codesourcery.com>
12514 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12515 found in a REG_EQUAL note, invalidate it.
12517 2013-05-17 Easwaran Raman <eraman@google.com>
12519 * tree-ssa-reassoc.c (find_insert_point): New function.
12520 (insert_stmt_after): Likewise.
12521 (get_def_stmt): Likewise.
12522 (ensure_ops_are_available): Likewise.
12523 (not_dominated_by): Likewise.
12524 (rewrite_expr_tree): Do not move statements beyond what is
12525 necessary. Remove call to swap_ops_for_binary_stmt...
12526 (reassociate_bb): ... and move it here.
12527 (build_and_add_sum): Assign UIDs for new statements.
12528 (linearize_expr): Likewise.
12529 (do_reassoc): Renumber gimple statement UIDs.
12531 2013-05-17 Jan Hubicka <jh@suse.cz>
12533 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12535 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12536 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12538 (output_weakrefs): Update.
12540 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12541 Martin Jambor <mjambor@suse.cz>
12543 PR middle-end/57276
12544 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12545 value that corresponds to the given aggval is found in values vector.
12547 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12549 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12550 sse, sse2, sse3, ssse3 and sse4a flags to options.
12552 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12554 * gengtype-state.c: (s_expr_writer): New class, to handle
12555 prettifying of output layout of s-expressions.
12556 (state_writer): New class, to write out gtype.state.
12557 (state_written_type_count): Move this variable into member data of
12559 (s_expr_writer::s_expr_writer): New code: constructor for new class
12560 (state_writer::state_writer(): ditto
12561 (s_expr_writer::write_new_line): New function
12562 (s_expr_writer::write_any_indent): ditto
12563 (s_expr_writer::begin_s_expr): ditto
12564 (s_expr_writer::end_s_expr): ditto
12565 (write_state_fileloc): convert to method of state_writer...
12566 (state_writer:: write_state_fileloc): ...and use methods of
12567 s_expr_writer to write indentation into the gtype.state output file
12568 to visually represent the hierarchical structure of the list
12570 (write_state_fields): ditto, renaming to...
12571 (state_writer::write_state_fields)
12572 (write_state_a_string): ditto, renaming to...
12573 (state_writer::write_state_a_string)
12574 (write_state_string_option): ditto, renaming to...
12575 (state_writer::write_state_string_option)
12576 (write_state_type_option): ditto, renaming to...
12577 (state_writer::write_state_type_option)
12578 (write_state_nested_option): ditto, renaming to...
12579 (state_writer::write_state_nested_option)
12580 (write_state_option): ditto, renaming to...
12581 (state_writer::write_state_option)
12582 (write_state_options): ditto, renaming to...
12583 (state_writer::write_state_options)
12584 (write_state_lang_bitmap): ditto, renaming to...
12585 (state_writer::write_state_lang_bitmap)
12586 (write_state_version): ditto, renaming to...
12587 (state_writer::write_state_version)
12588 (write_state_scalar_type): ditto, renaming to...
12589 (state_writer::write_state_scalar_type)
12590 (write_state_string_type): ditto, renaming to...
12591 (state_writer::write_state_string_type)
12592 (write_state_undefined_type): ditto, renaming to...
12593 (state_writer::write_state_undefined_type)
12594 (write_state_struct_union_type): ditto, renaming to...
12595 (state_writer::write_state_struct_union_type)
12596 (write_state_struct_type): ditto, renaming to...
12597 (state_writer::write_state_struct_type)
12598 (write_state_user_struct_type): ditto, renaming to...
12599 (state_writer::write_state_user_struct_type)
12600 (write_state_lang_struct_type): ditto, renaming to...
12601 (state_writer::write_state_lang_struct_type)
12602 (write_state_param_struct_type): ditto, renaming to...
12603 (state_writer::write_state_param_struct_type)
12604 (write_state_pointer_type): ditto, renaming to...
12605 (state_writer::write_state_pointer_type)
12606 (write_state_array_type): ditto, renaming to...
12607 (state_writer::write_state_array_type)
12608 (write_state_gc_used): ditto, renaming to...
12609 (state_writer::write_state_gc_used)
12610 (write_state_common_type_content): ditto, renaming to...
12611 (state_writer::write_state_common_type_content)
12612 (write_state_type): ditto, renaming to...
12613 (state_writer::write_state_type)
12614 (write_state_pair_list): ditto, renaming to...
12615 (state_writer::write_state_pair_list)
12616 (write_state_pair): ditto, renaming to...
12617 (state_writer::write_state_pair)
12618 (write_state_typedefs): ditto, renaming to...
12619 (state_writer::write_state_typedefs)
12620 (write_state_structures): ditto, renaming to...
12621 (state_writer::write_state_structures)
12622 (write_state_param_structs): ditto, renaming to...
12623 (state_writer::write_state_param_structs)
12624 (write_state_variables): ditto, renaming to...
12625 (state_writer::write_state_variables)
12626 (write_state_srcdir): ditto, renaming to...
12627 (state_writer::write_state_srcdir)
12628 (write_state_files_list): ditto, renaming to...
12629 (state_writer::write_state_files_list)
12630 (write_state_languages): ditto, renaming to...
12631 (state_writer::write_state_languages)
12632 (write_state): create a state_writer instance and use it when
12633 writing out the state file
12635 2013-05-17 Mike Stump <mikestump@comcast.net>
12637 PR rtl-optimization/57304
12638 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
12639 accessing DF_REF_REAL_LOC.
12641 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12643 PR rtl-optimization/57281
12644 PR rtl-optimization/57300
12645 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
12646 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
12647 what the other splitter did if the registers are dead.
12649 2013-05-17 Richard Biener <rguenther@suse.de>
12651 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
12654 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12656 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
12659 2013-05-17 Marek Polacek <polacek@redhat.com>
12661 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
12662 length when doing non-zero store of storing '\0' to '\0'.
12664 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12666 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
12667 vect_external_def oprnd1 with loop_vinfo, try to emit
12668 optional cast, negation and and stmts on the loop preheader
12669 edge instead of into the pattern def seq.
12671 PR tree-optimization/57051
12672 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
12673 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
12675 2013-05-16 Nick Clifton <nickc@redhat.com>
12677 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
12678 (rl78_is_naked_func): New function.
12679 (rl78_expand_prologue): Skip prologue generation for naked functions.
12680 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
12681 * doc/extend.texi (naked): Add RL78 to the list of processors
12682 that supports this attribute.
12684 2013-05-16 Jeff Law <law@redhat.com>
12686 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
12688 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12690 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
12691 cache parameters using detect_caches_amd also for CYRIX,
12692 NSC and TM2 signatures.
12694 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
12695 Dzianis Kahanovich <mahatma@eu.by>
12699 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12700 VIA/Centaur processors and determine their cache parameters
12701 using detect_caches_amd.
12703 2013-05-16 Teresa Johnson <tejohnson@google.com>
12705 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
12706 (rtl_verify_edges): New function.
12707 (rtl_verify_bb_insns): Ditto.
12708 (rtl_verify_bb_pointers): Ditto.
12709 (rtl_verify_bb_insn_chain): Ditto.
12710 (rtl_verify_fallthru): Ditto.
12711 (rtl_verify_bb_layout): Ditto.
12712 (rtl_verify_flow_info_1): Outline checks into new functions.
12713 (rtl_verify_flow_info): Ditto.
12715 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
12717 * cfghooks.c (copy_bbs): Add update_dominance argument.
12718 * cfghooks.h (copy_bbs): Update prototype.
12719 * tree-cfg.c (gimple_duplicate_sese_region):
12720 Add update_dominance argument.
12721 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
12722 * tree-ssa-loop-ch.c (copy_loop_headers): Update
12723 gimple_duplicate_sese_region call.
12724 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
12725 Update copy_bbs call.
12726 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
12727 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
12729 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12731 * tree-vectorizer.h (NUM_PATTERNS): Increment.
12732 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
12733 vect_recog_rotate_pattern.
12734 (vect_recog_rotate_pattern): New function.
12736 2013-05-16 Jason Merrill <jason@redhat.com>
12738 * Makefile.in (LLINKER): New variable.
12739 (mostlyclean): Remove link mutex.
12740 * configure.ac: Handle --enable-link-mutex.
12741 * lock-and-run.sh: New script.
12743 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12746 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
12749 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12751 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
12753 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
12755 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
12756 * config/arm/arm.c (next_consecutive_mem): New function.
12757 (gen_movmem_ldrd_strd): Likewise.
12758 * config/arm/arm.md (movmemqi): Update condition and code.
12759 (unaligned_loaddi, unaligned_storedi): New patterns.
12761 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12763 * config.gcc: Obsolete *-*-solaris2.9*.
12764 * doc/install.texi (Specific, *-*-solaris2*): Document it.
12766 2013-05-16 Richard Biener <rguenther@suse.de>
12768 * passes.c (init_optimization_passes): Move pass_parallelize_loops
12769 earlier, after GRAPHITE transforms and IV canonicalization.
12771 2013-05-16 Jakub Jelinek <jakub@redhat.com>
12773 * omp-low.c (extract_omp_for_data): For collapsed loops,
12774 if at least one of the loops is known at compile time to
12775 iterate zero times, set count to 0.
12776 (expand_omp_regimplify_p): New function.
12777 (expand_omp_for_generic): For collapsed loops, if at least
12778 one of the loops isn't known to iterate at least once,
12779 add runtime check with setting count to 0.
12780 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
12781 For unsigned types if it isn't known at compile time that
12782 the loop will iterate at least once, add runtime check to bypass
12783 the whole loop if initial condition isn't true.
12785 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
12787 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
12789 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
12791 PR middle-end/57286
12792 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
12793 transformations to avoid an infinite loop.
12795 2013-05-16 Marek Polacek <polacek@redhat.com>
12797 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
12799 2013-05-15 Leif Ekblad <leif@rdos.net>
12801 * config/i386/i386.c (ix86_decompose_address): Use
12802 DEFAULT_TLS_SEG_REG to access TLS segment register.
12803 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
12804 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
12805 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
12807 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
12810 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
12811 sibling calls to functions that would normally be lazily bound,
12812 unless $gp is call-clobbered.
12814 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
12816 * config/i386/i386.c (ix86_option_override_internal): Update
12817 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
12818 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
12819 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
12820 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
12822 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
12824 2013-05-15 Andreas Schwab <schwab@suse.de>
12826 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
12827 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
12830 2013-05-15 Teresa Johnson <tejohnson@google.com>
12832 * loop-unroll.c (report_unroll_peel): Check decision before
12833 emitting unroll/peel message.
12835 2013-05-15 Teresa Johnson <tejohnson@google.com>
12837 * function.h (has_bb_partition): New rtl_data flag.
12838 (bb_reorder_complete): Ditto.
12839 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
12840 instead of flag_reorder_blocks_and_partition.
12841 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
12842 with some enhancements.
12843 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
12844 * bb-reorder.c (connect_traces): Check for has_bb_partition
12845 instead of flag_reorder_blocks_and_partition.
12846 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
12847 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
12848 verify_hot_cold_block_grouping.
12849 (partition_hot_cold_basic_blocks): Set has_bb_partition.
12851 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12854 * config/arm/predicates.md (call_insn_operand): New predicate.
12855 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
12856 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
12857 if insn is not a tail call.
12858 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
12860 * config/arm/arm.h (enum reg_class): New caller save register class.
12861 (REG_CLASS_NAMES): Likewise.
12862 (REG_CLASS_CONTENTS): Likewise.
12863 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
12866 2013-05-15 Richard Biener <rguenther@suse.de>
12868 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
12869 of MSG_OPTIMIZED_LOCATIONS.
12870 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
12871 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
12873 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
12874 of MSG_OPTIMIZED_LOCATIONS.
12875 (execute_vect_slp): Likewise.
12876 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
12877 (vect_create_cond_for_alias_checks): Likewise.
12878 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
12879 (vect_recog_widen_mult_pattern): Likewise.
12880 (vect_recog_widen_sum_pattern): Likewise.
12881 (vect_recog_over_widening_pattern): Likewise.
12882 (vect_recog_widen_shift_pattern): Likewise.
12883 (vect_recog_vector_vector_shift_pattern): Likewise.
12884 (vect_recog_divmod_pattern): Likewise.
12885 (vect_recog_mixed_size_cond_pattern): Likewise.
12886 (vect_recog_bool_pattern): Likewise.
12887 (vect_pattern_recog_1): Likewise.
12889 2013-05-15 Martin Jambor <mjambor@suse.cz>
12891 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
12892 non-functions to builtin_unreachable.
12893 * ipa-inline-transform.c (inline_call): Do not assert estimates were
12894 correct when new direct edges were discovered.
12896 2013-05-15 Martin Jambor <mjambor@suse.cz>
12898 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
12899 header, print symbol order instead of node uid, print more information
12900 about indirect edge targets.
12901 (ipa_make_edge_direct_to_target): Print symbol order instead of node
12903 (ipa_make_edge_direct_to_target): Likewise.
12904 (remove_described_reference): Likewise.
12905 (propagate_controlled_uses): Likewise.
12906 (ipa_print_node_params): Also print symbol order.
12907 (ipcp_transform_function): Print symbol order instead of node uids.
12908 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
12909 (cgraph_get_create_real_symbol_node): Likewise.
12910 * ipa-cp.c (print_lattice): Likewise.
12911 (print_all_lattices): Likewise.
12912 (determine_versionability): Likewise.
12913 (initialize_node_lattices): Likewise.
12914 (estimate_local_effects): Likewise.
12915 (update_profiling_info): Likewise.
12916 (create_specialized_node): Likewise.
12917 (perhaps_add_new_callers): Likewise.
12918 (decide_about_value): Likewise.
12919 (decide_whether_version_node): Likewise.
12920 (identify_dead_nodes): Likewise.
12921 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
12922 (dump_inline_summary): Likewise.
12923 (estimate_node_size_and_time): Likewise.
12924 (inline_analyze_function): Likewise.
12925 * ipa-inline.c (report_inline_failed_reason): Likewise.
12926 (want_early_inline_function_p): Likewise.
12927 (edge_badness): Likewise.
12928 (update_edge_key): Likewise.
12929 (inline_small_functions): Likewise. Add dumping of order to two other
12931 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
12932 instead of node uids.
12933 (propagate_pure_const): Likewise.
12934 (propagate_pure_const): Likewise.
12935 * ipa-utils.c (dump_cgraph_node_set): Likewise.
12936 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
12937 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
12939 * tree-pretty-print.c (dump_function_header): Likewise.
12940 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
12941 Print symbol order instead of node uids.
12943 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12945 * config/s390/s390.c (s390_register_move_cost): Don't impose the
12946 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
12948 2013-05-15 Richard Biener <rguenther@suse.de>
12950 PR tree-optimization/57275
12951 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
12952 return value for fail to do runtime alias checks for gather loads.
12954 2013-05-15 Jan Hubicka <jh@suse.cz>
12958 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
12959 weakrefs are not external.
12960 (lto_symtab_merge_decls): Fix thinko when dealing with
12961 non-lto_symtab decls.
12962 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
12963 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
12964 * varpool.c (dump_varpool_node): Dump more flags.
12966 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12968 * config/i386/i386.c (processor_alias_table): Add instruction
12969 FSGSBASE for AMD bdver3 architecture.
12971 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12973 * tree.c (warn_deprecated_use): Print file:line using locus color.
12974 * diagnostic.c (diagnostic_report_current_module): Print file:line
12975 and file:line:column using locus color.
12977 2013-05-14 Mike Stump <mikestump@comcast.net>
12979 * gdbinit.in: Add __null.
12981 2013-05-14 Mike Stump <mikestump@comcast.net>
12983 * recog.h: Rename struct recog_data to Recog_data.
12984 * recog.c: Likewise.
12985 * reload.c (can_reload_into): Likewise.
12986 * config/picochip/picochip.c: Likewise.
12988 2013-05-14 Mike Stump <mikestump@comcast.net>
12990 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
12992 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
12994 * resource.h (struct resources): Remove unch_memory member.
12995 (CLEAR_RESOURCE): Don't clear unch_memory.
12996 * resource.c (mark_referenced_resources): Don't set it.
12997 (mark_set_resources): Likewise.
12998 (mark_target_live_regs): Don't clear it.
12999 (init_resource_info): Likewise.
13000 * reorg.c (resource_conflicts_p): Don't compare it.
13001 (redundant_insn): Don't set it.
13003 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13005 * emit-rtl.c (next_label): Remove unused function.
13006 (skip_consecutive_labels, link_cc0_insns): Move to ...
13007 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13008 only place where these functions are used, and make them static.
13010 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13012 * fold-const.c (fold_negate_expr): Handle vectors.
13013 (fold_truth_not_expr): Make it static.
13014 (fold_invert_truthvalue): New static function.
13015 (invert_truthvalue_loc): Handle vectors. Do not call
13016 fold_truth_not_expr directly.
13017 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13018 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13019 (fold_comparison): Handle vector constants.
13020 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13021 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13022 * tree.h (fold_truth_not_expr): Remove declaration.
13024 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13026 * config/aarch64/aarch64-simd.md
13027 (aarch64_vcond_internal<mode>): Rename to...
13028 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13029 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13030 float modes. Clarify all iterator modes.
13031 (vcond<mode><mode>): Use new name for vcond expanders.
13032 (vcond<v_cmp_result><mode>): Likewise.
13033 (vcondu<mode><mode>: Likewise.
13034 * config/aarch64/iterators.md (VDQF_COND): New.
13036 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13039 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13040 variable for the shift amount. Check that we shift by non-negative
13043 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13046 * config/arm/arm.h (EPILOGUE_USES): Only return true
13047 for LR_REGNUM after epilogue_completed.
13049 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13051 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13052 is error_mark_node.
13054 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13057 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13058 and Solaris 11+/x86 with gld.
13059 * configure: Regenerate.
13061 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13063 * expmed.c (expand_shift_1): Canonicalize rotates by
13064 constant bitsize / 2 to bitsize - 1.
13065 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13066 case ROTATERT>: Likewise.
13069 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13071 * config/i386/i386.md (rotateinv): New code attr.
13072 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13073 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13074 roll $31, %eax, etc.
13076 2013-05-14 Richard Biener <rguenther@suse.de>
13078 PR middle-end/57235
13079 * tree-eh.c (sink_clobbers): Give up for successors with
13080 multiple predecessors and no virtual uses.
13082 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13084 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13085 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13087 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13089 PR middle-end/57251
13090 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13091 the case when both op0 and op1 have VOIDmode.
13093 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13095 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13096 in multiply-accumulate mode.
13098 2013-05-13 Guozhi Wei <carrot@google.com>
13100 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13102 2013-05-13 Kai Tietz <ktietz@redhat.com>
13105 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13106 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13107 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13108 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13109 (ix86_expand_prologue): Likewise.
13110 (ix86_expand_split_stack_prologue): Likewise.
13111 (legitimate_pic_address_disp_p): Likewise.
13112 (legitimize_pic_address): Likewise.
13113 (legitimize_tls_address): Likewise.
13114 (legitimize_pe_coff_symbol): Likewise.
13115 (output_pic_addr_const): Likewise.
13116 (construct_plt_address): Likewise.
13117 (ix86_expand_call): Likewise.
13118 (x86_output_mi_thunk): Likewise.
13119 (x86_function_profiler): Likewise.
13121 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13123 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13124 similar switch cases.
13125 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13126 (aarch64_simd_mov_to_<mode>low): Delete.
13127 (aarch64_simd_mov_to_<mode>high): Delete.
13128 (move_lo_quad_<mode>): Add w<-r alternative.
13129 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13130 (aarch64_simd_mov_from_*): Update type attribute.
13131 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13134 2013-05-13 Jan Hubicka <jh@suse.cz>
13136 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13137 * config/i386/i386.c (ix86_compute_frame_layout,
13138 ix86_expand_epilogue, emit_i387_cw_initialization,
13139 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13140 ix86_local_alignment): Fix use of size/speed predicates.
13142 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13144 PR tree-optimization/45216
13145 PR tree-optimization/57157
13146 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13147 the (-Y) & (B - 1) variant if OP is |.
13148 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13149 return shifted. Use (-op1) & (prec - 1) as other_amount
13150 instead of prec - op1.
13152 2013-05-13 Martin Jambor <mjambor@suse.cz>
13154 PR middle-end/42371
13155 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13156 (ipa_constant_data): New type.
13157 (ipa_jump_func): Use ipa_constant_data to hold information about
13158 constant jump functions.
13159 (ipa_get_jf_constant): Adjust to jump function type changes.
13160 (ipa_get_jf_constant_rdesc): New function.
13161 (ipa_param_descriptor): New field controlled_uses.
13162 (ipa_get_controlled_uses): New function.
13163 (ipa_set_controlled_uses): Likewise.
13164 * ipa-ref.h (ipa_find_reference): Declare.
13165 * ipa-prop.c (ipa_cst_ref_desc): New type.
13166 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13168 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13169 New parameter cs. Adjust all callers.
13170 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13171 (remove_described_reference): New function.
13172 (jfunc_rdesc_usable): Likewise.
13173 (try_make_edge_direct_simple_call): Decrement controlled use count,
13174 attempt to remove reference if it hits zero.
13175 (combine_controlled_uses_counters): New function.
13176 (propagate_controlled_uses): Likewise.
13177 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13178 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13179 (ipa_print_node_params): Print described use counter.
13180 (ipa_write_jump_function): Adjust to jump function type changes.
13181 (ipa_read_jump_function): New parameter CS, pass it to
13182 ipa_set_jf_constant. Adjust caller.
13183 (ipa_write_node_info): Stream controlled use count
13184 (ipa_read_node_info): Likewise.
13185 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13187 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13188 count. Remove cloning-added reference if it reaches zero.
13189 * ipa-ref.c (ipa_find_reference): New function.
13191 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13193 * config/i386/i386.c (processor_target_table): Modified default
13194 alignment values for AMD BD and BT architectures.
13196 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13198 * tree-vect-generic.c (uniform_vector_p): Move ...
13199 * tree.c (uniform_vector_p): ... here.
13200 * tree.h (uniform_vector_p): Declare it.
13201 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13204 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13206 PR tree-optimization/57230
13207 * tree-ssa-strlen.c (handle_char_store): Record length for
13208 array store from STRING_CST.
13210 PR tree-optimization/57230
13211 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13214 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13216 * config/epiphany/epiphany.c (epiphany_init): Check size of
13217 NUM_MODES_FOR_MODE_SWITCHING.
13218 (epiphany_expand_prologue):
13219 Remove CONFIG_REGNUM initial value handling code.
13220 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13221 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13222 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13223 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13224 Don't return 1 for FP_MODE_NONE.
13225 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13226 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13227 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13228 * config/epiphany/epiphany.md (save_config): New pattern.
13230 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13232 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13234 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13236 * config/i386/i386.md (memory): Handle sseishft1.
13237 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13238 (*vec_extractv2di_1): Ditto.
13240 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13242 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13245 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13247 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13248 Add mthumb/march=armv7-a multilib.
13249 Add mthumb/march=armv7-r multilib.
13250 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13252 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13254 * config/v850/t-rtems: Add more multilibs.
13256 2013-05-10 Richard Biener <rguenther@suse.de>
13258 PR tree-optimization/57214
13259 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13260 not propagate from SSA names that occur in abnormal PHI nodes.
13262 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13264 * stor-layout.c (element_precision): New function.
13265 * machmode.h (element_precision): Declare it.
13266 * tree.c (build_minus_one_cst): New function.
13267 (element_precision): Likewise.
13268 * tree.h (build_minus_one_cst): Declare new function.
13269 (element_precision): Likewise.
13270 * fold-const.c (operand_equal_p): Use element_precision.
13271 (fold_binary_loc): Handle vector types.
13272 * convert.c (convert_to_integer): Use element_precision.
13273 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13276 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13278 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13279 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13280 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13281 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13282 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13283 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13284 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13285 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13286 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13287 (Uuw8): New constraints.
13288 (Usb4): Move into alphabetical order.
13289 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13290 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13291 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13292 previously unnamed patterns.
13293 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13294 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13295 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13296 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13297 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13298 of specific lengths.
13300 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13302 * config/i386/i386.md (rotateinv): New code attr.
13303 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13304 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13305 roll $31, %eax, etc.
13307 PR tree-optimization/45216
13308 PR tree-optimization/57157
13309 * tree-ssa-forwprop.c (simplify_rotate): New function.
13310 (ssa_forward_propagate_and_combine): Call it.
13312 2013-05-10 Richard Biener <rguenther@suse.de>
13314 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13315 disable peeling when we version for aliasing.
13316 (vector_alignment_reachable_p): Honor explicit user alignment.
13317 (vect_supportable_dr_alignment): Likewise.
13318 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13319 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13320 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13321 then peeling to arrange for the cost-model check to come first.
13323 2013-05-10 Alan Modra <amodra@gmail.com>
13325 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13326 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13327 * configure: Regenerate.
13329 2013-05-10 Alan Modra <amodra@gmail.com>
13332 * varasm.c (default_elf_select_section): Move !DECL_P check..
13333 (get_named_section): ..to here before calling get_section_name.
13335 (default_section_type_flags): Add DECL_P check.
13336 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13337 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13339 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13341 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13342 When using gen_stack_adjust_str with a register offset, add a
13343 REG_FRAME_RELATED_EXPR note.
13345 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13347 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13348 (*vec_extractv4si_zext_mem): Ditto.
13349 (*vec_extractv2di): Add 0->x and x->x alternatives.
13350 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13351 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13353 2013-05-09 Jason Merrill <jason@redhat.com>
13355 N3639 C++1y VLA support
13356 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13359 * tree.c (build_constructor_va): New.
13360 * tree.h: Declare it.
13362 2013-05-09 Martin Jambor <mjambor@suse.cz>
13365 * gimple-fold.c (canonicalize_constructor_val): Call
13366 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13368 2013-05-09 Jan Hubicka <jh@suse.cz>
13369 Richard Biener <rguenther@suse.de>
13372 * symtab.c (symtab_make_decl_local): Do not add private names.
13374 2013-05-09 Jan Hubicka <jh@suse.cz>
13377 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13378 (unlink_from_assembler_name_hash): Likewise.
13379 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13380 symtab_unregister_node, symtab_initialize_asm_name_hash,
13381 change_decl_assembler_name): Update.
13383 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13385 * config/aarch64/aarch64.md: New movtf split.
13386 (*movtf_aarch64): Update.
13387 (aarch64_movdi_tilow): Handle TF modes and rename to
13388 aarch64_movdi_<mode>low.
13389 (aarch64_movdi_tihigh): Handle TF modes and rename to
13390 aarch64_movdi_<mode>high
13391 (aarch64_movtihigh_di): Handle TF modes and rename to
13392 aarch64_mov<mode>high_di
13393 (aarch64_movtilow_di): Handle TF modes and rename to
13394 aarch64_mov<mode>low_di
13395 (aarch64_movtilow_tilow): Remove spurious whitespace.
13396 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13398 (aarch64_print_operand): Update.
13400 2013-05-09 Alan Modra <amodra@gmail.com>
13402 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13404 * configure: Regenerate.
13406 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13408 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13409 splitter preparation statements.
13410 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13411 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13414 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13416 * gimple-ssa-strength-reduction.c (count_candidates): Change
13417 return value to int.
13418 (analyze_candidates_and_replace): Change type of length to int.
13420 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13422 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13423 (*vec_extract<mode>): Use VI12_128 mode iterator.
13424 (*vec_extract<mode>_mem): Ditto.
13425 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13428 2013-05-08 Diego Novillo <dnovillo@google.com>
13433 2012-08-17 Diego Novillo <dnovillo@google.com>
13436 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13437 * config.in: Regenerate.
13438 * configure: Regenerate.
13439 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13441 2013-05-08 Jan Hubicka <jh@suse.cz>
13444 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13445 * cgraph.h (symtab_node_base): Add unique_name.
13446 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13447 input_overwrite_node, input_varpool_node): Stream unique_name.
13448 * cgraphclones.c (cgraph_create_virtual_clone,
13449 cgraph_function_versioning): Set unique_name.
13450 * ipa.c (function_and_variable_visibility): Set unique_name.
13452 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13454 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13455 (alloc_cand_and_find_basis): Restrict conditional candidate
13456 processing to CAND_MULTs.
13458 2013-05-08 Jan Hubicka <jh@suse.cz>
13461 lto-symtab.c (lto_symtab_symbol_p): New function.
13462 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13463 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13464 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13465 Skip static symbols.
13467 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13469 PR tree-optimization/57200
13470 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13471 Only call inform if the preceding warning_at returns true.
13473 2013-05-07 Han Shen <shenhan@google.com>
13475 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13476 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13477 * common.opt (fstack-protector-strong): New option.
13478 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13479 * doc/invoke.texi (Optimization Options): Document
13480 "-fstack-protector-strong".
13481 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13483 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13485 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13487 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13489 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13490 it is not lhs of the insn.
13491 (setup_reg_equiv): Remove insn having equiv memory which it is not
13493 * lra-constraints.c (process_address): Try to improve generation
13494 code for address base + disp.
13495 (lra_constraints): Make correct the code for checking insn setting
13496 up backward equivalence. Remove insn only if it is in the init
13498 * lra-eliminations.c (update_reg_eliminate): Change return value.
13499 (lra_eliminate): Use the result.
13501 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13503 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13504 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13505 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13506 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13507 register target operands.
13508 (*vec_extractv8hi_sse2): New pattern.
13509 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13510 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13511 (*vec_extract<mode>_mem): New insn and split pattern.
13513 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13515 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13516 (TARGET_ASAN_SHADOW_OFFSET): Define.
13517 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13518 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13520 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13522 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13523 (incr_vec_index): Return -1 if increment not found.
13524 (create_add_on_incoming_edge): Assert if increment not found.
13525 (record_increment): Limit number of increments recorded.
13526 (all_phi_incrs_profitable): Return false if an increment not found.
13527 (replace_profitable_candidates): Don't process increments that were
13529 (analyze_candidates_and_replace): Limit size of incr_vec.
13531 2013-05-07 Richard Biener <rguenther@suse.de>
13533 * calls.c (special_function_p): setjmp-like functions are leaf.
13534 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13535 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13537 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13539 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13540 (aarch64_simd_mov<mode>): New expander.
13541 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13542 (aarch64_simd_mov_to_<mode>high): Likewise.
13543 (aarch64_simd_mov_from_<mode>low): Likewise.
13544 (aarch64_simd_mov_from_<mode>high): Likewise.
13545 (aarch64_dup_lane<mode>): Update.
13546 (aarch64_dup_lanedi): New instruction pattern.
13547 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13548 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13550 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13552 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13553 (replace_mult_candidate): Remove unnecessary argument; remove
13554 unnecessary parameter from call to introduce_cast_before_cand.
13555 (replace_unconditional_candidate): Remove unnecessary parameter
13556 from call to replace_mult_candidate.
13557 (replace_conditional_candidate): Likewise.
13558 (insert_initializers): Use make_temp_ssa_name.
13559 (introduce_cast_before_cand): Remove unnecessary argument; use
13560 make_temp_ssa_name.
13561 (replace_one_candidate): Remove unnecessary argument; remove
13562 unnecessary parameter from calls to introduce_cast_before_cand.
13563 (replace_profitable_candidates): Remove unnecessary parameters
13564 from calls to replace_one_candidate.
13566 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13568 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13569 phi def as possibly hiding a basis for a CAND_ADD whose operands
13570 have been commuted in the analysis.
13571 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13573 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13575 * config/aarch64/aarch64.md
13576 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13577 shift value between 0-4.
13579 2013-05-07 Richard Biener <rguenther@suse.de>
13581 * double-int.h (rshift): New overload.
13582 * double-int.c (rshift): New function.
13583 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13584 (create_reference_ops_from_ref): Remove.
13585 (vn_reference_insert): Use shared ops for constructing the
13586 reference and copy it.
13588 2013-05-07 Richard Biener <rguenther@suse.de>
13590 PR middle-end/57190
13591 * tree-eh.c (sink_clobbers): Properly propagate
13592 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
13594 2013-05-07 Jakub Jelinek <jakub@redhat.com>
13596 PR tree-optimization/57149
13597 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
13598 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
13599 collect_phi_def_edges, execute_late_warn_uninitialized): Use
13600 uninit_undefined_value_p instead of ssa_undefined_value_p.
13603 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
13604 for modifier == EXPAND_INITIALIZER.
13606 2013-05-07 Anton Blanchard <anton@samba.org>
13608 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
13609 for powerpc64 little endian.
13610 * configure: Regenerate.
13612 2013-05-06 Graham Stott <grahams@btinternet.com>
13614 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
13615 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
13616 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
13617 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
13619 2013-05-06 Graham Stott <grahams@btinternet.com>
13621 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
13622 codes which allow non-lvalues.
13624 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
13626 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
13627 components are all 1s.
13628 (integer_minus_onep): New function.
13629 * tree.h (integer_minus_onep): Declare it.
13630 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
13631 integer_minus_onep instead of integer_all_onesp.
13633 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13636 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
13637 variations of these patterns.
13639 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13641 * config/i386/i386.md (isa): Add x64_sse4 member.
13642 (enabled): Handle x64_sse4.
13643 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
13644 instruction for 64bit SSE4_1 targets. Update insn attributes.
13645 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
13646 instruction for SSE4_1 targets. Update insn attributes.
13647 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13648 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
13649 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
13651 (*vec_extractv4si): Rename from *sse4_1_pextrd.
13652 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
13653 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
13655 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13658 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
13660 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
13662 * final.c (do_assembler_dialects): Don't handle curly braces and
13663 vertical bar escaped by % as dialect delimiters.
13664 (output_asm_insn): Print curly braces and vertical bar if escaped
13665 by % and ASSEMBLER_DIALECT defined.
13666 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
13667 * doc/tm.texi: Regenerated.
13669 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13671 * config/mips/mips.c: Include tree-pass.h.
13672 (mips_reorg): Split in pre- and post-dbr_schedule parts.
13673 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
13674 (pass_mips_machine_reorg2): New machine specific pass.
13675 (insert_pass_mips_machine_reorg2): New pass plugin definition.
13676 (mips_option_override): Register the new pass.
13677 * rtl.h (cleanup_barriers): Remove prototype.
13678 (dbr_schedule): Likewise.
13679 * jump.c (cleanup_barriers): Make static.
13680 * reorg.c (dbr_schedule): Likewise.
13682 2013-05-06 Richard Biener <rguenther@suse.de>
13684 PR tree-optimization/57185
13685 * tree-parloops.c (add_field_for_reduction): Handle anonymous
13686 SSA names properly.
13688 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13691 * config/i386/i386.c (add_parameter_dependencies): Add dependence
13692 between "first_arg" and "insn", not "last" and "insn".
13694 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13696 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
13697 (find_candidates_in_block): Re-enable slsr_process_phi.
13698 (create_phi_basis): Fix double counting of candidate adjustment.
13700 2013-05-06 Richard Biener <rguenther@suse.de>
13702 PR middle-end/57147
13703 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
13704 the edge is also fallthru, preserve it and just clear the
13706 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
13707 also complex, preserve that and just clear the fallthru flag.
13708 * tree-inline.c (update_ssa_across_abnormal_edges): Also
13709 update virtual operands.
13711 2013-05-06 Alan Modra <amodra@gmail.com>
13713 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
13714 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
13715 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
13716 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
13717 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13718 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
13720 2013-05-06 Alan Modra <amodra@gmail.com>
13722 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
13723 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
13724 (DEFAULT_ASM_ENDIAN): Define.
13725 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
13726 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
13727 Update -K PIC clause from sysv4.h.
13728 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
13729 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
13731 2013-05-06 Alan Modra <amodra@gmail.com>
13733 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
13734 twice for little-endian.
13735 (ashrdi3_no_power, ashrdi3): Support little-endian.
13737 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13740 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
13741 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
13743 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
13744 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
13745 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
13748 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
13749 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13751 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
13752 * config/pa/pa.opt: Make mbig-switch a no-op.
13753 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
13754 (CASE_VECTOR_MODE): Always return SImode.
13755 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
13756 for the !TARGET_BIG_SWITCH case.
13757 * config/pa/pa-linux.h: Likewise.
13758 * config/pa/pa-openbsd.h: Likewise.
13759 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
13760 * config/pa/pa.md (short_jump): Remove define_insn.
13761 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
13762 (casesi0): Remove define_insn.
13763 (type): Remove btable_branch.
13764 (pa_combine_type): Likewise.
13765 (in_nullified_branch_delay): Likewise.
13766 (in_call_delay): Likewise.
13767 (define_delay): Likewise.
13768 (define_insn_reservation "Z3"): Likewise.
13769 (define_insn_reservation "Z4"): Likewise.
13770 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
13771 (pa_adjust_insn_length): Remove adjustment for btable branches.
13772 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
13775 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
13777 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13778 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
13779 Add m->r,x alternatives.
13780 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
13781 splitters using SWI48x mode iterator.
13782 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
13783 TARGET_64BIT. Add m->x alternative.
13784 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
13785 Add o->x alternative. Enable for TARGET_SSE.
13786 (sse_storeq): Remove expander.
13787 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
13788 with memory input operand.
13789 (*vec_extractv2di_1 splitter): New.
13790 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
13791 * config/i386/i386.md (ssevecmodelower): New mode attribute.
13793 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13795 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
13796 (INT_LOWPART): Delete.
13797 (extract_MB): Adjust.
13798 (extract_ME): Adjust.
13799 (print_operand): Adjust.
13801 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13803 * config/rs6000/predicates.md (reg_or_add_cint_operand,
13804 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13805 (reg_or_logical_cint_operand, easy_fp_constant,
13806 logical_const_operand): Delete "CONST_DOUBLE" case.
13807 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
13808 "HOST_BITS_PER_WIDE_INT == 64" test.
13809 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
13810 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
13811 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
13812 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
13814 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
13815 CONST_DOUBLE DImode/VOIDmode case.
13816 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
13817 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13818 CONST_DOUBLE VOIDmode case.
13819 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13820 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
13821 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
13822 Delete CONST_DOUBLE case.
13823 (splitters for mov FMOVE64 const_double): Delete
13824 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13825 "HOST_BITS_PER_WIDE_INT >= 64" test.
13826 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
13828 (mov DI const_double): Delete.
13830 2013-05-04 Jakub Jelinek <jakub@redhat.com>
13832 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
13833 on op shows all bits zero in mode of a lowpart subreg, return zero.
13835 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
13838 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
13839 to save TFmode registers and DImode to save TImode registers for
13840 caller save operations.
13841 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
13842 mark being partially clobbered since they only use the first
13845 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
13846 and TDmode only use the upper 64-bits of each VSX register.
13848 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13850 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
13851 (find_candidates_in_block): Disable slsr_process_phi.
13853 2013-05-03 Guozhi Wei <carrot@google.com>
13855 * coverage.c (coverage_obj_init): Move the construction of gcov
13857 (build_init_ctor): ... here.
13859 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13861 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
13862 (slsr_cand_d): Redefine def_phi.
13863 (stride_status, phi_adjust_status, count_phis_status): New enums.
13864 (find_phi_def): New.
13865 (find_basis_for_base_expr): New.
13866 (find_basis_for_candidate): Handle hidden bases.
13867 (alloc_cand_and_find_basis): Handle phi candidates.
13868 (slsr_process_phi): New.
13869 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
13870 (create_mul_imm_cand): Likewise.
13871 (create_add_ssa_cand): Exclude phi base candidates.
13872 (create_add_imm_cand): Likewise.
13873 (slsr_process_cast): Likewise.
13874 (slsr_process_copy): Likewise.
13875 (find_candidates_in_block): Handle phi candidates.
13876 (dump_candidate): Likewise.
13877 (unconditional_cands): Delete.
13878 (unconditional_cands_with_known_stride_p): Delete.
13879 (phi_dependent_cand_p): New.
13880 (cand_increment): Handle phi-dependent candidates.
13881 (replace_dependent): Delete.
13882 (replace_mult_candidate): New.
13883 (replace_unconditional_candidate): New.
13884 (incr_vec_index): Move to avoid forward reference.
13885 (create_add_on_incoming_edge): New.
13886 (create_phi_basis): New.
13887 (replace_dependents): Delete.
13888 (replace_conditional_candidate): New.
13889 (phi_add_costs): New.
13890 (replace_uncond_cands_and_profitable_phis): New.
13891 (record_increment): Handle phi adjustments.
13892 (record_phi_increments): New.
13893 (record_increments): Handle phi adjustments.
13894 (phi_incr_cost): New.
13895 (lowest_cost_path): Handle phis.
13896 (total_savings): Likewise.
13897 (analyze_increments): Likewise.
13898 (ncd_with_phi): New.
13899 (ncd_of_cand_and_phis): New.
13900 (nearest_common_dominator_for_cands): Handle phi increments.
13901 (all_phi_incrs_profitable): New.
13902 (replace_profitable_candidates): Handle phi-dependent candidates.
13903 (analyze_candidates_and_replace): Likewise.
13905 2013-05-03 Teresa Johnson <tejohnson@google.com>
13908 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
13909 do not exceed REG_BR_PROB_BASE.
13911 2013-05-03 Jeff Law <law@redhat.com>
13913 PR tree-optimization/57144
13914 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
13915 operand of the condition will bit into the new type when eliminating
13916 a cast feeding a condition.
13918 2013-05-03 Jakub Jelinek <jakub@redhat.com>
13920 PR rtl-optimization/57130
13921 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
13922 of COMPARE as in_code to the recursive call if needed.
13924 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
13926 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
13927 (enabled): Handle new members.
13928 * config/i386/sse.md (*vec_concatv2si): Merge from
13929 *vec_concatv2si_sse2 and vec_concatv2si_sse.
13930 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
13932 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
13934 PR tree-optimization/57027
13935 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
13936 for fnms opportunity, check we got the prerequisite kind
13937 of tree / gimple before using accessor functions.
13939 2013-05-03 Richard Biener <rguenther@suse.de>
13941 * double-int.h (lshift): New overload without precision
13942 and arith argument.
13943 (operator *=, operator +=, operator -=): Move ...
13944 * double-int.c (operator *=, operator +=, operator -=): ... here
13945 and implement more efficiently.
13946 (mul_double_with_sign): Remove.
13947 (lshift_double): Adjust to take unsinged shift argument, push
13948 dispatching code to callers.
13949 (mul_double_wide_with_sign): Add early out for callers that
13950 are not interested in high parts or overflow.
13951 (lshift): New function.
13952 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
13953 dispatch code here.
13954 (lrotate, rrotate): Use logical shifts.
13955 * expr.c (get_inner_reference): Use lshift.
13956 * fixed-value.c (do_fixed_divide): Likewise.
13957 * tree-dfa.c (get_ref_base_and_extent): Likewise.
13958 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
13959 (indirect_refs_may_alias_p): Likewise.
13960 (stmt_kills_ref_p_1): Likewise.
13962 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13964 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
13966 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13968 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
13969 scalar form of FABD instruction.
13971 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
13973 * lra-constraints.c (process_alt_operands): Add checking alt
13974 number to choose the best alternative.
13976 2013-05-02 Richard Biener <rguenther@suse.de>
13978 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
13979 bitmap and its handling.
13980 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
13982 2013-05-02 Richard Biener <rguenther@suse.de>
13984 PR middle-end/57140
13985 * tree-inline.c (copy_loops): Properly handle removed loops.
13986 (copy_cfg_body): Mark destination loops for fixup if source
13987 loops needed fixup.
13989 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
13992 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
13993 generating simple_return for naked functions.
13995 2013-05-02 Martin Jambor <mjambor@suse.cz>
13997 PR middle-end/56988
13998 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
13999 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14001 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14002 ipa_agg_replacement_value structures.
14003 (known_aggs_to_agg_replacement_list): Likewise.
14004 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14005 (read_agg_replacement_chain): Likewise.
14006 (ipcp_transform_function): Also check that by_ref flags match.
14008 2013-05-02 Richard Biener <rguenther@suse.de>
14010 * graphds.h (struct graph): Add obstack member.
14011 * graphds.c (new_graph): Initialize obstack and allocate
14013 (add_edge): Allocate edge from the obstack.
14014 (free_graph): Free the obstack instead of all edges and vertices.
14016 2013-05-02 Teresa Johnson <tejohnson@google.com>
14018 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14020 * cfg.c (update_bb_profile_for_threading): Ditto.
14021 * tree-inline.c (copy_bb): Ditto.
14022 (copy_edges_for_bb): Ditto.
14023 (initialize_cfun): Ditto.
14024 (copy_cfg_body): Ditto.
14025 (expand_call_inline): Ditto.
14026 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14027 (estimate_node_size_and_time): Ditto.
14028 (inline_merge_summary): Ditto.
14029 * cgraphclones.c (cgraph_clone_edge): Ditto.
14030 (cgraph_clone_node): Ditto.
14031 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14032 (compute_trg_info): Ditto.
14034 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14036 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14037 S reg when fp attribute set.
14038 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14040 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14042 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14044 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14045 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14046 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14048 2013-05-02 Richard Biener <rguenther@suse.de>
14050 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14051 (struct instantiate_cache_entry): New type.
14052 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14053 (struct instantiate_cache_type): New type.
14054 (set_instantiated_value, get_instantiated_value): Remove.
14055 (get_instantiated_value_entry): New function.
14056 (instantiate_scev_name): Use the new cache and adjust.
14057 (instantiate_scev_poly): Adjust.
14058 (instantiate_scev_binary): Likewise.
14059 (instantiate_array_ref): Likewise.
14060 (instantiate_scev_convert): Likewise.
14061 (instantiate_scev_not): Likewise.
14062 (instantiate_scev_3): Likewise.
14063 (instantiate_scev_2): Likewise.
14064 (instantiate_scev_r): Likewise.
14065 (instantiate_scev): Likewise.
14066 (resolve_mixers): Likewise.
14068 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14071 * lra-constraints.c (best_small_class_operands_num): Remove.
14072 (process_alt_operands): Remove small_class_operands_num. Take
14073 small classes operands into losers and only if the operand is not
14074 matched. Modify debugging output.
14075 (curr_insn_transform): Remove best_small_class_operands_num.
14078 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14080 * config/aarch64/aarch64-builtins.c
14081 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14082 * config/aarch64/aarch64-simd-builtins.def
14083 (reduc_splus_): Add new modes.
14084 (reduc_uplus_): New.
14085 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14086 (reduc_uplus_v4sf): Likewise.
14087 (reduc_splus_v4sf): Likewise.
14088 (aarch64_addv<mode>): Likewise.
14089 (reduc_uplus_<mode>): Likewise.
14090 (reduc_splus_<mode>): Likewise.
14091 (aarch64_addvv2di): Likewise.
14092 (reduc_uplus_v2di): Likewise.
14093 (reduc_splus_v2di): Likewise.
14094 (aarch64_addvv2si): Likewise.
14095 (reduc_uplus_v2si): Likewise.
14096 (reduc_splus_v2si): Likewise.
14097 (reduc_<sur>plus_<mode>): New.
14098 (reduc_<sur>plus_v2di): Likewise.
14099 (reduc_<sur>plus_v2si): Likewise.
14100 (reduc_<sur>plus_v4sf): Likewise.
14101 (aarch64_addpv4sf): Likewise.
14102 * config/aarch64/arm_neon.h
14103 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14104 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14105 add UNSPEC_SADDV, UNSPEC_UADDV.
14107 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14109 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14111 * config/aarch64/arm_neon.h
14112 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14114 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14116 * config/aarch64/aarch64-builtins
14117 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14119 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14121 * config/aarch64/aarch64-simd-builtins.def
14122 (reduc_smax_): New.
14123 (reduc_smin_): Likewise.
14124 (reduc_umax_): Likewise.
14125 (reduc_umin_): Likewise.
14126 (reduc_smax_nan_): Likewise.
14127 (reduc_smin_nan_): Likewise.
14130 (smax): Update for V2SF, V4SF and V2DF modes.
14133 (smin_nan): Likewise.
14134 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14135 (<su><maxmin><mode>3): ...This, refactor.
14136 (s<maxmin><mode>3): New.
14137 (<maxmin_uns><mode>3): Likewise.
14138 (reduc_<maxmin_uns>_<mode>): Refactor.
14139 (reduc_<maxmin_uns>_v4sf): Likewise.
14140 (reduc_<maxmin_uns>_v2si): Likewise.
14141 (aarch64_<fmaxmin><mode>: Remove.
14142 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14144 (vmin<q>_f<32,64>): Likewise.
14145 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14147 (su): Add mappings for smax, smin, umax, umin.
14149 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14150 (FMAXMIN): Rename as...
14151 (FMAXMIN_UNS): ...This.
14153 (fmaxminv): Likewise.
14154 (fmaxmin): Likewise.
14156 (maxmin_uns_op): Likewise.
14158 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14160 * config/aarch64/arm_neon.h
14161 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14162 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14163 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14165 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14167 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14168 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14170 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14172 * config/aarch64/aarch64-simd.md
14173 (vcond<mode>_internal): Handle special cases for constant masks.
14174 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14175 (vcondu<mode><mode>): Likewise.
14176 (vcond<v_cmp_result><mode>): New.
14178 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14180 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14181 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14182 * config/aarch64/aarch64-simd-builtins.def
14183 (cmeq): Update to BUILTIN_VALLDI.
14188 * config/aarch64/arm_neon.h
14189 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14190 to builtins or C as appropriate.
14192 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14194 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14196 (cmhi): Rename to...
14198 * config/aarch64/aarch64-simd.md
14199 (simd_mode): Add SF.
14200 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14201 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14202 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14203 (cstore<mode>_neg): ...This.
14204 * config/aarch64/iterators.md
14206 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14207 (COMPARISONS): New.
14208 (UCOMPARISONS): Likewise.
14209 (optab): Add missing comparisons.
14215 (VCMP_S): Likewise.
14216 (VCMP_U): Likewise.
14217 (V_cmp_result): Add DF, SF modes.
14218 (v_cmp_result): Likewise.
14220 (vmtype): Likewise.
14221 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14223 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14225 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14226 define_insn to define_insn_and_split.
14227 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14228 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14229 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14230 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14231 (thumb2_negscc): Likewise.
14233 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14235 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14237 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14239 * config/arm/thumb2.md: Remove trailing whitespaces.
14241 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14243 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14244 Use gen_int_mode rather than GEN_INT.
14246 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14248 * value-prof.c (stream_in_histogram_value): Remove the strayed
14251 2013-04-30 Richard Biener <rguenther@suse.de>
14253 PR middle-end/57122
14254 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14256 2013-04-30 Richard Biener <rguenther@suse.de>
14258 PR middle-end/57107
14259 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14261 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14263 PR rtl-optimization/56957
14264 PR rtl-optimization/57105
14265 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14266 variable. Use just INSN_UID for determining whether an insn
14267 should be only disconnected from the insn stream.
14268 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14270 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14272 PR tree-optimization/57104
14273 * tsan.c (instrument_expr): Don't instrument accesses to
14274 DECL_HARD_REGISTER VAR_DECLs.
14276 2013-04-30 Richard Biener <rguenther@suse.de>
14278 * function.h (loops_for_fn): New inline function.
14279 (set_loops_for_fn): Likewise.
14280 * cfgloop.h (place_new_loop): Add struct function parameter.
14281 (get_loop): Likewise.
14282 (get_loops): Likewise.
14283 (number_of_loops): Likewise.
14284 (fel_next): Adjust.
14285 (fel_init): Likewise.
14286 * cfg.c (get_loop_copy): Adjust.
14287 * cfgloop.c (flow_loops_dump): Likewise.
14288 (record_loop_exits): Likewise.
14289 (verify_loop_structure): Likewise.
14290 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14291 (estimate_reg_pressure_cost): Likewise.
14292 (mark_loop_exit_edges): Likewise.
14293 * cfgloopmanip.c (place_new_loop): Likewise.
14294 (add_loop): Likewise.
14295 (duplicate_loop): Likewise.
14296 * graph.c (draw_cfg_nodes): Likewise.
14297 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14298 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14299 (extract_affine_chrec): Likewise.
14300 (build_scop_iteration_domain): Likewise.
14301 * graphite.c (graphite_initialize): Likewise.
14302 * ira-build.c (create_loop_tree_nodes): Likewise.
14303 (more_one_region_p): Likewise.
14304 (rebuild_regno_allocno_maps): Likewise.
14305 (mark_loops_for_removal): Likewise.
14306 (mark_all_loops_for_removal): Likewise.
14307 (remove_unnecessary_regions): Likewise.
14308 (ira_build): Likewise.
14309 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14310 * loop-init.c (fix_loop_structure): Likewise.
14311 (gate_rtl_move_loop_invariants): Likewise.
14312 (gate_rtl_unswitch): Likewise.
14313 (gate_rtl_unroll_and_peel_loops): Likewise.
14314 (rtl_doloop): Likewise.
14315 * lto-streamer-in.c (input_cfg): Likewise.
14316 * lto-streamer-out.c (output_cfg): Likewise.
14317 * modulo-sched.c (sms_schedule): Likewise.
14318 * predict.c (tree_estimate_probability): Likewise.
14319 (tree_estimate_probability_driver): Likewise.
14320 (estimate_loops): Likewise.
14321 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14322 (move_sese_region_to_fn): Likewise.
14323 (debug_loop_num): Likewise.
14324 * tree-chrec.c (chrec_evaluate): Likewise.
14325 (hide_evolution_in_other_loops_than_loop): Likewise.
14326 (chrec_component_in_loop_num): Likewise.
14327 (reset_evolution_in_loop): Likewise.
14328 (evolution_function_is_invariant_rec_p): Likewise.
14329 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14330 * tree-inline.c (copy_loops): Likewise.
14331 (copy_cfg_body): Likewise.
14332 (tree_function_versioning): Likewise.
14333 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14334 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14336 (add_to_evolution_1): Likewise.
14337 (scev_const_prop): Likewise.
14338 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14339 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14340 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14341 (tree_ssa_lim_initialize): Likewise.
14342 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14343 (verify_loop_closed_ssa): Likewise.
14344 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14345 (tree_ssa_loop_im): Likewise.
14346 (tree_ssa_loop_unswitch): Likewise.
14347 (tree_vectorize): Likewise.
14348 (check_data_deps): Likewise.
14349 (tree_ssa_loop_ivcanon): Likewise.
14350 (tree_ssa_loop_bounds): Likewise.
14351 (tree_complete_unroll): Likewise.
14352 (tree_complete_unroll_inner): Likewise.
14353 (tree_parallelize_loops): Likewise.
14354 (tree_ssa_loop_prefetch): Likewise.
14355 (tree_ssa_loop_ivopts): Likewise.
14356 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14357 * tree-vectorizer.c (vectorize_loops): Likewise.
14359 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14361 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14362 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14363 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14364 with EABI_LINK_SPEC.
14366 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14369 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14372 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14375 * lra-constraints.c (process_alt_operands): Discourage a bit more
14376 using memory for pseudos. Print cost dump for alternatives.
14377 Modify cost values for conflicts with early clobbers.
14378 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14380 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14383 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14385 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14387 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14388 from/to S register.
14389 (movdi_aarch64): Support LDR/STR from/to D register.
14391 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14393 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14394 or higher by default.
14396 2013-04-29 Richard Biener <rguenther@suse.de>
14398 PR middle-end/57075
14399 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14400 even if not adding abnormal edges for calls that can make
14403 2013-04-29 Richard Biener <rguenther@suse.de>
14405 PR middle-end/57103
14406 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14408 (move_stmt_r): Remove redundant checking.
14410 2013-04-29 Teresa Johnson <tejohnson@google.com>
14413 * basic-block.h (apply_scale): New function.
14414 (apply_probability): Use apply_scale.
14415 * gimple-streamer-in.c (input_bb): Ditto.
14416 * lto-streamer-in.c (input_cfg): Ditto.
14417 * lto-cgraph.c (merge_profile_summaries): Ditto.
14418 * tree-optimize.c (execute_fixup_cfg): Ditto.
14419 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14420 (copy_edges_for_bb): Ditto.
14421 (copy_cfg_body): Ditto.
14423 2013-04-29 Tom de Vries <tom@codesourcery.com>
14425 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14426 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14427 (tail_merge_optimize): Handle current_loops == NULL.
14429 2013-04-26 Jeff Law <law@redhat.com>
14431 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14432 (simplify_cond_using_ranges): Generalize code to simplify
14433 COND_EXPRs where one argument is a constant and the other
14434 is an SSA_NAME created by an integral type conversion.
14436 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14438 * config/arm/arm.md (store_minmaxsi): Use only when
14439 optimize_insn_for_size_p.
14441 2013-04-29 Christian Bruel <christian.bruel@st.com>
14444 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14446 2013-04-29 Richard Biener <rguenther@suse.de>
14448 PR middle-end/57089
14449 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14450 loop tree make sure to schedule a fixup for the child as well.
14451 (expand_omp_for_generic): Properly add loops.
14452 (expand_omp_for_static_nochunk): Likewise.
14453 (expand_omp_for_static_chunk): Likewise.
14454 (expand_omp_for): For the degenerate case fixup loops.
14455 (expand_omp_sections): Fix default bb placement in loops.
14456 (expand_omp_atomic_pipeline): Properly add loops.
14458 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14460 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14462 2013-04-29 Tom de Vries <tom@codesourcery.com>
14464 * tree-ssa-tail-merge.c: Update header comment.
14466 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14468 * config/aarch64/arm_neon.h
14469 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14470 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14471 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14472 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14473 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14474 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14475 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14476 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14478 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14480 * config/aarch64/aarch64-simd.md
14481 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14482 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14483 fix_trunc, fixuns_trunc.
14484 (ftrunc<VDQF:mode>2): New.
14485 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14486 (fix_trunc_optab): New.
14488 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14490 * config/aarch64/aarch64-builtins.c
14491 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14492 iceilf, lround, iroundf.
14494 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14497 * config/i386/i386.h (enum ix86_tune_indices)
14498 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14499 New, split from X86_TUNE_INTER_UNIT_MOVES.
14500 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14501 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14502 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14503 (TARGET_INTER_UNIT_MOVES): Remove.
14504 * config/i386/i386.c (initial_ix86_tune_features): Update.
14505 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14506 (ix86_expand_convert_uns_didf_sse): Use
14507 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14508 (ix86_expand_vector_init_one_nonzero): Ditto.
14509 (ix86_expand_vector_init_interleave): Ditto.
14510 (inline_secondary_memory_needed): Return true for moves from SSE class
14511 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14512 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14513 * config/i386/constraints.md (Yi, Ym): Depend on
14514 TARGET_INTER_UNIT_MOVES_TO_VEC.
14515 (Yj, Yn): New constraints.
14516 * config/i386/i386.md (*movdi_internal): Change constraints of
14517 operand 1 from Yi to Yj and from Ym to Yn.
14518 (*movsi_internal): Ditto.
14519 (*movdf_internal): Ditto.
14520 (*movsf_internal): Ditto.
14521 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14522 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14523 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14524 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14525 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14526 * config/i386/sse.md (movdi_to_sse): Ditto.
14527 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14528 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14529 TARGET_INTER_UNIT_MOVES.
14530 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14531 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14532 instead of TARGET_INTER_UNIT_MOVES.
14533 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14534 operand 1 from Yi to Yj and from Ym to Yn.
14536 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14538 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14539 (float_truncate_hi_): Likewise.
14540 (float_extend_lo_): Likewise.
14541 (float_truncate_lo_): Likewise.
14542 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14543 (aarch64_float_extend_lo_v2df): Likewise.
14544 (vec_unpacks_hi_v4sf): Likewise.
14545 (aarch64_float_truncate_lo_v2sf): Likewise.
14546 (aarch64_float_truncate_hi_v4sf): Likewise.
14547 (vec_pack_trunc_v2df): Likewise.
14548 (vec_pack_trunc_df): Likewise.
14550 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14552 * config/aarch64/aarch64-builtins.c
14553 (aarch64_fold_builtin): Fold float conversions.
14554 * config/aarch64/aarch64-simd-builtins.def
14555 (floatv2si, floatv4si, floatv2di): New.
14556 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14557 * config/aarch64/aarch64-simd.md
14558 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14559 * config/aarch64/iterators.md (FLOATUORS): New.
14560 (optab): Add float, floatuns.
14561 (su_optab): Likewise.
14563 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14565 * config/aarch64/aarch64-builtins.c
14566 (aarch64_builtin_vectorized_function): Use new names for
14568 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14569 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14570 (fcvtzu): Split as...
14571 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14572 (fcvtas): Split as...
14573 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14574 (fcvtau): Split as...
14575 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14576 (fcvtps): Split as...
14577 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14578 (fcvtpu): Split as...
14579 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14580 (fcvtms): Split as...
14581 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14582 (fcvtmu): Split as...
14583 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14584 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14585 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14586 (lfrintnusf, lfrintnudf): Likewise.
14587 * config/aarch64/aarch64-simd.md
14588 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
14590 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
14591 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
14592 (fcvt_pattern): Likewise.
14594 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14596 * config/aarch64/aarch64-simd.md
14597 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
14598 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
14600 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14602 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
14603 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
14604 (vrnd<a,m,n,p>_f32): Implement using builtins.
14605 (vrnd<i,x><q>_f<32, 64>): New.
14607 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14609 * config/aarch64/aarch64-builtins.c
14610 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
14611 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
14612 (frintz): Rename to...
14614 (frintp): Rename to...
14616 (frintm): Rename to...
14618 (frinti): Rename to...
14619 (nearbyint): ...this.
14620 (frintx): Rename to...
14622 (frinta): Rename to...
14624 * config/aarch64/aarch64-simd.md
14625 (aarch64_frint<frint_suffix><mode>): Delete.
14626 (<frint_pattern><mode>2): Convert to insn.
14627 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
14628 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
14629 (frint_pattern): Likewise.
14630 (frint_suffix): Likewise.
14632 2013-04-29 Richard Biener <rguenther@suse.de>
14634 PR tree-optimization/57081
14635 * loop-init.c: Include tree-flow.h.
14636 (loop_optimizer_finalize): Free number of iteration estimates.
14637 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
14639 2013-04-29 Jakub Jelinek <jakub@redhat.com>
14641 PR tree-optimization/57083
14642 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
14643 non-singleton shift count range, zero extend low_bound for uns case.
14645 * config/i386/predicates.md (general_vector_operand): New predicate.
14646 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
14647 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
14648 if they aren't nonimmediate operands. If their original values
14649 satisfy const_vector_equal_evenodd_p, don't shift them.
14650 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
14651 predicates. For the SSE4.1 case force operands[{1,2}] into registers
14652 if not nonimmediate_operand.
14653 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
14654 instead of register_operand.
14655 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
14657 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
14659 * stor-layout.c (finalize_size_functions): Allocate a structure and
14660 reset cfun before dumping the functions.
14662 2013-04-27 Jakub Jelinek <jakub@redhat.com>
14664 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
14667 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
14668 use xop_pmacsdqh if uns_p.
14669 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
14670 the immediate rotate count.
14672 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
14674 * rtl.h (struct rtx_def): Add comment for field jump.
14675 (LRA_SUBREG_P): New macro.
14676 * recog.c (register_operand): Check LRA_SUBREG_P.
14677 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14679 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14680 (lra_final_code_change): Skip subreg change for operators.
14681 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14682 if there are no operand changes.
14683 * lra-constraints.c (curr_insn_set): New.
14684 (match_reload): Set LRA_SUBREG_P.
14685 (emit_spill_move): Ditto.
14686 (check_and_process_move): Use curr_insn_set. Process only single
14687 set insns. Don't initialize sec_mem_p and change_p.
14688 (simplify_operand_subreg): Use LRA_SUBREG_P.
14689 (reg_in_class_p): New function.
14690 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14691 of #ifdef. Add code to remove cycling.
14692 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14693 non-null disp. Reload inner instead of disp when base and index
14694 are null. Try to put lo_sum into register.
14695 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14696 (check_and_process_move): Move code for move cost check to
14697 simple_move_p. Remove equiv_substitution.
14698 (simple_move_p): New function.
14699 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14700 curr_insn_set. Call check_and_process_move only for single set
14701 insns. Use the new function. Move call of check_and_process_move
14702 after operand equiv substitution and address process.
14704 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14707 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
14708 with nonlocal goto receivers or returns twice calls, ignore
14709 unininitialized values from abnormal edges to nl goto receiver
14710 or returns twice call.
14712 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14714 PR tree-optimization/57051
14715 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
14716 and VEC_RSHIFT_EXPR if shift count is a multiple of element
14719 2013-04-26 Richard Biener <rguenther@suse.de>
14721 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
14722 (expand_omp_taskreg): Likewise. Mark loops for fixup.
14723 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
14724 (fixup_loop_arrays_after_move): New function.
14725 (move_sese_region_to_fn): Properly outline the loop tree parts
14726 of the SESE region.
14728 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
14730 * config/i386/i386.md (type, unit): Fix long lines.
14732 2013-04-26 Richard Biener <rguenther@suse.de>
14734 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
14735 (lto-streamer-out.o): Likewise.
14736 * cfgloop.c (init_loops_structure): Export, add struct function
14737 argument and adjust.
14738 (flow_loops_find): Adjust.
14739 * cfgloop.h (enum loop_estimation): Add EST_LAST.
14740 (init_loops_structure): Declare.
14741 * lto-streamer-in.c: Include cfgloop.h.
14742 (input_cfg): Input the loop tree.
14743 * lto-streamer-out.c: Include cfgloop.h.
14744 (output_cfg): Output the loop tree.
14745 (output_struct_function_base): Do not drop PROP_loops.
14747 2013-03-26 Richard Biener <rguenther@suse.de>
14749 * tree-cfg.c (execute_build_cfg): Build the loop tree.
14750 (pass_build_cfg): Provide PROP_loops.
14751 (move_sese_region_to_fn): Remove loops that are outlined into fn
14753 * tree-inline.c: Include cfgloop.h.
14754 (initialize_cfun): Do not drop PROP_loops.
14755 (copy_loops): New function.
14756 (copy_cfg_body): Copy loop structure.
14757 (tree_function_versioning): Initialize destination loop tree.
14758 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
14759 (pass_parallelize_loops): Do IL verification.
14760 * loop-init.c (loop_optimizer_init): Fixup loops if required.
14761 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
14762 the CFG make sure we fixup loops as well.
14763 * tree-ssa-tail-merge.c: Include cfgloop.h.
14764 (replace_block_by): When merging loop latches mark loops for fixup.
14765 * lto-streamer-out.c (output_struct_function_base): Drop
14766 PROP_loops for now.
14767 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
14768 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
14769 * ipa-split.c: Include cfgloop.h.
14770 (split_function): Add the new return block to the loop tree root.
14771 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
14772 whether we have removed the forwarder block.
14773 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
14774 * cfgloop.h (place_new_loop): Declare.
14775 * cfgloopmanip.c (place_new_loop): Export.
14776 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
14777 (tree-switch-conversion.o): Likewise.
14778 (tree-complex.o): Likewise.
14779 (tree-inline.o): Likewise.
14780 (tree-ssa-tailmerge.o): Likewise.
14781 (ipa-split.o): Likewise.
14782 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
14783 (tree-ssa-copy.o): Likewise.
14784 * tree-switch-conversion.c: Include cfgloop.h
14785 (process_switch): If we emit a bit-test cascade, schedule loops
14787 * tree-complex.c: Include cfgloop.h.
14788 (expand_complex_div_wide): Properly add new basic-blocks to loops.
14789 * asan.c: Include cfgloop.h.
14790 (create_cond_insert_point): Properly add new basic-blocks to
14791 loops, schedule loop fixup.
14792 * cfgloop.c (verify_loop_structure): Check that looks are not
14794 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
14796 (expand_omp_for_generic): Likewise.
14797 (expand_omp_sections): Likewise.
14798 (expand_omp_atomic_pipeline): Schedule loops for fixup.
14799 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
14800 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
14801 is initialized, not when loops are present.
14802 * tree-parloops.c (parallelize_loops): Remove checking here.
14803 * passes.c (init_optimization_passes): Schedule a copy-propagation
14804 pass before complete unrolling of inner loops.
14806 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14808 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
14809 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
14810 (colorize_init): Add argument to _WIN32 version.
14811 * toplev.c: Include diagnostic-color.h.
14812 (process_options): Default to -fdiagnostics-color=auto if
14813 GCC_COLORS env var is in the environment.
14814 * common.opt (fdiagnostics-color=): Add Var and Init.
14815 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
14816 env var is in the environment, the default is auto rather than never.
14818 * diagnostic.h (file_name_as_prefix): Add context argument.
14819 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
14820 the string as locus.
14821 * langhooks.c (lhd_print_error_function): Adjust caller.
14823 2013-04-25 Lawrence Crowl <crowl@google.com>
14825 * var-tracking.c (shared_hash_def::htab):
14826 Change type to hash_table. Update dependent calls and types.
14828 2013-04-25 Lawrence Crowl <crowl@google.com>
14830 * Makefile.in: Update as needed below.
14832 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
14833 Move declaration to after the type's method definitons.
14835 * attribs.c (htab_t scoped_attributes::attribute_hash):
14836 Change type to hash_table. Update dependent calls and types.
14838 * bitmap.c (htab_t bitmap_desc_hash):
14839 Change type to hash_table. Update dependent calls and types.
14841 * cselib.c (htab_t cselib_hash_table):
14842 Change type to hash_table. Update dependent calls and types.
14844 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
14845 (hash_string_slot_node): Move implementation into lto-streamer.h
14846 struct string_slot_hasher.
14847 (eq_string_slot_node): Likewise.
14849 * data-streamer-out.c: Update output_block::string_hash_table
14850 dependent calls and types.
14852 * dwarf2cfi.c (htab_t trace_index):
14853 Change type to hash_table. Update dependent calls and types.
14855 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
14856 Change type to hash_table. Update dependent calls and types.
14857 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
14858 (htab_t optimize_external_refs::map): Likewise.
14859 (htab_t output_comp_unit::extern_map): Likewise.
14860 (htab_t output_comdat_type_unit::extern_map): Likewise.
14861 (htab_t output_macinfo::macinfo_htab): Likewise.
14862 (htab_t optimize_location_lists::htab): Likewise.
14863 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
14865 * except.c (htab_t ehspec_hash_type):
14866 Change type to hash_table. Update dependent calls and types.
14867 (assign_filter_values::ttypes): Likewise.
14868 (assign_filter_values::ehspec): Likewise.
14869 (sjlj_assign_call_site_values::ar_hash): Likewise.
14870 (convert_to_eh_region_ranges::ar_hash): Likewise.
14872 * gcse.c (htab_t pre_ldst_table):
14873 Change type to hash_table. Update dependent calls and types.
14875 * ggc-common.c (htab_t saving_htab):
14876 Change type to hash_table. Update dependent calls and types.
14877 (htab_t loc_hash): Likewise.
14878 (htab_t ptr_hash): Likewise.
14879 (call_count): Rename ggc_call_count.
14880 (call_alloc): Rename ggc_call_alloc.
14881 (loc_descriptor): Rename make_loc_descriptor.
14882 (add_statistics): Rename ggc_add_statistics.
14884 * ggc-common.c (saving_htab):
14885 Change type to hash_table. Update dependent calls and types.
14887 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
14888 (push_gimplify_context): Likewise.
14889 (pop_gimplify_context): Likewise.
14890 (struct gimple_temp_hash_elt): Added.
14891 (struct gimplify_hasher): Likewise.
14892 (struct gimplify_ctx.temp_htab):
14893 Change type to hash_table. Update dependent calls and types.
14895 * gimple-fold.c: Include gimplify-ctx.h.
14897 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
14898 Change type to hash_table. Update dependent calls and types.
14899 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
14900 avoid potential global name collision.
14902 * gimplify.c: Include gimplify-ctx.h.
14903 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
14904 (htab_t gimplify_ctx::temp_htab):
14905 Update dependent calls and types for new type hash_table.
14906 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
14907 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
14909 * gimplify-ctx.h: New.
14910 (struct gimple_temp_hash_elt): Move from gimplify.c.
14911 (class gimplify_hasher): New.
14912 (struct gimplify_ctx): Move from gimple.h.
14913 (htab_t gimplify_ctx::temp_htab):
14914 Change type to hash_table. Update dependent calls and types.
14916 * graphite-clast-to-gimple.c: Include graphite-htab.h.
14917 (htab_t ivs_params::newivs_index):
14918 Change type to hash_table. Update dependent calls and types.
14919 (htab_t ivs_params::params_index): Likewise.
14920 (htab_t print_generated_program::params_index): Likewise.
14921 (htab_t gloog::newivs_index): Likewise.
14922 (htab_t gloog::params_index): Likewise.
14924 * graphite.c: Include graphite-htab.h.
14925 4htab_t graphite_transform_loops::bb_pbb_mapping):
14926 Change type to hash_table. Update dependent calls and types.
14928 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
14929 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
14930 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
14932 * graphite-dependences.c: Include graphite-htab.h.
14933 (loop_is_parallel_p): Change hash table type of parameter.
14935 * graphite-htab.h: New.
14936 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
14937 (extern find_pbb_via_hash): Move from graphite-poly.h.
14938 (extern loop_is_parallel_p): Move from graphite-poly.h.
14939 (extern get_loop_body_pbbs): Move from graphite-poly.h.
14941 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
14942 (extern loop_is_parallel_p): Move to graphite-htab.h.
14943 (extern get_loop_body_pbbs): Move to graphite-htab.h.
14945 * haifa-sched.c (htab_t delay_htab):
14946 Change type to hash_table. Update dependent calls and types.
14947 (htab_t delay_htab_i2): Likewise.
14949 * ira-color.c (htab_t allocno_hard_regs_htab):
14950 Change type to hash_table. Update dependent calls and types.
14952 * ira-costs.c (htab_t cost_classes_htab):
14953 Change type to hash_table. Update dependent calls and types.
14955 * loop-invariant.c (htab_t merge_identical_invariants::eq):
14956 Change type to hash_table. Update dependent calls and types.
14958 * loop-iv.c (htab_t bivs):
14959 Change type to hash_table. Update dependent calls and types.
14961 * loop-unroll.c (htab_t opt_info::insns_to_split):
14962 Change type to hash_table. Update dependent calls and types.
14963 (htab_t opt_info::insns_with_var_to_expand): Likewise.
14965 * lto-streamer.h (struct string_slot): Move from data-streamer.h
14966 (struct string_slot_hasher): New.
14967 (htab_t output_block::string_hash_table):
14968 Change type to hash_table. Update dependent calls and types.
14970 * lto-streamer-in.c (freeing_string_slot_hasher): New.
14971 (htab_t file_name_hash_table):
14972 Change type to hash_table. Update dependent calls and types.
14974 * lto-streamer-out.c: Update output_block::string_hash_table dependent
14977 * lto-streamer.c (htab_t tree_htab):
14978 Change type to hash_table. Update dependent calls and types.
14980 * omp-low.c: Include gimplify-ctx.h.
14982 * passes.c (htab_t name_to_pass_map):
14983 Change type to hash_table. Update dependent calls and types.
14984 (pass_traverse): Rename to passes_pass_traverse.
14986 * plugin.c (htab_t event_tab):
14987 Change type to hash_table. Update dependent calls and types.
14989 * postreload-gcse.c (htab_t expr_table):
14990 Change type to hash_table. Update dependent calls and types.
14991 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
14993 * sese.c (debug_rename_map_1): Make extern.
14994 (htab_t copy_bb_and_scalar_dependences::rename_map):
14995 Change type to hash_table. Update dependent calls and types.
14997 * sese.h (extern debug_rename_map): Move to .c file.
14999 * store-motion.c (htab_t store_motion_mems_table):
15000 Change type to hash_table. Update dependent calls and types.
15002 * trans-mem.c (htab_t tm_new_mem_hash):
15003 Change type to hash_table. Update dependent calls and types.
15005 * tree-browser.c (htab_t TB_up_ht):
15006 Change type to hash_table. Update dependent calls and types.
15008 * tree-cfg.c (htab_t discriminator_per_locus):
15009 Change type to hash_table. Update dependent calls and types.
15011 * tree-complex.c: Include tree-hasher.h
15012 (htab_t complex_variable_components):
15013 Change type to hash_table. Update dependent calls and types.
15015 * tree-eh.c (htab_t finally_tree):
15016 Change type to hash_table. Update dependent calls and types.
15018 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15019 struct int_tree_hasher.
15020 (extern int_tree_map_eq): Likewise.
15021 (uid_decl_map_hash): Removed.
15022 (extern decl_tree_map_eq): Likewise.
15024 * tree-hasher.h: New.
15025 (struct int_tree_hasher): New.
15026 (typedef int_tree_htab_type): New.
15028 * tree-inline.c: Include gimplify-ctx.h.
15030 * tree-mudflap.c: Include gimplify-ctx.h.
15032 * tree-parloops.c: Include tree-hasher.h.
15033 (htab_t eliminate_local_variables_stmt::decl_address):
15034 Change type to hash_table. Update dependent calls and types.
15035 (htab_t separate_decls_in_region::decl_copies): Likewise.
15037 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15038 Change type to hash_table. Update dependent calls and types.
15040 * tree-sra.c (candidates):
15041 Change type to hash_table. Update dependent calls and types.
15043 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15045 (int_tree_map_hash): Likewise.
15047 * tree-ssa-dom.c (htab_t avail_exprs):
15048 Change type to hash_table. Update dependent calls and types.
15050 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15051 Change type to hash_table. Update dependent calls and types.
15053 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15054 Change type to hash_table. Update dependent calls and types.
15056 * tree-ssa-phiopt.c (seen_ssa_names):
15057 Change type to hash_table. Update dependent calls and types.
15059 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15060 Change type to hash_table. Update dependent calls and types.
15062 * tree-ssa-uncprop.c (equiv):
15063 Change type to hash_table. Update dependent calls and types.
15065 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15067 PR rtl-optimization/57003
15068 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15069 call note_stores with kill_clobbered_value callback again after
15070 killing regs_invalidated_by_call.
15072 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15074 * config/aarch64/aarch64-simd.md
15075 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15076 (aarch64_simd_bsl<mode>): Likewise.
15077 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15079 2013-04-25 Marek Polacek <polacek@redhat.com>
15081 PR tree-optimization/57066
15082 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15084 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15086 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15088 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15090 * config/aarch64/aarch64-builtins.c
15091 (aarch64_fold_builtin): New.
15092 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15093 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15094 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15095 * config/aarch64/arm_neon.h
15096 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15098 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15099 Tejas Belagod <tejas.belagod@arm.com>
15101 * config/aarch64/aarch64-builtins.c
15102 (aarch64_gimple_fold_builtin): New.
15103 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15104 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15105 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15106 (addvv4sf): Update.
15107 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15109 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15111 * config/aarch64/aarch64.md
15112 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15114 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15116 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15117 (*ngcsi_uxtw): New pattern.
15119 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15120 Julian Brown <julian@codesourcery.com>
15122 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15123 (TB_DREG): Add T_V4HF.
15124 (v4hf_UP): New macro.
15125 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15126 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15127 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15129 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15130 (arm_vector_mode_supported_p): Handle V4HF.
15131 (arm_mangle_map): Handle V4HFmode.
15132 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15133 * config/arm/arm_neon_builtins.def: Add entries for
15134 vcvtv4hfv4sf, vcvtv4sfv4hf.
15135 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15136 (neon_vcvtv4hfv4sf): Likewise.
15137 * config/arm/neon-gen.ml: Handle half-precision floating point
15139 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15140 * config/arm/arm_neon.h: Regenerate.
15141 * config/arm/neon.ml (type elts): Add F16.
15142 (type vectype): Add T_float16x4, T_floatHF.
15143 (type vecmode): Add V4HF.
15144 (type features): Add Requires_FP_bit feature.
15145 (elt_width): Handle F16.
15146 (elt_class): Likewise.
15147 (elt_of_class_width): Likewise.
15148 (mode_of_elt): Refactor.
15149 (type_for_elt): Handle F16, fix error messages.
15150 (vectype_size): Handle T_float16x4.
15151 (vcvt_sh): New function.
15152 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15153 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15154 (string_of_mode): Handle V4HF.
15155 * doc/arm-neon-intrinsics.texi: Regenerate.
15157 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15159 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15160 format specifier in 'X' case.
15162 2013-04-25 Alan Modra <amodra@gmail.com>
15165 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15166 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15167 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15168 Repeat for many other rotate/shift and mask patterns using subregs.
15169 Name lshiftrt insns.
15170 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15171 on WORDS_BIG_ENDIAN.
15173 2013-04-25 Alan Modra <amodra@gmail.com>
15175 * config.gcc: Support little-endian powerpc-linux targets.
15176 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15177 (LINK_OS_LINUX_SPEC): Define.
15178 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15179 Preserve MASK_LITTLE_ENDIAN.
15180 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15181 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15182 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15183 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15184 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15185 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15186 Correct fp word order for little-endian. Don't shift toc entries
15187 smaller than a word for little-endian.
15188 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15189 (bswapdi2 splits): Correct low-part subreg for little-endian.
15190 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15191 low/high where such is correct only for be.
15192 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15193 little-endian for -mcall-aixdesc.
15195 2013-04-25 Alan Modra <amodra@gmail.com>
15197 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15198 replace_equiv_address_nv.
15200 2013-04-25 Alan Modra <amodra@gmail.com>
15202 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15204 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15207 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15208 * rtl.h (struct rtx_def): ...
15210 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15212 PR rtl-optimizations/57046
15213 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15214 for multi-reg splits.
15216 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15218 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15220 2013-04-24 Sterling Augustine <saugustine@google.com>
15222 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15223 (comp_dir_string, debug_str_dwo_section): New.
15224 (DEBUG_STR_DWO_SECTION): Rename to ...
15225 (DEBUG_DWO_STR_SECTION): ... this.
15226 (DEBUG_NORM_STR_SECTION): Delete.
15227 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15228 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15229 (find_AT_string): Move most logic to ...
15230 (find_AT_string_in_table): ... here. New.
15231 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15232 add_skeleton_AT_string. Delete logic.
15233 (output_skeleton_debug_sections): Remove call to
15234 add_top_level_skeleton_die_attrs.
15235 (add_comp_dir_attribute): Move logic to comp_dir_string.
15236 (dwarf2out_init): Initialize debug_str_dwo_section.
15237 (output_indirect_string): Call find_string_form.
15238 (output_indirect_strings): Rewrite.
15239 (prune_unused_types): Empty skeleton_debug_str_hash.
15240 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15241 (dwarf2out_finish): Call output_indirect_strings.
15243 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15245 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15247 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15249 * rtl.h (struct rtx_def): Add comment for field jump.
15250 (LRA_SUBREG_P): New macro.
15251 * recog.c (register_operand): Check LRA_SUBREG_P.
15252 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15254 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15255 (lra_final_code_change): Skip subreg change for operators.
15256 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15257 if there are no operand changes.
15258 * lra-constraints.c (curr_insn_set): New.
15259 (match_reload): Set LRA_SUBREG_P.
15260 (emit_spill_move): Ditto.
15261 (check_and_process_move): Use curr_insn_set. Process only single
15262 set insns. Don't initialize sec_mem_p and change_p.
15263 (simplify_operand_subreg): Use LRA_SUBREG_P.
15264 (reg_in_class_p): New function.
15265 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15266 of #ifdef. Add code to remove cycling.
15267 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15268 non-null disp. Reload inner instead of disp when base and index
15269 are null. Try to put lo_sum into register.
15270 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15271 (check_and_process_move): Move code for move cost check to
15272 simple_move_p. Remove equiv_substitution.
15273 (simple_move_p): New function.
15274 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15275 curr_insn_set. Call check_and_process_move only for single set
15276 insns. Use the new function. Move call of check_and_process_move
15277 after operand equiv substitution and address process.
15279 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15281 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15282 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15283 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15285 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15287 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15289 2013-04-24 Marek Polacek <polacek@redhat.com>
15291 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15292 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15293 (select_loops_exit_conditions): Likewise.
15294 (number_of_iterations_for_all_loops): Likewise.
15295 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15296 (scev_analysis): Likewise.
15298 2013-04-02 Catherine Moore <clm@codesourcery.com>
15299 Chao-ying Fu <fu@mips.com>
15301 * config/mips/micromips.md (jraddiusp): New pattern.
15302 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15303 instruction if possible.
15305 2013-04-24 Alan Modra <amodra@gmail.com>
15307 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15309 2013-04-24 Julian Brown <julian@codesourcery.com>
15310 Chung-Lin Tang <cltang@codesourcery.com>
15312 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15313 dependency behavior in enumeration type DIE generation. Add TODO note
15314 to comments about future DW_FORM_sdata/udata re-work of related code.
15316 2013-04-23 Lawrence Crowl <crowl@google.com>
15318 * Makefile.in: Update as needed below.
15320 * hash-table.h (class hash_table):
15321 Correct many methods with parameter types compare_type to the correct
15322 value_type. (Correct code was unlikely to notice the change.)
15323 (hash_table::elements_with_deleted) New.
15324 (class hashtable::iterator): New.
15325 (hashtable::begin()): New.
15326 (hashtable::end()): New.
15327 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15329 * statistics.c (statistics_hashes):
15330 Change type to hash_table. Update dependent calls and types.
15332 * tree-into-ssa.c (var_infos):
15333 Change type to hash_table. Update dependent calls and types.
15335 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15336 Change type to hash_table. Update dependent calls and types.
15338 * tree-ssa-loop-im.c (struct mem_ref.refs):
15339 Change type to hash_table. Update dependent calls and types.
15341 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15342 Change type to hash_table. Update dependent calls and types.
15344 * tree-ssa-sccvn.c (vn_tables_s::nary):
15345 Change type to hash_table. Update dependent calls and types.
15346 (vn_tables_s::phis): Likewise.
15347 (vn_tables_s::references): Likewise.
15349 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15350 (vn_reference_eq): Update parameter and return types.
15352 * tree-ssa-structalias.c (pointer_equiv_class_table):
15353 Change type to hash_table. Update dependent calls and types.
15354 (location_equiv_class_table): Likewise.
15356 * tree-vect-data-refs.c: Consequential changes for making
15357 peeling a hash_table.
15359 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15360 (destroy_loop_vec_info): Dependent hash_table update.
15362 * tree-vectorizer.h (peeling_htab):
15363 Change type to hash_table. Update dependent calls and types.
15365 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15367 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15368 to check the register content is equal or not.
15369 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15370 to assign register content record.
15371 * lra-eliminations.c (update_reg_eliminate): Use
15372 lra_update_reg_val_offset to update register content offset.
15373 * lra-int.h (struct lra_reg): Add offset member.
15374 (lra_reg_val_equal_p): New static inline function.
15375 (lra_update_reg_val_offset): New static inline function.
15376 (lra_assign_reg_val): New static inline function.
15377 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15378 to assign register content record.
15379 (initialize_lra_reg_info_element): Initial offset to zero.
15381 2013-04-23 Catherine Moore <clm@codesourcery.com>
15383 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15384 operands. Record compression.
15386 2013-04-23 Xinliang David Li <davidxl@google.com>
15388 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15390 2013-04-23 Richard Biener <rguenther@suse.de>
15392 PR middle-end/57036
15393 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15394 parameter, only add abnormal goto edges from the copied body
15395 if the call could perform abnormal gotos.
15396 (copy_cfg_body): Adjust.
15398 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15400 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15402 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15404 * coretypes.h (gimple_stmt_iterator): Add struct to make
15407 2013-04-23 Richard Biener <rguenther@suse.de>
15409 PR tree-optimization/57026
15410 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15411 from SSA names occuring in abnormal PHI nodes.
15413 2013-04-22 Andi Kleen <ak@linux.intel.com>
15415 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15417 2013-04-22 Andi Kleen <ak@linux.intel.com>
15419 * lto/lto.c (print_lto_report_1): Declare early.
15420 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15422 2013-04-22 Andi Kleen <ak@linux.intel.com>
15424 * common.opt (-flto-report-wpa): Add.
15425 * doc/invoke.texi (-flto-report-wpa): Add.
15426 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15429 2013-04-22 Xinliang David Li <davidxl@google.com>
15431 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15432 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15433 * Makefile.in: New dependency
15435 David Daney <ddaney.cavm@gmail.com>
15437 * configure.ac (gcc_cv_as_micromips_support): Use the
15438 --fatal-warnings option.
15439 * configure: Regenerate.
15441 2013-04-22 Marek Polacek <polacek@redhat.com>
15444 * tsan.c (instrument_expr): Don't instrument expression
15445 in case its size is zero.
15447 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15451 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15453 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15455 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15457 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15458 (gimple_stmt_iterator): New typedef.
15459 * gimple.h (gimple_stmt_iterator): Rename to...
15460 (gimple_stmt_iterator_d): ... This.
15461 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15462 trees be valid for GIMPLE and GENERIC.
15463 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15464 * gimple-fold.c (gimple_fold_call): Call target hook
15465 gimple_fold_builtin.
15466 * hooks.c (hook_bool_gsiptr_false): New.
15467 * hooks.h (hook_bool_gsiptr_false): New.
15468 * target.def (fold_stmt): New.
15469 * doc/tm.texi: Regenerate.
15471 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15474 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15475 a set sp if no stack realignment.
15477 2013-04-22 Nick Clifton <nickc@redhat.com>
15479 * config.gcc (tilegx-linux): Extend extra_objs rather than
15481 (tilepro-linux): Likewise.
15483 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15485 * config/aarch64/aarch64-builtins.c
15487 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15488 (VAR<1-12>): Add MAP parameter.
15489 (BUILTIN_*): Likewise.
15490 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15491 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15492 (aarch64_ushl_n<mode>): Likewise.
15493 (aarch64_sshr_n<mode>): Likewise.
15494 (aarch64_ushr_n<mode>): Likewise.
15495 (aarch64_<maxmin><mode>): Likewise.
15496 (aarch64_sqrt<mode>): Likewise.
15497 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15498 (vshr<q>_n_*): Likewise.
15500 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15502 * config/aarch64/aarch64-builtins.c
15503 (aarch64_simd_builtin_type_mode): Handle SF types.
15505 (BUILTIN_GPF): Define.
15506 (aarch64_init_simd_builtins): Handle SF types.
15507 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15508 (frecps): Likewise.
15509 (frecpx): Likewise.
15510 * config/aarch64/aarch64-simd.md
15511 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15512 (aarch64_frecpe<mode>): New.
15513 (aarch64_frecps<mode>): Likewise.
15514 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15515 (v8type): Add frecp<esx>.
15516 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15517 (aarch64_frecps<mode>): Likewise.
15518 * config/aarch64/iterators.md (FRECP): New.
15519 (frecp_suffix): Likewise.
15520 * config/aarch64/arm_neon.h
15521 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15523 2013-04-22 Christian Bruel <christian.bruel@st.com>
15526 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15527 (REG_CLASS_NAMES): Idem.
15528 (REG_CLASS_CONTENTS): Idem.
15529 (REGCLASS_HAS_FP_REG): Idem.
15530 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15531 (sh_conditional_register_usage): Idem.
15533 2013-04-21 Jeff Law <law@redhat.com>
15535 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15536 (ssa_forward_propagate_and_combine): Use it.
15538 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15540 * lra.c: Update the flow chart diagram.
15542 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15544 PR rtl-optimization/56847
15545 * lra-constraints.c (process_alt_operands): Discourage alternative
15546 with non-matche doffsettable memory constraint fro memory with
15549 2013-04-19 Richard Biener <rguenther@suse.de>
15551 PR tree-optimization/56982
15552 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15554 * gimplify.c (gimplify_call_expr): Notice special calls.
15555 (gimplify_modify_expr): Likewise.
15556 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15557 abnormal control flow receivers.
15558 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15559 in the same way as cfun->has_nonlocal_labels.
15560 (gimple_purge_dead_abnormal_call_edges): Likewise.
15561 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15562 receivers start a basic-block.
15564 2013-04-19 Richard Biener <rguenther@suse.de>
15566 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15568 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15569 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15570 (SLP_TREE_LOAD_PERMUTATION): Add.
15571 (vect_transform_slp_perm_load): Adjust prototype.
15572 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15573 (vect_free_slp_instance): Likewise.
15574 (vect_create_new_slp_node): Likewise.
15575 (vect_supported_slp_permutation_p): Remove.
15576 (vect_slp_rearrange_stmts): Adjust.
15577 (vect_supported_load_permutation_p): Likewise. Inline
15578 vect_supported_slp_permutation_p here.
15579 (vect_analyze_slp_instance): Compute load permutations per
15580 slp node instead of per instance.
15581 (vect_get_slp_defs): Adjust.
15582 (vect_transform_slp_perm_load): Likewise.
15583 (vect_schedule_slp_instance): Remove redundant code.
15584 (vect_schedule_slp): Remove hack for PR56270, add it ...
15585 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15586 CSE loads for SLP. Adjust.
15588 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15590 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
15591 spelling in two comments.
15593 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15596 * config/arm/arm.c (load_multiple_sequence): Require SP
15597 as base register for loads if SP is in the register list.
15599 2013-04-19 Martin Jambor <mjambor@suse.cz>
15601 PR tree-optimization/56718
15602 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
15603 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
15604 and made public. Adjusted all callers.
15605 (ipa_intraprocedural_devirtualization): New function.
15606 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
15607 (ipa_intraprocedural_devirtualization): Likewise.
15608 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
15610 2013-04-19 Richard Biener <rguenther@suse.de>
15612 PR tree-optimization/57000
15613 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
15615 2013-04-19 Terry Guo <terry.guo@arm.com>
15617 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
15619 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
15620 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
15621 (cortex_m4_fmacs): Use new reservations.
15622 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
15624 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
15626 PR rtl-optimization/56999
15627 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
15629 (lra_coalesce): Remove split_origin_bitmap and related code.
15630 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
15631 ranges if necessary.
15633 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
15635 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
15637 (ix86_expand_call): Remove clobbered_registers array and use
15638 x86_64_ms_sysv_extra_clobbered_registers instead.
15639 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
15641 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
15643 * config/i386/i386.md (*call_rex64_ms_sysv): Use
15644 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
15645 (*call_value_rex64_ms_sysv): Ditto.
15647 2013-04-18 Cary Coutant <ccoutant@google.com>
15649 * dwarf2out.c (output_pubnames): Check die_perennial_p of
15650 parent instead of die_mark.
15652 2013-04-18 Diego Novillo <dnovillo@google.com>
15654 * gimple.c (create_gimple_tmp): New.
15655 (get_expr_type): New.
15656 (build_assign): New.
15657 (build_type_cast): New.
15658 * gimple.h (enum ssa_mode): Define.
15659 (gimple_seq_set_location): New.
15660 * asan.c (build_check_stmt): Change some gimple_build_* calls
15661 to use build_assign and build_type_cast.
15663 2013-04-18 Richard Biener <rguenther@suse.de>
15665 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
15666 handle negative step. Remove redundant checks.
15667 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
15668 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
15669 for negative step and grouped loads fail to vectorize.
15671 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15673 * emit-rtl.c (reset_insn_used_flags): New function.
15674 (reset_all_used_flags): Use it.
15675 (verify_insn_sharing): New function.
15676 (verify_rtl_sharing): Fix verification for SEQUENCEs.
15678 2013-04-18 Jakub Jelinek <jakub@redhat.com>
15680 PR tree-optimization/56984
15681 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
15682 and (x >> M) >= N don't register any assertion if N << M is the
15685 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
15687 * lower-subreg.c (resolve_simple_move): If called self-recursive,
15688 do not delete_insn insns that have not yet been emitted, only
15689 unlink them with remove_insn.
15690 * df-scan.c (df_insn_delete): Revert r197492.
15692 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
15694 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
15695 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
15697 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15699 * config/arm/arm.md (movsicc_insn): Convert define_insn into
15700 define_insn_and_split.
15701 (and_scc,ior_scc,negscc): Likewise.
15702 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
15704 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
15706 * config/arm/arm.c (use_return_insn): Return 0 for targets that
15707 can benefit from using a sequence of LDRD instructions in epilogue
15708 instead of a single LDM instruction.
15710 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
15713 * doc/extend.texi: Fix typo.
15715 2013-04-17 Richard Biener <rguenther@suse.de>
15717 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
15718 (vect_build_slp_tree): ... here.
15719 (vect_build_slp_tree_1): Compute which stmts of the SLP group
15720 match. Remove special-casing of mismatched complex loads.
15721 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
15722 re-try the match with swapped commutative operands.
15723 (vect_supported_load_permutation_p): Remove special-casing of
15724 mismatched complex loads.
15725 (vect_analyze_slp_instance): Adjust.
15727 2013-04-17 Richard Biener <rguenther@suse.de>
15729 PR rtl-optimization/56921
15730 * cfgloop.h (struct loop): Add simple_loop_desc member.
15731 (struct niter_desc): Mark with GTY(()).
15732 (simple_loop_desc): Do not use aux field but simple_loop_desc.
15733 * loop-iv.c (get_simple_loop_desc): Likewise.
15734 (free_simple_loop_desc): Likewise.
15737 2013-04-16 Richard Biener <rguenther@suse.de>
15739 PR rtl-optimization/56921
15740 * loop-init.c (pass_rtl_move_loop_invariants): Add
15741 TODO_do_not_ggc_collect to todo_flags_finish.
15742 (pass_rtl_unswitch): Same.
15743 (pass_rtl_unroll_and_peel_loops): Same.
15744 (pass_rtl_doloop): Same.
15746 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
15748 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
15749 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
15750 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
15752 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
15753 * tree-streamer.c (record_common_node): Adjust reference in comment.
15755 2013-04-17 Terry Guo <terry.guo@arm.com>
15757 * config/arm/cortex-m4.md: Add a new bypass.
15759 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15761 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
15763 (*subs_<optab><mode>_multp2): New pattern.
15764 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15765 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
15767 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15769 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
15770 (*subs_mul_imm_<mode>): New pattern.
15772 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
15775 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
15776 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
15777 (vsx_movti_32bit): Change j->wa to O->wa.
15779 2013-04-16 Richard Biener <rguenther@suse.de>
15781 PR rtl-optimization/56921
15782 * loop-init.c (pass_rtl_move_loop_invariants): Add
15783 TODO_do_not_ggc_collect to todo_flags_finish.
15784 (pass_rtl_unswitch): Same.
15785 (pass_rtl_unroll_and_peel_loops): Same.
15786 (pass_rtl_doloop): Same.
15788 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
15790 * config/arm/arm.c (emit_multi_reg_push): New declaration
15791 for an existing function.
15792 (arm_emit_strd_push): New function.
15793 (arm_expand_prologue): Used here.
15794 (arm_emit_ldrd_pop): New function.
15795 (arm_expand_epilogue): Used here.
15796 (arm_get_frame_offsets): Update condition.
15797 (arm_emit_multi_reg_pop): Add a special case for load of a single
15798 register with writeback.
15800 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
15802 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
15805 2013-04-16 Richard Biener <rguenther@suse.de>
15807 PR tree-optimization/56756
15808 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
15809 (first_mem_ref_loc): New.
15810 (execute_sm): Place the load temporarily before a previous
15811 access instead of in the latch edge to ensure its SSA dependencies
15812 are defined at points dominating the load.
15814 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
15816 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
15817 correct fix by moving header and footer insn to the footer of
15818 the merged basic block. Clear BB_END of the merged-away block.
15820 PR middle-end/43631
15821 * emit-rtl.c (make_note_raw): New function.
15822 (link_insn_into_chain): New static inline function.
15823 (add_insn): Use it.
15824 (add_insn_before, add_insn_after): Factor insn chain linking code...
15825 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
15826 using link_insn_into_chain.
15827 (note_outside_basic_block_p): New helper function for emit_note_after
15828 and emit_note_before.
15829 (emit_note_after): Use nobb variant of add_insn_after if the note
15830 should not be contained in a basic block.
15831 (emit_note_before): Use nobb variant of add_insn_before if the note
15832 should not be contained in a basic block.
15833 (emit_note_copy): Use make_note_raw.
15834 (emit_note): Likewise.
15835 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
15836 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
15837 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
15838 the moved barrier the tail of the basic block it follows.
15839 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
15841 2013-04-15 Jakub Jelinek <jakub@redhat.com>
15843 PR tree-optimization/56962
15844 * gimple-ssa-strength-reduction.c (record_increment): Only set
15845 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
15846 either rhs1 or rhs2 is equal to c->base_expr.
15848 2013-04-15 Richard Biener <rguenther@suse.de>
15850 PR tree-optimization/56933
15851 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
15853 (GROUP_READ_WRITE_DEPENDENCE): Remove.
15854 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
15855 * tree-vect-data-refs.c (vect_analyze_group_access): Move
15856 dependence check ...
15857 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
15859 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
15860 GROUP_READ_WRITE_DEPENDENCE.
15862 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15864 * emit-rtl.c (reset_all_used_flags): New function.
15865 (verify_rtl_sharing): Call reset_all_used_flags before and after
15866 performing the checks.
15868 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15870 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
15871 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
15872 * config/arm/constraints.md (De): New constraint.
15873 * config/arm/neon.md (anddi3_neon): Delete.
15874 (neon_vand<mode>): Expand to standard anddi3 pattern.
15875 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
15876 Move earlier in the file.
15877 (neon_inv_logic_op2): Likewise.
15878 (arm_anddi_operand_neon): New predicate.
15880 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15882 * configure.ac (gcc_cv_ld_as_needed): Set
15883 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
15884 Use -z ignore, -z record on *-*-solaris2*.
15885 (HAVE_LD_AS_NEEDED): Update comment.
15886 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
15887 * configure: Regenerate.
15888 * config.in: Regenerate.
15889 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
15890 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
15891 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
15892 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
15893 equivalents. Fix markup.
15894 * doc/tm.texi: Regenerate.
15896 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
15898 * config/i386/i386.opt: New option mstack-protector-guard=.
15899 * config/i386/i386-opts.h: Add enum stack_protector_guard.
15900 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
15901 TARGET_SSP_TLS_GUARD.
15902 * config/i386/i386.c (ix86_option_override_internal): Set
15903 ix86_stack_protector_guard.
15904 * config/i386/i386.md (stack_protect_set): Enable for
15905 TARGET_SSP_TLS_GUARD only.
15906 (stack_protect_set_<mode>): Ditto.
15907 (stack_protect_test): Ditto.
15908 (stack_protect_test_<mode>): Ditto.
15909 * doc/invoke.texi (i386 Option): Document.
15911 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
15914 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
15915 (S_MODES): Set H_MODE bit.
15916 (SF_MODES): Set only S_MODE and SF_MODE bits.
15917 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
15918 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
15919 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
15920 <MODE_FLOAT>: Likewise.
15922 2013-04-15 Joey Ye <joey.ye@arm.com>
15924 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
15926 2013-04-15 Joey Ye <joey.ye@arm.com>
15928 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
15930 (thumb_far_jump_used_p): Count instruction size and set
15933 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
15935 * reorg.c (fill_simple_delay_slots): Reindent block of code.
15936 * resource.c (mark_target_live_regs): Reformat conditional block.
15938 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15940 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
15941 notes, they are emitted only just before final.
15942 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
15944 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15946 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
15947 * cfgrtl.c (delete_insn): Call it here instead.
15948 * lra-spills.c (lra_final_code_change): Use delete_insn.
15949 * haifa-sched.c (sched_remove_insn): Likewise.
15950 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
15951 returning to the nop pool.
15952 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
15953 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
15955 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15957 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
15958 * doc/tm.texi: Regenerated.
15960 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
15962 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
15965 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15967 * df-core.c (df_find_def): Compare register numbers.
15968 (df_find_use): Likewise.
15970 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
15973 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
15974 lra_in_progress for return.
15976 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15978 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
15979 define_insn into define_insn_and_split and emit movsicc patterns.
15981 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15983 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
15985 2013-04-12 Richard Biener <rguenther@suse.de>
15987 * tree-pass.h (TODO_do_not_ggc_collect): New.
15988 * passes.c (execute_one_ipa_transform_pass): Honor
15989 TODO_do_not_ggc_collect.
15990 (execute_one_pass): Likewise.
15993 2013-04-10 Richard Biener <rguenther@suse.de>
15995 * passes.c (init_optimization_passes): Remove reload pass.
15996 * ira.c (do_reload): Merge into ...
15998 (rest_of_handle_reload): Remove.
15999 (pass_reload): Likewise.
16000 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16001 of reload for vzeroupper pass placement.
16003 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16005 PR tree-optimization/56918
16006 PR tree-optimization/56920
16007 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16008 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16009 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16010 use wide_mul_with_sign method.
16012 2013-04-12 Richard Biener <rguenther@suse.de>
16014 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16015 not be considered a gimple constant.
16017 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16019 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16020 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16023 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16024 Jakub Jelinek <jakub@redhat.com>
16026 * opts.c: Include diagnostic-color.h.
16027 (common_handle_option): Handle OPT_fdiagnostics_color_.
16028 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16029 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16030 (diagnostic-color.o): New.
16031 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16032 (diagnostic_color_rule): New enum.
16033 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16034 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16035 the location string.
16036 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16037 either NULL, or color kind.
16038 * diagnostic-color.c: New file.
16039 * diagnostic-color.h: New file.
16040 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16042 * doc/invoke.texi (-fdiagnostics-color): Document.
16043 * pretty-print.h (pp_show_color): Define.
16044 (struct pretty_print_info): Add show_color field.
16045 * diagnostic.c: Include diagnostic-color.h.
16046 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16047 macros. Colorize error:, warning: etc. strings and also the location
16049 (diagnostic_show_locus): Colorize the caret line.
16050 * pretty-print.c: Include diagnostic-color.h.
16051 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16052 inside of %< %> quotes or quoted through q format modifier.
16054 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16056 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16058 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16060 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16061 code in CC_NZ mode.
16062 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16065 2013-04-11 Marek Polacek <polacek@redhat.com>
16067 PR tree-optimization/48184
16068 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16070 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16072 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16073 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16074 (skip_simple_arithmetic): Tidy up.
16075 * tree.h (skip_simple_constant_arithmetic): Declare.
16077 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16079 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16081 2013-04-11 Richard Biener <rguenther@suse.de>
16083 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16084 generate vector constants.
16086 2013-04-11 Richard Biener <rguenther@suse.de>
16088 PR tree-optimization/56878
16089 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16090 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16092 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16093 Prefer to align the DR with the most invariant base address.
16095 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16097 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16100 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16102 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16103 floating-point vector comparisons against 0.
16105 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16107 PR tree-optimization/56899
16108 * fold-const.c (extract_muldiv_1): Apply distributive law
16109 only if TYPE_OVERFLOW_WRAPS (ctype).
16111 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16114 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16115 loading has side effect.
16117 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16119 * config/sparc/sparc.c: Include tree-pass.h.
16120 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16121 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16122 head of file. Change return type. Split off gate function.
16123 (sparc_gate_work_around_errata): New function.
16124 (pass_work_around_errata): New pass definition.
16125 (insert_pass_work_around_errata) New pass insert definition to
16126 insert pass_work_around_errata just after delayed-branch scheduling.
16127 (sparc_option_override): Insert the pass.
16128 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16130 2013-04-10 David S. Miller <davem@davemloft.net>
16132 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16133 or -mcpu=hypersparc.
16135 * target.def (cstore_mode): New hook.
16136 * target.h: Include insn-codes.h
16137 * targhooks.c: Likewise.
16138 (default_cstore_mode): New function.
16139 * targhooks.h: Declare it.
16140 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16141 * doc/tm.texi: Rebuild.
16142 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16143 target hook, rather than inspecting the insn_data.
16144 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16145 (TARGET_CSTORE_MODE): Redefine.
16146 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16148 * config/sparc/predicates.md (cstore_result_operand): New special
16150 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16151 Use it for operand 0.
16152 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16153 (*snesi_special): Likewise.
16154 (*snesi_zero): Likewise.
16155 (*seqsi_zero): Likewise.
16156 (*sltu_insn): Likewise.
16157 (*sgeu_insn): Likewise.
16158 (*seqdi_special): Make operand 0 and comparison operation be of
16160 (*snedi_special): Likewise.
16161 (*snedi_special_vis3): Likewise.
16162 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16163 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16164 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16165 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16166 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16167 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16168 (*sltu_extend_sp64): Likewise.
16169 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16170 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16171 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16172 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16173 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16175 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16177 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16178 (aarch64_start_file): Use the new function.
16180 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16181 Jason Merrill <jason@redhat.com>
16183 * common.opt: Add -gdwarf.
16184 * opts.c (common_handle_option): Handle it.
16185 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16187 2013-04-10 Richard Biener <rguenther@suse.de>
16189 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16190 (execute_one_ipa_transform_pass): But unconditionally here.
16191 (execute_one_pass): And here.
16192 (init_optimization_passes): Remove reload pass.
16193 * tree-pass.h (TODO_ggc_collect): Remove.
16194 (pass_reload): Likewise.
16195 * ira.c (do_reload): Merge into ...
16197 (rest_of_handle_reload): Remove.
16198 (pass_reload): Likewise.
16199 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16200 of reload for vzeroupper pass placement.
16201 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16202 and todo_flags_finish of all passes.
16204 2013-04-10 Richard Biener <rguenther@suse.de>
16206 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16207 first_const_oprnd field, rename first_def_type to first_op_type.
16208 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16209 (vect_get_and_check_slp_defs): Always use the type of the
16210 operand. Allow mixed vect_external_def, vect_constant_def types.
16211 (vect_get_constant_vectors): Handle mixed vect_external_def,
16212 vect_constant_def types.
16214 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16216 PR tree-optimization/55524
16217 * tree-ssa-math-opts.c
16218 (convert_mult_to_fma): Don't use an fms construct
16219 when we don't have an fms operation, but fnma, and it looks
16220 likely that we'll be able to use the latter.
16222 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16224 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16226 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16227 inline fail caused by overwritable functions.
16229 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16231 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16232 unnecessary bits in the constant power of two case.
16234 2013-04-10 Richard Biener <rguenther@suse.de>
16236 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16237 broken code swapping operands.
16238 (vect_build_slp_tree): Do not compute load permutations here.
16239 (vect_analyze_slp_instance): Compute load permutations here,
16240 after building the SLP tree.
16242 2013-04-09 Christian Bruel <christian.bruel@st.com>
16244 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16245 of next/prev_real_insn.
16247 2013-04-09 Jan Hubicka <jh@suse.cz>
16249 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16250 Drop aliased parameter.
16251 (function_and_variable_visibility): Do not handle alias pairs.
16252 * cgraph.c (varpool_externally_visible_p): Update prototype.
16253 * varpool.c (varpool_add_new_variable): Update.
16255 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16257 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16259 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16261 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16263 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16265 2013-04-09 Marek Polacek <polacek@redhat.com>
16267 PR tree-optimization/48762
16268 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16270 2013-04-09 Richard Biener <rguenther@suse.de>
16272 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16274 (vect_build_slp_tree): Likewise.
16275 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16276 calculating the cost of a SLP instance.
16277 (vect_analyze_slp_instance): Use it from here, after building
16280 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16282 PR middle-end/56883
16283 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16284 expand_omp_for_static_chunk): Use simple_p = true in
16285 force_gimple_operand_gsi calls when assigning to addressable decls.
16287 2013-04-09 Jeff Law <law@redhat.com>
16289 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16290 when the boolean was created by converting a wider object which
16291 had a boolean range.
16293 2013-04-09 Richard Biener <rguenther@suse.de>
16295 * tree-vectorizer.h (slp_void_p): Remove.
16296 (slp_tree): Typedef before _slp_tree declaration.
16297 (struct _slp_tree): Use a vector of slp_tree as children.
16298 (vect_get_place_in_interleaving_chain): Remove.
16299 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16301 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16303 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16304 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16305 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16306 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16307 Use slp_node instead of slp_void_p and adjust.
16309 2013-04-09 Richard Biener <rguenther@suse.de>
16311 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16312 work that is not necessary.
16314 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16316 PR tree-optimization/56854
16317 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16318 forward into clobber stmts if it would change MEM_REF lhs into
16321 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16323 * tree.c (type_hash_lookup, type_hash_add): Make static.
16324 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16326 2013-04-09 Richard Biener <rguenther@suse.de>
16328 * tree.h (unsave_expr_now): Remove.
16329 * tree-inline.c (mark_local_for_remap_r): Remove.
16330 (unsave_expr_1): Likewise.
16331 (unsave_r): Likewise.
16332 (unsave_expr_now): Likewise.
16333 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16334 (propagate_tree_value): Likewise.
16336 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16338 * doc/rtl.texi (sequence): Rewrite documentation to match the
16339 current use of SEQUENCE rtl objects.
16340 * rtl.def (SEQUENCE): Likewise.
16342 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16343 Update documentation.
16344 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16345 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16347 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16349 2013-04-08 Teresa Johnson <tejohnson@google.com>
16351 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16352 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16354 (estimate_edge_size_and_time): Add comment to suggest using rounding
16356 (estimate_node_size_and_time): Ditto.
16357 (remap_edge_change_prob): Use helper rounding divide methods.
16358 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16359 (gimple_mod_pow2_value_transform): Ditto.
16360 (gimple_mod_subtract_transform): Ditto.
16361 (gimple_ic_transform): Ditto.
16362 (gimple_stringops_transform): Ditto.
16363 * stmt.c (conditional_probability): Ditto.
16364 (emit_case_dispatch_table): Ditto.
16365 * lto-cgraph.c (merge_profile_summaries): Ditto.
16366 * tree-optimize.c (execute_fixup_cfg): Ditto.
16367 * cfgcleanup.c (try_forward_edges): Ditto.
16368 * cfgloopmanip.c (scale_loop_profile): Ditto.
16370 (duplicate_loop_to_header_edge): Ditto.
16371 (lv_adjust_loop_entry_edge): Ditto.
16372 * tree-vect-loop.c (vect_transform_loop): Ditto.
16373 * profile.c (compute_branch_probabilities): Ditto.
16374 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16375 * lto-streamer-in.c (input_cfg): Ditto.
16376 * gimple-streamer-in.c (input_bb): Ditto.
16377 * ipa-cp.c (update_profiling_info): Ditto.
16378 (update_specialized_profile): Ditto.
16379 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16380 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16382 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16383 (compute_trg_info): Ditto.
16384 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16385 (purge_dead_edges): Ditto.
16386 * loop-unswitch.c (unswitch_loop): Ditto.
16387 * cgraphclones.c (cgraph_clone_edge): Ditto.
16388 (cgraph_clone_node): Ditto.
16389 * tree-inline.c (copy_bb): Ditto.
16390 (copy_edges_for_bb): Ditto.
16391 (initialize_cfun): Ditto.
16392 (copy_cfg_body): Ditto.
16393 (expand_call_inline): Ditto.
16395 2013-04-08 Kai Tietz <ktietz@redhat.com>
16397 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16398 TARGET_CYGWIN64 by TARGET_64BIT.
16400 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16402 * config/epiphany/epiphany.md (GPR_1): New constant.
16403 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16404 * config/epiphany/epiphany.c (gen_compare_reg):
16405 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16406 is already in place.
16407 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16408 Don't require being called during rtl expansion; If y operlaps r0,
16410 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16411 (epiphany_expand_epilogue): Likewise.
16413 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16414 Don't use CC_FPmode for ORDERED / UNORDERED.
16415 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16417 * config/epiphany/constraints.md (CnL): New constraint.
16418 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16419 * config/epiphany/predicates.md (add_operand): Allow 1024.
16421 * config/epiphany/epiphany.md (logical_op): New code iterator.
16422 (op_mnc): New code attribute.
16423 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16424 (mov_f+1, mov_f+2): New peephole2 patterns.
16426 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16427 (cstoresi4): Also allow re-use of zero result when doing a NE
16428 comparison to a non-zero operand.
16429 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16431 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16432 Use gen_addsi3_i / gen_subsi3_i.
16434 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16438 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16439 contain anything but clobbers, at most one __builtin_stack_restore,
16440 optionally debug stmts and final resx, and if it has at least one
16441 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16442 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16443 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16444 which isn't defaut definition, remove them.
16445 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16446 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16447 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16448 with MEM_REF LHS with SSA_NAME address.
16450 2013-04-08 Jeff Law <law@redhat.com>
16452 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16454 2013-04-08 Richard Biener <rguenther@suse.de>
16456 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16458 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16459 determined vector type.
16460 (vect_analyze_data_refs): Likewise.
16461 (vect_get_new_vect_var): Adjust.
16462 (vect_create_destination_var): Preserve SSA name versions.
16463 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16464 not dump anything here.
16466 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16468 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16469 Add member lr_slot_known.
16470 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16472 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16473 Remove code that sets lr_slot_offset according to what a previous
16474 version of epiphany_emit_save_restore used to do.
16475 (epiphany_emit_save_restore): When doing an lr save or restore,
16476 set/verify lr_slot_known and lr_slot_offset.
16478 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16481 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16484 2013-04-08 Richard Biener <rguenther@suse.de>
16486 * alias.c (find_base_term): Fix thinko in previous change.
16488 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16490 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16491 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16492 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16493 if possible to compute val.
16494 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16495 For QImode integers don't require anything about precision. Use
16496 const_with_all_bytes_same to find out if the constant doesn't have
16497 repeated bytes in it.
16499 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16501 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16504 2013-04-08 Marek Polacek <polacek@redhat.com>
16506 PR rtl-optimization/48182
16507 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16510 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16513 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16514 nuses, make sure we have a label.
16516 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16519 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16520 (rs6000_emit_swdiv_low_precision): Remove.
16521 (rs6000_emit_swdiv): Rewrite to handle between one and four
16522 iterations of Newton-Raphson generally; modify required number of
16523 iterations for some cases.
16524 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16526 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16528 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16529 set-but-unused variable.
16531 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16532 basic blocks of released function bodies garbage-collectable.
16534 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16535 (struct rtl_opt_pass): Add TODO_df_finish.
16537 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16539 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16541 * config/arm/constraints.md (q): New constraint.
16542 * config/arm/ldrdstrd.md: New file.
16543 * config/arm/arm.md (ldrdstrd.md) New include.
16544 (arm_movdi): Use "q" instead of "r" constraint
16545 for double-word memory access.
16546 (movdf_soft_insn): Likewise.
16547 * config/arm/vfp.md (movdi_vfp): Likewise.
16548 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16549 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16550 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16551 (mem_ok_for_ldrd_strd): Likewise.
16552 (output_move_double): Update assertion.
16554 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16556 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16558 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16560 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16561 define_insn_and_split.
16562 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16564 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16566 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16567 define_insn_and_split.
16568 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16569 (shiftsi3_compare): New pattern.
16570 (rrx): New pattern.
16571 * config/arm/unspecs.md (UNSPEC_RRX): New.
16573 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16575 * config/arm/arm.md (negdi_extendsidi): New pattern.
16576 (negdi_zero_extendsidi): Likewise.
16578 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16580 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16581 define_insn_and_split.
16582 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16583 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16585 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16587 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16588 define_insn_and_split.
16589 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
16590 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
16592 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16594 * config/arm/arm.md (subsi3_carryin): New pattern.
16595 (subsi3_carryin_const): Likewise.
16596 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
16597 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
16599 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16601 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
16603 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16605 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
16606 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
16608 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16610 * config/arm/arm.c (arm_expand_builtin): Change fcode
16611 type to unsigned int.
16613 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16615 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
16617 2013-04-04 Ian Lance Taylor <iant@google.com>
16619 * doc/standards.texi (Standards): The Go frontend supports the Go 1
16622 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
16624 PR middle-end/56729
16625 * df-scan.c (df_insn_delete): Disable failing assert.
16627 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16629 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
16630 New function prototype.
16631 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
16632 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
16633 (arm_builtin_vectorized_function): New function.
16635 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16637 * config/arm/arm_neon_builtins.def: New file.
16638 * config/arm/arm.c (neon_builtin_data): Move contents to
16639 arm_neon_builtins.def.
16640 (enum arm_builtins): Include neon builtin definitions.
16641 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
16642 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
16644 2013-04-04 Marek Polacek <polacek@redhat.com>
16646 PR tree-optimization/48186
16647 * predict.c (maybe_hot_frequency_p): Return false if
16648 HOT_BB_FREQUENCY_FRACTION is 0.
16649 (cgraph_maybe_hot_edge_p): Likewise.
16651 2013-04-04 Richard Biener <rguenther@suse.de>
16653 PR tree-optimization/56826
16654 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
16657 2013-04-04 Richard Biener <rguenther@suse.de>
16659 PR tree-optimization/56213
16660 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
16661 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
16663 2013-04-04 Richard Biener <rguenther@suse.de>
16665 PR tree-optimization/56837
16666 * tree-loop-distribution.c (classify_partition): For non-zero
16667 values require that the value has the same precision as its
16668 mode to be useful as memset value.
16670 2013-04-03 Nick Clifton <nickc@redhat.com>
16672 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
16673 (fmssf4): Use fmsf.s on E3V5 architectures.
16674 (fnmasf4): Use fnmaf.s on E3V5 architectures.
16675 (fnmssf4): Use fnmsf.s on E3V5 architectures.
16677 2013-04-03 Jeff Law <law@redhat.com>
16679 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
16680 (lra-eliminations.o): Likewise.
16682 2013-04-03 Teresa Johnson <tejohnson@google.com>
16684 * gcov-io.c (compute_working_sets): Moved most of body of old
16685 compute_working_sets here from profile.c.
16686 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
16687 (gcov_working_set_t): Moved typedef here from basic-block.h
16688 (compute_working_set): Declare.
16689 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
16690 (get_working_sets): Renamed from compute_working_set,
16691 replace most of body with call to new compute_working_sets.
16692 (get_exec_counts): Replace call to compute_working_sets
16693 to get_working_sets.
16694 * profile.h (get_working_sets): Renamed from compute_working_set.
16695 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
16696 to get_working_sets.
16697 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
16698 * gcov-dump.c (dump_working_sets): New function.
16700 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
16702 * hwint.c (sext_hwi, zext_hwi): New functions.
16703 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
16704 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
16705 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
16706 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
16707 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
16708 (sext_hwi, zext_hwi): New functions.
16710 2013-04-03 Jeff Law <law@redhat.com>
16712 PR tree-optimization/56799
16713 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
16714 back test for widening conversion erroneously dropped in prior change.
16716 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16719 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
16720 instead of next_real_insn.
16722 2013-04-03 Marek Polacek <polacek@redhat.com>
16725 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
16727 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16730 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
16732 (thumb1_output_casesi): Likewise.
16733 (thumb2_output_casesi): Likewise.
16735 2013-04-03 Richard Biener <rguenther@suse.de>
16737 PR tree-optimization/56817
16738 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
16740 (tree_unroll_loops_completely_1): ... new function to manually
16741 walk the loop tree, properly defering outer loops of unrolled
16742 loops to later iterations.
16744 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16746 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
16747 (vectorizable_load): Likewise.
16748 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
16749 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
16751 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16753 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
16756 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16758 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
16760 2013-04-03 Bin Cheng <bin.cheng@arm.com>
16762 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
16764 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16766 PR tree-optimization/56790
16767 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
16770 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
16772 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
16774 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
16775 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
16778 2013-04-03 Jakub Jelinek <jakub@redhat.com>
16781 * tree.h (force_folding_builtin_constant_p): New decl.
16782 * builtins.c (force_folding_builtin_constant_p): New variable.
16783 (fold_builtin_constant_p): Fold immediately also if
16784 force_folding_builtin_constant_p.
16786 2013-04-03 Richard Biener <rguenther@suse.de>
16788 PR tree-optimization/56812
16789 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
16790 DRs of the same interleaving chain are independent.
16792 2013-04-02 Jason Merrill <jason@redhat.com>
16794 * gdbinit.in (pbb): Use debug fn.
16796 2013-04-02 Lawrence Crowl <crowl@google.com>
16798 * sese.h (struct ivtype_map_elt_s): Remove unused.
16799 (extern debug_ivtype_map): Remove unused.
16800 (extern eq_ivtype_map_elts): Remove unused.
16801 * sese.c (debug_ivtype_map): Removed unused.
16802 (debug_ivtype_map_1): Removed unused.
16803 (debug_ivtype_elt): Remove unused.
16804 (eq_ivtype_map_elts): Remove unused.
16807 2013-04-02 Kai Tietz <ktietz@redhat.com>
16810 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
16811 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
16812 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
16814 (legitimize_pe_coff_symbol): Likewise.
16815 (is_imported_p): New helper-function.
16816 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
16817 for Windows x64 targets.
16818 (ix86_expand_prologue): Optimize for pe-coff targets.
16819 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
16820 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
16821 medium/large code-model.
16822 (legitimize_pic_address): Likewise.
16823 (legitimize_tls_address): Likewise.
16824 (ix86_expand_call): Likewise.
16825 (x86_output_mi_thunk): Likewise.
16826 (get_dllimport_decl): Add new beimport argument.
16827 (construct_plt_address): Don't assert for x64 pe-coff targets.
16828 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
16830 (SYMBOL_FLAG_STUBVAR): New macro.
16831 (SYMBOL_REF_STUBVAR_P): Likewise.
16832 * config/i386/winnt.c (stub_list): New structure.
16833 (stub_head): New local variable.
16834 (i386_pe_record_stub): New function.
16835 (i386_pe_file_end): Emit refptr-stubs.
16837 2013-04-02 Jakub Jelinek <jakub@redhat.com>
16839 PR rtl-optimization/56745
16840 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
16841 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
16844 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
16845 and both of them are MEM_REFs, just compare first argument for
16846 equality and attempt to deal even with differing offsets.
16849 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
16850 of gimple_clobber_p to be MEM_REF.
16851 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
16852 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
16853 after gimplification.
16854 * asan.c (get_mem_ref_of_assignment): Don't instrument
16855 gimple_clobber_p stmts.
16856 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
16857 gimple_clobber_p stmt if they have MEM_REF lhs and
16858 are dead because of another gimple_clobber_p stmt.
16859 * tree-ssa-live.c (clear_unused_block_pointer): Treat
16860 gimple_clobber_p stmts like debug stmts.
16861 (remove_unused_locals): Remove clobbers with MEM_REF lhs
16862 that refer to unused VAR_DECLs or uninitialized values.
16863 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
16864 gimple_clobber_p stmts if they refer to removed parameters.
16865 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
16868 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
16870 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
16871 using SWI48 mode attribute.
16873 2013-04-02 Wei Mi <wmi@google.com>
16875 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
16876 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
16877 *<rotate_insn><mode>3_mask in i386.md.
16879 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
16881 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
16883 2013-04-02 Richard Biener <rguenther@suse.de>
16885 PR tree-optimization/56778
16886 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
16887 Runtime alias tests are not supported for gather loads.
16888 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
16889 stmts referenced from SSA operands before updating SSA form.
16891 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
16892 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16894 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
16895 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
16896 * config/arm/cortex-a53.md: New file.
16897 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
16898 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
16899 * config/arm/arm.c (arm_issue_rate): Likewise.
16900 * config/arm/arm-tune.md: Regenerate
16901 * config/arm/arm-tables.opt: Regenerate.
16902 * config/arm/arm-cores.def: Add cortex-a53.
16904 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
16906 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
16909 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16911 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
16912 scalar load/store operations using B/H registers.
16913 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
16915 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16917 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
16919 * config/aarch64/aarch64.c
16920 (aarch64_simd_scalar_immediate_valid_for_move): New.
16921 * config/aarch64/aarch64-protos.h
16922 (aarch64_simd_scalar_immediate_valid_for_move): New.
16923 * config/aarch64/constraints.md (Dh, Dq): New.
16924 * config/aarch64/iterators.md (hq): New.
16926 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
16928 * reorg.c (get_branch_condition): Deal with conditional returns.
16929 (fill_simple_delay_slots): Remove dead code dealing with jumps.
16931 2013-04-01 Wei Mi <wmi@google.com>
16933 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
16934 Truncate operand 2 using %b asm operand modifier.
16935 (*<shift_insn><mode>3_mask): Ditto.
16936 (*<rotate_insn><mode>3_mask): Ditto.
16938 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
16940 PR middle-end/56798
16941 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
16943 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
16945 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
16947 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
16949 2013-03-30 Lawrence Crowl <crowl@google.com>
16951 * dse.c (clear_alias_sets): Remove never set.
16952 (disqualified_clear_alias_sets): Remove never set.
16953 (clear_alias_mode_pool): Remove never set.
16954 (dse_step0): Remove condition that is never true.
16955 (canon_address): Remove condition that is never true.
16956 (dse_step7): Remove condition that is never true.
16957 (rest_of_handle_dse): Remove condition that is never true.
16958 (rest_of_handle_dse::did_global): Remove never read from above.
16959 (dse_step2_spill): Remove never called from above.
16960 (dse_step5_spill): Remove never called from above.
16962 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
16964 * doc/md.texi (Standard Names) <casesi>: Update documentation for
16965 JUMP_TABLE_DATA changes.
16966 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
16967 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
16968 (Insns) <jump_table_data>: New entry.
16969 * doc/tm.texi: Regenerate.
16971 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
16973 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
16974 for table jump at the end of a basic block using tablejump_p.
16975 * targhooks.c (default_invalid_within_doloop): Likewise.
16976 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
16977 target hook implementation that is identical to the default hook.
16978 (rs6000_invalid_within_doloop): Remove.
16980 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
16981 unused variable from tablejump_p call.
16983 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
16984 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
16985 (INSN_DELETED_P): Likewise.
16986 (emit_jump_table_data): New prototype.
16987 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
16988 after 4th as unused.
16989 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
16990 * sched-vis.c (print_insn): Likewise.
16991 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
16992 insn for compatibility with back ends that use next_active_insn to
16993 identify jump table data.
16994 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
16995 (remove_insn): Likewise.
16996 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
16998 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
16999 (emit_jump_table_data): New function.
17001 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17002 basic block, a JUMP_TABLE_DATA never is.
17003 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17004 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17005 off from code handling real insns.
17006 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17007 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17008 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17009 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17010 is not a NONDEBUG_INSN_P.
17011 * ira-costs.c (scan_one_insn): Likewise.
17012 * jump.c (mark_all_labels): Likewise.
17013 (mark_jump_label_1): Likewise.
17014 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17015 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17017 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17018 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17019 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17020 JUMP_TABLE_DATA_P insns.
17021 (calculate_elim_costs_all_insns): Likewise.
17022 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17023 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17024 (delete_output_reload): Code style fixups.
17025 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17026 insn flags on this non-insn.
17027 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17028 as scheduling barriers, for pre-change compatibility.
17029 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17030 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17032 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17033 redundant JUMP_TABLE_DATA_P test.
17034 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17035 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17036 (frv_for_each_packet): Likewise.
17037 * config/i386/i386.c (min_insn_size): Likewise.
17038 (ix86_avoid_jump_mispredicts): Likewise.
17039 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17040 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17041 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17042 (mips16_insn_length): Robustify.
17043 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17044 (mips16_split_long_branches): Likewise.
17045 * config/pa/pa.c (pa_combine_instructions): Likewise.
17046 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17047 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17048 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17049 as contributing to pool range lengths.
17050 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17051 Remove redundant JUMP_TABLE_DATA_P test.
17052 (sh_loop_align): Likewise.
17053 (split_branches): Likewise.
17054 (sh_insn_length_adjustment): Likewise.
17055 * config/spu/spu.c (get_branch_target): Likewise.
17057 2013-03-29 Jan Hubicka <jh@suse.cz>
17059 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17060 gcov streaming; stream hot bb threshold to ltrans.
17061 * predict.c (get_hot_bb_threshold): Break out from ....
17062 (maybe_hot_count_p): ... here.
17063 (set_hot_bb_threshold): New function.
17064 * lto-section-in.c (lto_section_name): Add profile.
17065 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17066 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17067 and data-streamer.h
17068 (histogram_entry): New structure.
17069 (histogram, histogram_pool): New global vars.
17070 (histogram_hash): New structure.
17071 (histogram_hash::hash): New method.
17072 (histogram_hash::equal): Likewise.
17073 (account_time_size): New function.
17074 (cmp_counts): New function.
17075 (dump_histogram): New function.
17076 (ipa_profile_generate_summary): New function.
17077 (ipa_profile_write_summary): New function.
17078 (ipa_profile_read_summary): New function.
17079 (ipa_profile): Decide on threshold.
17080 (pass_ipa_profile): Add ipa_profile_write_summary and
17081 ipa_profile_read_summary.
17082 * Makefile.in (ipa.o): Update dependencies.
17083 * lto-streamer.h (LTO_section_ipa_profile): New section.
17085 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17087 * tree.h (VAR_P): New.
17089 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17092 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17094 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17096 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17097 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17098 (control_flow_insn_p): Likewise.
17099 * cfgrtl.c (duplicate_insn_chain): Likewise.
17100 * final.c (get_attr_length_1): Likewise.
17101 (shorten_branches): Likewise.
17102 (final_scan_insn): Likewise.
17103 * function.c (instantiate_virtual_regs): Likewise.
17104 * gcse.c (insert_insn_end_basic_block): Likewise.
17105 * ira-costs.c (scan_one_insn): Likewise.
17106 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17107 * lra.c (check_rtl): Likewise.
17108 * reload1.c (elimination_costs_in_insn): Likewise.
17109 * reorg.c (follow_jumps): Likewise.
17111 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17112 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17113 (thumb_far_jump_used_p): Likewise.
17114 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17115 (workaround_speculation): Likewise.
17116 (add_sched_insns_for_speculation): Likewise.
17117 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17118 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17119 (frv_for_each_packet): Likewise.
17120 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17121 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17122 (final_emit_insn_group_barriers): Likewise.
17123 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17124 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17125 (mips16_insn_length): Likewise.
17126 * config/pa/pa.c (pa_reorg): Likewise.
17127 (pa_combine_instructions): Likewise.
17128 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17129 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17130 (sh_reorg): Likewise.
17131 (split_branches): Likewise.
17132 * config/spu/spu.c (get_branch_target): Likewise.
17134 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17137 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17139 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17140 Fix declaration name.
17142 2013-03-28 Lawrence Crowl <crowl@google.com>
17144 * graphds.h (struct graph.indicies): Remove unused.
17145 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17146 (SCOP_ORIGINAL_PDDRS): Remove unused.
17147 * sese.h (extern insert_loop_close_phis): Removed unused.
17148 (extern insert_guard_phis): Removed unused.
17149 (extern ivtype_map_elt_info): Removed unused.
17150 (new_ivtype_map_elt): Removed unused.
17151 * sese.c (ivtype_map_elt_info): Removed unused.
17153 2013-03-28 Lawrence Crowl <crowl@google.com>
17155 * Makefile.in: Add several missing include dependences.
17157 (test-dump.o): New. This object is not added to any executable,
17158 but is present for ad-hoc testing.
17160 (debug (const bitmap_head_def &)): New.
17161 (debug (const bitmap_head_def *)): New.
17163 (extern debug (const bitmap_head_def &)): New.
17164 (extern debug (const bitmap_head_def *)): New.
17166 (debug (edge_def &)): New.
17167 (debug (edge_def *)): New.
17169 (debug (basic_block_def &)): New.
17170 (debug (basic_block_def *)): New.
17172 (dump_node (const_tree, int, FILE *)): Correct source file.
17174 (debug (die_struct &)): New.
17175 (debug (die_struct *)): New.
17177 (extern debug (die_struct &)): New.
17178 (extern debug (die_struct *)): New.
17179 * gimple-pretty-print.c
17180 (debug (gimple_statement_d &)): New.
17181 (debug (gimple_statement_d *)): New.
17182 * gimple-pretty-print.h
17183 (extern debug (gimple_statement_d &)): New.
17184 (extern debug (gimple_statement_d *)): New.
17186 (debug (ira_allocno_copy &)): New.
17187 (debug (ira_allocno_copy *)): New.
17188 (debug (ira_allocno &)): New.
17189 (debug (ira_allocno *)): New.
17191 (extern debug (ira_allocno_copy &)): New.
17192 (extern debug (ira_allocno_copy *)): New.
17193 (extern debug (ira_allocno &)): New.
17194 (extern debug (ira_allocno *)): New.
17196 (debug (live_range &)): New.
17197 (debug (live_range *)): New.
17199 (debug (lra_live_range &)): New.
17200 (debug (lra_live_range *)): New.
17202 (debug (lra_live_range &)): New.
17203 (debug (lra_live_range *)): New.
17205 (debug (omega_pb_d &)): New.
17206 (debug (omega_pb_d *)): New.
17208 (extern debug (omega_pb_d &)): New.
17209 (extern debug (omega_pb_d *)): New.
17211 (debug (const rtx_def &)): New.
17212 (debug (const rtx_def *)): New.
17214 (debug_tree (tree): Move within file.
17215 (debug_raw (const tree_node &)): New.
17216 (debug_raw (const tree_node *)): New.
17217 (dump_tree_via_hooks (const tree_node *, int)): New.
17218 (debug (const tree_node &)): New.
17219 (debug (const tree_node *)): New.
17220 (debug_verbose (const tree_node &)): New.
17221 (debug_verbose (const tree_node *)): New.
17222 (debug_head (const tree_node &)): New.
17223 (debug_head (const tree_node *)): New.
17224 (debug_body (const tree_node &)): New.
17225 (debug_body (const tree_node *)): New.
17226 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17227 (debug (vec<tree, va_gc> &)): New.
17228 (debug (vec<tree, va_gc> *)): New.
17230 (extern debug (const rtx_def &)): New.
17231 (extern debug (const rtx_def *)): New.
17233 (debug_raw (simple_bitmap_def &)): New.
17234 (debug_raw (simple_bitmap_def *)): New.
17235 (debug (simple_bitmap_def &)): New.
17236 (debug (simple_bitmap_def *)): New.
17238 (extern debug (simple_bitmap_def &)): New.
17239 (extern debug (simple_bitmap_def *)): New.
17240 (extern debug_raw (simple_bitmap_def &)): New.
17241 (extern debug_raw (simple_bitmap_def *)): New.
17243 (debug (vinsn_def &)): New.
17244 (debug (vinsn_def *)): New.
17245 (debug_verbose (vinsn_def &)): New.
17246 (debug_verbose (vinsn_def *)): New.
17247 (debug (expr_def &)): New.
17248 (debug (expr_def *)): New.
17249 (debug_verbose (expr_def &)): New.
17250 (debug_verbose (expr_def *)): New.
17251 (debug (vec<rtx> &)): New.
17252 (debug (vec<rtx> *)): New.
17254 (extern debug (vinsn_def &)): New.
17255 (extern debug (vinsn_def *)): New.
17256 (extern debug_verbose (vinsn_def &)): New.
17257 (extern debug_verbose (vinsn_def *)): New.
17258 (extern debug (expr_def &)): New.
17259 (extern debug (expr_def *)): New.
17260 (extern debug_verbose (expr_def &)): New.
17261 (extern debug_verbose (expr_def *)): New.
17262 (extern debug (vec<rtx> &)): New.
17263 (extern debug (vec<rtx> *)): New.
17265 (_list_iter_cond_expr): Make inline instead of static.
17267 (debug (sreal &)): New.
17268 (debug (sreal *)): New.
17270 (extern debug (sreal &)): New.
17271 (extern debug (sreal *)): New.
17273 (extern debug_raw (const tree_node &)): New.
17274 (extern debug_raw (const tree_node *)): New.
17275 (extern debug (const tree_node &)): New.
17276 (extern debug (const tree_node *)): New.
17277 (extern debug_verbose (const tree_node &)): New.
17278 (extern debug_verbose (const tree_node *)): New.
17279 (extern debug_head (const tree_node &)): New.
17280 (extern debug_head (const tree_node *)): New.
17281 (extern debug_body (const tree_node &)): New.
17282 (extern debug_body (const tree_node *)): New.
17283 (extern debug (vec<tree, va_gc> &)): New.
17284 (extern debug (vec<tree, va_gc> *)): New.
17286 (debug (struct loop &)): New.
17287 (debug (struct loop *)): New.
17288 (debug_verbose (struct loop &)): New.
17289 (debug_verbose (struct loop *)): New.
17290 * tree-dump.c: Add header dependence.
17292 (extern debug (struct loop &)): New.
17293 (extern debug (struct loop *)): New.
17294 (extern debug_verbose (struct loop &)): New.
17295 (extern debug_verbose (struct loop *)): New.
17297 (debug (data_reference &)): New.
17298 (debug (data_reference *)): New.
17299 (debug (vec<data_reference_p> &)): New.
17300 (debug (vec<data_reference_p> *)): New.
17301 (debug (vec<ddr_p> &)): New.
17302 (debug (vec<ddr_p> *)): New.
17304 (extern debug (data_reference &)): New.
17305 (extern debug (data_reference *)): New.
17306 (extern debug (vec<data_reference_p> &)): New.
17307 (extern debug (vec<data_reference_p> *)): New.
17308 (extern debug (vec<ddr_p> &)): New.
17309 (extern debug (vec<ddr_p> *)): New.
17311 (debug (pt_solution &)): New.
17312 (debug (pt_solution *)): New.
17314 (extern debug (pt_solution &)): New.
17315 (extern debug (pt_solution *)): New.
17317 (debug (_var_map &)): New.
17318 (debug (_var_map *)): New.
17319 (debug (tree_live_info_d &)): New.
17320 (debug (tree_live_info_d *)): New.
17322 (extern debug (_var_map &)): New.
17323 (extern debug (_var_map *)): New.
17324 (extern debug (tree_live_info_d &)): New.
17325 (extern debug (tree_live_info_d *)): New.
17327 2013-03-28 Jan Hubicka <jh@suse.cz>
17329 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17331 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17333 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17334 record only when desired or required.
17336 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17338 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17339 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17341 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17343 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17344 (*andsi3_compare0_uxtw): New pattern.
17345 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17346 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17348 2013-03-28 Jan Hubicka <jh@suse.cz>
17350 * data-streamer-in.c (streamer_read_gcov_count): New function.
17351 * gimple-streamer-out.c: Include value-prof.h.
17352 (output_gimple_stmt): Output histogram.
17353 (output_bb): Use streamer_write_gcov_count.
17354 * value-prof.c: Include data-streamer.h
17355 (dump_histogram_value): Add HIST_TYPE_MAX.
17356 (stream_out_histogram_value): New function.
17357 (stream_in_histogram_value): New function.
17358 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17359 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17360 * data-streamer-out.c (streamer_write_gcov_count): New function.
17361 (streamer_write_gcov_count_stream): New function.
17362 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17363 (lto_output_node): Likewise.
17364 (input_node, input_edge): Likewise.
17365 * lto-streamer-out.c (output_cfg): Update streaming.
17366 * lto-streamer-in.c (input_cfg): Likewise.
17367 * data-streamer.h (streamer_write_gcov_count,
17368 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17369 * gimple-streamer-in.c: Include value-prof.h
17370 (input_gimple_stmt): Input histograms.
17371 (input_bb): Update profile streaming.
17373 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17375 * genmodes.c (emit_max_int): New function.
17376 (emit_insn_modes_h): Added call to emit_max_function.
17377 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17379 * machmode.def: Fixed comment.
17381 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17383 * combine.c (try_combine): Removed useless assert.
17384 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17386 2013-03-28 Marek Polacek <polacek@redhat.com>
17387 Richard Biener <rguenther@suse.de>
17389 PR tree-optimization/56695
17390 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17391 build signed result of a vector comparison.
17392 * tree-cfg.c (verify_gimple_comparison): Check that a result
17393 of a vector comparison has signed type.
17395 2013-03-28 Richard Biener <rguenther@suse.de>
17397 PR tree-optimization/37021
17398 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17399 do not restrict gaps between groups.
17400 * tree-vect-stmts.c (vectorizable_load): Properly account for
17401 a gap between groups.
17403 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17405 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17406 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17409 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17411 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17412 * doc/extend.texi (Named Address Spaces): Ditto.
17413 (Variable Attributes): Ditto.
17415 2013-03-27 Kai Tietz <ktietz@redhat.com>
17417 * config.build: Add support for cygwin x64 target.
17418 * config.gcc: Likewise.
17419 * config.host: Likewise.
17420 * configure.ac: Likewise
17421 * configure: Regenerated.
17423 2013-03-27 Kai Tietz <ktietz@redhat.com>
17425 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17426 * config/i386/t-cygwin-w64: New file.
17427 * config/i386/cygwin-w64.h: New file.
17428 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17429 and add support for x64-cygwin target.
17430 (CPP_SPEC): Likewise.
17431 (CXX_WRAP_SPEC_LIST): Undefine before define.
17432 (LIBGCJ_SONAME): Use 15 as version.
17434 2013-03-27 Richard Biener <rguenther@suse.de>
17436 PR tree-optimization/56716
17437 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17438 dumping for ref nodes.
17440 2013-03-27 Martin Jambor <mjambor@suse.cz>
17442 PR tree-optimization/55334
17443 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17444 restricted pointers to arrays.
17446 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17448 * Makefile.in (.SUFFIXES): Add .cc.
17449 (.c.o): Apply same recipe for implicit rule .cc.o.
17451 2013-03-27 Richard Biener <rguenther@suse.de>
17453 PR tree-optimization/37021
17454 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17455 REALPART/IMAGPART_EXPRs around the supported refs.
17456 * tree-ssa-structalias.c (find_func_aliases): Assume that
17457 floating-point values are not used to transfer pointers.
17459 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17461 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17462 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17463 * doc/tm.texi: Regenerate.
17464 * targhooks.h (default_has_ifunc_p): New.
17465 * targhooks.c (default_has_ifunc_p): Ditto.
17466 * config/linux-protos.h: New file.
17467 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17468 hook for linux which disables support of indirect functions in android.
17469 * config/linux-android.c: New file.
17470 * config/t-linux-android.c: Ditto.
17471 * config.gcc: Added new object file linux-android.o.
17472 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17473 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17474 * varasm.c (do_assemble_alias): Likewise.
17475 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17476 doesn't support indirect functions.
17477 * configure: Regenerate.
17479 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17482 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17483 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17486 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17488 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17490 2013-03-27 Terry Guo <terry.guo@arm.com>
17492 * config/arm/arm-cores.def: Added core cortex-r7.
17493 * config/arm/arm-tune.md: Regenerated.
17494 * config/arm/arm-tables.opt: Regenerated.
17495 * doc/invoke.texi: Added entry for core cortex-r7.
17497 2013-03-27 Walter Lee <walt@tilera.com>
17499 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17500 double-decrement of next_scratch_regno.
17502 2013-03-27 Walter Lee <walt@tilera.com>
17504 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17506 (insn_v1mulus): Ditto.
17507 (insn_v2muls): Ditto.
17509 2013-03-27 Walter Lee <walt@tilera.com>
17511 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17512 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17514 2013-03-27 Walter Lee <walt@tilera.com>
17516 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17517 (*sibcall_value): Ditto.
17519 2013-03-27 Walter Lee <walt@tilera.com>
17521 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17522 (insn_mnz_v8qi): ... this ...
17523 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17525 (insn_v<n>mnz): Replaced by ...
17526 (insn_v1mnz): ... this ...
17527 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17529 (insn_mz_<mode>): Replaced by ...
17530 (insn_mz_v8qi): ... this ...
17531 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17533 (insn_v<n>mz): Replaced by ...
17534 (insn_v1mz): ... this ...
17535 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17538 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17540 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17542 2013-03-26 Roland McGrath <mcgrathr@google.com>
17544 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17545 than fprintf with a non-constant, non-format string.
17547 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17549 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17550 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17551 operand 0 predicate.
17552 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17553 attribute. Use general_x64nomem_operand as operand 1 predicate.
17554 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17555 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17556 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17557 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17558 (mov<mode>_insv_1): Remove expander. Merge insn with
17559 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17560 Use general_x64nomem_operand as operand 1 predicate.
17561 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17562 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17563 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17564 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17565 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17566 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17567 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17568 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17569 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17570 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17571 (general_x64nomem_operand): Ditto.
17573 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17575 * config/rtems.opt: Add -pthread option.
17577 2013-03-26 Richard Biener <rguenther@suse.de>
17579 * alias.c (find_base_term): Avoid redundant and not used recursion.
17580 (base_alias_check): Get the initial base term from the caller.
17581 (true_dependence_1): Compute and pass base terms to base_alias_check.
17582 (write_dependence_p): Likewise.
17583 (may_alias_p): Likewise.
17585 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17587 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17588 PC-relative load in SI modes and above only.
17590 2013-03-26 Xinyu Qi <xyqi@marvell.com>
17592 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
17593 * config/arm/iwmmxt.md (WCGR0): Update.
17594 (WCGR1, WCGR2, WCGR3): Likewise.
17596 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17598 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
17599 Use x64 and nox64 isa attributes.
17601 2013-03-26 Richard Biener <rguenther@suse.de>
17603 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
17604 alignment computations and rely on get_object_alignment_1
17605 for the !TYPE_P case.
17606 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
17608 2013-03-26 Walter Lee <walt@tilera.com>
17610 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
17611 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
17613 2013-03-25 Jeff Law <law@redhat.com>
17615 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
17616 check for INTEGRAL_TYPE_P that was missing due to checking in
17617 wrong version of prior patch.
17619 2013-03-25 Walter Lee <walt@tilera.com>
17621 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
17622 TILEGX_INSN_SHUFFLEBYTES1.
17623 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
17625 (tilegx_builtins): Ditto.
17626 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
17628 2013-03-25 Walter Lee <walt@tilera.com>
17630 * config/tilegx/tilegx.md (floatsisf2): New pattern.
17631 (floatunssisf2): New pattern.
17632 (floatsidf2): New pattern.
17633 (floatunssidf2): New pattern.
17635 2013-03-25 Walter Lee <walt@tilera.com>
17637 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
17638 tests for constraint J, K, N, P.
17640 2013-03-25 Walter Lee <walt@tilera.com>
17642 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
17643 Use indirect/pcrel encoding.
17644 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
17647 2013-03-25 Steve Ellcey <sellcey@mips.com>
17649 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
17650 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
17651 * config/mips/mips.c (mips_option_override): Set IMADD default.
17652 * config/mips/mips.h (PTF_AVOID_IMADD): New.
17653 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
17654 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
17655 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
17656 * doc/invoke.texi (-mimadd/-mno-imadd): New.
17658 2013-03-25 Jeff Law <law@redhat.com>
17660 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
17661 slightly to avoid creating and folding useless trees. Simplify
17662 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
17664 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
17666 * config/i386/i386.md (*zero_extendsidi2): Merge with
17667 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
17668 * config/i386/predicates.md (x86_64_zext_operand): Rename from
17669 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
17670 targets. Clarify comment.
17672 2013-03-25 Martin Jambor <mjambor@suse.cz>
17674 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
17675 pass-through jump functions differently.
17676 (ipa_read_jump_function): Likewise. Also use setter functions to set
17679 2013-03-25 Martin Jambor <mjambor@suse.cz>
17681 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
17682 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
17684 (ipa_get_indirect_edge_target): New function.
17685 (devirtualization_time_bonus): New parameter known_aggs, pass it to
17686 ipa_get_indirect_edge_target. Update all callers.
17687 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
17688 ipa_get_indirect_edge_target_1 instead of calling
17689 ipa_get_indirect_edge_target.
17690 (create_specialized_node): Pass aggvlas to
17691 ipcp_discover_new_direct_edges.
17693 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17695 * config/arm/arm.md (f_sels, f_seld): New types.
17696 (*cmov<mode>): New pattern.
17697 * config/arm/predicates.md (arm_vsel_comparison_operator): New
17700 2013-03-25 Kai Tietz <ktietz@redhat.com>
17702 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
17703 POSIX-printf for mingw-hosted builds.
17705 2013-03-25 Richard Biener <rguenther@suse.de>
17707 PR middle-end/56694
17708 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
17709 must-not-throw stmt location.
17711 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17713 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
17714 Emit load-acquire versions when acq is true.
17715 (arm_emit_store_exclusive): Add rel parameter.
17716 Emit store-release versions when rel is true.
17717 (arm_split_compare_and_swap): Use acquire-release instructions
17719 of barriers when appropriate.
17720 (arm_split_atomic_op): Likewise.
17721 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
17722 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
17723 (VUNSPEC_SLX): Likewise.
17724 (VUNSPEC_LDA): Likewise.
17725 (VUNSPEC_STL): Likewise.
17726 * config/arm/sync.md (atomic_load<mode>): New pattern.
17727 (atomic_store<mode>): Likewise.
17728 (arm_load_acquire_exclusive<mode>): Likewise.
17729 (arm_load_acquire_exclusivesi): Likewise.
17730 (arm_load_acquire_exclusivedi): Likewise.
17731 (arm_store_release_exclusive<mode>): Likewise.
17733 2013-03-25 Catherine Moore <clm@codesourcery.com>
17735 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
17736 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
17737 * config/mip/predicates.md (lwsp_swsp_operand,
17738 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
17739 sb16_operand, db4_operand, db7_operand, ib3_operand,
17740 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
17741 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
17742 andi16_operand): New predicates.
17743 * config/mips/mips.md (compression): New attribute.
17744 (enabled): New attribute.
17745 (length): Consider compression in computing length.
17746 (shift_compression): New code attribute.
17747 (*add<mode>3): New operands. Record compression.
17748 (sub<mode>3): Likewise.
17749 (one_cmpl<mode>2): Likewise.
17750 (*and<mode>3): Likewise.
17751 (*ior<mode>3): Likewise.
17752 (unnamed pattern for xor): Likewise.
17753 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
17754 (*<optab><mode>3): Likewise.
17755 (*mov<mode>_internal: Likewise.
17756 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
17757 (mips_unsigned_immediate_p): New.
17758 (umips_lwsp_swsp_address_p): New.
17759 (m16_based_address_p): New.
17760 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
17761 (mips_unsigned_immediate_p): New prototype.
17762 (lwsp_swsp_address_p): New prototype.
17763 (m16_based_address_p): New prototype.
17764 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
17765 (mips_signed_immediate_p): New function.
17766 (m16_based_address_p): New function.
17767 (lwsp_swsp_address_p): New function.
17768 (mips_print_operand_punctuation): Recognize short delay slot insns
17769 for microMIPS.add<mode>3"
17771 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17774 * config/arm/iterators.md (v_cmp_result): New mode attribute.
17775 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
17777 2013-03-25 Richard Biener <rguenther@suse.de>
17779 PR tree-optimization/56689
17780 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
17783 2013-03-25 Richard Biener <rguenther@suse.de>
17785 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
17787 (memory_references): Likewise.
17788 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
17789 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
17790 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
17791 (gather_mem_refs_in_loops): Fold into ...
17792 (analyze_memory_references): ... this. Move initialization
17793 to tree_ssa_lim_initialize.
17794 (fill_always_executed_in): Rename to ...
17795 (fill_always_executed_in_1): ... this.
17796 (fill_always_executed_in): Move contains_call computation to
17797 this new function from ...
17798 (tree_ssa_lim_initialize): ... here.
17799 (tree_ssa_lim): Call fill_always_executed_in.
17801 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17803 * postreload.c (reload_combine): Fix code detecting returns.
17805 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17807 * function.c (emit_use_return_register_into_block): On cc0 targets,
17808 do not emit the sequence between cc0 setter and user.
17810 2013-03-25 Kai Tietz <ktietz@redhat.com>
17812 * config/i386/predicates.md (local_symbolic_operand): Interpret
17813 dll-imported symbols as none-local.
17815 2013-03-25 Richard Biener <rguenther@suse.de>
17817 * tree-ssa-loop-im.c (struct depend): Remove.
17818 (struct lim_aux_data): Make depends a vec of gimples.
17819 (free_lim_aux_data): Adjust.
17820 (add_dependency): Likewise.
17821 (set_level): Likewise.
17823 2013-03-25 Richard Biener <rguenther@suse.de>
17825 PR middle-end/56434
17826 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
17827 the pointer returned by calls with ECF_MALLOC set.
17829 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17831 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
17833 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17835 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
17836 using MMXMODE mode iterator.
17837 (*move<mode>_internal): Merge with *movv2sf_internal and
17838 *movv2sf_internal_rex64 using MMXMODE mode iterator.
17840 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17842 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
17843 (record_last_mem_set_info): Likewise.
17845 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
17846 of XNEWVEC followed by memset.
17847 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
17849 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17851 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
17852 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
17853 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
17854 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
17855 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
17856 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
17857 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
17858 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
17859 BARRIER_P instead of GET_CODE.
17861 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17863 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
17864 inaccuracy in the probing code.
17866 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
17867 (ctrapdi4): Likewise.
17869 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17871 * calls.c (expand_call): Add missing guard to code handling return
17872 of non-BLKmode structures in MSB.
17873 * function.c (expand_function_end): Likewise.
17875 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17877 * combine.c (try_combine): Adjust comment. Do not add the set of
17878 insn #0 if the destination indirectly is set or dies in insn #2.
17879 Tidy up code to distribute a new note.
17881 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17883 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
17884 also for alternatives 16 and 17.
17886 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17888 * config/i386/sse.md (*mov<mode>_internal): Merge with
17889 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
17890 Emit insn template depending on type attribute. Use
17891 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17892 movd instead of movq mnemonic for interunit moves. Rewrite mode
17893 attribute calculation. Remove unit attribute calculation.
17894 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
17895 Set prefix_data16 attribute for DImode ssemov types.
17896 Use Ym instead of y for SSE-MMX conversion alternatives.
17897 Reorder operand constraints.
17899 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
17901 * df.h (df_insn_delete): Adjust prototype.
17902 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
17903 and let it decide whether mark the basic block dirty.
17904 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
17905 * df-scan.c (df_insn_info_delete): New helper function, split
17906 off from df_insn_delete.
17907 (df_scan_free_bb_info): Use it.
17908 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
17910 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
17911 that the insn is actually an insn and it has a non-NULL basic block.
17912 Do not mark basic block dirty if only deleting a DEBUG_INSN.
17914 2013-03-22 Richard Biener <rguenther@suse.de>
17916 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
17918 (mem_ref_alloc): Do not allocate them.
17919 (refs_independent_p): Do not query or maintain a cache.
17921 2013-03-22 Richard Biener <rguenther@suse.de>
17923 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
17924 (gather_mem_refs_in_loops): Do not compute it.
17925 (analyze_memory_references): Do not allocate it.
17926 (tree_ssa_lim_finalize): Do not free it.
17927 (for_all_locs_in_loop): Do not query all_refs_in_loop.
17929 2013-03-22 Richard Biener <rguenther@suse.de>
17931 * is-a.h (as_a): Use gcc_checking_assert.
17933 2013-03-22 Ian Bolton <ian.bolton@arm.com>
17935 * config/aarch64/aarch64.c (aarch64_print_operand): New
17936 format specifier for printing a constant in hex.
17937 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
17938 format specifier for printing second operand.
17940 2013-03-22 Richard Biener <rguenther@suse.de>
17942 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
17944 (gather_mem_refs_in_loops): Perform store accumulation here.
17945 (create_vop_ref_mapping_loop): Remove.
17946 (create_vop_ref_mapping): Likewise.
17947 (analyze_memory_references): Initialize refs_stored_in_loop.
17948 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
17949 (record_indep_loop): Remove.
17950 (record_dep_loop): New function.
17951 (ref_indep_loop_p_1): Adjust to only walk over references
17952 in the loop, not its subloops.
17953 (ref_indep_loop_p): Rename to ...
17954 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
17955 maintaining a more fine-grained cache.
17956 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
17957 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
17959 2013-03-22 Richard Biener <rguenther@suse.de>
17961 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
17962 (struct mem_ref): Make accesses_in_loop a vec of a vec of
17963 aggregate mem_ref_loc.
17964 (free_mem_ref_locs): Inline into ...
17965 (memref_free): ... this and adjust.
17966 (mem_ref_alloc): Adjust.
17967 (mem_ref_locs_alloc): Remove.
17968 (record_mem_ref_loc): Adjust.
17969 (get_all_locs_in_loop): Rewrite into ...
17970 (for_all_locs_in_loop): ... this iterator.
17971 (rewrite_mem_ref_loc): New functor.
17972 (rewrite_mem_refs): Use for_all_locs_in_loop.
17973 (sm_set_flag_if_changed): New functor.
17974 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
17975 (ref_always_accessed): New functor.
17976 (ref_always_accessed_p): Use for_all_locs_in_loop.
17978 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17980 * tree-pass.h (PROP_gimple_lvec): New.
17981 * passes.c (dump_properties): Handle PROP_gimple_lvec.
17982 (init_optimization_passes): Move pass_lower_vector.
17983 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
17985 (pass_lower_vector): Provide PROP_gimple_lvec.
17986 (pass_lower_vector_ssa): Likewise.
17987 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
17989 2013-03-21 Mark Wielaard <mjw@redhat.com>
17991 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
17993 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17995 * config/i386/i386.md (*movdi_internal): Disparage slightly
17996 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
17997 conversion alternatives.
17999 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18001 PR middle-end/48087
18002 * diagnostic.def (DK_WERROR): New kind.
18003 * diagnostic.h (werrorcount): Define.
18004 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18005 promoted to DK_ERROR, increment DK_WERROR counter instead of
18007 * toplev.c (toplev_main): Call print_ignored_options even if
18008 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18009 even if just werrorcount is non-zero.
18012 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18014 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18015 (string_cst_pool_decl): New function.
18016 (optimize_one_addr_into_implicit_ptr): New function.
18017 (resolve_addr_in_expr): Optimize DWARF location expression
18018 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18019 which doesn't live in memory, but has DW_AT_location or
18020 DW_AT_const_value, or refers to a string literal, into
18021 DW_OP_GNU_implicit_pointer.
18022 (optimize_location_into_implicit_ptr): New function.
18023 (resolve_addr): If removing DW_AT_location of a variable because
18024 it was DW_OP_addr of address of the variable, but the variable doesn't
18025 live in memory, try to emit const value attribute for the initializer.
18027 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18029 * tree.h (VECTOR_TYPE_P): New macro.
18030 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18031 TYPE_MODE): Use it.
18032 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18033 VEC_COND_EXPR cannot be lvalues.
18034 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18036 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18038 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18039 Restrict the transformation to equal modes.
18041 2013-03-21 Richard Biener <rguenther@suse.de>
18043 PR tree-optimization/39326
18044 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18045 (MEM_ANALYZABLE): Adjust.
18046 (record_mem_ref_loc): Move bitmap ops ...
18047 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18048 unanalyzable refs, do not record locations for it.
18049 (analyze_memory_references): Allocate ref zero as shared
18051 (refs_independent_p): Do not test for unanalyzed mems here.
18052 (ref_indep_loop_p_1): Special-case disambiguation against
18053 the unanalyzed ref.
18054 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18056 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18058 * config/arm/arm-protos.h (tune_params): Add
18059 prefer_neon_for_64bits field.
18060 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18061 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18062 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18063 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18064 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18065 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18066 (arm_option_override): Handle -mneon-for-64bits new option.
18067 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18068 (prefer_neon_for_64bits): Declare new variable.
18069 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18070 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18071 (arch_enabled): Handle new arch types. Remove support for onlya8
18073 (one_cmpldi2): Use new arch names.
18074 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18075 * config/arm/arm.opt (mneon-for-64bits): Add option.
18076 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18077 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18078 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18080 * doc/invoke.texi (-mneon-for-64bits): Document.
18082 2013-03-21 Richard Biener <rguenther@suse.de>
18084 PR tree-optimization/39326
18085 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18086 (sort_bbs_in_loop_postorder_cmp): New function.
18087 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18089 2013-03-21 Richard Biener <rguenther@suse.de>
18091 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18092 (vect_insert_into_interleaving_chain): Likewise.
18093 (vect_drs_dependent_in_basic_block): Inline ...
18094 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18096 (vect_analyze_data_ref_dependence): ... here. Simplify.
18097 (vect_check_interleaving): Simplify.
18098 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18099 (vect_slp_analyze_data_ref_dependences): ... this new function.
18100 (dr_group_sort_cmp): New function.
18101 (vect_analyze_data_ref_accesses): Compute data-reference groups
18102 here instead of in vect_analyze_data_ref_dependence. Use
18103 a more efficient algorithm.
18104 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18105 vect_slp_analyze_data_ref_dependences. Call
18106 vect_analyze_data_ref_accesses earlier.
18107 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18108 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18109 (vect_slp_analyze_data_ref_dependences): New prototype.
18111 2013-03-21 Richard Biener <rguenther@suse.de>
18113 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18114 ref is stored in the loop.
18115 (find_refs_for_sm): Walk only over all stores.
18116 (store_motion_loop): Allocate from lim_bitmap_obstack.
18117 (store_motion): Likewise.
18119 2013-03-21 Richard Biener <rguenther@suse.de>
18121 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18122 Update virtual SSA form.
18124 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18126 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18127 * configure: Regenerate.
18128 * config.in: Regenerate.
18129 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18130 if !HAVE_LD_EH_FRAME_CIEV3.
18132 2013-03-21 Richard Biener <rguenther@suse.de>
18134 * tree-cfg.c (verify_expr_no_block): New function.
18135 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18136 nor DECL_VALUE_EXPR have locations with associated blocks.
18137 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18138 (clear_unused_block_pointer): Remove code dealing with
18139 blocks in DECL_DEBUG_EXPR locations.
18141 2013-03-21 Richard Biener <rguenther@suse.de>
18143 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18144 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18145 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18146 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18147 instead of DECL_DEBUG_EXPR_IS_FROM.
18148 * gimplify.c (gimplify_modify_expr): Likewise.
18149 * tree-cfg.c (verify_expr_location_1): Likewise.
18150 * tree-complex.c (create_one_component_var): Likewise.
18151 * tree-sra.c (create_access_replacement): Likewise.
18152 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18153 (clear_unused_block_pointer): Likewise.
18154 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18155 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18156 * var-tracking.c (var_debug_decl): Likewise.
18157 (track_expr_p): Likewise.
18158 * tree-inline.c (add_local_variables): Likewise. Set
18159 DECL_HAS_DEBUG_EXPR_P after copying it.
18160 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18161 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18163 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18166 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18167 * configure: Regenerate.
18168 * config.in: Regenerate.
18169 * config/i386/i386.md (*movdf_internal): Use
18170 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18171 movd instead of movq mnemonic for interunit moves.
18172 (*movdi_internal): Ditto.
18174 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18176 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18177 (abd<mode>_3): New pattern.
18178 (aba<mode>_3): New pattern.
18179 (fabd<mode>_3): New pattern.
18181 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18183 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18184 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18185 occurrence of REGISTER_PREFIX as its empty string.
18187 2013-03-20 Jeff Law <law@redhat.com>
18189 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18190 addititional equivalences for equality comparisons between an SSA_NAME
18191 and a constant where the SSA_NAME was set from a widening conversion.
18193 2013-03-20 Walter Lee <walt@tilera.com>
18195 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18197 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18199 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18200 depending on type attribute.
18201 (*movti_internal): Ditto.
18202 (*movtf_internal): Ditto.
18203 (*movxf_internal): Ditto.
18204 (*movdf_internal): Ditto.
18205 (*movsf_internal): Ditto.
18207 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18209 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18210 maybe_vex for sselog1 and ssemov types.
18211 (*movdi_internal): Reorder operand constraints.
18212 (*movsi_internal): Ditto. Set prefix attribute to
18213 maybe_vex for sselog1 and ssemov types.
18214 (*movtf_internal): Set prefix attribute to maybe_vex
18215 for sselog1 and ssemov types.
18216 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18217 DImode ssemov types. Reorder operand constraints.
18218 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18219 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18220 attribute for SImode ssemov types. Reorder operand constraints.
18222 2013-03-20 Martin Jambor <mjambor@suse.cz>
18224 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18225 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18227 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18229 * config/rs6000/predicates.md (indexed_address, update_address_mem
18230 update_indexed_address_mem): New predicates.
18231 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18232 attribute for load/store instructions.
18233 * config/rs6000/dfp.md (movsd_store): Likewise.
18234 (movsd_load): Likewise.
18235 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18236 (unnamed HI->DI extend define_insn): Likewise.
18237 (unnamed SI->DI extend define_insn): Likewise.
18238 (unnamed QI->SI extend define_insn): Likewise.
18239 (unnamed QI->HI extend define_insn): Likewise.
18240 (unnamed HI->SI extend define_insn): Likewise.
18241 (unnamed HI->SI extend define_insn): Likewise.
18242 (extendsfdf2_fpr): Likewise.
18243 (movsi_internal1): Likewise.
18244 (movsi_internal1_single): Likewise.
18245 (movhi_internal): Likewise.
18246 (movqi_internal): Likewise.
18247 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18248 attribute for load/store instructions.
18249 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18251 (mov<mode>_softfloat): Likewise.
18252 (mov<mode>_hardfloat32): Likewise.
18253 (mov<mode>_hardfloat64): Likewise.
18254 (mov<mode>_softfloat64): Likewise.
18255 (movdi_internal32): Likewise.
18256 (movdi_internal64): Likewise.
18257 (probe_stack_<mode>): Likewise.
18259 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18261 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18262 floating point, and decimal floating point to reload iterator.
18264 * config/rs6000/constraints.md (wl constraint): New constraints to
18265 return FLOAT_REGS if certain options are used to reduce the number
18266 of separate patterns that exist in the file.
18267 (wx constraint): Likewise.
18268 (wz constraint): Likewise.
18270 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18271 -mdebug=reg, print wg, wl, wx, and wz constraints.
18272 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18273 Initialize the reload functions for 64-bit binary/decimal floating
18275 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18276 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18277 create the buffer on the stack to overcome not having a 32-bit
18279 (rs6000_emit_move): Likewise.
18280 (rs6000_secondary_memory_needed_rtx): Likewise.
18281 (rs6000_alloc_sdmode_stack_slot): Likewise.
18282 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18283 via xxlxor, just like DFmode 0.0.
18285 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18286 define as 1 if we are running on a power7 or newer.
18287 (enum r6000_reg_class_enum): Add new constraints.
18289 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18290 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18291 with other moves by using conditional constraits (wg). Use LFIWZX
18292 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18293 (movsd splitter): Likewise.
18294 (movsd_hardfloat): Likewise.
18295 (movsd_softfloat): Likewise.
18297 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18298 binary and decimal floating point moves.
18299 (fmove_ok): New attributes to combine binary and decimal floating
18300 point moves, and to combine power6x (mfpgpr) moves along normal
18302 (real_value_to_target): Likewise.
18303 (f32_lr): Likewise.
18304 (f32_lm): Likewise.
18305 (f32_li): Likewise.
18306 (f32_sr): Likewise.
18307 (f32_sm): Likewise.
18308 (f32_si): Likewise.
18309 (movsf): Combine binary and decimal floating point moves. Combine
18310 power6x (mfpgpr) moves with other moves by using conditional
18311 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18312 (mov<mode> for SFmode/SDmode); Likewise.
18313 (SFmode/SDmode splitters): Likewise.
18314 (movsf_hardfloat): Likewise.
18315 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18316 (movsf_softfloat): Likewise.
18317 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18319 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18320 wx and wz constraints.
18322 * config/rs6000/constraints.md (wg constraint): New constraint to
18323 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18325 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18328 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18329 -mdebug=reg, print wg, wl, wx, and wz constraints.
18330 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18331 Initialize the reload functions for 64-bit binary/decimal floating
18333 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18334 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18335 create the buffer on the stack to overcome not having a 32-bit
18337 (rs6000_emit_move): Likewise.
18338 (rs6000_secondary_memory_needed_rtx): Likewise.
18339 (rs6000_alloc_sdmode_stack_slot): Likewise.
18340 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18341 via xxlxor, just like DFmode 0.0.
18343 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18344 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18345 with other moves by using conditional constraits (wg). Use LFIWZX
18346 and STFIWX for loading SDmode on power7.
18347 (movdd splitters): Likewise.
18348 (movdd_hardfloat32): Likewise.
18349 (movdd_softfloat32): Likewise.
18350 (movdd_hardfloat64_mfpgpr): Likewise.
18351 (movdd_hardfloat64): Likewise.
18352 (movdd_softfloat64): Likewise.
18354 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18355 64-bit binary and decimal floating point moves.
18356 (FMOVE64X): Likewise.
18357 (movdf): Combine 64-bit binary and decimal floating point moves.
18358 Combine power6x (mfpgpr) moves with other moves by using
18359 conditional constraits (wg).
18360 (mov<mode> for DFmode/DDmode): Likewise.
18361 (DFmode/DDmode splitters): Likewise.
18362 (movdf_hardfloat32): Likewise.
18363 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18364 (movdf_softfloat32): Likewise.
18365 (movdf_hardfloat64_mfpgpr): Likewise.
18366 (movdf_hardfloat64): Likewise.
18367 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18368 (movdf_softfloat64): Likewise.
18369 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18370 (reload_<mode>_load): Move to later in the file so they aren't in
18371 the middle of the floating point move insns.
18372 (reload_<mode>_store): Likewise.
18374 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18377 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18378 constraint if -mdebug=reg.
18379 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18380 Enable using dd reload support if needed.
18382 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18383 binary and decimal floating point moves in rs6000.md.
18384 (movtd_internal): Likewise.
18386 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18387 decimal floating point moves.
18389 (movtf_internal): Likewise.
18390 (mov<mode>_internal, TDmode/TFmode): Likewise.
18391 (movtf_softfloat): Likewise.
18392 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18394 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18395 movdi_internal64, using wg constraint for move direct operations.
18396 (movdi_internal64): Likewise.
18398 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18399 MODES_TIEABLE_P for selected modes. Print the numerical value of
18400 the various virtual registers. Use GPR/FPR first/last values,
18401 instead of hard coding the register numbers. Print which modes
18402 have reload functions registered.
18403 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18404 settings before/after setting cpu, target and subtarget settings.
18405 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18406 and for secondary reload failures in rs6000_secondary_reload_inner.
18407 (rs6000_secondary_reload_fail): Likewise.
18408 (rs6000_secondary_reload_inner): Likewise.
18410 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18411 macros for first/last GPR and FPR registers.
18412 (LAST_GPR_REGNO): Likewise.
18413 (FIRST_FPR_REGNO): Likewise.
18414 (LAST_FPR_REGNO): Likewise.
18416 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18417 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18418 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18419 (vcond<mode><mode>): Likewise.
18420 (vcondu<mode><mode>): Likewise.
18421 (vector_gtu<mode>): Likewise.
18422 (vector_gte<mode>): Likewise.
18423 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18424 to prevent the compiler from converting DImode operations to TImode.
18425 (ior<mode>3): Likewise.
18426 (and<mode>3): Likewise.
18427 (one_cmpl<mode>2): Likewise.
18428 (nor<mode>3): Likewise.
18429 (andc<mode>3): Likewise.
18431 * config/rs6000/constraints.md (wt constraint): New constraint
18432 that returns VSX_REGS if TImode is allowed in VSX registers.
18434 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18435 constant under VSX.
18437 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18438 similar to TImode, but it is restricted to being in the GPRs.
18440 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18441 TImode to occupy a single VSX register.
18443 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18444 -mvsx-timode for power7/power8.
18445 (power7 cpu): Likewise.
18446 (power8 cpu): Likewise.
18448 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18449 sure that TFmode/TDmode take up two registers if they are ever
18450 allowed in the upper VSX registers.
18451 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18453 (rs6000_init_hard_regno_mode_ok): Likewise.
18454 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18455 constraint. Print if LRA is turned on.
18456 (rs6000_option_override_internal): Give an error if -mvsx-timode
18457 and VSX is not enabled.
18458 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18459 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18460 to reg+offset addressing. Use PTImode when checking offset
18461 addresses for validity.
18462 (reg_offset_addressing_ok_p): Likewise.
18463 (rs6000_legitimate_offset_address_p): Likewise.
18464 (rs6000_legitimize_address): Likewise.
18465 (rs6000_legitimize_reload_address): Likewise.
18466 (rs6000_legitimate_address_p): Likewise.
18467 (rs6000_eliminate_indexed_memrefs): Likewise.
18468 (rs6000_emit_move): Likewise.
18469 (rs6000_secondary_reload): Likewise.
18470 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18471 reloads to fpr registers to continue to use reg+offset addressing,
18472 but 64-bit reloads to altivec registers need reg+reg addressing.
18473 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18474 it. Treat LO_SUM like a PLUS operation.
18475 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18476 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18477 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18478 registers to share a register with a smaller sized type, since VSX
18479 puts scalars in the upper 64-bits.
18480 (print_operand): Add support for PTImode.
18481 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18482 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18483 registers, but don't have arithmetic support.
18484 (rs6000_memory_move_cost): Add test for VSX.
18485 (rs6000_opt_masks): Add -mvsx-timode.
18487 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18490 (VSr): Use wt constraint for TImode.
18491 (VSv): Drop TImode support.
18492 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18493 (vsx_movti_64bit): Likewise.
18494 (vsx_movti_32bit): Likewise.
18495 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18496 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18497 one '?' on the appropriate output constraint. Do not allow TImode
18498 logical operations on 32-bit systems.
18499 (vsx_ior<mode>3): Likewise.
18500 (vsx_xor<mode>3): Likewise.
18501 (vsx_one_cmpl<mode>2): Likewise.
18502 (vsx_nor<mode>3): Likewise.
18503 (vsx_andc<mode>3): Likewise.
18504 (vsx_concat_<mode>): Likewise.
18505 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18507 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18508 OPTION_MASK_VSX_TIMODE.
18509 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18510 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18512 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18513 (TI2 iterator): New iterator for TImode, PTImode.
18514 (wd mode attribute): Add values for vector types.
18515 (movti_string): Replace TI move operations with operations for TImode
18516 and PTImode. Add support for TImode being allowed in VSX registers.
18517 (mov<mode>_string, TImode/PTImode): Likewise.
18518 (movti_ppc64): Likewise.
18519 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18520 (TI mode splitters): Likewise.
18522 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18525 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18527 PR tree-optimization/56355
18528 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18529 Also handle integers with undefined overflow.
18531 2013-03-20 Catherine Moore <clm@codesourcery.com>
18532 Maciej W. Rozycki <macro@codesourcery.com>
18533 Tom de Vries <tom@codesourcery.com>
18534 Nathan Sidwell <nathan@codesourcery.com>
18535 Iain Sandoe <iain@codesourcery.com>
18536 Nathan Froyd <froydnj@codesourcery.com>
18537 Chao-ying Fu <fu@mips.com>
18539 * doc/extend.texi: (micromips, nomicromips, nocompression):
18540 Document new function attributes.
18541 * doc/invoke.texi (minterlink-compressed, mmicromips,
18542 m14k, m14ke, m14kec): Document new options.
18543 (minterlink-mips16): Update documentation.
18544 * doc/md.texi (ZC, ZD): Document new constraints.
18545 * configure.ac (gcc_cv_as_micromips): Check if linker
18546 supports the .set micromips directive.
18547 * configure: Regenerate.
18548 * config.in: Regenerate.
18549 * config/mips/mips-tables.opt: Regenerate.
18550 * config/mips/micromips.md: New file.
18551 * constraints.md (ZC, ZD): New constraints.
18552 * config/mips/predicates.md (movep_src_register): New predicate.
18553 (movep_src_operand): New predicate.
18554 (non_volatile_mem_operand): New predicate.
18555 * config/mips/mips.md (multimem): New type.
18556 (length): Differentiate between 17-bit and 18-bit branch offsets.
18557 (MOVEP1, MOVEP2): New mode iterator.
18558 (mov_<load>l): Use ZC constraint.
18559 (mov_<load>r): Likewise.
18560 (mov_<store>l): Likewise.
18561 (mov_<store>r): Likewise.
18562 (*branch_equality<mode>_inverted): Add microMIPS support.
18563 (*branch_equality<mode>): Likewise.
18564 (*jump_absolute): Likewise.
18565 (indirect_jump_<mode>): Likewise.
18566 (tablejump_<mode>): Likewise.
18567 (<optab>_internal): Likewise.
18568 (sibcall_internal): Likewise.
18569 (sibcall_value_internal): Likewise.
18570 (prefetch): Use constraint ZD.
18571 * config/mips/mips.opt (minterlink-compressed): New option.
18572 (minterlink-mips16): Now an alias for minterlink-compressed.
18573 (mmicromips): New option.
18574 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18575 (compare_and_swap_12): Likewise.
18576 (sync_add<mode>): Likewise.
18577 (sync_<optab>_12): Likewise.
18578 (sync_old_<optab>_12): Likewise.
18579 (sync_new_<optab>_12): Likewise.
18580 (sync_nand_12): Likewise.
18581 (sync_old_nand_12): Likewise.
18582 (sync_new_nand_12): Likewise.
18583 (sync_sub<mode>): Likewise.
18584 (sync_old_add<mode>): Likewise.
18585 (sync_old_sub<mode>): Likewise.
18586 (sync_new_add<mode>): Likewise.
18587 (sync_new_sub<mode>): Likewise.
18588 (sync_<optab><mode>): Likewise.
18589 (sync_old_<optab><mode>): Likewise.
18590 (sync_new_<optab><mode>): Likewise.
18591 (sync_nand<mode>): Likewise.
18592 (sync_old_nand<mode>): Likewise.
18593 (sync_new_nand<mode>): Likewise.
18594 (sync_lock_test_and_set<mode>): Likewise.
18595 (test_and_set_12): Likewise.
18596 (atomic_compare_and_swap<mode>): Likewise.
18597 (atomic_exchange<mode>_llsc): Likewise.
18598 (atomic_fetch_add<mode>_llsc): Likewise.
18599 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18600 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18601 (umips_save_restore_pattern_p): Likewise.
18602 (umips_load_store_pair_p): Likewise.
18603 (umips_output_load_store_pair): Likewise.
18604 (umips_movep_target_p): Likewise.
18605 (umips_12bit_offset_address_p): Likewise.
18606 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18607 (mips_base_mips16): Rename this...
18608 (mips_base_compression_flags): ...to this. Update all uses.
18609 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18610 (mips_mips16_decl_p): Delete.
18611 (mips_nomips16_decl_p): Delete.
18612 (mips_get_compress_on_flags): New function.
18613 (mips_get_compress_off_flags): New function.
18614 (mips_get_compress_mode): New function.
18615 (mips_get_compress_on_name): New function.
18616 (mips_get_compress_off_name): New function.
18617 (mips_insert_attributes): Support multiple compression types.
18618 (mips_merge_decl_attributes): Likewise.
18619 (umips_12bit_offset_address_p): New function.
18620 (mips_start_function_definition): Emit .set micromips directive.
18621 (mips_call_may_need_jalx_p): New function.
18622 (mips_function_ok_for_sibcall): Add microMIPS support.
18623 (mips_print_operand_punctuation): Support short delay slots and
18625 (umips_swm_mask, umips_swm_encoding): New.
18626 (umips_build_save_restore): New function.
18627 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18628 (was_mips16_p): Remove.
18629 (old_compression_mode): New.
18630 (mips_set_compression_mode): New function.
18631 (mips_set_current_function): Add microMIPS support.
18632 (mips_option_override): Likewise.
18633 (umips_save_restore_pattern_p): New function.
18634 (umips_output_save_restore): New function.
18635 (umips_load_store_pair_p_1): New function.
18636 (umips_load_store_pair_p): New function.
18637 (umips_output_load_store_pair_1): New function.
18638 (umips_output_load_store_pair): New function.
18639 (umips_movep_target_p) New function.
18640 (mips_prepare_pch_save): Add microMIPS support.
18641 * config/mips/mips.h (TARGET_COMPRESSION): New.
18642 (TARGET_CPU_CPP_BUILTINS): Update macro
18643 to use new compression flags and to support microMIPS.
18644 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18645 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18646 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18647 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18648 (ASM_SPEC): Support mmicromips and mno-micromips.
18649 (M16STORE_REG_P): New macro.
18650 (MIPS_CALL): Support TARGET_MICROMIPS.
18651 (MICROMIPS_J): New macro.
18652 (mips_base_mips16): Rename this...
18653 (mips_base_compression_flags): ...to this.
18654 (UMIPS_12BIT_OFFSET_P): New macro.
18655 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18656 (MULTILIB_DIRNAMES): Likewise.
18657 2013-03-20 Richard Biener <rguenther@suse.de>
18659 PR tree-optimization/56661
18660 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
18661 the result does not have to be distinct.
18663 2013-03-20 Richard Biener <rguenther@suse.de>
18665 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
18668 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18669 Steven Bosscher <steven@gcc.gnu.org>
18671 PR rtl-optimization/56605
18672 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
18674 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18677 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
18678 that require movd instead of movq.
18680 2013-03-20 Richard Biener <rguenther@suse.de>
18682 * tree-ssa-structalias.c (struct variable_info): Add pointer
18683 to the first field of an aggregate with sub-vars. Make
18684 this and the pointer to the next subfield its ID.
18685 (vi_next): New function.
18686 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
18687 storedanything_id, integer_id): Increment by one.
18688 (new_var_info, get_call_vi, lookup_call_clobber_vi,
18689 get_call_clobber_vi): Adjust.
18690 (solution_set_expand): Simplify and speedup.
18691 (solution_set_add): Inline into ...
18692 (set_union_with_increment): ... this. Adjust accordingly.
18693 (do_sd_constraint): Likewise.
18694 (do_ds_constraint): Likewise.
18695 (do_complex_constraint): Simplify.
18696 (build_pred_graph): Adjust.
18697 (solve_graph): Likewise. Simplify and speedup.
18698 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
18699 get_constraint_for_component_ref, get_constraint_for_1,
18700 first_vi_for_offset, first_or_preceding_vi_for_offset,
18701 create_function_info_for, create_variable_info_for_1,
18702 create_variable_info_for, intra_create_variable_infos): Adjust.
18703 (init_base_vars): Push NULL for ID zero.
18704 (compute_points_to_sets): Adjust.
18706 2013-03-20 Richard Biener <rguenther@suse.de>
18708 * cfgloop.c (verify_loop_structure): Streamline and avoid
18709 ICEing on corrupt loop tree.
18710 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
18713 2013-03-20 Richard Biener <rguenther@suse.de>
18715 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
18716 check whether an SSA update is needed.
18718 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
18720 * config/mips/constraints.md (T): Rename to...
18724 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
18725 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
18727 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18729 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
18730 (*subsi3_carryin_uxtw): Likewise.
18732 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18734 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
18735 (*rorsi3_insn_uxtw): Likewise.
18737 2013-03-19 Ian Bolton <ian.bolton@arm.com>
18739 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
18740 (*extrsi5_insn_uxtw): Likewise.
18742 2013-03-19 Richard Biener <rguenther@suse.de>
18744 PR tree-optimization/56273
18745 * passes.c (init_optimization_passes): Move second VRP after DOM.
18747 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
18749 * config/i386/i386.md (*movti_internal): Merge from
18750 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
18751 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
18752 nox64 isa attributes.
18754 2013-03-18 Richard Biener <rguenther@suse.de>
18756 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
18758 (merge_node_constraints): Likewise.
18759 (build_succ_graph): Likewise.
18760 (valid_graph_edge): Inline into single caller.
18761 (unify_nodes): Likewise. Use bitmap_set_bit return value
18763 (scc_visit): Fix formatting and variable use.
18764 (do_sd_constraint): Use gcc_checking_assert.
18765 (do_ds_constraint): Likewise.
18766 (do_complex_constraint): Likewise.
18767 (condense_visit): Likewise. Cleanup.
18768 (dump_pred_graph): New function.
18769 (perform_var_substitution): Dump the pred-graph before
18770 variable substitution.
18771 (find_equivalent_node): Use gcc_checking_assert.
18772 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
18774 2013-03-18 Richard Biener <rguenther@suse.de>
18776 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
18777 Remove cond_expr_stmt_list argument and do not gimplify the
18779 (vect_loop_versioning): Adjust.
18780 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
18781 Cleanup to use less temporaries.
18782 (vect_create_data_ref_ptr): Cleanup.
18784 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18786 PR tree-optimization/56635
18787 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
18788 require types_compatible_p types.
18790 2013-03-18 Nick Clifton <nickc@redhat.com>
18792 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
18793 spurious backslash.
18795 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
18796 Add missing line to comment describing function.
18798 2013-03-18 Richard Biener <rguenther@suse.de>
18800 PR tree-optimization/56210
18801 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18802 Handle string / character search functions.
18803 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
18805 2013-03-18 Richard Biener <rguenther@suse.de>
18807 PR middle-end/56483
18808 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
18809 and implement properly.
18810 * gimple.h (gimple_cond_single_var_p): Remove.
18812 2013-03-18 Richard Biener <rguenther@suse.de>
18814 * tree-data-ref.h (find_data_references_in_loop): Declare.
18815 * tree-data-ref.c (get_references_in_stmt): Use a stack
18816 vector pre-allocated in the callers.
18817 (find_data_references_in_stmt): Adjust.
18818 (graphite_find_data_references_in_stmt): Likewise.
18819 (create_rdg_vertices): Likewise.
18820 (find_data_references_in_loop): Export.
18821 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
18822 Compute dependences here...
18823 (vect_analyze_data_refs): ...not here. When we encounter
18824 a non-vectorizable data reference in basic-block vectorization
18825 truncate the data reference vector. Do not bother to
18826 fixup data-dependence information for gather loads.
18827 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
18828 of data references, as reported.
18830 2013-03-18 Richard Biener <rguenther@suse.de>
18832 PR tree-optimization/3713
18833 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
18834 has_constants and expr.
18835 (stmt_has_constants): Properly valueize SSA names when deciding
18836 whether the stmt has constants.
18838 2013-03-18 Richard Biener <rguenther@suse.de>
18840 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
18841 whole function when there is nothing to do.
18842 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
18843 * tree-vectorizer.c (vectorize_loops): Update virtual and
18844 loop-closed SSA once.
18845 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
18847 2013-03-18 Richard Biener <rguenther@suse.de>
18849 PR middle-end/56113
18850 * domwalk.c (bb_postorder): New global static.
18851 (cmp_bb_postorder): New function.
18852 (walk_dominator_tree): Replace scheme imposing an order for
18853 visiting dominator sons by one sorting them at the time they
18854 are pushed on the stack.
18856 2013-03-18 Richard Biener <rguenther@suse.de>
18858 PR tree-optimization/39326
18859 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
18860 (struct mem_ref): Replace mem member with ao_ref typed member.
18861 (MEM_ANALYZABLE): Adjust.
18862 (memref_eq): Likewise.
18863 (mem_ref_alloc): Likewise.
18864 (gather_mem_refs_stmt): Likewise.
18865 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
18866 (execute_sm_if_changed_flag_set): Adjust.
18867 (execute_sm): Likewise.
18868 (ref_always_accessed_p): Likewise.
18869 (refs_independent_p): Likewise.
18870 (can_sm_ref_p): Likewise.
18872 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18875 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
18876 return 1 even for !unsignedp.
18878 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18880 * config/i386/i386.md (isa): Add x64 and nox64.
18881 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
18882 (*pushtf): Enable *roF alternative for x64 isa only.
18883 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
18884 mode attribute of integer alternatives to DImode for TARGET_64BIT.
18885 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
18886 (*movtf_internal): Merge from *movtf_internal_rex64 and
18887 *movtf_internal_sse. Use x64 and nox64 isa attributes.
18888 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
18889 nox64 isa attributes.
18890 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
18891 nox64 isa attributes.
18892 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
18894 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18896 * config/alpha/alpha.c (TARGET_LRA_P): New define.
18898 2013-03-17 Jakub Jelinek <jakub@redhat.com>
18901 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
18902 class names. Remove trailing comma after "ALL_REGS".
18904 2013-03-16 Jan Hubicka <jh@suse.cz>
18906 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
18907 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
18908 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
18909 of cgraph_get_create_node.
18910 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
18912 2013-03-16 Jason Merrill <jason@redhat.com>
18915 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
18916 with DW_AT_default_value.
18918 2013-03-16 Jakub Jelinek <jakub@redhat.com>
18920 * BASE-VER: Set to 4.9.0.
18922 2013-03-14 Andi Kleen <ak@linux.intel.com>
18925 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
18926 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
18927 Document _x* TSX intrinsics.
18929 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
18930 David Holsgrove <david.holsgrove@xilinx.com>
18932 * configure.ac: Add MicroBlaze TLS support detection.
18933 * configure: Regenerate.
18934 * config/microblaze/microblaze-protos.h
18935 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
18936 symbol_mentioned_p, label_mentioned_p): Add prototypes.
18937 * config/microblaze/microblaze.c (microblaze_address_type): Add
18938 ADDRESS_TLS and tls_reloc address types.
18939 (microblaze_address_info): Add tls_reloc.
18940 (TARGET_HAVE_TLS): Define.
18941 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
18942 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
18943 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
18944 load_tls_operand, microblaze_call_tls_get_addr,
18945 microblaze_legitimize_tls_address): New functions.
18946 (microblaze_classify_unspec): Handle UNSPEC_TLS.
18947 (get_base_reg): Use microblaze_tls_symbol_p.
18948 (microblaze_classify_address): Handle TLS.
18949 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
18950 label_mentioned_p and microblaze_tls_referenced_p.
18951 (microblaze_legitimize_address): Handle TLS.
18952 (microblaze_address_insns): Handle ADDRESS_TLS.
18953 (pic_address_needs_scratch): Handle TLS.
18954 (print_operand_address): Handle TLS.
18955 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
18956 (microblaze_expand_move): Handle TLS.
18957 (microblaze_legitimate_constant_p): Check
18958 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
18959 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
18960 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
18961 (PIC_OFFSET_TABLE_REGNUM): Set.
18962 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
18963 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
18964 (addsi3, movsi_internal2, movdf_internal): Update constraints
18965 * config/microblaze/predicates.md (arith_plus_operand): Define
18966 (move_operand): Redefine as move_src_operand,
18967 check microblaze_tls_referenced_p.
18969 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18971 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
18972 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
18974 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18976 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
18979 2013-03-14 Jakub Jelinek <jakub@redhat.com>
18981 PR tree-optimization/53265
18982 * common.opt (Waggressive-loop-optimizations): New option.
18983 * tree-ssa-loop-niter.c: Include tree-pass.h.
18984 (do_warn_aggressive_loop_optimizations): New function.
18985 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
18986 if number_of_latch_executions returned constant.
18987 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
18988 early. If number_of_latch_executions returned constant, set
18989 nb_iterations_upper_bound back to it.
18990 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
18992 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
18993 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
18995 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
18996 (MULTILIB_OSDIRNAMES): Set.
18997 * genmultilib: If defaultosdirname doesn't start with :: , set
18998 defaultosdirname2 instead, clear it and emit two . multilib_raw
18999 entries instead of just one.
19001 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19003 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19004 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19005 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19006 (SUBTARGET_OVERRIDE_OPTIONS): New.
19008 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19011 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19012 (musermode): Convert to Var(TARGET_USERMODE).
19013 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19014 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19015 * config/sh/sh.c (sh_option_override): Use
19016 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19017 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19019 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19021 (udivsi3_i4_single, divsi3_i4_single): Use
19022 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19024 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19026 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19029 2013-03-13 Richard Biener <rguenther@suse.de>
19031 PR tree-optimization/56608
19032 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19033 calls when vectorizing basic-blocks.
19035 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19038 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19041 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19043 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19045 2013-03-11 Jan Hubicka <jh@suse.cz>
19048 * lto-streamer-out.c (output_symbol_p): Skip references from
19049 constructors of external variables.
19051 2013-03-11 Jan Hubicka <jh@suse.cz>
19053 PR middle-end/56571
19054 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19056 * emit-rtl.c (verify_rtx_sharing): Likewise.
19057 (copy_insn_1): Likewise.
19058 * rtl.c (copy_rtx): Likewise.
19060 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19063 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19064 output_operand_lossage message.
19066 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19069 * arm.c (shift_op): Validate RTL pattern on the fly.
19070 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19073 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19076 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19077 new call patterns that clobber %fr12.
19078 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19079 split and postreload patterns.
19080 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19081 registers %fr12 and %fr12R as call used.
19083 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19085 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19086 (canon_address, record_store, replace_read, check_mem_read_rtx,
19087 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19088 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19089 rest_of_handle_dse): Likewise.
19091 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19093 PR middle-end/56524
19094 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19096 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19097 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19098 (save_optabs_if_changed): Replace with...
19099 (init_tree_optimization_optabs): ...this.
19100 * optabs.c (save_optabs_if_changed): Rename to...
19101 (init_tree_optimization_optabs): ...this. Take the optimization node
19102 as argument. Do nothing if the base optabs are already correct.
19103 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19104 to recompute optabs.
19105 * function.h (function): Remove optabs field.
19106 * function.c (invoke_set_current_function_hook): Call
19107 init_tree_optimization_optabs. Use the result to initialize
19110 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19112 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19113 if GTMA_HAS_NO_INSTRUMENTATION.
19114 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19115 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19116 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19117 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19118 GTMA_HAS_NO_INSTRUMENTATION.
19120 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19122 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19125 2013-03-08 Marek Polacek <polacek@redhat.com>
19126 Jakub Jelinek <jakub@redhat.com>
19128 PR tree-optimization/56478
19129 * predict.c (is_comparison_with_loop_invariant_p): Change the
19130 type of loop_step to tree.
19131 (predict_loops): Adjust.
19132 (predict_iv_comparison): Perform the computations on double_ints.
19134 2013-03-08 Richard Biener <rguenther@suse.de>
19136 PR tree-optimization/56570
19137 * tree-cfg.c (verify_expr_location_1): Verify locations for
19139 * tree-sra.c (create_access_replacement): Strip locations
19140 from DECL_DEBUG_EXPRs.
19142 2013-03-08 Richard Biener <rguenther@suse.de>
19144 * tree-inline.c (expand_call_inline): Do not associate
19145 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19146 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19148 2013-03-08 Richard Biener <rguenther@suse.de>
19150 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19151 or block changes with -Og. Fix for location / block encoding
19152 changes and PHI arguments with locations.
19154 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19156 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19158 (struct output_info): Likewise.
19159 (register_overhead): Remove bad gcc_assert.
19160 (bitmap_find_bit): If there is only a single bitmap element, do not
19161 count a miss as a search.
19162 (print_statistics): Update for counter type changes.
19163 (dump_bitmap_statistics): Likewise. Print headers such that they
19164 are properly lined up with the printed counters.
19166 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19168 PR tree-optimization/56559
19169 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19170 check that it has only a single use.
19172 2013-03-07 Richard Biener <rguenther@suse.de>
19174 * doc/invoke.texi (fwhole-program): Discourage use in combination
19177 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19179 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19181 PR tree-optimization/56539
19182 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19183 instead of GSI_CONTINUE_LINKING as last argument to
19184 force_gimple_operand_gsi. Adjust function comment.
19186 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19189 PR middle-end/56548
19190 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19191 promoted mode, convert the result back to the original mode.
19193 2013-03-06 Richard Biener <rguenther@suse.de>
19195 PR middle-end/56294
19196 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19197 (insert_updated_phi_nodes_compare_uids): New function.
19198 (update_ssa): Sort symbols_to_rename after UID before
19199 traversing it to insert PHI nodes.
19201 2013-03-06 Richard Biener <rguenther@suse.de>
19203 PR middle-end/50494
19204 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19205 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19208 2013-02-13 Richard Biener <rguenther@suse.de>
19211 * varasm.c (output_constant_def_1): Get the decl representing
19212 the constant as argument.
19213 (output_constant_def): Wrap output_constant_def_1.
19214 (make_decl_rtl): Use output_constant_def_1 with the decl
19215 representing the constant.
19216 (build_constant_desc): Optionally re-use a decl already
19217 representing the constant.
19218 (tree_output_constant_def): Adjust.
19220 2013-03-06 Joey Ye <joey.ye@arm.com>
19223 * gcc.c (convert_white_space): New function.
19224 (main): Handles white space in function name.
19226 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19229 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19230 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19231 to SH_DIV_CALL_TABLE for TARGET_SH2.
19232 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19234 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19235 call-table options.
19237 2013-03-05 Sterling Augustine <saugustine@google.com>
19238 Cary Coutant <ccoutant@google.com>
19241 * dwarf2out.c (resolve_addr): Don't call
19242 remove_loc_list_addr_table_entries a second time for the same
19245 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19248 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19249 (avoid_complex_debug_insns): New function.
19250 (expand_debug_locations): Call it.
19252 PR rtl-optimization/56484
19253 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19254 lifetimes of hard registers on small register class machines.
19256 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19258 * config/microblaze/microblaze-protos.h: Rename
19259 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19260 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19262 (microblaze_fast_interrupt_function_p): New function.
19263 (microblaze_is_interrupt_handler): Rename to
19264 microblaze_is_interrupt_variant and add fast_interrupt check.
19265 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19266 (save_restore_insns): Likewise.
19267 (compute_frame_size): Likewise.
19268 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19269 (microblaze_globalize_label): Likewise.
19270 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19271 * config/microblaze/microblaze.md: Use wrapper
19272 microblaze_is_interrupt_variant.
19274 2013-03-05 Kai Tietz <ktietz@redhat.com>
19276 * sdbout.c (sdbout_one_type): Switch to current function's section
19277 supporting cold/hot.
19279 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19281 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19284 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19286 PR middle-end/56461
19287 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19288 if VALGRIND_GET_VBITS is defined, temporarily make object
19289 memory all defined, and restore previous valgrind addressability
19290 and definability afterwards. Free this_object at the end.
19292 PR middle-end/56461
19293 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19294 right before calling lra_create_live_ranges, also call it
19295 when clearing live_p. Only call lra_clear_live_ranges
19296 at the end if live_p.
19298 PR middle-end/56461
19299 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19301 2013-03-05 Richard Biener <rguenther@suse.de>
19303 PR tree-optimization/56521
19304 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19307 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19310 * except.h (remove_unreachable_eh_regions): New prototype.
19311 * except.c (remove_eh_handler_splicer): New function, split out
19312 of remove_eh_handler.
19313 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19314 warning about running it on many EH regions one at a time.
19315 (remove_unreachable_eh_regions_worker): New function, walk the
19316 EH tree in depth-first order and remove non-marked regions.
19317 (remove_unreachable_eh_regions): New function.
19318 * tree-eh.c (mark_reachable_handlers): New function, split out
19319 from remove_unreachable_handlers.
19320 (remove_unreachable_handlers): Use mark_reachable_handlers and
19321 remove_unreachable_eh_regions.
19322 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19323 and remove_unreachable_eh_regions.
19325 2013-03-05 Richard Biener <rguenther@suse.de>
19327 PR middle-end/56525
19328 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19329 not freeing them until the end.
19331 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19333 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19335 2013-03-05 Richard Biener <rguenther@suse.de>
19337 PR tree-optimization/56270
19338 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19339 of loads after scheduling an SLP instance.
19341 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19343 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19345 (check_gcc_parallelize): Run guality.exp as a separate job from
19346 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19347 struct-layout-1.exp with stackalign.exp.
19349 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19351 PR middle-end/56461
19352 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19353 load_index sbitmap even if some bit in it isn't set.
19355 PR middle-end/56461
19356 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19357 (discover_iteration_bound_by_body_walk): Change queues to
19358 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19359 spelling in comment. Call safe_push on queues[bound_index] directly.
19360 Release queues[queue_index] in every iteration unconditionally.
19361 Release bounds vector.
19363 PR middle-end/56461
19364 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19365 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19366 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19369 2013-03-05 Richard Biener <rguenther@suse.de>
19372 * tree-inline.c (remap_blocks_to_null): New function.
19373 (expand_call_inline): When expanding a call stmt without
19374 an associated block inline remap all callee blocks to NULL.
19376 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19378 PR rtl-optimization/56494
19379 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19380 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19381 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19383 PR middle-end/56461
19384 * sel-sched-ir.c (free_sched_pools): Release
19385 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19386 if succs_info_pool.max_top isn't -1.
19389 * opts.c (opts_obstack, opts_concat): Moved to...
19390 * opts-common.c (opts_obstack, opts_concat): ... here.
19392 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19394 PR middle-end/56461
19395 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19397 2013-03-04 Martin Jambor <mjambor@suse.cz>
19399 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19400 all appropriate places.
19402 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19404 PR tree-optimization/56424
19405 * ipa-split.c (split_function): Do not set the RSO flag if result is
19406 not by reference and its type is a register type.
19408 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19410 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19411 (microblaze_legitimate_pic_operand): Likewise
19412 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19413 new function microblaze_legitimate_pic_operand
19414 * config/microblaze/microblaze-protos.h
19415 (microblaze_legitimate_pic_operand): Declare.
19417 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19419 * config/microblaze/predicates.md (call_insn_simple_operand):
19420 New predicate for supported rtx code types.
19421 * config/microblaze/microblaze.md (call_internal1): Use
19422 call_insn_simple_operand predicate.
19424 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19426 PR middle-end/56461
19427 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19428 partitions.ordered_remove.
19430 PR middle-end/56461
19431 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19432 vec_oprnds0.create (1) for modifier == NONE.
19434 PR middle-end/56461
19435 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19436 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19437 vec_oprnds1 right before pushing anything to it for
19440 PR middle-end/56461
19441 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19442 set nbbs to 0 instead of having separate code path.
19443 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19444 instead of false as last argument if returning NULL.
19446 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19448 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19449 the attribute is now called "target" instead of "option".
19450 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19451 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19452 attribute/pragma name for TARGET_OPTION_VALID_P and
19453 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19454 * doc/tm.texi: Regenerated.
19456 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19458 * config/microblaze/microblaze.c:
19459 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19460 * config/microblaze/microblaze.h: Add -mxl-reorder to
19462 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19463 instructions emitted if TARGET_REORDER.
19464 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19465 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19468 2013-03-01 Xinliang David Li <davidxl@google.com>
19470 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19473 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19475 PR middle-end/56461
19476 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19477 vector even when returning true. Fix up function comment formatting.
19479 PR middle-end/56461
19480 * ira-build.c (ira_loop_nodes_count): New variable.
19481 (create_loop_tree_nodes): Initialize it.
19482 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19484 PR middle-end/56461
19485 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19486 method on dr_chain and result_chain.
19487 * tree-vect-stmts.c (vectorizable_store): Only call
19488 result_chain.create if j == 0.
19490 PR middle-end/56461
19491 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19492 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19493 before overwriting it.
19495 2013-03-01 Tobias Burnus <burnus@net-b.de>
19497 * doc/extended.texi (C Extensions): Change order in @menu
19499 (Other MIPS Built-in Functions): Move last MIPS entry before
19500 "picoChip Built-in Functions".
19501 (SH Built-in Functions): Move after RX Built-in Functions.
19502 * doc/gcc.texi (Introduction): Change order in @menu
19504 * doc/md.texi (Constraints): Ditto.
19505 * gty.texi (Type Information): Ditto.
19506 (User-provided marking routines for template types): Make
19508 * doc/invoke.texi (AArch64 Options): Move before
19509 "Adapteva Epiphany Options".
19511 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19512 Jakub Jelinek <jakub@redhat.com>
19515 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19516 no_address_safety_analysis attribute.
19517 * doc/extend.texi (no_address_safety_attribute): Rename to
19518 no_sanitize_address attribute, mention no_address_safety_analysis
19519 attribute as deprecated alias.
19521 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19523 PR middle-end/56461
19524 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19525 type to vec<vec<tree> > *.
19526 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19527 to be vec<tree> instead of vec<tree> *, set vec_defs
19528 to vNULL and call vec_defs.create (number_of_vects), adjust other
19530 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19531 vectorizable_condition): Adjust vect_get_slp_defs callers.
19533 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19535 * config/aarch64/aarch64.c
19536 (aarch64_float_const_representable): Remove unused variable.
19538 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19540 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19542 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19544 * config/aarch64/aarch64-builtins.c
19545 (aarch64_init_simd_builtins): Make static.
19547 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19549 * config/aarch64/aarch64.c
19550 (aarch64_simd_make_constant): Make static.
19552 2013-02-28 Martin Jambor <mjambor@suse.cz>
19554 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19555 with no initialization to the RHS of debug statements.
19557 2013-02-28 Martin Jambor <mjambor@suse.cz>
19559 PR tree-optimization/56294
19560 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19562 (get_access_replacement): Do not call create_access_replacement.
19563 Assert a replacement exists.
19564 (get_repl_default_def_ssa_name): Create the replacement declaration
19567 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19569 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19570 final_end_function.
19572 2013-02-28 Marek Polacek <polacek@redhat.com>
19574 PR rtl-optimization/56466
19575 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19576 if we're changing a loop.
19577 (peel_loops_completely): Likewise.
19579 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19582 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19584 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19587 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19588 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19589 INTX_FTYPE_FX, FX_FTYPE_INTX.
19590 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
19592 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19594 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
19595 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
19596 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19597 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
19598 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
19599 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
19600 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
19601 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
19602 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
19603 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
19604 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
19605 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
19606 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
19607 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
19608 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
19609 (avrxmega6): Increase max flash segments from 5 to 6.
19610 * config/avr/t-multilib: Regenerate.
19611 * config/avr/avr-tables.opt: Regenerate.
19612 * doc/avr-mmcu.texi: Regenerate.
19614 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19616 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
19617 (avr_device_to_arch): Rename to avr_device_to_ld.
19618 (avr_device_to_as): New prototype.
19619 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
19620 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
19621 * config/avr/driver-avr.c (avr_device_to_as): New.
19622 (avr_device_to_arch): Rename to avr_device_to_ld.
19624 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19626 PR middle-end/56461
19627 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
19628 method on dr_chain and result_chain.
19630 PR middle-end/56461
19631 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
19632 pointer_set_destroy on not_executed_last_iteration.
19634 PR middle-end/56461
19635 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
19637 PR middle-end/56461
19638 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
19639 FOR_EACH_DEFINED_FUNCTION when freeing state.
19641 PR middle-end/56461
19642 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
19644 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
19647 PR middle-end/56461
19648 * ipa-cp.c (decide_whether_version_node): Call vec_free on
19649 known_aggs[i].items and release known_aggs vector.
19651 PR middle-end/56461
19652 * ipa-reference.c (propagate): Free node_info even for alias nodes.
19654 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19656 * config/microblaze/microblaze.c (microblaze_emit_compare):
19657 Use xor for EQ/NE comparisions.
19658 * config/microblaze/microblaze.md (cstoresf4): Add constraints
19659 (cbranchsf4): Adjust operator to comparison_operator.
19661 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19663 PR middle-end/56461
19664 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
19666 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
19667 vec_safe_push, always update *slot.
19668 (redirect_edge_var_map_clear): Use vec_free.
19669 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
19670 (free_var_map_entry): Use vec_free.
19671 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
19672 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
19674 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
19676 PR middle-end/45472
19677 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
19678 when the may_trap_p bit of the exprs being merged differs.
19679 Reorder tests for speculativeness in the logical and operator.
19681 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19683 * incpath.c (add_standard_paths): Use reconcat instead of concat
19684 where appropriate and avoid leaking memory.
19686 * opts.h: Include obstack.h.
19687 (opts_concat): New prototype.
19688 (opts_obstack): New declaration.
19689 * opts.c (opts_concat): New function.
19690 (opts_obstack): New variable.
19691 (init_options_struct): Call gcc_init_obstack on opts_obstack.
19692 (finish_options): Use opts_concat instead of concat
19693 and XOBNEWVEC instead of XNEWVEC.
19694 * opts-common.c (generate_canonical_option, decode_cmdline_option,
19695 generate_option): Likewise.
19696 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
19697 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
19700 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
19701 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
19703 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19705 PR middle-end/56461
19706 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
19708 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19710 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
19711 (arm_block_move_unaligned_straight): Likewise.
19712 (arm_adjust_block_mem): Likewise.
19714 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
19717 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
19718 temp, cond and label.
19719 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
19722 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
19723 * config/c6x/c6x.h (dbx_register_map): Update declaration.
19726 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
19727 of prologue/epilogue functions.
19730 * config/tilegx/tilegx.c (tilegx_expand_prologue):
19731 Remove unused variable cfa_offset.
19732 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
19735 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
19736 type promotion to unsigned.
19739 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
19740 for HOST_WIDE_INT of 32 bit / same size as int.
19741 (arm_block_move_unaligned_straight): Likewise.
19742 (arm_adjust_block_mem): Likewise.
19745 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
19748 2013-02-26 Marek Polacek <polacek@redhat.com>
19750 PR tree-optimization/56426
19751 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
19753 2013-02-26 Richard Biener <rguenther@suse.de>
19756 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
19757 unused variable loops.
19759 2013-02-26 Jakub Jelinek <jakub@redhat.com>
19761 PR tree-optimization/56448
19762 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
19763 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
19764 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
19765 later operands of the references, or even first operand for
19766 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
19768 PR tree-optimization/56443
19769 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
19770 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
19771 to type_for_mode langhook.
19773 2013-02-25 Matt Turner <mattst88@gmail.com>
19775 * doc/invoke.texi: Document r4700.
19777 2013-02-25 Richard Biener <rguenther@suse.de>
19779 PR tree-optimization/56175
19780 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
19782 (simplify_bitwise_binary): ... here. Also guard the conversion
19783 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
19785 2013-02-25 Catherine Moore <clm@codesourcery.com>
19788 2013-02-24 Catherine Moore <clm@codesourcery.com>
19789 Maciej W. Rozycki <macro@codesourcery.com>
19790 Tom de Vries <tom@codesourcery.com>
19791 Nathan Sidwell <nathan@codesourcery.com>
19792 Iain Sandoe <iain@codesourcery.com>
19793 Nathan Froyd <froydnj@codesourcery.com>
19794 Chao-ying Fu <fu@mips.com>
19796 * doc/extend.texi: (micromips, nomicromips, nocompression):
19797 Document new function attributes.
19798 * doc/invoke.texi (minterlink-compressed, mmicromips,
19799 m14k, m14ke, m14kec): Document new options.
19800 (minterlink-mips16): Update documentation.
19801 * doc/md.texi (ZC, ZD): Document new constraints.
19802 * configure.ac (gcc_cv_as_micromips): Check if linker
19803 supports the .set micromips directive.
19804 * configure: Regenerate.
19805 * config.in: Regenerate.
19806 * config/mips/mips-tables.opt: Regenerate.
19807 * config/mips/micromips.md: New file.
19808 * constraints.md (ZC, AD): New constraints.
19809 * config/mips/predicates.md (movep_src_register): New predicate.
19810 (movep_src_operand): New predicate.
19811 (non_volatile_mem_operand): New predicate.
19812 * config/mips/mips.md (multimem): New type.
19813 (length): Differentiate between 17-bit and 18-bit branch offsets.
19814 (MOVEP1, MOVEP2): New mode iterator.
19815 (mov_<load>l): Use ZC constraint.
19816 (mov_<load>r): Likewise.
19817 (mov_<store>l): Likewise.
19818 (mov_<store>r): Likewise.
19819 (*branch_equality<mode>_inverted): Add microMIPS support.
19820 (*branch_equality<mode>): Likewise.
19821 (*jump_absolute): Likewise.
19822 (indirect_jump_<mode>): Likewise.
19823 (tablejump_<mode>): Likewise.
19824 (<optab>_internal): Likewise.
19825 (sibcall_internal): Likewise.
19826 (sibcall_value_internal): Likewise.
19827 (prefetch): Use constraint ZD.
19828 * config/mips/mips.opt (minterlink-compressed): New option.
19829 (minterlink-mips16): Now an alias for minterlink-compressed.
19830 (mmicromips): New option.
19831 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19832 (compare_and_swap_12): Likewise.
19833 (sync_add<mode>): Likewise.
19834 (sync_<optab>_12): Likewise.
19835 (sync_old_<optab>_12): Likewise.
19836 (sync_new_<optab>_12): Likewise.
19837 (sync_nand_12): Likewise.
19838 (sync_old_nand_12): Likewise.
19839 (sync_new_nand_12): Likewise.
19840 (sync_sub<mode>): Likewise.
19841 (sync_old_add<mode>): Likewise.
19842 (sync_old_sub<mode>): Likewise.
19843 (sync_new_add<mode>): Likewise.
19844 (sync_new_sub<mode>): Likewise.
19845 (sync_<optab><mode>): Likewise.
19846 (sync_old_<optab><mode>): Likewise.
19847 (sync_new_<optab><mode>): Likewise.
19848 (sync_nand<mode>): Likewise.
19849 (sync_old_nand<mode>): Likewise.
19850 (sync_new_nand<mode>): Likewise.
19851 (sync_lock_test_and_set<mode>): Likewise.
19852 (test_and_set_12): Likewise.
19853 (atomic_compare_and_swap<mode>): Likewise.
19854 (atomic_exchange<mode>_llsc): Likewise.
19855 (atomic_fetch_add<mode>_llsc): Likewise.
19856 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19857 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19858 (umips_save_restore_pattern_p): Likewise.
19859 (umips_load_store_pair_p): Likewise.
19860 (umips_output_load_store_pair): Likewise.
19861 (umips_movep_target_p): Likewise.
19862 (umips_12bit_offset_address_p): Likewise.
19863 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19864 (mips_base_mips16): Rename this...
19865 (mips_base_compression_flags): ...to this. Update all uses.
19866 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19867 (mips_mips16_decl_p): Delete.
19868 (mips_nomips16_decl_p): Delete.
19869 (mips_get_compress_on_flags): New function.
19870 (mips_get_compress_off_flags): New function.
19871 (mips_get_compress_mode): New function.
19872 (mips_get_compress_on_name): New function.
19873 (mips_get_compress_off_name): New function.
19874 (mips_insert_attributes): Support multiple compression types.
19875 (mips_merge_decl_attributes): Likewise.
19876 (umips_12bit_offset_address_p): New function.
19877 (mips_start_function_definition): Emit .set micromips directive.
19878 (mips_call_may_need_jalx_p): New function.
19879 (mips_function_ok_for_sibcall): Add microMIPS support.
19880 (mips_print_operand_punctuation): Support short delay slots and
19882 (umips_swm_mask, umips_swm_encoding): New.
19883 (umips_build_save_restore): New function.
19884 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19885 (was_mips16_p): Remove.
19886 (old_compression_mode): New.
19887 (mips_set_compression_mode): New function.
19888 (mips_set_current_function): Add microMIPS support.
19889 (mips_option_override): Likewise.
19890 (umips_save_restore_pattern_p): New function.
19891 (umips_output_save_restore): New function.
19892 (umips_load_store_pair_p_1): New function.
19893 (umips_load_store_pair_p): New function.
19894 (umips_output_load_store_pair_1): New function.
19895 (umips_output_load_store_pair): New function.
19896 (umips_movep_target_p) New function.
19897 (mips_prepare_pch_save): Add microMIPS support.
19898 * config/mips/mips.h (TARGET_COMPRESSION): New.
19899 (TARGET_CPU_CPP_BUILTINS): Update macro
19900 to use new compression flags and to support microMIPS.
19901 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19902 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19903 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19904 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19905 (ASM_SPEC): Support mmicromips and mno-micromips.
19906 (M16STORE_REG_P): New macro.
19907 (MIPS_CALL): Support TARGET_MICROMIPS.
19908 (MICROMIPS_J): New macro.
19909 (mips_base_mips16): Rename this...
19910 (mips_base_compression_flags): ...to this.
19911 (UMIPS_12BIT_OFFSET_P): New macro.
19912 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19913 (MULTILIB_DIRNAMES): Likewise.
19915 2013-02-25 Tom de Vries <tom@codesourcery.com>
19917 PR rtl-optimization/56131
19918 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
19919 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
19920 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
19922 2013-02-25 Tobias Burnus <burnus@net-b.de>
19924 * doc/invoke.texi (-fsanitize=): Move from optimization
19925 to debugging options.
19927 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19929 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
19931 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19932 Alexander Monakov <amonakov@ispras.ru>
19934 PR middle-end/56077
19935 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
19936 flush pending lists also on non-jumps. Adjust comment.
19938 2013-02-24 Catherine Moore <clm@codesourcery.com>
19939 Maciej W. Rozycki <macro@codesourcery.com>
19940 Tom de Vries <tom@codesourcery.com>
19941 Nathan Sidwell <nathan@codesourcery.com>
19942 Iain Sandoe <iain@codesourcery.com>
19943 Nathan Froyd <froydnj@codesourcery.com>
19944 Chao-ying Fu <fu@mips.com>
19946 * doc/extend.texi: (micromips, nomicromips, nocompression):
19947 Document new function attributes.
19948 * doc/invoke.texi (minterlink-compressed, mmicromips,
19949 m14k, m14ke, m14kec): Document new options.
19950 (minterlink-mips16): Update documentation.
19951 * doc/md.texi (ZC, ZD): Document new constraints.
19952 * configure.ac (gcc_cv_as_micromips): Check if linker
19953 supports the .set micromips directive.
19954 * configure: Regenerate.
19955 * config.in: Regenerate.
19956 * config/mips/mips-tables.opt: Regenerate.
19957 * config/mips/micromips.md: New file.
19958 * constraints.md (ZC, AD): New constraints.
19959 * config/mips/predicates.md (movep_src_register): New predicate.
19960 (movep_src_operand): New predicate.
19961 (non_volatile_mem_operand): New predicate.
19962 * config/mips/mips.md (multimem): New type.
19963 (length): Differentiate between 17-bit and 18-bit branch offsets.
19964 (MOVEP1, MOVEP2): New mode iterator.
19965 (mov_<load>l): Use ZC constraint.
19966 (mov_<load>r): Likewise.
19967 (mov_<store>l): Likewise.
19968 (mov_<store>r): Likewise.
19969 (*branch_equality<mode>_inverted): Add microMIPS support.
19970 (*branch_equality<mode>): Likewise.
19971 (*jump_absolute): Likewise.
19972 (indirect_jump_<mode>): Likewise.
19973 (tablejump_<mode>): Likewise.
19974 (<optab>_internal): Likewise.
19975 (sibcall_internal): Likewise.
19976 (sibcall_value_internal): Likewise.
19977 (prefetch): Use constraint ZD.
19978 * config/mips/mips.opt (minterlink-compressed): New option.
19979 (minterlink-mips16): Now an alias for minterlink-compressed.
19980 (mmicromips): New option.
19981 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19982 (compare_and_swap_12): Likewise.
19983 (sync_add<mode>): Likewise.
19984 (sync_<optab>_12): Likewise.
19985 (sync_old_<optab>_12): Likewise.
19986 (sync_new_<optab>_12): Likewise.
19987 (sync_nand_12): Likewise.
19988 (sync_old_nand_12): Likewise.
19989 (sync_new_nand_12): Likewise.
19990 (sync_sub<mode>): Likewise.
19991 (sync_old_add<mode>): Likewise.
19992 (sync_old_sub<mode>): Likewise.
19993 (sync_new_add<mode>): Likewise.
19994 (sync_new_sub<mode>): Likewise.
19995 (sync_<optab><mode>): Likewise.
19996 (sync_old_<optab><mode>): Likewise.
19997 (sync_new_<optab><mode>): Likewise.
19998 (sync_nand<mode>): Likewise.
19999 (sync_old_nand<mode>): Likewise.
20000 (sync_new_nand<mode>): Likewise.
20001 (sync_lock_test_and_set<mode>): Likewise.
20002 (test_and_set_12): Likewise.
20003 (atomic_compare_and_swap<mode>): Likewise.
20004 (atomic_exchange<mode>_llsc): Likewise.
20005 (atomic_fetch_add<mode>_llsc): Likewise.
20006 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20007 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20008 (umips_save_restore_pattern_p): Likewise.
20009 (umips_load_store_pair_p): Likewise.
20010 (umips_output_load_store_pair): Likewise.
20011 (umips_movep_target_p): Likewise.
20012 (umips_12bit_offset_address_p): Likewise.
20013 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20014 (mips_base_mips16): Rename this...
20015 (mips_base_compression_flags): ...to this. Update all uses.
20016 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20017 (mips_mips16_decl_p): Delete.
20018 (mips_nomips16_decl_p): Delete.
20019 (mips_get_compress_on_flags): New function.
20020 (mips_get_compress_off_flags): New function.
20021 (mips_get_compress_mode): New function.
20022 (mips_get_compress_on_name): New function.
20023 (mips_get_compress_off_name): New function.
20024 (mips_insert_attributes): Support multiple compression types.
20025 (mips_merge_decl_attributes): Likewise.
20026 (umips_12bit_offset_address_p): New function.
20027 (mips_start_function_definition): Emit .set micromips directive.
20028 (mips_call_may_need_jalx_p): New function.
20029 (mips_function_ok_for_sibcall): Add microMIPS support.
20030 (mips_print_operand_punctuation): Support short delay slots and
20032 (umips_swm_mask, umips_swm_encoding): New.
20033 (umips_build_save_restore): New function.
20034 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20035 (was_mips16_p): Remove.
20036 (old_compression_mode): New.
20037 (mips_set_compression_mode): New function.
20038 (mips_set_current_function): Add microMIPS support.
20039 (mips_option_override): Likewise.
20040 (umips_save_restore_pattern_p): New function.
20041 (umips_output_save_restore): New function.
20042 (umips_load_store_pair_p_1): New function.
20043 (umips_load_store_pair_p): New function.
20044 (umips_output_load_store_pair_1): New function.
20045 (umips_output_load_store_pair): New function.
20046 (umips_movep_target_p) New function.
20047 (mips_prepare_pch_save): Add microMIPS support.
20048 * config/mips/mips.h (TARGET_COMPRESSION): New.
20049 (TARGET_CPU_CPP_BUILTINS): Update macro
20050 to use new compression flags and to support microMIPS.
20051 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20052 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20053 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20054 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20055 (ASM_SPEC): Support mmicromips and mno-micromips.
20056 (M16STORE_REG_P): New macro.
20057 (MIPS_CALL): Support TARGET_MICROMIPS.
20058 (MICROMIPS_J): New macro.
20059 (mips_base_mips16): Rename this...
20060 (mips_base_compression_flags): ...to this.
20061 (UMIPS_12BIT_OFFSET_P): New macro.
20062 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20063 (MULTILIB_DIRNAMES): Likewise.
20065 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20068 * target-globals.c (save_target_globals): For init_reg_sets and
20069 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20071 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20073 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20074 * config/aarch64/t-aarch64
20075 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20077 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20079 PR inline-asm/56148
20080 * lra-constraints.c (process_alt_operands): Reload operand
20081 conflicting with earlier clobber only if no more other conflicting
20084 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20087 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20088 if not linking a shared library.
20090 2013-02-22 Seth LaForge <sethml@google.com>
20092 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20094 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20096 * config/arm/arm.md (split for extendsidi): Update condition.
20097 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20098 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20099 (qhs_zextenddi_cstr): Likewise.
20101 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20103 PR middle-end/56420
20104 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20105 avoid signed wrapping.
20106 (expand_mult): Handle properly multiplication by
20107 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20108 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20109 in the compiler if coeff is HOST_WIDE_INT_MIN.
20110 (expand_divmod): Don't make ext_op1 static, change it's type to
20111 uhwi. Avoid undefined behavior in -INTVAL (op1).
20113 PR rtl-optimization/50339
20114 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20116 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20117 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20118 into splitting_ashiftrt field.
20119 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20121 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20124 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20126 PR middle-end/56108
20127 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20128 are sure to go irrevocable.
20130 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20132 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20133 scalars are valid operands.
20135 2013-02-21 Martin Jambor <mjambor@suse.cz>
20137 PR tree-optimization/56310
20138 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20139 only matching indices and non-negative final offsets.
20140 (intersect_aggregates_with_edge): Pass src_idx to
20141 agg_replacements_to_vector. Pass src_idx insstead of index to
20142 intersect_with_agg_replacements.
20144 2013-02-21 Martin Jambor <mjambor@suse.cz>
20146 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20147 instead of hard-wired defaults.
20149 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20151 * doc/invoke.texi (MIPS Options): Update documentation of the
20152 floating-point multiply-accumulate instruction restrictions.
20154 2013-02-21 Kostya Serebryany <kcc@google.com>
20156 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20157 asan_shadow_offset on x86_64 linux.
20159 2013-02-21 Richard Biener <rguenther@suse.de>
20161 PR tree-optimization/56415
20163 2013-02-11 Richard Biener <rguenther@suse.de>
20165 PR tree-optimization/56273
20166 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20169 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20172 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20175 PR inline-asm/56405
20176 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20177 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20179 2013-02-20 Jan Hubicka <jh@suse.cz>
20181 PR tree-optimization/56265
20182 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20183 when target is referenced for first time.
20185 2013-02-20 Richard Biener <rguenther@suse.de>
20187 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20188 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20189 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20190 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20191 not return anything.
20192 (rename_ssa_copies): Do not remove unused locals.
20193 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20194 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20195 * passes.c (execute_function_todo): Do not schedule unused locals
20196 removal if cleanup_tree_cfg did something.
20197 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20198 about the number of removed locals.
20200 2013-02-20 Richard Biener <rguenther@suse.de>
20202 PR tree-optimization/56398
20203 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20205 2013-02-20 Martin Jambor <mjambor@suse.cz>
20207 PR tree-optimization/55334
20208 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20209 restricted pointers to arrays.
20211 2013-02-20 Richard Biener <rguenther@suse.de>
20212 Jakub Jelinek <jakub@redhat.com>
20214 PR tree-optimization/56396
20215 * tree-ssa-ccp.c (n_const_val): New static variable.
20216 (get_value): Return NULL for SSA names we don't have a lattice
20218 (ccp_initialize): Initialize n_const_val.
20219 * tree-ssa-copy.c (n_copy_of): New static variable.
20220 (init_copy_prop): Initialize n_copy_of.
20221 (get_value): Return NULL_TREE for SSA names we don't have a
20224 2013-02-20 Martin Jambor <mjambor@suse.cz>
20226 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20228 2013-02-20 Richard Biener <rguenther@suse.de>
20230 * genpreds.c (write_lookup_constraint): Do not compare first
20231 letter of the constraint again.
20233 2013-02-20 Richard Biener <rguenther@suse.de>
20235 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20237 (get_use_iv_cost): Terminate hashtable walk when coming across
20240 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20242 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20243 reassociation for avx2 targets.
20245 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20247 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20248 Add version check for v8.10.a to enable microblaze_has_clz
20249 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20250 version and TARGET_PATTERN_COMPARE check
20251 * config/microblaze/microblaze.md: New clzsi2 instruction
20253 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20255 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20256 function before branching.
20258 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20260 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20262 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20264 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20266 PR middle-end/55889
20267 * sel-sched.c: Include ira.h.
20268 (implicit_clobber_conflict_p): New function.
20269 (moveup_expr): Use it.
20270 * Makefile.in (sel-sched.o): Depend on ira.h.
20272 2013-02-19 Richard Biener <rguenther@suse.de>
20274 PR tree-optimization/56384
20275 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20276 (vn_hash_type): Split out from ...
20277 (vn_hash_constant_with_type): ... here.
20278 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20279 (vn_phi_eq): Compare types from vn_phi_s structure.
20280 (vn_phi_lookup): Populate vn_phi_s type.
20281 (vn_phi_insert): Likewise.
20283 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20285 PR tree-optimization/56350
20286 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20287 if haven't found reduction or nested cycle operand, rather than
20288 asserting we must find it.
20290 PR tree-optimization/56381
20291 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20294 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20295 Jakub Jelinek <jakub@redhat.com>
20298 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20299 (swap_optab_enable): Same.
20300 (init_all_optabs): Use argument instead of global.
20301 * tree.h (struct tree_optimization_option): New field target_optabs.
20302 * expr.h (init_all_optabs): Add argument to prototype.
20303 (TREE_OPTIMIZATION_OPTABS): New.
20304 (save_optabs_if_changed): Protoize.
20305 * optabs.h: Declare this_fn_optabs.
20306 * optabs.c (save_optabs_if_changed): New.
20307 Declare this_fn_optabs.
20308 (init_optabs): Add argument to init_all_optabs() call.
20309 * function.c (invoke_set_current_function_hook): Handle per
20311 * function.h (struct function): New field optabs.
20312 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20313 optimization_current_node has changed.
20314 * target-globals.h (save_target_globals_default_opts): Protoize.
20315 * target-globals.c (save_target_globals_default_opts): New.
20317 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20320 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20321 registers %fr12 and %fr12R as call used.
20324 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20325 and HImode, require all displacements to be an integer multiple of
20327 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20328 only allow QImode and HImode when reload is in progress and strict is
20329 true. Likewise for symbolic addresses. Use base14_operand to check
20330 displacements in REG+BASE addresses.
20332 2013-02-18 Richard Biener <rguenther@suse.de>
20334 PR tree-optimization/56366
20335 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20336 handle sign-conversion of outer-loop initial induction value.
20338 2013-02-18 Richard Biener <rguenther@suse.de>
20340 PR middle-end/56349
20341 * cfghooks.c (merge_blocks): If we merge a latch into another
20342 block adjust references to it.
20343 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20344 (verify_loop_structure): Verify that a recorded latch is in fact
20347 2013-02-18 Richard Biener <rguenther@suse.de>
20349 PR tree-optimization/56321
20350 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20351 order SSA name release and virtual operand unlinking.
20353 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20355 * config/microblaze/microblaze.md (save_stack_block): Define.
20356 (restore_stack_block): Likewise.
20358 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20360 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20361 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20362 * config/microblaze/microblaze.c (microblaze_option_override):
20363 Bail out early for PIC modes when target does not support PIC.
20365 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20367 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20368 Replace with a microblaze version.
20369 (microblaze_trampoline_init): Adapt for microblaze.
20370 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20373 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20374 Dodji Seketeli <dodji@redhat.com>
20377 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20378 (instrument_mem_region_access): Do not forget to always put
20379 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20380 0) statement, even for cases where either 'base' or 'base + len'
20381 are not instrumented -- because they have been previously
20382 instrumented. Simplify the logic by putting all the statements
20383 instrument 'base + len' inside a sequence, and then insert that
20384 sequence right before the current insertion point. Then, to
20385 instrument 'base + len', just get an iterator on that statement.
20386 And do not forget to update the pointer to iterator the function
20387 received as argument.
20389 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20391 PR rtl-optimization/56348
20392 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20394 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20396 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20397 (clean_graph_dump_file): Pass base to start_graph_dump.
20399 2013-02-14 Richard Henderson <rth@redhat.com>
20402 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20404 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20406 * collect2-aix.h: Define F_LOADONLY.
20408 2013-02-14 Richard Biener <rguenther@suse.de>
20411 * varasm.c (output_constant_def_1): Get the decl representing
20412 the constant as argument.
20413 (output_constant_def): Wrap output_constant_def_1.
20414 (make_decl_rtl): Use output_constant_def_1 with the decl
20415 representing the constant.
20416 (build_constant_desc): Optionally re-use a decl already
20417 representing the constant.
20418 (tree_output_constant_def): Adjust.
20420 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20423 * asan.c (instrument_builtin_call): Really put the length of the
20424 second source argument into src1_len.
20426 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20428 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20429 argument. If it is false, don't create edge from then_bb to
20431 (insert_if_then_before_iter): Pass true to it.
20432 (build_check_stmt): Pass false to it.
20433 (transform_statements): Flush hash table only on extended basic
20434 block boundaries, rather than at the beginning of every bb.
20435 Don't flush hash table on nonfreeing_call_p calls.
20436 * tree-flow.h (nonfreeing_call_p): New prototype.
20437 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20439 2013-02-13 David S. Miller <davem@davemloft.net>
20441 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20443 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20446 * ira.c (max_regno_before_ira): Move from ...
20448 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20451 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20453 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20455 2013-02-13 Richard Biener <rguenther@suse.de>
20458 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20459 globals in MEM_REFs.
20461 2013-02-13 Richard Biener <rguenther@suse.de>
20463 * loop-init.c (loop_optimizer_init): Clear loop state when
20464 re-initializing preserved loops.
20465 * loop-unswitch.c (unswitch_single_loop): Return whether
20466 we unswitched the loop. Do not verify loop state here.
20467 (unswitch_loops): When we unswitched a loop discover new loops.
20469 2013-02-13 Kostya Serebryany <kcc@google.com>
20471 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20473 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20475 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20477 Avoid instrumenting duplicated memory access in the same basic block
20478 * Makefile.in (asan.o): Add new dependency on hash-table.h
20479 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20480 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20481 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20482 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20483 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20484 (get_mem_ref_of_assignment): New functions.
20485 (get_mem_refs_of_builtin_call): Extract from
20486 instrument_builtin_call and tweak a little bit to make it fit with
20488 (instrument_builtin_call): Use the new
20489 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20490 of is_gimple_builtin_call.
20491 (instrument_derefs, instrument_mem_region_access): Insert the
20492 instrumented memory reference into the hash table.
20493 (maybe_instrument_assignment): Renamed instrument_assignment into
20494 this, and change it to advance the iterator when instrumentation
20495 actually happened and return true in that case. This makes it
20496 homogeneous with maybe_instrument_assignment, and thus give a
20497 chance to callers to be more 'regular'.
20498 (transform_statements): Clear the memory reference hash table
20499 whenever we enter a new BB, when we cross a function call, or when
20500 we are done transforming statements. Use
20501 maybe_instrument_assignment instead of instrumentation. No more
20502 need to special case maybe_instrument_assignment and advance the
20503 iterator after calling it; it's now handled just like
20504 maybe_instrument_call. Update comment.
20506 2013-02-13 Richard Biener <rguenther@suse.de>
20508 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20509 Fix loop discovery code.
20511 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20513 PR inline-asm/56148
20514 * lra-constraints.c (process_alt_operands): Match early clobber
20515 operand with itself. Check conflicts with earlyclobber only if
20516 the operand is not reloaded. Prefer to reload conflicting operand
20517 if earlyclobber and matching operands are the same.
20519 2013-02-12 Richard Biener <rguenther@suse.de>
20522 * lto-streamer-out.c (write_symbol): Do not output symbols
20523 for hard register variables.
20525 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20528 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20529 (umulsidi3_insn, mulsidi3_insn): New insns.
20531 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20533 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20534 (struct tune_params): Add vec_costs field.
20535 * config/arm/arm.c (arm_builtin_vectorization_cost)
20536 (arm_add_stmt_cost): New functions.
20537 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20538 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20539 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20540 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20541 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20542 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20543 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20545 2013-02-12 Richard Biener <rguenther@suse.de>
20548 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20549 decls again if possible.
20551 2013-02-12 Richard Biener <rguenther@suse.de>
20553 PR middle-end/56288
20554 * tree-ssa.c (verify_ssa_name): Fix check, move
20555 SSA_NAME_IN_FREE_LIST check up.
20557 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20558 Steven Bosscher <steven@gcc.gnu.org>
20560 PR rtl-optimization/56151
20561 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20562 equal to op0 or op1, and last_insn pattern is CODE operation
20563 with MEM dest and one of the operands matches that MEM.
20565 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20567 * doc/extend.texi: Document Function Multiversioning and "default"
20568 parameter string to target attribute.
20569 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20570 target attribute parameter is "default".
20571 (ix86_compare_version_priority): Remove checks for target attribute.
20572 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20573 Remove check for target attribute equal to NULL. Add assert.
20574 (ix86_generate_version_dispatcher_body): Change error to sorry.
20576 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20577 Jack Howarth <howarth@bromo.med.uc.edu>
20578 Patrick Marlier <patrick.marlier@gmail.com>
20581 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20582 define ENDFILE_SPEC as TM_DESTRUCTOR.
20583 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20585 2013-02-11 Alexander Potapenko <glider@google.com>
20586 Jack Howarth <howarth@bromo.med.uc.edu>
20587 Jakub Jelinek <jakub@redhat.com>
20590 * config/darwin.c (cdtor_record): Rename ctor_record.
20591 (sort_cdtor_records): Rename sort_ctor_records.
20592 (finalize_dtors): New routine to sort destructors by
20593 priority before use in assemble_integer.
20594 (machopic_asm_out_destructor): Use finalize_dtors if needed.
20596 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
20598 PR rtl-optimization/56275
20599 * simplify-rtx.c (avoid_constant_pool_reference): Check that
20600 offset is non-negative and less than cmode size before
20601 calling simplify_subreg.
20603 2013-02-11 Richard Biener <rguenther@suse.de>
20605 PR tree-optimization/56264
20606 * cfgloop.h (fix_loop_structure): Adjust prototype.
20607 * loop-init.c (fix_loop_structure): Return the number of
20608 newly discovered loops.
20609 * tree-cfgcleanup.c (repair_loop_structures): When new loops
20610 are discovered, do a full loop-closed SSA rewrite.
20612 2013-02-11 Richard Biener <rguenther@suse.de>
20614 PR tree-optimization/56273
20615 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20617 (check_array_ref): Fix missing newline in dumps.
20618 (search_for_addr_array): Likewise.
20620 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
20622 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
20624 2013-02-09 Jakub Jelinek <jakub@redhat.com>
20627 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
20629 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
20631 PR rtl-optimization/56246
20632 * lra-constraints.c (simplify_operand_subreg): Try to reuse
20634 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
20635 constraints are satisfied.
20637 2013-02-08 Jeff Law <law@redhat.com>
20640 * emit-rtl.c (reg_is_parm_p): New function.
20641 * regs.h (reg_is_parm_p): New prototype.
20642 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
20643 callee-clobbered registers.
20645 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
20648 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
20649 If there is no implicit builtin declaration, just return NULL.
20651 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20653 * config/i386/sse.md (FMAMODEM): New mode iterator.
20654 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
20655 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
20657 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20659 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
20660 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
20661 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
20663 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20665 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
20666 (microblaze*-*-elf): Likewise.
20667 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
20669 * config/microblaze/microblaze-c.c: Add builtin defines for
20670 _LITTLE_ENDIAN and _BIG_ENDIAN.
20671 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
20672 add to TARGET_DEFAULT flags.
20673 Expand ASM_SPEC and LINK_SPEC.
20674 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
20675 * config/microblaze/microblaze.md: Update extendsidi2 and
20676 movdi_internal instructions to use low-order / high-order reg
20678 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
20679 options and inversemask / mask of LITTLE_ENDIAN.
20680 * config/microblaze/t-microblaze: Expand multilib options to
20681 include mlittle-endian (le) and update exceptions patterns.
20683 2013-02-08 Jakub Jelinek <jakub@redhat.com>
20685 PR rtl-optimization/56195
20686 * lra-constraints.c (get_reload_reg): Don't reuse regs
20687 if they have smaller mode than requested, if they have
20688 wider mode than requested, try to return a SUBREG.
20690 PR tree-optimization/56250
20691 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
20692 if type is unsigned and code isn't MULT_EXPR.
20694 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20696 PR tree-optimization/56064
20697 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
20698 bits according to mode.
20699 * fixed-value.h (fixed_from_double_int)
20700 (const_fixed_from_double_int): Adjust comments.
20702 2013-02-08 Richard Biener <rguenther@suse.de>
20705 * lto-streamer.h (struct data_in): Remove current_file, current_line
20706 and current_col members.
20707 * lto-streamer-out.c (lto_output_location): Stream changed bits
20708 en-block for efficiency.
20709 * lto-streamer-in.c (clear_line_info): Remove.
20710 (lto_input_location): Cache current file, line and column
20711 globally via local statics. Read changed bits en-block.
20712 (input_function): Do not call clear_line_info.
20713 (lto_read_body): Likewise.
20714 (lto_input_toplevel_asms): Likewise.
20716 2013-02-08 Michael Matz <matz@suse.de>
20718 PR tree-optimization/52448
20719 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
20720 (nt_call_phase): New static.
20721 (add_or_mark_expr): Only mark accesses with newer phase than any
20723 (nonfreeing_call_p): New.
20724 (nt_init_block): Update nt_call_phase, mark blocks as visited.
20725 (nt_fini_block): Keep blocks marked as visited.
20726 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
20728 2013-02-08 Richard Biener <rguenther@suse.de>
20730 * ira.c (ira): Free broken dominator information.
20732 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20734 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
20736 2013-02-08 Marek Polacek <polacek@redhat.com>
20738 * cfgloop.c (verify_loop_structure): Add more checking of headers.
20740 2013-02-08 Richard Biener <rguenther@suse.de>
20742 PR middle-end/56181
20743 * cfgloop.h (flow_loops_find): Adjust.
20744 (bb_loop_header_p): Declare.
20745 * cfgloop.c (bb_loop_header_p): New function split out from ...
20746 (flow_loops_find): ... here. Adjust function signature,
20747 support incremental loop structure update.
20748 (verify_loop_structure): Cleanup. Verify a loop is a loop.
20749 * cfgloopmanip.c (fix_loop_structure): Move ...
20750 * loop-init.c (fix_loop_structure): ... here.
20751 (apply_loop_flags): Split out from ...
20752 (loop_optimizer_init): ... here.
20753 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
20754 in incremental mode, only remove dead loops here.
20756 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
20759 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
20760 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
20761 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
20762 (*round<mode>3.libgcc): New insns for fixed-modes.
20763 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
20764 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
20765 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
20766 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
20767 implementations. Define to __builtin_avr_absFX,
20768 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
20769 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
20770 __builtin_avr_countlsFX, respectively.
20771 * config/avr/avr-c.c (target.h): Include it.
20772 (enum avr_builtin_id): New enum.
20773 (avr_resolve_overloaded_builtin): New static function.
20774 (avr_register_target_pragmas): Use it to set
20775 targetm.resolve_overloaded_builtin.
20776 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
20777 tree nodes used by DEF_BUILTIN.
20778 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
20779 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
20780 <AVR_BUILTIN_xxBITS>: Same.
20782 2013-02-08 Richard Biener <rguenther@suse.de>
20784 * cfgloop.c (verify_loop_structure): Properly handle
20785 a loop exiting to another loop header.
20786 * ira-int.h (ira_loops): Remove.
20787 * ira.c (ira_loops): Remove.
20788 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
20789 (do_reload): Use loop_optimizer_finalize.
20790 * ira-build.c (create_loop_tree_nodes): Use get_loops and
20791 number_of_loops to access the loop tree.
20792 (more_one_region_p): Likewise.
20793 (finish_loop_tree_nodes): Likewise.
20794 (rebuild_regno_allocno_maps): Likewise.
20795 (mark_loops_for_removal): Likewise.
20796 (mark_all_loops_for_removal): Likewise.
20797 (remove_unnecessary_regions): Likewise.
20798 (ira_build): Likewise.
20799 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20801 2013-02-08 Richard Biener <rguenther@suse.de>
20803 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
20804 * ipa-pure-const.c (analyze_function): Avoid calling
20805 mark_irreducible_loops twice.
20806 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
20808 2013-02-07 David S. Miller <davem@davemloft.net>
20810 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
20812 * var-tracking.c (vt_add_function_parameter): Test the presence of
20813 HAVE_window_save properly and do not remap argument registers when
20814 we have a leaf function.
20816 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
20819 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
20821 * config/i386/i386.c (ix86_print_operand): Ditto.
20823 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20825 * lra-constraints.c (process_alt_operands): Fix recently added comment.
20827 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20829 PR rtl-optimization/56225
20830 * lra-constraints.c (process_alt_operands): Check that reload hard
20831 reg can hold value for strict_low_part.
20833 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20836 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
20837 dwarf2out_end_function.
20838 (in_first_function_p, maybe_at_text_label_p,
20839 first_loclabel_num_not_at_text_label): New variables.
20840 (dwarf2out_var_location): In the first function find out
20841 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
20842 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
20845 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
20847 PR rtl-optimization/56178
20848 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
20849 SUBREG of a register. Tidy up related block of code.
20850 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
20851 note if the source is a register or a SUBREG of a register.
20853 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20856 * config/rs6000/rs6000.md (ptrm): New mode attr.
20857 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
20858 call_value_indirect_aix<pttrsize>,
20859 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
20862 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20864 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
20865 if -bnortl. Convert to strcmp and strncmp.
20867 2013-02-07 Alan Modra <amodra@gmail.com>
20870 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
20871 addresses won't wrap when offsetting.
20872 (rs6000_secondary_reload): Provide secondary reloads needed for
20873 wrapping LO_SUM addresses.
20875 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
20877 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
20878 MACH, just __MACH__.
20880 2013-02-06 Richard Biener <rguenther@suse.de>
20882 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
20883 instead of calling fix_loop_structure.
20885 2013-02-06 Jakub Jelinek <jakub@redhat.com>
20887 PR middle-end/56217
20888 * omp-low.c (use_pointer_for_field): Return false if
20889 lower_send_shared_vars doesn't generate any copy-out code.
20891 2013-02-06 Tom de Vries <tom@codesourcery.com>
20893 PR rtl-optimization/56131
20894 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
20895 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
20896 of the label is NULL. Add comment.
20898 2013-02-05 Jakub Jelinek <jakub@redhat.com>
20900 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
20903 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
20904 (STATIC_LIBTSAN_LIBS): Likewise.
20905 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
20906 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
20907 is defined, don't add anything else beyond that.
20908 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
20909 (LINK_COMMAND_SPEC): Use them.
20911 PR tree-optimization/56205
20912 * tree-stdarg.c (check_all_va_list_escapes): Return true if
20913 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
20914 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
20916 2013-02-05 Richard Biener <rguenther@suse.de>
20918 PR tree-optimization/53342
20919 PR tree-optimization/53185
20920 * tree-vectorizer.h (vect_check_strided_load): Remove.
20921 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
20922 not disallow peeling for vectorized strided loads.
20923 (vect_check_strided_load): Make static and simplify.
20924 (vect_analyze_data_refs): Adjust.
20925 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
20926 correctly when vectorizing strided loads.
20928 2013-02-05 Richard Biener <rguenther@suse.de>
20930 * doc/install.texi: Refer to ISL, not PPL.
20932 2013-02-05 Jan Hubicka <jh@suse.cz>
20934 PR tree-optimization/55789
20935 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
20937 2013-02-05 Jan Hubicka <jh@suse.cz>
20939 PR tree-optimization/55789
20940 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
20941 the dead call anyway.
20943 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
20946 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
20948 2013-02-04 Alexander Potapenko <glider@google.com>
20949 Jack Howarth <howarth@bromo.med.uc.edu>
20950 Jakub Jelinek <jakub@redhat.com>
20953 * config/darwin.c (sort_ctor_records): Stabilized qsort
20954 on constructor priority by using original position.
20955 (finalize_ctors): New routine to sort constructors by
20956 priority before use in assemble_integer.
20957 (machopic_asm_out_constructor): Use finalize_ctors if needed.
20959 2013-02-04 Jakub Jelinek <jakub@redhat.com>
20962 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
20963 about visibility on artificial decls.
20964 * config/sol2.c (solaris_assemble_visibility): Likewise.
20966 2013-02-04 Kai Tietz <ktietz@redhat.com>
20969 * config/i386/i386.c (function_value_ms_64): Add additional valtype
20970 argument and improve checking of return-argument types for 16-byte
20972 (ix86_function_value_1): Add additional valtype argument on call
20973 of function_value_64.
20974 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
20975 handling infunction_value_64 function.
20977 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
20979 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
20981 2013-02-04 Richard Biener <rguenther@suse.de>
20983 PR tree-optimization/56188
20984 * tree-ssa-structalias.c (label_visit): Consider case with
20985 initially non-empty points-to set.
20986 (perform_var_substitution): Dump node mapping and clean up.
20988 2013-02-04 Richard Guenther <rguenther@suse.de>
20991 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
20992 node prevail as last resort.
20993 (lto_symtab_merge_decls): Remove guard on LTRANS here.
20994 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
20996 2013-02-04 Richard Biener <rguenther@suse.de>
20998 PR tree-optimization/56113
20999 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21001 (equiv_class_lookup_or_add): ... this.
21002 (label_visit): Adjust and fix error in previous patch.
21003 (perform_var_substitution): Adjust.
21005 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21007 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21008 * config/sh/sh4-300.md: Likewise.
21009 * config/sh/sh4a.md: Likewise.
21010 * config/sh/constraints.md: Likewise.
21011 * config/sh/sh.md: Likewise.
21012 * config/sh/netbsd-elf.h: Likewise.
21013 * config/sh/predicates.md: Likewise.
21014 * config/sh/sh-protos.h: Likewise.
21015 * config/sh/ushmedia.h: Likewise.
21016 * config/sh/linux.h: Likewise.
21017 * config/sh/sh.c: Likewise.
21018 * config/sh/superh.h: Likewise.
21019 * config/sh/elf.h: Likewise.
21020 * config/sh/sh4.md: Likewise.
21021 * config/sh/sh.h: Likewise.
21023 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21025 * config/pa/constraints.md: Adjust unused letters. Change "T"
21026 constraint to match_test floating_point_store_memory_operand().
21027 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21028 (base14_operand): New.
21029 (floating_point_store_memory_operand): New.
21030 (integer_store_memory_operand): Revise to use base14_operand and
21031 reg_plus_base_memory_operand.
21032 (move_dest_operand): Allow symbolic_memory_operands.
21033 (symbolic_memory_operand): Check for LO_SOM.
21034 (symbolic_operand): Change default case to break.
21035 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21036 CONST_DOUBLE values to be reloaded by putting them into memory when
21037 the destination is a floating point register.
21038 (movdf): Remove code to handle CONST_DOUBLE.
21040 (reload_indf_r1): New.
21041 (reload_insf_r1): New.
21042 Consistently use "Q" and "T" constraints with integer and floating
21043 point move instructions, respectively.
21044 (movdi): Remove FAIL.
21045 Change predicate for source operand unamed DImode move from
21046 general_operand to move_src_operand.
21047 (umulsidi3): Change predicate for destination operand to
21049 Likewise for similar unamed patterns.
21050 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21051 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21052 (hppa_legitimize_address): Simplify mask calculation.
21053 (pa_emit_move_sequence): Revised handling of secondary reloads from
21054 REG+D addresses for floating point loads and stores. Directly handle
21055 loading CONST0_RTX (mode) to a floating point register.
21056 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21057 to floating point registers. Don't restrict secondary reloads to
21058 floating point registers to integer modes. Revise some comments and
21060 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21061 (pa_legitimate_address_p): New.
21062 (pa_legitimize_reload_address): New.
21063 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21064 (STRICT_REG_OK_FOR_BASE_P): New.
21065 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21066 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21068 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21069 Andrew Dixie <andrewd@gentrack.com>
21071 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21074 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21076 * expmed.c (extract_bit_field_1): Pass the full width of the
21077 structure to get_best_reg_extraction_insn.
21079 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21082 * configure.ac (use_cxa_atexit): Add AIX.
21083 * configure: Regenerate.
21085 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21087 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21090 * final.c (need_profile_function): New variable.
21091 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21092 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21093 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21094 notes, targetm.asm_out.function_prologue doesn't emit anything,
21095 HAVE_prologue and profiler should be emitted before prologue,
21096 set need_profile_function instead of emitting it.
21097 (final_scan_insn): If need_profile_function, emit
21098 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21099 NOTE_INSN_FUNCTION_BEG note.
21101 2013-02-01 Richard Henderson <rth@redhat.com>
21103 * config/rs6000/rs6000.md (smulditi3): New.
21106 * config/alpha/alpha.md (umulditi3): New.
21108 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21110 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21111 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21113 2013-02-01 Richard Biener <rguenther@suse.de>
21115 PR tree-optimization/56113
21116 * tree-ssa-structalias.c (label_visit): Reduce work for
21117 single-predecessor nodes.
21119 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21121 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21122 range isn't testing for zero.
21124 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21126 PR middle-end/56113
21127 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21129 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21130 Nick Clifton <nickc@redhat.com>
21132 * config/v850/constraints.md (Q): Define as a memory constraint.
21133 * config/v850/predicates.md (label_ref_operand): New predicate.
21134 (e3v5_shift_operand): New predicate.
21135 (ior_operator): New predicate.
21136 * config/v850/t-v850: Add e3v5 multilib.
21137 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21138 (v850_gen_movdi): Prototype.
21139 * config/v850/v850.c: Add support for e3v5 architecture.
21140 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21142 (construct_save_jarl): Add e3v5 long JARL support.
21143 (v850_adjust_insn_length): New function. Adjust length of call
21144 insns when using e3v5 instructions.
21145 (v850_gen_movdi): New function: Generate instructions to move a
21147 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21148 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21149 (TARGET_USE_FPU): Enable for e3v5.
21150 (CONST_OK_FOR_W): New macro.
21151 (ADJUST_INSN_LENGTH): Define.
21152 * config/v850/v850.md (UNSPEC_LOOP): Define.
21153 (attr cpu): Add v850e3v5.
21154 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21155 (movdi): New pattern.
21156 (movdi_internal): New pattern.
21157 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21158 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21159 (cstoresf4): Likewise.
21160 (cstoredf4): Likewise.
21161 (insv): New pattern.
21162 (rotlso3_a): New pattern.
21163 (rotlsi3_b): New pattern
21164 (rotlsi3_v850e3v5): New pattern.
21165 (doloop_begin): New pattern.
21166 (fix_loop_counter): New pattern.
21167 (doloop_end): New pattern.
21168 (branch_normal): Add e3v5 long branch support.
21169 (branch_invert): Likewise.
21170 (branch_z_normal): Likewise.
21171 (branch_z_invert): Likewise.
21172 (branch_nz_normal): Likewise.
21173 (branch_nz_invert): Likewise.
21174 (call_internal_short): Add e3v5 register-indirect JARL support.
21175 (call_internal_long): Likewise.
21176 (call_value_internal_short): Likewise.
21177 (call_value_internal_long): Likewise.
21178 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21179 (mloop): New option.
21180 * config.gcc: Add support for configuring v840e3v5 target.
21181 * doc/invoke.texi: Document new v850 specific command line options.
21183 2013-01-31 Paul Koning <ni1d@arrl.net>
21187 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21188 children if parent is a class.
21189 (prune_unused_types_prune): Don't add DW_AT_declaration.
21191 2013-01-31 Richard Biener <rguenther@suse.de>
21193 PR tree-optimization/56157
21194 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21195 match up operand with SLP child.
21197 2013-01-31 Jason Merrill <jason@redhat.com>
21200 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21201 parameters the first time.
21202 (gen_scheduled_generic_parms_dies): Check completeness here.
21204 2013-01-31 Richard Biener <rguenther@suse.de>
21206 PR middle-end/53073
21207 * common.opt (faggressive-loop-optimizations): New flag,
21208 enabled by default.
21209 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21210 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21211 infer_loop_bounds_from_undefined by it.
21213 2013-01-31 Richard Biener <rguenther@suse.de>
21215 PR tree-optimization/56150
21216 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21217 visit virtual operands.
21218 (find_uses_to_rename_bb): Likewise.
21220 2013-01-31 Richard Biener <rguenther@suse.de>
21222 PR tree-optimization/56150
21223 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21224 mixed store non-store stmts.
21226 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21229 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21230 LIBASAN_EARLY_SPEC is defined.
21231 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21232 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21234 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21237 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21238 invalid args instead of ICEing on it.
21239 (ix86_valid_target_attribute_tree): Return error_mark_node if
21240 ix86_valid_target_attribute_inner_p failed.
21241 (ix86_valid_target_attribute_p): Return false only if
21242 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21243 target("default") attribute.
21244 (sorted_attr_string): Change argument from const char * to tree,
21245 merge in all target attribute arguments rather than just one.
21246 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21247 instead of free. Avoid using strcat.
21248 (ix86_mangle_function_version_assembler_name): Mangle
21249 target("default") as if no target attribute is present. Adjust
21250 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21251 instead of xmalloc and XDELETEVEC instead of free.
21252 (ix86_function_versions): Don't return true if one of the decls
21253 doesn't have target attribute. If they don't and one of the decls
21254 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21255 sorted_attr_string caller. Use XDELETEVEC instead of free.
21256 (ix86_supports_function_versions): Remove.
21257 (make_name): Fix up formatting.
21258 (make_dispatcher_decl): Remove resolver_name and its initialization.
21259 Avoid leaking memory.
21260 (is_function_default_version): Return true if there is
21261 target("default") attribute rather than no target attribute at all.
21262 (make_resolver_func): Avoid leaking memory.
21263 (ix86_generate_version_dispatcher_body): Likewise.
21264 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21265 * target.def (supports_function_versions): Remove.
21266 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21267 * doc/tm.texi: Regenerated.
21269 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21271 PR rtl-optimization/56144
21272 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21273 for values with side effects.
21275 2013-01-30 Richard Biener <rguenther@suse.de>
21277 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21278 (sparseset_pop): Likewise.
21279 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21280 to be able to use quick_push in the worker loop.
21282 2013-01-30 Marek Polacek <polacek@redhat.com>
21284 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21286 2013-01-30 Richard Biener <rguenther@suse.de>
21289 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21291 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21293 PR tree-optimization/56064
21294 * fixed-value.c (fixed_from_double_int): New function.
21295 * fixed-value.h (fixed_from_double_int): New prototype.
21296 (const_fixed_from_double_int): New static inline function.
21297 * fold-const.c (native_interpret_fixed): New static function.
21298 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21299 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21300 (native_encode_fixed): New static function.
21301 (native_encode_expr) <FIXED_CST>: Use it.
21302 (native_interpret_int): Move double_int worker code to...
21303 * double-int.c (double_int::from_buffer): ...this new static method.
21304 * double-int.h (double_int::from_buffer): Prototype it.
21306 2013-01-30 Richard Biener <rguenther@suse.de>
21308 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21309 New pointer-map and obstack.
21310 (init_alias_vars): Allocate pointer-map and obstack.
21311 (delete_points_to_sets): Free them.
21312 (find_what_var_points_to): Cache result.
21313 (find_what_p_points_to): Adjust for changed interface of
21314 find_what_var_points_to.
21315 (compute_points_to_sets): Likewise.
21316 (ipa_pta_execute): Likewise.
21318 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21320 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21321 * configure: Regenerate.
21322 * config.in: Regenerate.
21323 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21324 #nobits/#progbits if supported.
21326 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21329 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21330 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21331 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21333 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21335 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21336 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21338 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21340 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21341 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21343 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21345 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21347 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21348 * config/arm/cortex-a7.md: New bypasses using
21349 arm_mac_accumulator_is_result.
21351 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21353 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21354 (cortex_a7_neon_mla): Likewise.
21355 (cortex_a7_fpfmad): New reservation.
21356 (cortex_a7_fpmacs): Use ffmas and update required units.
21357 (cortex_a7_fpmuld): Update required units and latency.
21358 (cortex_a7_fpmacd): Likewise.
21359 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21360 (cortex_a7_neon). Likewise.
21361 (bypass) Update participating units.
21363 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21365 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21366 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21368 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21369 (vfp_fmul): Use ffmad.
21370 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21371 (cortex_r4_fmacd): Use ffmad.
21372 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21373 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21374 (cortex_a9_fmacd): Use ffmad.
21375 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21376 (cortex_a8_vfp_macd): Use ffmad.
21377 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21378 (cortex_a5_fpmacd): Use ffmad.
21379 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21380 (cortex_a15_vfp_macd): Use ffmad.
21381 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21383 2013-01-29 Jason Merrill <jason@redhat.com>
21386 * varasm.c (default_assemble_visibility): Don't warn about
21387 visibility on artificial decls.
21389 2013-01-29 Richard Biener <rguenther@suse.de>
21391 PR tree-optimization/56113
21392 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21394 (label_visit): Free duplicate bitmaps and record the leader instead.
21395 (perform_var_substitution): Adjust.
21397 2013-01-29 Richard Biener <rguenther@suse.de>
21399 PR tree-optimization/55270
21400 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21401 the CFG, schedule loops for fixup.
21403 2013-01-29 Nick Clifton <nickc@redhat.com>
21405 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21408 2013-01-28 Leif Ekblad <leif@rdos.net>
21410 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21411 * config/i386/i386.h (TARGET_RDOS): New macro.
21412 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21413 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21414 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21415 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21416 DEFAULT_LARGE_SECTION_THRESHOLD.
21417 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21418 * config/i386/rdos.h: New file.
21419 * config/i386/rdos64.h: New file.
21421 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21424 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21427 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21429 PR rtl-optimization/56117
21430 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21431 call cselib_lookup_from_insn on the MEM before calling
21432 add_insn_mem_dependence.
21434 2013-01-28 Richard Biener <rguenther@suse.de>
21436 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21437 to a stmt that didn't have one.
21438 (copy_phis_for_bb): Likewise for PHI arguments.
21439 (copy_debug_stmt): Likewise for debug stmts.
21441 2013-01-28 Richard Biener <rguenther@suse.de>
21443 PR tree-optimization/56034
21444 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21445 (partition_builtin_p): Adjust.
21446 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21447 it is the last partition.
21448 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21449 up the vertex for the definition.
21450 (classify_partition): Classify whether a partition is a
21451 PKIND_REDUCTION, thus has uses outside of the loop.
21452 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21453 Merge all PKIND_REDUCTION partitions into the last partition.
21454 (tree_loop_distribution): Seed partitions from reductions as well.
21456 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21458 PR tree-optimization/56125
21459 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21460 pow(x,c) into sqrt(x) * powi(x, n/2) or
21461 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21462 optimizing for size.
21463 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21464 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21467 PR tree-optimization/56094
21468 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21469 to UNKNOWN_LOCATION while gimplifying expr.
21471 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21474 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21475 operand 0 in movabs insn template for -masm=intel asm alternative.
21476 (*movabs<mode>_2): Ditto for operand 1.
21478 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21481 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21484 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21486 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21489 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21491 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21492 Undef to avoid warning.
21494 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21496 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21497 * configure: Regenerate.
21499 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21501 PR tree-optimization/56098
21502 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21503 for stmts with volatile ops.
21504 (cond_store_replacement): Don't optimize if assign has volatile ops.
21505 (cond_if_else_store_replacement_1): Don't optimize if either
21506 then_assign or else_assign have volatile ops.
21507 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21510 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21512 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21514 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21516 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21517 missing ':' in asm example.
21519 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21521 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21522 entries into lane and laneq entries.
21523 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21524 Remove AdvSIMD scalar modes.
21525 (aarch64_sq<r>dmulh_laneq<mode>): New.
21526 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21528 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21529 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21530 * config/aarch64/iterators.md (VCOND): New.
21533 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21536 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21537 Add NULL LIBNAME argument to existing definitions.
21538 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21539 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21540 * config/avr/avr.c (DEF_BUILTIN): Same.
21541 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21542 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21543 implementation is available (DECL_ASSEMBLER_NAME is set).
21544 (avr_fold_absfx): New static function.
21545 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21546 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21547 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21548 AVR_BUILTIN_ABSLLK.
21549 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21550 (abshk, absk, abslk, absllk): Provide as static inline functions.
21552 2013-01-25 Marek Polacek <polacek@redhat.com>
21554 PR tree-optimization/56035
21555 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21557 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21559 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21560 (*movtf_internal_rex64): Add (!o,C) alternative
21561 (*movxf_internal_rex64): Ditto.
21562 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21564 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21566 * doc/invoke.texi: fix typo.
21567 * doc/objc.texi: fix typo.
21569 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21571 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21572 for the first two alternatives.
21574 2013-01-24 Diego Novillo <dnovillo@google.com>
21576 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21577 (ggc-zone.o): Remove.
21578 * configure.ac: Remove option --with-gc.
21579 * configure: Re-generate.
21580 * doc/install.texi: Remove documentation for --with-gc.
21581 * gengtype.c (write_enum_defn): Remove. Update all users.
21582 (write_Types_process_field): Remove generation of gt_e_* argument.
21583 (output_type_enum): Remove. Update all users.
21584 (write_enum_defn): Remove. Update all users.
21585 (enum alloc_zone): Remove. Update all users.
21586 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21587 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21588 Update all callers.
21589 (struct ptr_data): Remove field TYPE. Update all users.
21590 (gt_pch_note_object): Remove argument TYPE. Update all users.
21591 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
21593 * ggc-none.c (ggc_alloc_typed_stat): Remove.
21594 (struct alloc_zone): Remove.
21595 (ggc_internal_alloc_zone_stat): Remove.
21596 (ggc_internal_cleared_alloc_zone_stat): Remove.
21597 * ggc-page.c (ggc_alloc_typed_stat): Remove.
21598 (ggc_pch_count_object): Remove last argument. Update all users.
21599 (ggc_pch_alloc_object): Remove last argument. Update all users.
21600 (struct alloc_zone): Remove.
21601 * ggc-zone.c: Remove.
21602 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
21603 (struct alloc_zone): Remove.
21604 (ggc_alloc_typed_stat): Remove.
21605 (ggc_alloc_typed): Remove.
21606 (ggc_splay_alloc): Remove first argument.
21607 (rtl_zone): Remove. Update all users.
21608 (tree_zone): Remove. Update all users.
21609 (tree_id_zone): Remove. Update all users.
21610 (ggc_internal_zone_alloc_stat): Remove. Update all users.
21611 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
21612 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
21613 * tree-ssanames.c: Remove references to zone allocator in comments.
21615 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
21617 * config/avr/avr.c (avr_out_fract): Make register numbers that
21618 might be outside of source operand signed.
21620 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
21622 * config/i386/constraints.md (Yf): New constraint.
21623 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
21624 of f constraint to conditionaly disable x87 register preferences.
21625 (*movdf_internal): Ditto.
21626 (*movsf_internal): Ditto.
21628 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
21630 PR inline-asm/55934
21631 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
21632 that have operands with impossible constraints.
21633 Add a FIXME for a speed-up opportunity.
21634 * lra-constraints.c (process_alt_operands): Verify that a class
21635 selected from constraints on asms is valid for the operand mode.
21636 (curr_insn_transform): Remove incorrect comment.
21638 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
21640 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
21641 TOC operand is a valid symbol ref in the constant pool.
21643 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21645 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
21647 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
21650 * config/avr/stdfix.h: New file.
21651 * t-avr (stdfix-gcc.h): New rule to build it.
21652 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
21654 2013-01-23 Kostya Serebryany <kcc@google.com>
21656 * config/darwin.h: remove dependency on
21657 CoreFoundation (asan on Mac OS).
21659 2013-01-23 Jakub Jelinek <jakub@redhat.com>
21662 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
21663 instead of cmpdi_operand for first comparison operand.
21664 Don't assert that comparison operands aren't both constants.
21666 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21668 * doc/install.texi (Downloading the Source): Update references to
21669 downloading separate components.
21671 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
21673 * doc/extend.texi (__int128): Improve grammar.
21675 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
21678 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
21679 alternative to (o,r).
21680 (*movdi_internal_rex64): Remove (!o,n) alternative.
21681 (DImode immediate->memory splitter): Remove.
21682 (DImode immediate->memory peephole2): Remove.
21683 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
21684 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
21685 alternative to (!o,*r).
21686 (*movtf_internal_sse): New pattern.
21687 (*movxf_internal_rex64): New pattern.
21688 (*movxf_internal): Disable for TARGET_64BIT.
21689 (*movdf_internal_rex64): Remove (!o,F) alternative.
21691 2013-01-22 Jakub Jelinek <jakub@redhat.com>
21693 PR middle-end/56074
21694 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
21695 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
21696 * tree-vect-loop-manip.c (find_loop_location): Also ignore
21697 stmt locations where LOCATION_LOCUS of the stmt location is
21698 UNKNOWN_LOCATION or BUILTINS_LOCATION.
21701 * config/i386/i386.md (UNSPEC_STOS): New.
21702 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
21703 *strsetqi_1): Add UNSPEC_STOS.
21705 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
21708 * doc/invoke.texi: Remove left over -Wsynth example.
21710 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21712 PR tree-optimization/56051
21713 * fold-const.c (fold_binary_loc): Don't fold
21714 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
21715 a narrowing conversion, or widening conversion from signed
21718 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21720 PR rtl-optimization/56023
21721 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
21722 dependent on debug instruction.
21724 2013-01-21 Martin Jambor <mjambor@suse.cz>
21726 PR middle-end/56022
21727 * function.c (allocate_struct_function): Call
21728 invoke_set_current_function_hook earlier.
21730 2013-01-21 Jakub Jelinek <jakub@redhat.com>
21732 * reload1.c (init_reload): Only initialize reload_obstack
21733 during the first call.
21735 2013-01-21 Marek Polacek <polacek@redhat.com>
21737 * cfgloop.c (verify_loop_structure): Fix up grammar.
21739 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
21741 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
21742 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
21744 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21747 * config/arm/marvell-pj4.md: Update copyright year.
21748 Fix up use of alu to alu_reg and simple_alu_imm.
21750 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
21752 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
21754 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
21757 * lra-constraints.c (curr_insn_transform): Don't reuse original
21758 insn for secondary memory move when memory mode should be different.
21760 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21762 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
21763 atomic_storedi_1): New patterns.
21765 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21767 btver2 pipeline descriptions.
21768 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
21770 * config/i386/i386.md (btver2_decode): New type attributes.
21771 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
21773 * config/i386/btver2.md: New file describing btver2 pipelines.
21775 2013-01-19 Andrew Pinski <apinski@cavium.com>
21777 PR tree-optimization/52631
21778 * tree-ssa-sccvn (visit_use): Before looking up the original
21779 statement, try looking up the simplified expression.
21781 2013-01-19 Anthony Green <green@moxielogic.com>
21783 * config/moxie/moxie.c (moxie_expand_prologue): Set
21784 current_function_static_stack_size.
21786 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21788 PR tree-optimization/56029
21789 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
21790 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
21792 2013-01-18 Sharad Singhai <singhai@google.com>
21794 PR tree-optimization/55995
21795 * dumpfile.c (dump_loc): Print location only if available.
21796 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
21798 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
21801 * lra-constraints.c (curr_insn_transform): Reuse original insn for
21802 secondary memory move.
21803 (inherit_reload_reg): Use rclass instead of cl for
21804 check_secondary_memory_needed_p.
21806 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21808 PR middle-end/56015
21809 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
21810 the case where writing real complex part of target modifies op1.
21812 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
21814 * config/aarch64/aarch64-simd.md
21815 (aarch64_vcond_internal<mode>): Handle unordered cases.
21816 * config/aarch64/iterators.md (v_cmp_result): New.
21818 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
21819 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21821 * config/arm/marvell-pj4.md: New file.
21822 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
21823 * config/arm/arm.md (generic_sched): Add marvell_pj4.
21824 (generic_vfp): Likewise.
21825 * config/arm/arm-cores.def: Add marvell-pj4.
21826 * config/arm/arm-tune.md: Regenerate.
21827 * config/arm/arm-tables.opt: Regenerate.
21828 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
21829 * doc/invoke.texi: Document marvell-pj4.
21831 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
21833 * config/aarch64/arm_neon.h: Map scalar types to standard types.
21835 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21840 * var-tracking.c (negative_power_of_two_p): New.
21841 (global_get_addr_cache, local_get_addr_cache): New.
21842 (get_addr_from_global_cache, get_addr_from_local_cache): New.
21843 (vt_canonicalize_addr): Rewrite using the above. Adjust the
21845 (vt_stack_offset_p): Remove.
21846 (vt_canon_true_dep): Always canonicalize loc's address.
21847 (clobber_overlapping_mems): Make sure we have a MEM.
21848 (local_get_addr_clear_given_value): New.
21849 (val_reset): Clear local cached entries.
21850 (compute_bb_dataflow): Create and release the local cache.
21851 Disable duplicate MEMs clobbering.
21852 (emit_notes_in_bb): Clobber MEMs likewise.
21853 (vt_emit_notes): Create and release the local cache.
21854 (vt_initialize, vt_finalize): Create and release the global
21855 cache, respectively.
21856 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
21858 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21860 PR libmudflap/53359
21861 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
21862 not found in the symtab.
21864 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21867 PR rtl-optimization/55547
21868 PR rtl-optimization/53827
21871 * alias.c (offset_overlap_p): New, factored out of...
21872 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
21873 the conservative special case for symbolic constants. Don't
21874 adjust zero sizes on alignment.
21876 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
21878 PR rtl-optimization/52573
21879 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
21880 REG_UNUSED for the same register.
21882 2013-01-17 Richard Biener <rguenther@suse.de>
21883 Marek Polacek <polacek@redhat.com>
21885 PR rtl-optimization/55833
21886 * loop-unswitch.c (unswitch_loops): Move loop verification...
21887 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
21888 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
21889 Set it to true when we're removing a loop from hierarchy tree in
21890 an irreducible region.
21891 (fix_bb_placements): Adjust caller.
21892 (fix_loop_placements): Likewise.
21894 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
21896 * config/avr/builtins.def (DEF_BUILTIN): Factor out
21897 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
21898 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
21899 Remove ID. Adjust comments.
21900 * config/avr/avr-c.c (avr_builtin_name): Remove.
21901 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
21902 * config/avr/avr.c (avr_tolower): New static function.
21903 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
21904 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
21905 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
21908 2013-01-17 Jan Hubicka <jh@suse.cz>
21910 PR tree-optimization/55273
21911 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
21913 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
21916 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
21917 store through atomic_store<mode>_1.
21918 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
21920 2013-01-17 Martin Jambor <mjambor@suse.cz>
21922 PR tree-optimizations/55264
21923 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
21924 for virtual methods.
21925 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
21926 virtual methods before inlining is over.
21927 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
21929 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
21932 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
21934 PR rtl-optimization/56005
21935 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
21936 pending reads for prefetch.
21938 2013-01-16 Ian Bolton <ian.bolton@arm.com>
21940 * config/aarch64/aarch64.md
21941 (*cstoresi_neg_uxtw): New pattern.
21942 (*cmovsi_insn_uxtw): New pattern.
21943 (*<optab>si3_uxtw): New pattern.
21944 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
21945 (*<optab>si3_insn_uxtw): New pattern.
21946 (*bswapsi2_uxtw): New pattern.
21948 2013-01-16 Richard Biener <rguenther@suse.de>
21950 * tree-inline.c (tree_function_versioning): Remove set but
21951 never used variable.
21953 2013-01-16 Richard Biener <rguenther@suse.de>
21955 PR tree-optimization/55964
21956 * tree-flow.h (rename_variables_in_loop): Remove.
21957 (rename_variables_in_bb): Likewise.
21958 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
21959 (copy_loop_before): Adjust and delete update-ssa status.
21960 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
21961 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
21962 (rename_variables_in_loop): Remove.
21963 (slpeel_update_phis_for_duplicate_loop): Likewise.
21964 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
21965 use available cfg machinery instead of duplicating it.
21966 Update PHI nodes and perform poor-mans SSA update here.
21967 (slpeel_tree_peel_loop_to_edge): Adjust.
21969 2013-01-16 Richard Biener <rguenther@suse.de>
21971 PR tree-optimization/54767
21972 PR tree-optimization/53465
21973 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
21974 (vrp_visit_phi_node): For PHI arguments coming via backedges
21975 drop all symbolical range information.
21976 (execute_vrp): Compute backedges.
21978 2013-01-16 Richard Biener <rguenther@suse.de>
21980 * doc/install.texi: Update CLooG and ISL requirements to
21983 2013-01-16 Christian Bruel <christian.bruel@st.com>
21986 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
21987 (broken_move): Handle UNSPECV_SP_SWITCH_B.
21988 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
21990 2013-01-16 DJ Delorie <dj@redhat.com>
21992 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
21993 (UNSPECV_SP_SWITCH_E): New.
21994 (sp_switch_1): Change to an unspec.
21995 (sp_switch_2): Change to an unspec. Don't use post-inc when we
21998 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22000 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22001 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22002 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22003 (expand_mem_thread_fence): Ditto.
22004 (expand_mem_signal_fence): Ditto.
22005 (expand_atomic_load): Ditto.
22006 (expand_atomic_store): Ditto.
22008 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22010 PR rtl-optimization/55547
22011 PR rtl-optimization/53827
22014 * alias.c (memrefs_conflict_p): Set sizes to negative after
22017 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22020 * function.c (thread_prologue_and_epilogue_insns): Always
22021 add crtl->drap_reg to set_up_by_prologue.set, even if
22022 stack_realign_drap is false.
22024 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22026 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22027 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22028 *call): Fix indention.
22030 2013-01-15 Tom de Vries <tom@codesourcery.com>
22033 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22036 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22038 PR rtl-optimization/55153
22039 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22041 2013-01-15 Martin Jambor <mjambor@suse.cz>
22043 PR tree-optimization/55920
22044 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22045 accesses as grp_to_be_debug_replaced.
22047 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22049 PR tree-optimization/55920
22050 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22051 there is non-useless type conversion needed from debug rhs to lhs,
22052 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22054 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22055 Mikael Pettersson <mikpe@it.uu.se>
22058 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22060 (ASM_OUTPUT_CASE_LABEL): Remove.
22061 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22062 * final.c (shorten_branches): Update alignment of labels before
22063 jump tables if CASE_VECTOR_SHORTEN_MODE.
22065 2013-01-15 Richard Biener <rguenther@suse.de>
22068 * system.h: Do not include gmp.h for building host tools.
22070 2013-01-15 Richard Biener <rguenther@suse.de>
22072 PR middle-end/55882
22073 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22074 account for bitpos when computing alignment.
22076 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22078 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22079 (ix86_target_macros_internal): Likewise.
22081 * config/i386/i386.c (m_CORE2I7): Removed.
22082 (m_CORE_HASWELL): New macro.
22083 (m_CORE_ALL): Likewise.
22084 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22085 (initial_ix86_arch_features): Likewise.
22086 (processor_target_table): Initializations for Core avx2.
22087 (cpu_names): New names "core-avx2".
22088 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22089 PROCESSOR_CORE_HASWELL.
22090 (ix86_issue_rate): New case.
22091 (ia32_multipass_dfa_lookahead): Likewise.
22092 (ix86_sched_init_global): Likewise.
22094 * config/i386/i386.h (TARGET_HASWELL): New macro.
22095 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22096 (processor_type): New PROCESSOR_HASWELL.
22098 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22100 PR tree-optimization/55955
22101 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22102 *SHIFT_EXPR and *ROTATE_EXPR codes.
22104 PR tree-optimization/48766
22105 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22106 -ftrapv disable -fwrapv.
22108 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22111 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22112 etc. to 1 and not to __flash.
22113 Use LL suffix for __INT24_MAX__ with -mint8.
22114 Use ULL suffix for __UINT24_MAX__ with -mint8.
22116 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22118 * config/avr/avr-arch.h
22119 (struct base_arch_s): Use typedef avr_arch_t instead.
22120 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22121 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22122 * config/avr/avr.c: Same.
22123 * config/avr/avr-devices.c: Same.
22124 * config/avr/driver-avr.c: Same.
22125 * config/avr/gen-avr-mmcu-texi.c: Same.
22126 * config/avr/avr-mcus.def: Adjust comment.
22128 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22130 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22131 * config/aarch64/iterators.md (VALLDI): New.
22133 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22134 Andi Kleen <ak@linux.intel.com>
22137 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22138 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22141 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22143 * config/avr/avr-stdint.h: Remove trailing blanks.
22144 * config/avr/avr-log.h: Same.
22145 * config/avr/avr-arch.h: Same.
22146 * config/avr/avr-devices.c: Same.
22147 * config/avr/avr-dimode.md: Same.
22148 * config/avr/predicates.md: Same.
22149 * config/avr/avr-c.c: Same. And fix typo.
22151 * config/avr/avr-protos.h: Same. And:
22152 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22153 (init_cumulative_args): Rename to avr_init_cumulative_args.
22154 (expand_prologue): Rename to avr_expand_prologue.
22155 (expand_epilogue): Rename to avr_expand_epilogue.
22156 (adjust_insn_length): Rename to avr_adjust_insn_length.
22157 (notice_update_cc): Rename to avr_notice_update_cc.
22158 (final_prescan_insn): Rename to avr_final_prescan_insn.
22159 * config/avr/avr.c: Same.
22160 * config/avr/avr.h: Same.
22161 * config/avr/avr.md: Remove trailing blanks.
22162 (prologue): Use avr_expand_prologue.
22163 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22165 2013-01-14 Richard Biener <rguenther@suse.de>
22167 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22168 verify_location, collect_subblocks): New functions.
22169 (verify_gimple_in_cfg): Verify that locations only reference
22170 BLOCKs in the functions BLOCK tree.
22172 2013-01-14 Richard Biener <rguenther@suse.de>
22174 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22176 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22178 (insert_out_of_ssa_copy_on_edge): Likewise.
22179 (rewrite_close_phi_out_of_ssa): Likewise.
22180 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22182 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22183 propagated constants.
22184 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22187 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22189 * config/avr/avr-modes.def: Add GPL copyright notice.
22191 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22193 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22194 MEMMODEL_MASK to determine memory model.
22195 (atomic_store<mode>): Ditto from operands[2].
22196 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22198 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22201 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22202 (fold_gimple_assign): Don't call unshare_expr here.
22203 (fold_ctor_reference): Call unshare_expr.
22205 2013-01-13 Terry Guo <terry.guo@arm.com>
22207 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22208 * doc/fragments.texi: Document MULTILIB_REUSE.
22209 * gcc.c (multilib_reuse): New internal spec.
22210 (set_multilib_dir): Also search multilib from multilib_reuse.
22211 * genmultilib (tmpmultilib3): Refactor code.
22212 (tmpmultilib4): Ditto.
22213 (multilib_reuse): New multilib argument.
22215 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22217 * Makefile.in: Update copyright.
22219 2013-01-12 Tom de Vries <tom@codesourcery.com>
22221 PR middle-end/55890
22222 * calls.c (expand_call): Check if arg_nr is valid.
22224 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22226 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22227 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22228 documentation. Add missing '__' in front of
22229 __builtin_ia32_packssdw256.
22231 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22234 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22235 NO_REGS for larl operands.
22236 (s390_reload_larl_operand): Use s390_load_address instead of
22239 2013-01-11 Richard Biener <rguenther@suse.de>
22241 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22242 (verify_node_sharing): ... here.
22243 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22245 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22247 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22248 Substitute TREECHECKING.
22249 * configure: Regenerate.
22250 * Makefile.in (TREECHECKING): New.
22252 2013-01-11 Richard Guenther <rguenther@suse.de>
22254 PR tree-optimization/44061
22255 * tree-vrp.c (extract_range_basic): Compute zero as
22256 value-range for __builtin_constant_p of function parameters.
22258 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22260 Update copyright years.
22262 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22264 PR rtl-optimization/55672
22265 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22266 const to be eliminable.
22268 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22270 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22271 * configure: Regenerate.
22273 2013-01-10 Richard Biener <rguenther@suse.de>
22275 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22277 2013-01-10 Richard Biener <rguenther@suse.de>
22280 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22281 locations for virtual PHI arguments.
22282 (rewrite_update_phi_arguments): Likewise.
22284 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22286 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22289 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22291 PR tree-optimization/55921
22292 * tree-complex.c (expand_complex_asm): New function.
22293 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22295 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22298 * config/s390/s390.c (s390_symref_operand_p)
22299 (s390_loadrelative_operand_p): Merge the two functions.
22300 (s390_check_qrst_address, print_operand_address): Add parameters
22301 to s390_loadrelative_operand_p invokation.
22302 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22303 (s390_reload_larl_operand, s390_secondary_reload): Use
22304 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22305 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22307 2013-01-09 Mike Stump <mikestump@comcast.net>
22309 * dse.c (record_store): Remove unnecessary assert.
22311 2013-01-09 Jan Hubicka <jh@suse.cz>
22313 PR tree-optimization/55569
22314 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22315 * cfgloop.h (scale_loop_profile): Likewise.
22317 2013-01-09 Jan Hubicka <jh@suse.cz>
22320 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22322 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22324 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22326 PR middle-end/55114
22327 * expr.h (maybe_emit_group_store): Declare.
22328 * expr.c (maybe_emit_group_store): New function.
22329 * builtins.c (expand_builtin_int_roundingfn): Call it.
22330 (expand_builtin_int_roundingfn_2): Likewise.
22332 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22334 PR rtl-optimization/55829
22335 * lra-constraints.c (match_reload): Add code for absent output.
22336 (curr_insn_transform): Add code for reloads of matched inputs
22339 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22341 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22342 attribute of movddup insn to DF.
22343 (*vec_interleave_lowv2df): Ditto.
22344 (vec_dupv2df): Ditto.
22346 2013-01-09 Jan Hubicka <jh@suse.cz>
22348 PR tree-optimiation/55875
22349 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22350 EVERY_ITERATION parameter.
22351 (number_of_iterations_exit): Check if exit is executed every iteration.
22352 (idx_infer_loop_bounds): Similarly here.
22353 (n_of_executions_at_most): Simplify
22354 to only test for cases where statement is dominated by the
22355 particular bound; handle correctly the "postdominance" test.
22356 (scev_probably_wraps_p): Use max loop iterations info
22357 as a global bound first.
22359 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22360 Nick Clifton <nickc@redhat.com>
22362 * config/v850/v850.md (cbranchsf4): New pattern.
22363 (cstoresf4): New pattern.
22364 (cbranchdf4): New pattern.
22365 (cstoredf4): New pattern.
22366 (movsicc): Disallow floating point comparisons.
22367 (cmpsf_le_insn): Fix order of operators.
22368 (cmpsf_lt_insn): Likewise.
22369 (cmpsf_eq_insn): Likewise.
22370 (cmpdf_le_insn): Likewise.
22371 (cmpdf_lt_insn): Likewise.
22372 (cmpdf_eq_insn): Likewise.
22373 (cmpsf_ge_insn): Use LE comparison.
22374 (cmpdf_ge_insn): Likewise.
22375 (cmpsf_gt_insn): Use LT comparison.
22376 (cmpdf_gt_insn): Likewise.
22377 (cmpsf_ne_insn): Delete pattern.
22378 (cmpdf_ne_insn): Delete pattern.
22379 * config/v850/v850.c (v850_gen_float_compare): Use
22380 gen_cmpdf_eq_insn for NE comparison.
22381 (v850_float_z_comparison_operator)
22382 (v850_float_nz_comparison_operator): Move from here ...
22383 * config/v850/predicates.md: ... to here. Move GT and GE
22384 comparisons into v850_float_z_comparison_operator.
22385 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22387 (v850_float_nz_comparison_operator): Likewise.
22389 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22391 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22392 with calls to gen_insvsi/gen_insvdi.
22394 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22396 * config/i386/i386.c (initial_ix86_tune_features): Set up
22397 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22399 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22400 Jakub Jelinek <jakub@redhat.com>
22402 PR tree-optimization/48189
22403 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22404 If nitercst is 0, don't predict the exit edge.
22406 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22408 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22409 in asm_fprintf with reg_names.
22410 (aarch64_print_operand_address): Likewise.
22411 (aarch64_return_addr): Likewise.
22412 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22414 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22416 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22417 (INT_U6_BITS): Likewise.
22418 * config/pa/predicates.md (uint6_operand): New predicate.
22419 (shift5_operand, shift6_operand): Likewise.
22420 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22422 (lshrdi3): Use shift6_operand.
22423 (shrpsi4, shrpdi4): New insn patterns.
22424 (extzv): Delete expander.
22425 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22426 predicates in unamed zero extract patterns. Tighten common constraint.
22427 (extv): Delete expander.
22428 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22429 predicates in unamed sign extract patterns. Tighten common constraint.
22430 (insv): Delete expander.
22431 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22432 predicates in unamed insert patterns. Tighten common constraint.
22433 Change uint32_operand predicate to uint6_operand predicate in unamed
22434 DImode pattern to insert constant values of type 1...1xxxx.
22436 2013-01-04 Jan Hubicka <jh@suse.cz>
22438 PR tree-optimization/55823
22439 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22442 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22443 Uros Bizjak <ubizjak@gmail.com>
22445 PR rtl-optimization/55845
22446 * df-problems.c (can_move_insns_across): Stop scanning at
22447 volatile_insn_p source instruction or give up if
22448 across_from .. across_to range contains any volatile_insn_p
22451 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22453 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22454 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22456 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22457 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22459 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22462 * asan.c (asan_clear_shadow): New function.
22463 (asan_emit_stack_protection): Use it.
22465 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22467 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22468 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22469 with tab instead of space.
22471 2013-01-08 Nick Clifton <nickc@redhat.com>
22473 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22474 register bank 0 at the start of an interrupt handler.
22475 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22478 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22480 * config/aarch64/aarch64-simd.md
22481 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22482 (aarch64_simd_bsl): Likewise.
22483 (aarch64_vcond_internal<mode>): Likewise.
22484 (vcond<mode><mode>): Likewise.
22485 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22486 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22488 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22490 * config/aarch64/aarch64-builtins.c
22491 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22493 2013-01-08 Martin Jambor <mjambor@suse.cz>
22496 * tree-sra.c (analyze_access_subtree): Return true also after
22497 potentially creating a debug-only replacement.
22499 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22501 PR middle-end/55890
22502 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22504 PR tree-optimization/54120
22505 * tree-vrp.c (range_fits_type_p): Don't allow
22506 src_precision < precision from signed vr to unsigned_p
22507 if vr->min or vr->max is negative.
22508 (simplify_float_conversion_using_ranges): Test can_float_p
22509 against CODE_FOR_nothing.
22511 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22512 Richard Biener <rguenther@suse.de>
22514 PR middle-end/55851
22515 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22516 types instead of just INTEGER_TYPE types.
22518 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22520 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22523 2013-01-07 Steve Ellcey <sellcey@mips.com>
22526 * config/mips/mips.opt: Change mad to mmad to match documentation.
22528 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22531 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22532 .progmemx.data now.
22534 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22537 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22538 (avr_addrspace_t): Add .section_name field.
22539 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22541 (avr_addrspace): Same. Initialize .section_name. Remove last
22542 NULL entry. Put __memx into .progmemx.data.
22543 (progmem_section_prefix): Remove.
22544 (avr_asm_init_sections): No need to initialize progmem_section.
22545 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22546 section name prefix.
22547 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22548 retrieve the progmem section.
22549 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22550 boundary to run over avr_addrspace[].
22551 (avr_register_target_pragmas): Ditto.
22553 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22555 * varasm.c (output_constant_def_contents): For asan_protect_global
22556 protected strings, adjust DECL_ALIGN if needed, before testing for
22558 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22559 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22561 (output_object_block): For asan protected decls, emit asan padding
22562 after their contents.
22563 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22564 (asan_finish_file): Test it here instead.
22566 2013-01-07 Nick Clifton <nickc@redhat.com>
22567 Matthias Klose <doko@debian.org>
22568 Doug Kwan <dougkwan@google.com>
22569 H.J. Lu <hongjiu.lu@intel.com>
22572 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22574 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22576 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22578 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22580 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22582 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22585 * doc/install.texi (Cross-Compiler-Specific Options): Document
22588 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
22590 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
22591 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
22592 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
22593 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
22594 vqmovun_high_s64): Fix source operand number and update copyright.
22596 2013-01-07 Richard Biener <rguenther@suse.de>
22598 PR middle-end/55890
22599 * gimple.h (gimple_call_builtin_p): New overload.
22600 * gimple.c (validate_call): New function.
22601 (gimple_call_builtin_p): Likewise.
22602 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22603 Use gimple_call_builtin_p.
22604 (find_func_clobbers): Likewise.
22605 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
22606 (strlen_optimize_stmt): Likewise.
22608 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22610 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
22611 (vld1q_dup_*): Likewise.
22612 (vld1_*): Likewise.
22613 (vld1q_*): Likewise.
22614 (vld1_lane_*): Likewise.
22615 (vld1q_lane_*): Likewise.
22617 2013-01-07 Richard Biener <rguenther@suse.de>
22619 * lto-streamer.h (LTO_minor_version): Bump to 2.
22621 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22623 * config/aarch64/aarch64-protos.h
22624 (aarch64_const_double_zero_rtx_p): Rename to...
22625 (aarch64_float_const_zero_rtx_p): ...this.
22626 (aarch64_float_const_representable_p): New.
22627 (aarch64_output_simd_mov_immediate): Likewise.
22628 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
22629 move immediate case.
22630 * config/aarch64/aarch64.c
22631 (aarch64_const_double_zero_rtx_p): Rename to...
22632 (aarch64_float_const_zero_rtx_p): ...this.
22633 (aarch64_print_operand): Allow printing of new constants.
22634 (aarch64_valid_floating_const): New.
22635 (aarch64_legitimate_constant_p): Check for valid floating-point
22637 (aarch64_simd_valid_immediate): Likewise.
22638 (aarch64_vect_float_const_representable_p): New.
22639 (aarch64_float_const_representable_p): Likewise.
22640 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
22641 (aarch64_output_simd_mov_immediate): New.
22642 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
22643 (*movdf_aarch64): Likewise.
22644 * config/aarch64/constraints.md (Ufc): New.
22645 (Y): call aarch64_float_const_zero_rtx.
22646 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
22648 2013-01-07 Richard Biener <rguenther@suse.de>
22650 PR tree-optimization/55888
22651 PR tree-optimization/55862
22652 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
22653 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
22654 not if it is contained therein.
22656 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22658 * config/avr/t-avr: Typo.
22660 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22663 * config/avr/t-avr: Don't automatically rebuild
22664 $(srcdir)/config/avr/t-multilib
22665 $(srcdir)/config/avr/avr-tables.opt
22666 $(srcdir)/doc/avr-mmcu.texi
22667 (avr-mcus): New phony target to build them on request.
22668 (s-avr-mlib, s-avr-mmcu-texi): Remove.
22669 * avr/avr-mcus.def: Adjust comments.
22671 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
22673 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
22675 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22677 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
22679 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
22681 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
22683 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
22685 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
22686 to generate profiling.
22687 * config/rs6000/aix64.h (LIB_SPEC): Same.
22689 2013-01-04 Andrew Pinski <apinski@cavium.com>
22691 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
22693 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
22695 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
22697 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
22699 (ix86_expand_move): Ditto.
22700 (ix86_zero_extend_to_Pmode): Ditto.
22701 (ix86_expand_call): Ditto.
22702 (ix86_expand_special_args_builtin): Ditto.
22703 (ix86_expand_builtin): Ditto.
22705 2013-01-04 Richard Biener <rguenther@suse.de>
22707 PR tree-optimization/55862
22708 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
22709 translating them through PHI nodes.
22711 2013-01-04 Martin Jambor <mjambor@suse.cz>
22713 PR tree-optimization/55755
22714 * tree-sra.c (sra_modify_assign): Do not check that an access has no
22715 children when trying to avoid producing a VIEW_CONVERT_EXPR.
22717 2013-01-04 Marek Polacek <polacek@redhat.com>
22719 PR middle-end/55859
22720 * opts.c (default_options_optimization): Clarify error message.
22722 2013-01-04 Richard Biener <rguenther@suse.de>
22724 PR middle-end/55863
22725 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
22728 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22731 * config/pa/pa.md (movsi): Revert previous change.
22732 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
22735 2013-01-03 Richard Henderson <rth@redhat.com>
22737 * config/i386/i386.c (ix86_expand_move): Always assign to op1
22738 after eliminating TLS symbols.
22740 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
22743 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
22744 * graphite-poly.c (debug_gmp_value): Likewise.
22746 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
22749 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
22750 selected code model, define __code_mode_small__, __code_model_medium__,
22751 __code_model_large__, __code_model_32__ or __code_model_kernel__.
22752 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
22753 xchg temporary register with %k. Declare temporary register as
22755 [__x86_64__]: For medium and large code models, preserve %rbx register.
22757 2013-01-03 Richard Biener <rguenther@suse.de>
22759 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
22760 (dump_subscript): Adjust.
22761 (finalize_ddr_dependent): Do not dump redundant info.
22762 (analyze_siv_subscript): Adjust.
22763 (subscript_dependence_tester): Likewise.
22764 (compute_affine_dependence): Likewise.
22766 2013-01-03 Richard Biener <rguenther@suse.de>
22769 2013-01-03 Richard Biener <rguenther@suse.de>
22771 PR tree-optimization/55857
22772 * tree-vect-stmts.c (vectorizable_load): Do not setup
22773 re-alignment for invariant loads.
22775 2013-01-02 Richard Biener <rguenther@suse.de>
22777 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22778 invariant load do not generate a vector load from the scalar location.
22780 2013-01-03 Richard Biener <rguenther@suse.de>
22782 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
22783 for not vectorizing.
22784 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
22785 not build INDIRECT_REFs, call get_name once only.
22786 (vect_create_data_ref_ptr): Likewise. Dump base object kind
22787 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
22789 2013-01-03 Richard Biener <rguenther@suse.de>
22791 PR tree-optimization/55857
22792 * tree-vect-stmts.c (vectorizable_load): Do not setup
22793 re-alignment for invariant loads.
22795 2013-01-03 Richard Biener <rguenther@suse.de>
22798 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
22799 prefer a built-in decl.
22801 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22803 * gcc.c (process_command): Update copyright notice dates.
22804 * gcov.c (print_version): Likewise.
22805 * gcov-dump.c (print_version): Likewise.
22807 PR rtl-optimization/55838
22808 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
22809 iv0.step, iv1.step and step.
22811 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22812 Marc Glisse <marc.glisse@inria.fr>
22814 PR tree-optimization/55832
22815 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
22816 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
22817 integer_{one,zero}_node.
22819 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22822 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
22823 * var-tracking.c (reverse_op): Don't add reverse ops to
22824 VALUEs that have already
22825 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
22827 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
22829 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
22831 2013-01-02 Teresa Johnson <tejohnson@google.com>
22833 * dumpfile.c (dump_loc): Print filename with location.
22834 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
22835 new location_t parameter to emit complete unroll message with
22836 new dump framework.
22837 (canonicalize_loop_induction_variables): Compute loops location
22838 and pass to try_unroll_loop_completely.
22839 * loop-unroll.c (report_unroll_peel): New function.
22840 (peel_loops_completely): Use new dump format with location
22841 for main dumpfile message, and invoke report_unroll_peel on success.
22842 (decide_unrolling_and_peeling): Ditto.
22843 (decide_peel_once_rolling): Remove old dumpfile message subsumed
22844 by report_unroll_peel.
22845 (decide_peel_completely): Ditto.
22846 (decide_unroll_constant_iterations): Ditto.
22847 (decide_unroll_runtime_iterations): Ditto.
22848 (decide_peel_simple): Ditto.
22849 (decide_unroll_stupid): Ditto.
22850 * cfgloop.c (get_loop_location): New function.
22851 * cfgloop.h (get_loop_location): Declare.
22853 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22855 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
22858 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22860 PR middle-end/55198
22861 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
22862 BLKmode objects when EXPAND_MEMORY is specified.
22864 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22866 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
22868 (fold_builtin_cpu): Do not share cpu model decls across statements.
22870 2013-01-02 Jason Merrill <jason@redhat.com>
22873 * tree.c (build_array_type_1): Revert earlier change.
22875 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
22877 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
22879 * config/aarch64/aarch64-tune.md: Re-generate.
22881 2013-01-02 Richard Biener <rguenther@suse.de>
22883 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22884 invariant load do not generate a vector load from the scalar location.
22886 2013-01-02 Richard Biener <rguenther@suse.de>
22889 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
22890 * configure: Regenerate.
22892 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22894 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
22895 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
22896 (expand_builtin_int_roundingfn_2): Keep the original target around
22897 for the fallback case.
22899 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22901 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
22902 to be clear for sign changes.
22904 2013-01-01 Jan Hubicka <jh@suse.cz>
22906 * ipa-inline-analysis.c: Fix formatting.
22908 2013-01-01 Jakub Jelinek <jakub@redhat.com>
22910 PR tree-optimization/55831
22911 * tree-vect-loop.c (get_initial_def_for_induction): Use
22912 gsi_after_labels instead of gsi_start_bb.
22914 Copyright (C) 2013 Free Software Foundation, Inc.
22916 Copying and distribution of this file, with or without modification,
22917 are permitted in any medium without royalty provided the copyright
22918 notice and this notice are preserved.