| /* |
| * Copyright (c) 2005, 2016, Oracle and/or its affiliates. All rights reserved. |
| * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
| * |
| * This code is free software; you can redistribute it and/or modify it |
| * under the terms of the GNU General Public License version 2 only, as |
| * published by the Free Software Foundation. |
| * |
| * This code is distributed in the hope that it will be useful, but WITHOUT |
| * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
| * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
| * version 2 for more details (a copy is included in the LICENSE file that |
| * accompanied this code). |
| * |
| * You should have received a copy of the GNU General Public License version |
| * 2 along with this work; if not, write to the Free Software Foundation, |
| * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. |
| * |
| * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
| * or visit www.oracle.com if you need additional information or have any |
| * questions. |
| * |
| */ |
| |
| #ifndef SHARE_VM_C1_C1_LINEARSCAN_HPP |
| #define SHARE_VM_C1_C1_LINEARSCAN_HPP |
| |
| #include "c1/c1_FpuStackSim.hpp" |
| #include "c1/c1_FrameMap.hpp" |
| #include "c1/c1_IR.hpp" |
| #include "c1/c1_Instruction.hpp" |
| #include "c1/c1_LIR.hpp" |
| #include "c1/c1_LIRGenerator.hpp" |
| #include "utilities/align.hpp" |
| #include "utilities/macros.hpp" |
| |
| class DebugInfoCache; |
| class FpuStackAllocator; |
| class IRScopeDebugInfo; |
| class Interval; |
| class IntervalWalker; |
| class LIRGenerator; |
| class LinearScan; |
| class MoveResolver; |
| class Range; |
| |
| typedef GrowableArray<Interval*> IntervalArray; |
| typedef GrowableArray<Interval*> IntervalList; |
| typedef GrowableArray<IntervalList*> IntervalsList; |
| typedef GrowableArray<ScopeValue*> ScopeValueArray; |
| typedef GrowableArray<LIR_OpList*> LIR_OpListStack; |
| |
| enum IntervalUseKind { |
| // priority of use kinds must be ascending |
| noUse = 0, |
| loopEndMarker = 1, |
| shouldHaveRegister = 2, |
| mustHaveRegister = 3, |
| |
| firstValidKind = 1, |
| lastValidKind = 3 |
| }; |
| |
| enum IntervalKind { |
| fixedKind = 0, // interval pre-colored by LIR_Generator |
| anyKind = 1, // no register/memory allocated by LIR_Generator |
| nofKinds, |
| firstKind = fixedKind |
| }; |
| |
| |
| // during linear scan an interval is in one of four states in |
| enum IntervalState { |
| unhandledState = 0, // unhandled state (not processed yet) |
| activeState = 1, // life and is in a physical register |
| inactiveState = 2, // in a life time hole and is in a physical register |
| handledState = 3, // spilled or not life again |
| invalidState = -1 |
| }; |
| |
| |
| enum IntervalSpillState { |
| noDefinitionFound, // starting state of calculation: no definition found yet |
| oneDefinitionFound, // one definition has already been found. |
| // Note: two consecutive definitions are treated as one (e.g. consecutive move and add because of two-operand LIR form) |
| // the position of this definition is stored in _definition_pos |
| oneMoveInserted, // one spill move has already been inserted. |
| storeAtDefinition, // the interval should be stored immediately after its definition because otherwise |
| // there would be multiple redundant stores |
| startInMemory, // the interval starts in memory (e.g. method parameter), so a store is never necessary |
| noOptimization // the interval has more then one definition (e.g. resulting from phi moves), so stores to memory are not optimized |
| }; |
| |
| |
| #define for_each_interval_kind(kind) \ |
| for (IntervalKind kind = firstKind; kind < nofKinds; kind = (IntervalKind)(kind + 1)) |
| |
| #define for_each_visitor_mode(mode) \ |
| for (LIR_OpVisitState::OprMode mode = LIR_OpVisitState::firstMode; mode < LIR_OpVisitState::numModes; mode = (LIR_OpVisitState::OprMode)(mode + 1)) |
| |
| |
| class LinearScan : public CompilationResourceObj { |
| // declare classes used by LinearScan as friends because they |
| // need a wide variety of functions declared here |
| // |
| // Only the small interface to the rest of the compiler is public |
| friend class Interval; |
| friend class IntervalWalker; |
| friend class LinearScanWalker; |
| friend class FpuStackAllocator; |
| friend class MoveResolver; |
| friend class LinearScanStatistic; |
| friend class LinearScanTimers; |
| friend class RegisterVerifier; |
| |
| public: |
| enum { |
| any_reg = -1, |
| nof_cpu_regs = pd_nof_cpu_regs_linearscan, |
| nof_fpu_regs = pd_nof_fpu_regs_linearscan, |
| nof_xmm_regs = pd_nof_xmm_regs_linearscan, |
| nof_regs = nof_cpu_regs + nof_fpu_regs + nof_xmm_regs |
| }; |
| |
| private: |
| Compilation* _compilation; |
| IR* _ir; |
| LIRGenerator* _gen; |
| FrameMap* _frame_map; |
| |
| BlockList _cached_blocks; // cached list with all blocks in linear-scan order (only correct if original list keeps unchanged) |
| int _num_virtual_regs; // number of virtual registers (without new registers introduced because of splitting intervals) |
| bool _has_fpu_registers; // true if this method uses any floating point registers (and so fpu stack allocation is necessary) |
| int _num_calls; // total number of calls in this method |
| int _max_spills; // number of stack slots used for intervals allocated to memory |
| int _unused_spill_slot; // unused spill slot for a single-word value because of alignment of a double-word value |
| |
| IntervalList _intervals; // mapping from register number to interval |
| IntervalList* _new_intervals_from_allocation; // list with all intervals created during allocation when an existing interval is split |
| IntervalArray* _sorted_intervals; // intervals sorted by Interval::from() |
| bool _needs_full_resort; // set to true if an Interval::from() is changed and _sorted_intervals must be resorted |
| |
| LIR_OpArray _lir_ops; // mapping from LIR_Op id to LIR_Op node |
| BlockBeginArray _block_of_op; // mapping from LIR_Op id to the BlockBegin containing this instruction |
| ResourceBitMap _has_info; // bit set for each LIR_Op id that has a CodeEmitInfo |
| ResourceBitMap _has_call; // bit set for each LIR_Op id that destroys all caller save registers |
| BitMap2D _interval_in_loop; // bit set for each virtual register that is contained in each loop |
| |
| // cached debug info to prevent multiple creation of same object |
| // TODO: cached scope values for registers could be static |
| ScopeValueArray _scope_value_cache; |
| |
| static ConstantOopWriteValue* _oop_null_scope_value; |
| static ConstantIntValue* _int_m1_scope_value; |
| static ConstantIntValue* _int_0_scope_value; |
| static ConstantIntValue* _int_1_scope_value; |
| static ConstantIntValue* _int_2_scope_value; |
| |
| // accessors |
| IR* ir() const { return _ir; } |
| Compilation* compilation() const { return _compilation; } |
| LIRGenerator* gen() const { return _gen; } |
| FrameMap* frame_map() const { return _frame_map; } |
| |
| // unified bailout support |
| void bailout(const char* msg) const { compilation()->bailout(msg); } |
| bool bailed_out() const { return compilation()->bailed_out(); } |
| |
| // access to block list (sorted in linear scan order) |
| int block_count() const { assert(_cached_blocks.length() == ir()->linear_scan_order()->length(), "invalid cached block list"); return _cached_blocks.length(); } |
| BlockBegin* block_at(int idx) const { assert(_cached_blocks.at(idx) == ir()->linear_scan_order()->at(idx), "invalid cached block list"); return _cached_blocks.at(idx); } |
| |
| int num_virtual_regs() const { return _num_virtual_regs; } |
| // size of live_in and live_out sets of BasicBlocks (BitMap needs rounded size for iteration) |
| int live_set_size() const { return align_up(_num_virtual_regs, BitsPerWord); } |
| bool has_fpu_registers() const { return _has_fpu_registers; } |
| int num_loops() const { return ir()->num_loops(); } |
| bool is_interval_in_loop(int interval, int loop) const { return _interval_in_loop.at(interval, loop); } |
| |
| // handling of fpu stack allocation (platform dependent, needed for debug information generation) |
| #ifdef X86 |
| FpuStackAllocator* _fpu_stack_allocator; |
| bool use_fpu_stack_allocation() const { return UseSSE < 2 && has_fpu_registers(); } |
| #else |
| bool use_fpu_stack_allocation() const { return false; } |
| #endif |
| |
| |
| // access to interval list |
| int interval_count() const { return _intervals.length(); } |
| Interval* interval_at(int reg_num) const { return _intervals.at(reg_num); } |
| |
| IntervalList* new_intervals_from_allocation() const { return _new_intervals_from_allocation; } |
| |
| // access to LIR_Ops and Blocks indexed by op_id |
| int max_lir_op_id() const { assert(_lir_ops.length() > 0, "no operations"); return (_lir_ops.length() - 1) << 1; } |
| LIR_Op* lir_op_with_id(int op_id) const { assert(op_id >= 0 && op_id <= max_lir_op_id() && op_id % 2 == 0, "op_id out of range or not even"); return _lir_ops.at(op_id >> 1); } |
| BlockBegin* block_of_op_with_id(int op_id) const { assert(_block_of_op.length() > 0 && op_id >= 0 && op_id <= max_lir_op_id() + 1, "op_id out of range"); return _block_of_op.at(op_id >> 1); } |
| |
| bool is_block_begin(int op_id) { return op_id == 0 || block_of_op_with_id(op_id) != block_of_op_with_id(op_id - 1); } |
| bool covers_block_begin(int op_id_1, int op_id_2) { return block_of_op_with_id(op_id_1) != block_of_op_with_id(op_id_2); } |
| |
| bool has_call(int op_id) { assert(op_id % 2 == 0, "must be even"); return _has_call.at(op_id >> 1); } |
| bool has_info(int op_id) { assert(op_id % 2 == 0, "must be even"); return _has_info.at(op_id >> 1); } |
| |
| |
| // functions for converting LIR-Operands to register numbers |
| static bool is_valid_reg_num(int reg_num) { return reg_num >= 0; } |
| static int reg_num(LIR_Opr opr); |
| static int reg_numHi(LIR_Opr opr); |
| |
| // functions for classification of intervals |
| static bool is_precolored_interval(const Interval* i); |
| static bool is_virtual_interval(const Interval* i); |
| |
| static bool is_precolored_cpu_interval(const Interval* i); |
| static bool is_virtual_cpu_interval(const Interval* i); |
| static bool is_precolored_fpu_interval(const Interval* i); |
| static bool is_virtual_fpu_interval(const Interval* i); |
| |
| static bool is_in_fpu_register(const Interval* i); |
| static bool is_oop_interval(const Interval* i); |
| |
| |
| // General helper functions |
| int allocate_spill_slot(bool double_word); |
| void assign_spill_slot(Interval* it); |
| void propagate_spill_slots(); |
| |
| Interval* create_interval(int reg_num); |
| void append_interval(Interval* it); |
| void copy_register_flags(Interval* from, Interval* to); |
| |
| // platform dependent functions |
| static bool is_processed_reg_num(int reg_num); |
| static int num_physical_regs(BasicType type); |
| static bool requires_adjacent_regs(BasicType type); |
| static bool is_caller_save(int assigned_reg); |
| |
| // spill move optimization: eliminate moves from register to stack if |
| // stack slot is known to be correct |
| void change_spill_definition_pos(Interval* interval, int def_pos); |
| void change_spill_state(Interval* interval, int spill_pos); |
| static bool must_store_at_definition(const Interval* i); |
| void eliminate_spill_moves(); |
| |
| // Phase 1: number all instructions in all blocks |
| void number_instructions(); |
| |
| // Phase 2: compute local live sets separately for each block |
| // (sets live_gen and live_kill for each block) |
| // |
| // helper methods used by compute_local_live_sets() |
| void set_live_gen_kill(Value value, LIR_Op* op, BitMap& live_gen, BitMap& live_kill); |
| |
| void compute_local_live_sets(); |
| |
| // Phase 3: perform a backward dataflow analysis to compute global live sets |
| // (sets live_in and live_out for each block) |
| void compute_global_live_sets(); |
| |
| |
| // Phase 4: build intervals |
| // (fills the list _intervals) |
| // |
| // helper methods used by build_intervals() |
| void add_use (Value value, int from, int to, IntervalUseKind use_kind); |
| |
| void add_def (LIR_Opr opr, int def_pos, IntervalUseKind use_kind); |
| void add_use (LIR_Opr opr, int from, int to, IntervalUseKind use_kind); |
| void add_temp(LIR_Opr opr, int temp_pos, IntervalUseKind use_kind); |
| |
| void add_def (int reg_num, int def_pos, IntervalUseKind use_kind, BasicType type); |
| void add_use (int reg_num, int from, int to, IntervalUseKind use_kind, BasicType type); |
| void add_temp(int reg_num, int temp_pos, IntervalUseKind use_kind, BasicType type); |
| |
| // Add platform dependent kills for particular LIR ops. Can be used |
| // to add platform dependent behaviour for some operations. |
| void pd_add_temps(LIR_Op* op); |
| |
| IntervalUseKind use_kind_of_output_operand(LIR_Op* op, LIR_Opr opr); |
| IntervalUseKind use_kind_of_input_operand(LIR_Op* op, LIR_Opr opr); |
| void handle_method_arguments(LIR_Op* op); |
| void handle_doubleword_moves(LIR_Op* op); |
| void add_register_hints(LIR_Op* op); |
| |
| void build_intervals(); |
| |
| |
| // Phase 5: actual register allocation |
| // (Uses LinearScanWalker) |
| // |
| // helper functions for building a sorted list of intervals |
| NOT_PRODUCT(bool is_sorted(IntervalArray* intervals);) |
| static int interval_cmp(Interval** a, Interval** b); |
| void add_to_list(Interval** first, Interval** prev, Interval* interval); |
| void create_unhandled_lists(Interval** list1, Interval** list2, bool (is_list1)(const Interval* i), bool (is_list2)(const Interval* i)); |
| |
| void sort_intervals_before_allocation(); |
| void sort_intervals_after_allocation(); |
| void allocate_registers(); |
| |
| |
| // Phase 6: resolve data flow |
| // (insert moves at edges between blocks if intervals have been split) |
| // |
| // helper functions for resolve_data_flow() |
| Interval* split_child_at_op_id(Interval* interval, int op_id, LIR_OpVisitState::OprMode mode); |
| Interval* interval_at_block_begin(BlockBegin* block, int reg_num); |
| Interval* interval_at_block_end(BlockBegin* block, int reg_num); |
| Interval* interval_at_op_id(int reg_num, int op_id); |
| void resolve_collect_mappings(BlockBegin* from_block, BlockBegin* to_block, MoveResolver &move_resolver); |
| void resolve_find_insert_pos(BlockBegin* from_block, BlockBegin* to_block, MoveResolver &move_resolver); |
| void resolve_data_flow(); |
| |
| void resolve_exception_entry(BlockBegin* block, int reg_num, MoveResolver &move_resolver); |
| void resolve_exception_entry(BlockBegin* block, MoveResolver &move_resolver); |
| void resolve_exception_edge(XHandler* handler, int throwing_op_id, int reg_num, Phi* phi, MoveResolver &move_resolver); |
| void resolve_exception_edge(XHandler* handler, int throwing_op_id, MoveResolver &move_resolver); |
| void resolve_exception_handlers(); |
| |
| // Phase 7: assign register numbers back to LIR |
| // (includes computation of debug information and oop maps) |
| // |
| // helper functions for assign_reg_num() |
| VMReg vm_reg_for_interval(Interval* interval); |
| VMReg vm_reg_for_operand(LIR_Opr opr); |
| |
| static LIR_Opr operand_for_interval(Interval* interval); |
| static LIR_Opr calc_operand_for_interval(const Interval* interval); |
| LIR_Opr canonical_spill_opr(Interval* interval); |
| |
| LIR_Opr color_lir_opr(LIR_Opr opr, int id, LIR_OpVisitState::OprMode); |
| |
| // methods used for oop map computation |
| IntervalWalker* init_compute_oop_maps(); |
| OopMap* compute_oop_map(IntervalWalker* iw, LIR_Op* op, CodeEmitInfo* info, bool is_call_site); |
| void compute_oop_map(IntervalWalker* iw, const LIR_OpVisitState &visitor, LIR_Op* op); |
| |
| // methods used for debug information computation |
| void init_compute_debug_info(); |
| |
| MonitorValue* location_for_monitor_index(int monitor_index); |
| LocationValue* location_for_name(int name, Location::Type loc_type); |
| void set_oop(OopMap* map, VMReg name) { |
| if (map->legal_vm_reg_name(name)) { |
| map->set_oop(name); |
| } else { |
| bailout("illegal oopMap register name"); |
| } |
| } |
| |
| int append_scope_value_for_constant(LIR_Opr opr, GrowableArray<ScopeValue*>* scope_values); |
| int append_scope_value_for_operand(LIR_Opr opr, GrowableArray<ScopeValue*>* scope_values); |
| int append_scope_value(int op_id, Value value, GrowableArray<ScopeValue*>* scope_values); |
| |
| IRScopeDebugInfo* compute_debug_info_for_scope(int op_id, IRScope* cur_scope, ValueStack* cur_state, ValueStack* innermost_state); |
| void compute_debug_info(CodeEmitInfo* info, int op_id); |
| |
| void assign_reg_num(LIR_OpList* instructions, IntervalWalker* iw); |
| void assign_reg_num(); |
| |
| |
| // Phase 8: fpu stack allocation |
| // (Used only on x86 when fpu operands are present) |
| void allocate_fpu_stack(); |
| |
| |
| // helper functions for printing state |
| #ifndef PRODUCT |
| static void print_bitmap(BitMap& bitmap); |
| void print_intervals(const char* label); |
| void print_lir(int level, const char* label, bool hir_valid = true); |
| #endif |
| |
| #ifdef ASSERT |
| // verification functions for allocation |
| // (check that all intervals have a correct register and that no registers are overwritten) |
| void verify(); |
| void verify_intervals(); |
| void verify_no_oops_in_fixed_intervals(); |
| void verify_constants(); |
| void verify_registers(); |
| #endif |
| |
| public: |
| // creation |
| LinearScan(IR* ir, LIRGenerator* gen, FrameMap* frame_map); |
| |
| // main entry function: perform linear scan register allocation |
| void do_linear_scan(); |
| |
| // accessors used by Compilation |
| int max_spills() const { return _max_spills; } |
| int num_calls() const { assert(_num_calls >= 0, "not set"); return _num_calls; } |
| |
| // entry functions for printing |
| #ifndef PRODUCT |
| static void print_statistics(); |
| static void print_timers(double total); |
| #endif |
| }; |
| |
| |
| // Helper class for ordering moves that are inserted at the same position in the LIR |
| // When moves between registers are inserted, it is important that the moves are |
| // ordered such that no register is overwritten. So moves from register to stack |
| // are processed prior to moves from stack to register. When moves have circular |
| // dependencies, a temporary stack slot is used to break the circle. |
| // The same logic is used in the LinearScanWalker and in LinearScan during resolve_data_flow |
| // and therefore factored out in a separate class |
| class MoveResolver: public StackObj { |
| private: |
| LinearScan* _allocator; |
| |
| LIR_List* _insert_list; |
| int _insert_idx; |
| LIR_InsertionBuffer _insertion_buffer; // buffer where moves are inserted |
| |
| IntervalList _mapping_from; |
| LIR_OprList _mapping_from_opr; |
| IntervalList _mapping_to; |
| bool _multiple_reads_allowed; |
| int _register_blocked[LinearScan::nof_regs]; |
| |
| int register_blocked(int reg) { assert(reg >= 0 && reg < LinearScan::nof_regs, "out of bounds"); return _register_blocked[reg]; } |
| void set_register_blocked(int reg, int direction) { assert(reg >= 0 && reg < LinearScan::nof_regs, "out of bounds"); assert(direction == 1 || direction == -1, "out of bounds"); _register_blocked[reg] += direction; } |
| |
| void block_registers(Interval* it); |
| void unblock_registers(Interval* it); |
| bool save_to_process_move(Interval* from, Interval* to); |
| |
| void create_insertion_buffer(LIR_List* list); |
| void append_insertion_buffer(); |
| void insert_move(Interval* from_interval, Interval* to_interval); |
| void insert_move(LIR_Opr from_opr, Interval* to_interval); |
| |
| DEBUG_ONLY(void verify_before_resolve();) |
| void resolve_mappings(); |
| public: |
| MoveResolver(LinearScan* allocator); |
| |
| DEBUG_ONLY(void check_empty();) |
| void set_multiple_reads_allowed() { _multiple_reads_allowed = true; } |
| void set_insert_position(LIR_List* insert_list, int insert_idx); |
| void move_insert_position(LIR_List* insert_list, int insert_idx); |
| void add_mapping(Interval* from, Interval* to); |
| void add_mapping(LIR_Opr from, Interval* to); |
| void resolve_and_append_moves(); |
| |
| LinearScan* allocator() { return _allocator; } |
| bool has_mappings() { return _mapping_from.length() > 0; } |
| }; |
| |
| |
| class Range : public CompilationResourceObj { |
| friend class Interval; |
| |
| private: |
| static Range* _end; // sentinel (from == to == max_jint) |
| |
| int _from; // from (inclusive) |
| int _to; // to (exclusive) |
| Range* _next; // linear list of Ranges |
| |
| // used only by class Interval, so hide them |
| bool intersects(Range* r) const { return intersects_at(r) != -1; } |
| int intersects_at(Range* r) const; |
| |
| public: |
| Range(int from, int to, Range* next); |
| |
| static void initialize(Arena* arena); |
| static Range* end() { return _end; } |
| |
| int from() const { return _from; } |
| int to() const { return _to; } |
| Range* next() const { return _next; } |
| void set_from(int from) { _from = from; } |
| void set_to(int to) { _to = to; } |
| void set_next(Range* next) { _next = next; } |
| |
| // for testing |
| void print(outputStream* out = tty) const PRODUCT_RETURN; |
| }; |
| |
| |
| // Interval is an ordered list of disjoint ranges. |
| |
| // For pre-colored double word LIR_Oprs, one interval is created for |
| // the low word register and one is created for the hi word register. |
| // On Intel for FPU double registers only one interval is created. At |
| // all times assigned_reg contains the reg. number of the physical |
| // register. |
| |
| // For LIR_Opr in virtual registers a single interval can represent |
| // single and double word values. When a physical register is |
| // assigned to the interval, assigned_reg contains the |
| // phys. reg. number and for double word values assigned_regHi the |
| // phys. reg. number of the hi word if there is any. For spilled |
| // intervals assigned_reg contains the stack index. assigned_regHi is |
| // always -1. |
| |
| class Interval : public CompilationResourceObj { |
| private: |
| static Interval* _end; // sentinel (interval with only range Range::end()) |
| |
| int _reg_num; |
| BasicType _type; // valid only for virtual registers |
| Range* _first; // sorted list of Ranges |
| intStack _use_pos_and_kinds; // sorted list of use-positions and their according use-kinds |
| |
| Range* _current; // interval iteration: the current Range |
| Interval* _next; // interval iteration: sorted list of Intervals (ends with sentinel) |
| IntervalState _state; // interval iteration: to which set belongs this interval |
| |
| |
| int _assigned_reg; |
| int _assigned_regHi; |
| |
| int _cached_to; // cached value: to of last range (-1: not cached) |
| LIR_Opr _cached_opr; |
| VMReg _cached_vm_reg; |
| |
| Interval* _split_parent; // the original interval where this interval is derived from |
| IntervalList _split_children; // list of all intervals that are split off from this interval (only available for split parents) |
| Interval* _current_split_child; // the current split child that has been active or inactive last (always stored in split parents) |
| |
| int _canonical_spill_slot; // the stack slot where all split parts of this interval are spilled to (always stored in split parents) |
| bool _insert_move_when_activated; // true if move is inserted between _current_split_child and this interval when interval gets active the first time |
| IntervalSpillState _spill_state; // for spill move optimization |
| int _spill_definition_pos; // position where the interval is defined (if defined only once) |
| Interval* _register_hint; // this interval should be in the same register as the hint interval |
| |
| int calc_to(); |
| Interval* new_split_child(); |
| public: |
| Interval(int reg_num); |
| |
| static void initialize(Arena* arena); |
| static Interval* end() { return _end; } |
| |
| // accessors |
| int reg_num() const { return _reg_num; } |
| void set_reg_num(int r) { assert(_reg_num == -1, "cannot change reg_num"); _reg_num = r; } |
| BasicType type() const { assert(_reg_num == -1 || _reg_num >= LIR_OprDesc::vreg_base, "cannot access type for fixed interval"); return _type; } |
| void set_type(BasicType type) { assert(_reg_num < LIR_OprDesc::vreg_base || _type == T_ILLEGAL || _type == type, "overwriting existing type"); _type = type; } |
| |
| Range* first() const { return _first; } |
| int from() const { return _first->from(); } |
| int to() { if (_cached_to == -1) _cached_to = calc_to(); assert(_cached_to == calc_to(), "invalid cached value"); return _cached_to; } |
| int num_use_positions() const { return _use_pos_and_kinds.length() / 2; } |
| |
| Interval* next() const { return _next; } |
| Interval** next_addr() { return &_next; } |
| void set_next(Interval* next) { _next = next; } |
| |
| int assigned_reg() const { return _assigned_reg; } |
| int assigned_regHi() const { return _assigned_regHi; } |
| void assign_reg(int reg) { _assigned_reg = reg; _assigned_regHi = LinearScan::any_reg; } |
| void assign_reg(int reg,int regHi) { _assigned_reg = reg; _assigned_regHi = regHi; } |
| |
| Interval* register_hint(bool search_split_child = true) const; // calculation needed |
| void set_register_hint(Interval* i) { _register_hint = i; } |
| |
| int state() const { return _state; } |
| void set_state(IntervalState s) { _state = s; } |
| |
| // access to split parent and split children |
| bool is_split_parent() const { return _split_parent == this; } |
| bool is_split_child() const { return _split_parent != this; } |
| Interval* split_parent() const { assert(_split_parent->is_split_parent(), "must be"); return _split_parent; } |
| Interval* split_child_at_op_id(int op_id, LIR_OpVisitState::OprMode mode); |
| Interval* split_child_before_op_id(int op_id); |
| bool split_child_covers(int op_id, LIR_OpVisitState::OprMode mode); |
| DEBUG_ONLY(void check_split_children();) |
| |
| // information stored in split parent, but available for all children |
| int canonical_spill_slot() const { return split_parent()->_canonical_spill_slot; } |
| void set_canonical_spill_slot(int slot) { assert(split_parent()->_canonical_spill_slot == -1, "overwriting existing value"); split_parent()->_canonical_spill_slot = slot; } |
| Interval* current_split_child() const { return split_parent()->_current_split_child; } |
| void make_current_split_child() { split_parent()->_current_split_child = this; } |
| |
| bool insert_move_when_activated() const { return _insert_move_when_activated; } |
| void set_insert_move_when_activated(bool b) { _insert_move_when_activated = b; } |
| |
| // for spill optimization |
| IntervalSpillState spill_state() const { return split_parent()->_spill_state; } |
| int spill_definition_pos() const { return split_parent()->_spill_definition_pos; } |
| void set_spill_state(IntervalSpillState state) { assert(state >= spill_state(), "state cannot decrease"); split_parent()->_spill_state = state; } |
| void set_spill_definition_pos(int pos) { assert(spill_definition_pos() == -1, "cannot set the position twice"); split_parent()->_spill_definition_pos = pos; } |
| // returns true if this interval has a shadow copy on the stack that is always correct |
| bool always_in_memory() const { return split_parent()->_spill_state == storeAtDefinition || split_parent()->_spill_state == startInMemory; } |
| |
| // caching of values that take time to compute and are used multiple times |
| LIR_Opr cached_opr() const { return _cached_opr; } |
| VMReg cached_vm_reg() const { return _cached_vm_reg; } |
| void set_cached_opr(LIR_Opr opr) { _cached_opr = opr; } |
| void set_cached_vm_reg(VMReg reg) { _cached_vm_reg = reg; } |
| |
| // access to use positions |
| int first_usage(IntervalUseKind min_use_kind) const; // id of the first operation requiring this interval in a register |
| int next_usage(IntervalUseKind min_use_kind, int from) const; // id of next usage seen from the given position |
| int next_usage_exact(IntervalUseKind exact_use_kind, int from) const; |
| int previous_usage(IntervalUseKind min_use_kind, int from) const; |
| |
| // manipulating intervals |
| void add_use_pos(int pos, IntervalUseKind use_kind); |
| void add_range(int from, int to); |
| Interval* split(int split_pos); |
| Interval* split_from_start(int split_pos); |
| void remove_first_use_pos() { _use_pos_and_kinds.trunc_to(_use_pos_and_kinds.length() - 2); } |
| |
| // test intersection |
| bool covers(int op_id, LIR_OpVisitState::OprMode mode) const; |
| bool has_hole_between(int from, int to); |
| bool intersects(Interval* i) const { return _first->intersects(i->_first); } |
| int intersects_at(Interval* i) const { return _first->intersects_at(i->_first); } |
| |
| // range iteration |
| void rewind_range() { _current = _first; } |
| void next_range() { assert(this != _end, "not allowed on sentinel"); _current = _current->next(); } |
| int current_from() const { return _current->from(); } |
| int current_to() const { return _current->to(); } |
| bool current_at_end() const { return _current == Range::end(); } |
| bool current_intersects(Interval* it) { return _current->intersects(it->_current); }; |
| int current_intersects_at(Interval* it) { return _current->intersects_at(it->_current); }; |
| |
| // printing |
| void print(outputStream* out = tty) const PRODUCT_RETURN; |
| }; |
| |
| |
| class IntervalWalker : public CompilationResourceObj { |
| protected: |
| Compilation* _compilation; |
| LinearScan* _allocator; |
| |
| Interval* _unhandled_first[nofKinds]; // sorted list of intervals, not life before the current position |
| Interval* _active_first [nofKinds]; // sorted list of intervals, life at the current position |
| Interval* _inactive_first [nofKinds]; // sorted list of intervals, intervals in a life time hole at the current position |
| |
| Interval* _current; // the current interval coming from unhandled list |
| int _current_position; // the current position (intercept point through the intervals) |
| IntervalKind _current_kind; // and whether it is fixed_kind or any_kind. |
| |
| |
| Compilation* compilation() const { return _compilation; } |
| LinearScan* allocator() const { return _allocator; } |
| |
| // unified bailout support |
| void bailout(const char* msg) const { compilation()->bailout(msg); } |
| bool bailed_out() const { return compilation()->bailed_out(); } |
| |
| void check_bounds(IntervalKind kind) { assert(kind >= fixedKind && kind <= anyKind, "invalid interval_kind"); } |
| |
| Interval** unhandled_first_addr(IntervalKind kind) { check_bounds(kind); return &_unhandled_first[kind]; } |
| Interval** active_first_addr(IntervalKind kind) { check_bounds(kind); return &_active_first[kind]; } |
| Interval** inactive_first_addr(IntervalKind kind) { check_bounds(kind); return &_inactive_first[kind]; } |
| |
| void append_unsorted(Interval** first, Interval* interval); |
| void append_sorted(Interval** first, Interval* interval); |
| void append_to_unhandled(Interval** list, Interval* interval); |
| |
| bool remove_from_list(Interval** list, Interval* i); |
| void remove_from_list(Interval* i); |
| |
| void next_interval(); |
| Interval* current() const { return _current; } |
| IntervalKind current_kind() const { return _current_kind; } |
| |
| void walk_to(IntervalState state, int from); |
| |
| // activate_current() is called when an unhandled interval becomes active (in current(), current_kind()). |
| // Return false if current() should not be moved the the active interval list. |
| // It is safe to append current to any interval list but the unhandled list. |
| virtual bool activate_current() { return true; } |
| |
| // interval_moved() is called whenever an interval moves from one interval list to another. |
| // In the implementation of this method it is prohibited to move the interval to any list. |
| virtual void interval_moved(Interval* interval, IntervalKind kind, IntervalState from, IntervalState to); |
| |
| public: |
| IntervalWalker(LinearScan* allocator, Interval* unhandled_fixed_first, Interval* unhandled_any_first); |
| |
| Interval* unhandled_first(IntervalKind kind) { check_bounds(kind); return _unhandled_first[kind]; } |
| Interval* active_first(IntervalKind kind) { check_bounds(kind); return _active_first[kind]; } |
| Interval* inactive_first(IntervalKind kind) { check_bounds(kind); return _inactive_first[kind]; } |
| |
| // active contains the intervals that are live after the lir_op |
| void walk_to(int lir_op_id); |
| // active contains the intervals that are live before the lir_op |
| void walk_before(int lir_op_id) { walk_to(lir_op_id-1); } |
| // walk through all intervals |
| void walk() { walk_to(max_jint); } |
| |
| int current_position() { return _current_position; } |
| }; |
| |
| |
| // The actual linear scan register allocator |
| class LinearScanWalker : public IntervalWalker { |
| enum { |
| any_reg = LinearScan::any_reg |
| }; |
| |
| private: |
| int _first_reg; // the reg. number of the first phys. register |
| int _last_reg; // the reg. nmber of the last phys. register |
| int _num_phys_regs; // required by current interval |
| bool _adjacent_regs; // have lo/hi words of phys. regs be adjacent |
| |
| int _use_pos[LinearScan::nof_regs]; |
| int _block_pos[LinearScan::nof_regs]; |
| IntervalList* _spill_intervals[LinearScan::nof_regs]; |
| |
| MoveResolver _move_resolver; // for ordering spill moves |
| |
| // accessors mapped to same functions in class LinearScan |
| int block_count() const { return allocator()->block_count(); } |
| BlockBegin* block_at(int idx) const { return allocator()->block_at(idx); } |
| BlockBegin* block_of_op_with_id(int op_id) const { return allocator()->block_of_op_with_id(op_id); } |
| |
| void init_use_lists(bool only_process_use_pos); |
| void exclude_from_use(int reg); |
| void exclude_from_use(Interval* i); |
| void set_use_pos(int reg, Interval* i, int use_pos, bool only_process_use_pos); |
| void set_use_pos(Interval* i, int use_pos, bool only_process_use_pos); |
| void set_block_pos(int reg, Interval* i, int block_pos); |
| void set_block_pos(Interval* i, int block_pos); |
| |
| void free_exclude_active_fixed(); |
| void free_exclude_active_any(); |
| void free_collect_inactive_fixed(Interval* cur); |
| void free_collect_inactive_any(Interval* cur); |
| void free_collect_unhandled(IntervalKind kind, Interval* cur); |
| void spill_exclude_active_fixed(); |
| void spill_block_unhandled_fixed(Interval* cur); |
| void spill_block_inactive_fixed(Interval* cur); |
| void spill_collect_active_any(); |
| void spill_collect_inactive_any(Interval* cur); |
| |
| void insert_move(int op_id, Interval* src_it, Interval* dst_it); |
| int find_optimal_split_pos(BlockBegin* min_block, BlockBegin* max_block, int max_split_pos); |
| int find_optimal_split_pos(Interval* it, int min_split_pos, int max_split_pos, bool do_loop_optimization); |
| void split_before_usage(Interval* it, int min_split_pos, int max_split_pos); |
| void split_for_spilling(Interval* it); |
| void split_stack_interval(Interval* it); |
| void split_when_partial_register_available(Interval* it, int register_available_until); |
| void split_and_spill_interval(Interval* it); |
| |
| int find_free_reg(int reg_needed_until, int interval_to, int hint_reg, int ignore_reg, bool* need_split); |
| int find_free_double_reg(int reg_needed_until, int interval_to, int hint_reg, bool* need_split); |
| bool alloc_free_reg(Interval* cur); |
| |
| int find_locked_reg(int reg_needed_until, int interval_to, int hint_reg, int ignore_reg, bool* need_split); |
| int find_locked_double_reg(int reg_needed_until, int interval_to, int hint_reg, bool* need_split); |
| void split_and_spill_intersecting_intervals(int reg, int regHi); |
| void alloc_locked_reg(Interval* cur); |
| |
| bool no_allocation_possible(Interval* cur); |
| void update_phys_reg_range(bool requires_cpu_register); |
| void init_vars_for_alloc(Interval* cur); |
| bool pd_init_regs_for_alloc(Interval* cur); |
| |
| void combine_spilled_intervals(Interval* cur); |
| bool is_move(LIR_Op* op, Interval* from, Interval* to); |
| |
| bool activate_current(); |
| |
| public: |
| LinearScanWalker(LinearScan* allocator, Interval* unhandled_fixed_first, Interval* unhandled_any_first); |
| |
| // must be called when all intervals are allocated |
| void finish_allocation() { _move_resolver.resolve_and_append_moves(); } |
| }; |
| |
| |
| |
| /* |
| When a block has more than one predecessor, and all predecessors end with |
| the same sequence of move-instructions, than this moves can be placed once |
| at the beginning of the block instead of multiple times in the predecessors. |
| |
| Similarly, when a block has more than one successor, then equal sequences of |
| moves at the beginning of the successors can be placed once at the end of |
| the block. But because the moves must be inserted before all branch |
| instructions, this works only when there is exactly one conditional branch |
| at the end of the block (because the moves must be inserted before all |
| branches, but after all compares). |
| |
| This optimization affects all kind of moves (reg->reg, reg->stack and |
| stack->reg). Because this optimization works best when a block contains only |
| few moves, it has a huge impact on the number of blocks that are totally |
| empty. |
| */ |
| class EdgeMoveOptimizer : public StackObj { |
| private: |
| // the class maintains a list with all lir-instruction-list of the |
| // successors (predecessors) and the current index into the lir-lists |
| LIR_OpListStack _edge_instructions; |
| intStack _edge_instructions_idx; |
| |
| void init_instructions(); |
| void append_instructions(LIR_OpList* instructions, int instructions_idx); |
| LIR_Op* instruction_at(int edge); |
| void remove_cur_instruction(int edge, bool decrement_index); |
| |
| bool operations_different(LIR_Op* op1, LIR_Op* op2); |
| |
| void optimize_moves_at_block_end(BlockBegin* cur); |
| void optimize_moves_at_block_begin(BlockBegin* cur); |
| |
| EdgeMoveOptimizer(); |
| |
| public: |
| static void optimize(BlockList* code); |
| }; |
| |
| |
| |
| class ControlFlowOptimizer : public StackObj { |
| private: |
| BlockList _original_preds; |
| |
| enum { |
| ShortLoopSize = 5 |
| }; |
| void reorder_short_loop(BlockList* code, BlockBegin* header_block, int header_idx); |
| void reorder_short_loops(BlockList* code); |
| |
| bool can_delete_block(BlockBegin* cur); |
| void substitute_branch_target(BlockBegin* cur, BlockBegin* target_from, BlockBegin* target_to); |
| void delete_empty_blocks(BlockList* code); |
| |
| void delete_unnecessary_jumps(BlockList* code); |
| void delete_jumps_to_return(BlockList* code); |
| |
| DEBUG_ONLY(void verify(BlockList* code);) |
| |
| ControlFlowOptimizer(); |
| public: |
| static void optimize(BlockList* code); |
| }; |
| |
| |
| #ifndef PRODUCT |
| |
| // Helper class for collecting statistics of LinearScan |
| class LinearScanStatistic : public StackObj { |
| public: |
| enum Counter { |
| // general counters |
| counter_method, |
| counter_fpu_method, |
| counter_loop_method, |
| counter_exception_method, |
| counter_loop, |
| counter_block, |
| counter_loop_block, |
| counter_exception_block, |
| counter_interval, |
| counter_fixed_interval, |
| counter_range, |
| counter_fixed_range, |
| counter_use_pos, |
| counter_fixed_use_pos, |
| counter_spill_slots, |
| blank_line_1, |
| |
| // counter for classes of lir instructions |
| counter_instruction, |
| counter_label, |
| counter_entry, |
| counter_return, |
| counter_call, |
| counter_move, |
| counter_cmp, |
| counter_cond_branch, |
| counter_uncond_branch, |
| counter_stub_branch, |
| counter_alu, |
| counter_alloc, |
| counter_sync, |
| counter_throw, |
| counter_unwind, |
| counter_typecheck, |
| counter_fpu_stack, |
| counter_misc_inst, |
| counter_other_inst, |
| blank_line_2, |
| |
| // counter for different types of moves |
| counter_move_total, |
| counter_move_reg_reg, |
| counter_move_reg_stack, |
| counter_move_stack_reg, |
| counter_move_stack_stack, |
| counter_move_reg_mem, |
| counter_move_mem_reg, |
| counter_move_const_any, |
| |
| number_of_counters, |
| invalid_counter = -1 |
| }; |
| |
| private: |
| int _counters_sum[number_of_counters]; |
| int _counters_max[number_of_counters]; |
| |
| void inc_counter(Counter idx, int value = 1) { _counters_sum[idx] += value; } |
| |
| const char* counter_name(int counter_idx); |
| Counter base_counter(int counter_idx); |
| |
| void sum_up(LinearScanStatistic &method_statistic); |
| void collect(LinearScan* allocator); |
| |
| public: |
| LinearScanStatistic(); |
| void print(const char* title); |
| static void compute(LinearScan* allocator, LinearScanStatistic &global_statistic); |
| }; |
| |
| |
| // Helper class for collecting compilation time of LinearScan |
| class LinearScanTimers : public StackObj { |
| public: |
| enum Timer { |
| timer_do_nothing, |
| timer_number_instructions, |
| timer_compute_local_live_sets, |
| timer_compute_global_live_sets, |
| timer_build_intervals, |
| timer_sort_intervals_before, |
| timer_allocate_registers, |
| timer_resolve_data_flow, |
| timer_sort_intervals_after, |
| timer_eliminate_spill_moves, |
| timer_assign_reg_num, |
| timer_allocate_fpu_stack, |
| timer_optimize_lir, |
| |
| number_of_timers |
| }; |
| |
| private: |
| elapsedTimer _timers[number_of_timers]; |
| const char* timer_name(int idx); |
| |
| public: |
| LinearScanTimers(); |
| |
| void begin_method(); // called for each method when register allocation starts |
| void end_method(LinearScan* allocator); // called for each method when register allocation completed |
| void print(double total_time); // called before termination of VM to print global summary |
| |
| elapsedTimer* timer(int idx) { return &(_timers[idx]); } |
| }; |
| |
| |
| #endif // ifndef PRODUCT |
| |
| // Pick up platform-dependent implementation details |
| #include CPU_HEADER(c1_LinearScan) |
| |
| #endif // SHARE_VM_C1_C1_LINEARSCAN_HPP |