|
/*
* Copyright (c) 2000, 2022, 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_OOPS_METHODDATA_HPP
#define SHARE_OOPS_METHODDATA_HPP
#include "interpreter/bytecodes.hpp"
#include "oops/metadata.hpp"
#include "oops/method.hpp"
#include "oops/oop.hpp"
#include "runtime/atomic.hpp"
#include "runtime/deoptimization.hpp"
#include "runtime/mutex.hpp"
#include "utilities/align.hpp"
#include "utilities/copy.hpp"
class BytecodeStream;
// The MethodData object collects counts and other profile information
// during zeroth-tier (interpreter) and third-tier (C1 with full profiling)
// execution.
//
// The profile is used later by compilation heuristics. Some heuristics
// enable use of aggressive (or "heroic") optimizations. An aggressive
// optimization often has a down-side, a corner case that it handles
// poorly, but which is thought to be rare. The profile provides
// evidence of this rarity for a given method or even BCI. It allows
// the compiler to back out of the optimization at places where it
// has historically been a poor choice. Other heuristics try to use
// specific information gathered about types observed at a given site.
//
// All data in the profile is approximate. It is expected to be accurate
// on the whole, but the system expects occasional inaccuraces, due to
// counter overflow, multiprocessor races during data collection, space
// limitations, missing MDO blocks, etc. Bad or missing data will degrade
// optimization quality but will not affect correctness. Also, each MDO
// is marked with its birth-date ("creation_mileage") which can be used
// to assess the quality ("maturity") of its data.
//
// Short (<32-bit) counters are designed to overflow to a known "saturated"
// state. Also, certain recorded per-BCI events are given one-bit counters
// which overflow to a saturated state which applied to all counters at
// that BCI. In other words, there is a small lattice which approximates
// the ideal of an infinite-precision counter for each event at each BCI,
// and the lattice quickly "bottoms out" in a state where all counters
// are taken to be indefinitely large.
//
// The reader will find many data races in profile gathering code, starting
// with invocation counter incrementation. None of these races harm correct
// execution of the compiled code.
// forward decl
class ProfileData;
// DataLayout
//
// Overlay for generic profiling data.
class DataLayout {
friend class VMStructs;
friend class JVMCIVMStructs;
private:
// Every data layout begins with a header. This header
// contains a tag, which is used to indicate the size/layout
// of the data, 8 bits of flags, which can be used in any way,
// 32 bits of trap history (none/one reason/many reasons),
// and a bci, which is used to tie this piece of data to a
// specific bci in the bytecodes.
union {
u8 _bits;
struct {
u1 _tag;
u1 _flags;
u2 _bci;
u4 _traps;
} _struct;
} _header;
// The data layout has an arbitrary number of cells, each sized
// to accommodate a pointer or an integer.
intptr_t _cells[1];
// Some types of data layouts need a length field.
static bool needs_array_len(u1 tag);
public:
enum {
counter_increment = 1
};
enum {
cell_size = sizeof(intptr_t)
};
// Tag values
enum {
no_tag,
bit_data_tag,
counter_data_tag,
jump_data_tag,
receiver_type_data_tag,
virtual_call_data_tag,
ret_data_tag,
branch_data_tag,
multi_branch_data_tag,
arg_info_data_tag,
call_type_data_tag,
virtual_call_type_data_tag,
parameters_type_data_tag,
speculative_trap_data_tag
};
enum {
// The trap state breaks down as [recompile:1 | reason:31].
// This further breakdown is defined in deoptimization.cpp.
// See Deoptimization::trap_state_reason for an assert that
// trap_bits is big enough to hold reasons < Reason_RECORDED_LIMIT.
//
// The trap_state is collected only if ProfileTraps is true.
trap_bits = 1+31, // 31: enough to distinguish [0..Reason_RECORDED_LIMIT].
trap_mask = -1,
first_flag = 0
};
// Size computation
static int header_size_in_bytes() {
return header_size_in_cells() * cell_size;
}
static int header_size_in_cells() {
return LP64_ONLY(1) NOT_LP64(2);
}
static int compute_size_in_bytes(int cell_count) {
return header_size_in_bytes() + cell_count * cell_size;
}
// Initialization
void initialize(u1 tag, u2 bci, int cell_count);
// Accessors
u1 tag() {
return _header._struct._tag;
}
// Return 32 bits of trap state.
// The state tells if traps with zero, one, or many reasons have occurred.
// It also tells whether zero or many recompilations have occurred.
// The associated trap histogram in the MDO itself tells whether
// traps are common or not. If a BCI shows that a trap X has
// occurred, and the MDO shows N occurrences of X, we make the
// simplifying assumption that all N occurrences can be blamed
// on that BCI.
uint trap_state() const {
return _header._struct._traps;
}
void set_trap_state(uint new_state) {
assert(ProfileTraps, "used only under +ProfileTraps");
uint old_flags = _header._struct._traps;
_header._struct._traps = new_state | old_flags;
}
u1 flags() const {
return _header._struct._flags;
}
u2 bci() const {
return _header._struct._bci;
}
void set_header(u8 value) {
_header._bits = value;
}
u8 header() {
return _header._bits;
}
void set_cell_at(int index, intptr_t value) {
_cells[index] = value;
}
void release_set_cell_at(int index, intptr_t value);
intptr_t cell_at(int index) const {
return _cells[index];
}
void set_flag_at(u1 flag_number) {
_header._struct._flags |= (0x1 << flag_number);
}
bool flag_at(u1 flag_number) const {
return (_header._struct._flags & (0x1 << flag_number)) != 0;
}
// Low-level support for code generation.
static ByteSize header_offset() {
return byte_offset_of(DataLayout, _header);
}
static ByteSize tag_offset() {
return byte_offset_of(DataLayout, _header._struct._tag);
}
static ByteSize flags_offset() {
return byte_offset_of(DataLayout, _header._struct._flags);
}
static ByteSize bci_offset() {
return byte_offset_of(DataLayout, _header._struct._bci);
}
static ByteSize cell_offset(int index) {
return byte_offset_of(DataLayout, _cells) + in_ByteSize(index * cell_size);
}
// Return a value which, when or-ed as a byte into _flags, sets the flag.
static u1 flag_number_to_constant(u1 flag_number) {
DataLayout temp; temp.set_header(0);
temp.set_flag_at(flag_number);
return temp._header._struct._flags;
}
// Return a value which, when or-ed as a word into _header, sets the flag.
static u8 flag_mask_to_header_mask(uint byte_constant) {
DataLayout temp; temp.set_header(0);
temp._header._struct._flags = byte_constant;
return temp._header._bits;
}
ProfileData* data_in();
int size_in_bytes() {
int cells = cell_count();
assert(cells >= 0, "invalid number of cells");
return DataLayout::compute_size_in_bytes(cells);
}
int cell_count();
// GC support
void clean_weak_klass_links(bool always_clean);
};
// ProfileData class hierarchy
class ProfileData;
class BitData;
class CounterData;
class ReceiverTypeData;
class VirtualCallData;
class VirtualCallTypeData;
class RetData;
class CallTypeData;
class JumpData;
class BranchData;
class ArrayData;
class MultiBranchData;
class ArgInfoData;
class ParametersTypeData;
class SpeculativeTrapData;
// ProfileData
//
// A ProfileData object is created to refer to a section of profiling
// data in a structured way.
class ProfileData : public ResourceObj {
friend class TypeEntries;
friend class ReturnTypeEntry;
friend class TypeStackSlotEntries;
private:
enum {
tab_width_one = 16,
tab_width_two = 36
};
// This is a pointer to a section of profiling data.
DataLayout* _data;
char* print_data_on_helper(const MethodData* md) const;
protected:
DataLayout* data() { return _data; }
const DataLayout* data() const { return _data; }
enum {
cell_size = DataLayout::cell_size
};
public:
// How many cells are in this?
virtual int cell_count() const {
ShouldNotReachHere();
return -1;
}
// Return the size of this data.
int size_in_bytes() {
return DataLayout::compute_size_in_bytes(cell_count());
}
protected:
// Low-level accessors for underlying data
void set_intptr_at(int index, intptr_t value) {
assert(0 <= index && index < cell_count(), "oob");
data()->set_cell_at(index, value);
}
void release_set_intptr_at(int index, intptr_t value);
intptr_t intptr_at(int index) const {
assert(0 <= index && index < cell_count(), "oob");
return data()->cell_at(index);
}
void set_uint_at(int index, uint value) {
set_intptr_at(index, (intptr_t) value);
}
void release_set_uint_at(int index, uint value);
uint uint_at(int index) const {
return (uint)intptr_at(index);
}
void set_int_at(int index, int value) {
set_intptr_at(index, (intptr_t) value);
}
void release_set_int_at(int index, int value);
int int_at(int index) const {
return (int)intptr_at(index);
}
int int_at_unchecked(int index) const {
return (int)data()->cell_at(index);
}
void set_oop_at(int index, oop value) {
set_intptr_at(index, cast_from_oop<intptr_t>(value));
}
oop oop_at(int index) const {
return cast_to_oop(intptr_at(index));
}
void set_flag_at(int flag_number) {
data()->set_flag_at(flag_number);
}
bool flag_at(int flag_number) const {
return data()->flag_at(flag_number);
}
// two convenient imports for use by subclasses:
static ByteSize cell_offset(int index) {
return DataLayout::cell_offset(index);
}
static int flag_number_to_constant(int flag_number) {
return DataLayout::flag_number_to_constant(flag_number);
}
ProfileData(DataLayout* data) {
_data = data;
}
public:
// Constructor for invalid ProfileData.
ProfileData();
u2 bci() const {
return data()->bci();
}
address dp() {
return (address)_data;
}
int trap_state() const {
return data()->trap_state();
}
void set_trap_state(int new_state) {
data()->set_trap_state(new_state);
}
// Type checking
virtual bool is_BitData() const { return false; }
virtual bool is_CounterData() const { return false; }
virtual bool is_JumpData() const { return false; }
virtual bool is_ReceiverTypeData()const { return false; }
virtual bool is_VirtualCallData() const { return false; }
virtual bool is_RetData() const { return false; }
virtual bool is_BranchData() const { return false; }
virtual bool is_ArrayData() const { return false; }
virtual bool is_MultiBranchData() const { return false; }
virtual bool is_ArgInfoData() const { return false; }
virtual bool is_CallTypeData() const { return false; }
virtual bool is_VirtualCallTypeData()const { return false; }
virtual bool is_ParametersTypeData() const { return false; }
virtual bool is_SpeculativeTrapData()const { return false; }
BitData* as_BitData() const {
assert(is_BitData(), "wrong type");
return is_BitData() ? (BitData*) this : NULL;
}
CounterData* as_CounterData() const {
assert(is_CounterData(), "wrong type");
return is_CounterData() ? (CounterData*) this : NULL;
}
JumpData* as_JumpData() const {
assert(is_JumpData(), "wrong type");
return is_JumpData() ? (JumpData*) this : NULL;
}
ReceiverTypeData* as_ReceiverTypeData() const {
assert(is_ReceiverTypeData(), "wrong type");
return is_ReceiverTypeData() ? (ReceiverTypeData*)this : NULL;
}
VirtualCallData* as_VirtualCallData() const {
assert(is_VirtualCallData(), "wrong type");
return is_VirtualCallData() ? (VirtualCallData*)this : NULL;
}
RetData* as_RetData() const {
assert(is_RetData(), "wrong type");
return is_RetData() ? (RetData*) this : NULL;
}
BranchData* as_BranchData() const {
assert(is_BranchData(), "wrong type");
return is_BranchData() ? (BranchData*) this : NULL;
}
ArrayData* as_ArrayData() const {
assert(is_ArrayData(), "wrong type");
return is_ArrayData() ? (ArrayData*) this : NULL;
}
MultiBranchData* as_MultiBranchData() const {
assert(is_MultiBranchData(), "wrong type");
return is_MultiBranchData() ? (MultiBranchData*)this : NULL;
}
ArgInfoData* as_ArgInfoData() const {
assert(is_ArgInfoData(), "wrong type");
return is_ArgInfoData() ? (ArgInfoData*)this : NULL;
}
CallTypeData* as_CallTypeData() const {
assert(is_CallTypeData(), "wrong type");
return is_CallTypeData() ? (CallTypeData*)this : NULL;
}
VirtualCallTypeData* as_VirtualCallTypeData() const {
assert(is_VirtualCallTypeData(), "wrong type");
return is_VirtualCallTypeData() ? (VirtualCallTypeData*)this : NULL;
}
ParametersTypeData* as_ParametersTypeData() const {
assert(is_ParametersTypeData(), "wrong type");
return is_ParametersTypeData() ? (ParametersTypeData*)this : NULL;
}
SpeculativeTrapData* as_SpeculativeTrapData() const {
assert(is_SpeculativeTrapData(), "wrong type");
return is_SpeculativeTrapData() ? (SpeculativeTrapData*)this : NULL;
}
// Subclass specific initialization
virtual void post_initialize(BytecodeStream* stream, MethodData* mdo) {}
// GC support
virtual void clean_weak_klass_links(bool always_clean) {}
// CI translation: ProfileData can represent both MethodDataOop data
// as well as CIMethodData data. This function is provided for translating
// an oop in a ProfileData to the ci equivalent. Generally speaking,
// most ProfileData don't require any translation, so we provide the null
// translation here, and the required translators are in the ci subclasses.
virtual void translate_from(const ProfileData* data) {}
virtual void print_data_on(outputStream* st, const char* extra = NULL) const {
ShouldNotReachHere();
}
void print_data_on(outputStream* st, const MethodData* md) const;
void print_shared(outputStream* st, const char* name, const char* extra) const;
void tab(outputStream* st, bool first = false) const;
};
// BitData
//
// A BitData holds a flag or two in its header.
class BitData : public ProfileData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
enum {
// null_seen:
// saw a null operand (cast/aastore/instanceof)
null_seen_flag = DataLayout::first_flag + 0
#if INCLUDE_JVMCI
// bytecode threw any exception
, exception_seen_flag = null_seen_flag + 1
#endif
};
enum { bit_cell_count = 0 }; // no additional data fields needed.
public:
BitData(DataLayout* layout) : ProfileData(layout) {
}
virtual bool is_BitData() const { return true; }
static int static_cell_count() {
return bit_cell_count;
}
virtual int cell_count() const {
return static_cell_count();
}
// Accessor
// The null_seen flag bit is specially known to the interpreter.
// Consulting it allows the compiler to avoid setting up null_check traps.
bool null_seen() { return flag_at(null_seen_flag); }
void set_null_seen() { set_flag_at(null_seen_flag); }
#if INCLUDE_JVMCI
// true if an exception was thrown at the specific BCI
bool exception_seen() { return flag_at(exception_seen_flag); }
void set_exception_seen() { set_flag_at(exception_seen_flag); }
#endif
// Code generation support
static int null_seen_byte_constant() {
return flag_number_to_constant(null_seen_flag);
}
static ByteSize bit_data_size() {
return cell_offset(bit_cell_count);
}
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// CounterData
//
// A CounterData corresponds to a simple counter.
class CounterData : public BitData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
enum {
count_off,
counter_cell_count
};
public:
CounterData(DataLayout* layout) : BitData(layout) {}
virtual bool is_CounterData() const { return true; }
static int static_cell_count() {
return counter_cell_count;
}
virtual int cell_count() const {
return static_cell_count();
}
// Direct accessor
int count() const {
intptr_t raw_data = intptr_at(count_off);
if (raw_data > max_jint) {
raw_data = max_jint;
} else if (raw_data < min_jint) {
raw_data = min_jint;
}
return int(raw_data);
}
// Code generation support
static ByteSize count_offset() {
return cell_offset(count_off);
}
static ByteSize counter_data_size() {
return cell_offset(counter_cell_count);
}
void set_count(int count) {
set_int_at(count_off, count);
}
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// JumpData
//
// A JumpData is used to access profiling information for a direct
// branch. It is a counter, used for counting the number of branches,
// plus a data displacement, used for realigning the data pointer to
// the corresponding target bci.
class JumpData : public ProfileData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
enum {
taken_off_set,
displacement_off_set,
jump_cell_count
};
void set_displacement(int displacement) {
set_int_at(displacement_off_set, displacement);
}
public:
JumpData(DataLayout* layout) : ProfileData(layout) {
assert(layout->tag() == DataLayout::jump_data_tag ||
layout->tag() == DataLayout::branch_data_tag, "wrong type");
}
virtual bool is_JumpData() const { return true; }
static int static_cell_count() {
return jump_cell_count;
}
virtual int cell_count() const {
return static_cell_count();
}
// Direct accessor
uint taken() const {
return uint_at(taken_off_set);
}
void set_taken(uint cnt) {
set_uint_at(taken_off_set, cnt);
}
// Saturating counter
uint inc_taken() {
uint cnt = taken() + 1;
// Did we wrap? Will compiler screw us??
if (cnt == 0) cnt--;
set_uint_at(taken_off_set, cnt);
return cnt;
}
int displacement() const {
return int_at(displacement_off_set);
}
// Code generation support
static ByteSize taken_offset() {
return cell_offset(taken_off_set);
}
static ByteSize displacement_offset() {
return cell_offset(displacement_off_set);
}
// Specific initialization.
void post_initialize(BytecodeStream* stream, MethodData* mdo);
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// Entries in a ProfileData object to record types: it can either be
// none (no profile), unknown (conflicting profile data) or a klass if
// a single one is seen. Whether a null reference was seen is also
// recorded. No counter is associated with the type and a single type
// is tracked (unlike VirtualCallData).
class TypeEntries {
public:
// A single cell is used to record information for a type:
// - the cell is initialized to 0
// - when a type is discovered it is stored in the cell
// - bit zero of the cell is used to record whether a null reference
// was encountered or not
// - bit 1 is set to record a conflict in the type information
enum {
null_seen = 1,
type_mask = ~null_seen,
type_unknown = 2,
status_bits = null_seen | type_unknown,
type_klass_mask = ~status_bits
};
// what to initialize a cell to
static intptr_t type_none() {
return 0;
}
// null seen = bit 0 set?
static bool was_null_seen(intptr_t v) {
return (v & null_seen) != 0;
}
// conflicting type information = bit 1 set?
static bool is_type_unknown(intptr_t v) {
return (v & type_unknown) != 0;
}
// not type information yet = all bits cleared, ignoring bit 0?
static bool is_type_none(intptr_t v) {
return (v & type_mask) == 0;
}
// recorded type: cell without bit 0 and 1
static intptr_t klass_part(intptr_t v) {
intptr_t r = v & type_klass_mask;
return r;
}
// type recorded
static Klass* valid_klass(intptr_t k) {
if (!is_type_none(k) &&
!is_type_unknown(k)) {
Klass* res = (Klass*)klass_part(k);
assert(res != NULL, "invalid");
return res;
} else {
return NULL;
}
}
static intptr_t with_status(intptr_t k, intptr_t in) {
return k | (in & status_bits);
}
static intptr_t with_status(Klass* k, intptr_t in) {
return with_status((intptr_t)k, in);
}
static void print_klass(outputStream* st, intptr_t k);
protected:
// ProfileData object these entries are part of
ProfileData* _pd;
// offset within the ProfileData object where the entries start
const int _base_off;
TypeEntries(int base_off)
: _pd(NULL), _base_off(base_off) {}
void set_intptr_at(int index, intptr_t value) {
_pd->set_intptr_at(index, value);
}
intptr_t intptr_at(int index) const {
return _pd->intptr_at(index);
}
public:
void set_profile_data(ProfileData* pd) {
_pd = pd;
}
};
// Type entries used for arguments passed at a call and parameters on
// method entry. 2 cells per entry: one for the type encoded as in
// TypeEntries and one initialized with the stack slot where the
// profiled object is to be found so that the interpreter can locate
// it quickly.
class TypeStackSlotEntries : public TypeEntries {
private:
enum {
stack_slot_entry,
type_entry,
per_arg_cell_count
};
// offset of cell for stack slot for entry i within ProfileData object
int stack_slot_offset(int i) const {
return _base_off + stack_slot_local_offset(i);
}
const int _number_of_entries;
// offset of cell for type for entry i within ProfileData object
int type_offset_in_cells(int i) const {
return _base_off + type_local_offset(i);
}
public:
TypeStackSlotEntries(int base_off, int nb_entries)
: TypeEntries(base_off), _number_of_entries(nb_entries) {}
static int compute_cell_count(Symbol* signature, bool include_receiver, int max);
void post_initialize(Symbol* signature, bool has_receiver, bool include_receiver);
int number_of_entries() const { return _number_of_entries; }
// offset of cell for stack slot for entry i within this block of cells for a TypeStackSlotEntries
static int stack_slot_local_offset(int i) {
return i * per_arg_cell_count + stack_slot_entry;
}
// offset of cell for type for entry i within this block of cells for a TypeStackSlotEntries
static int type_local_offset(int i) {
return i * per_arg_cell_count + type_entry;
}
// stack slot for entry i
uint stack_slot(int i) const {
assert(i >= 0 && i < _number_of_entries, "oob");
return _pd->uint_at(stack_slot_offset(i));
}
// set stack slot for entry i
void set_stack_slot(int i, uint num) {
assert(i >= 0 && i < _number_of_entries, "oob");
_pd->set_uint_at(stack_slot_offset(i), num);
}
// type for entry i
intptr_t type(int i) const {
assert(i >= 0 && i < _number_of_entries, "oob");
return _pd->intptr_at(type_offset_in_cells(i));
}
// set type for entry i
void set_type(int i, intptr_t k) {
assert(i >= 0 && i < _number_of_entries, "oob");
_pd->set_intptr_at(type_offset_in_cells(i), k);
}
static ByteSize per_arg_size() {
return in_ByteSize(per_arg_cell_count * DataLayout::cell_size);
}
static int per_arg_count() {
return per_arg_cell_count;
}
ByteSize type_offset(int i) const {
return DataLayout::cell_offset(type_offset_in_cells(i));
}
// GC support
void clean_weak_klass_links(bool always_clean);
void print_data_on(outputStream* st) const;
};
// Type entry used for return from a call. A single cell to record the
// type.
class ReturnTypeEntry : public TypeEntries {
private:
enum {
cell_count = 1
};
public:
ReturnTypeEntry(int base_off)
: TypeEntries(base_off) {}
void post_initialize() {
set_type(type_none());
}
intptr_t type() const {
return _pd->intptr_at(_base_off);
}
void set_type(intptr_t k) {
_pd->set_intptr_at(_base_off, k);
}
static int static_cell_count() {
return cell_count;
}
static ByteSize size() {
return in_ByteSize(cell_count * DataLayout::cell_size);
}
ByteSize type_offset() {
return DataLayout::cell_offset(_base_off);
}
// GC support
void clean_weak_klass_links(bool always_clean);
void print_data_on(outputStream* st) const;
};
// Entries to collect type information at a call: contains arguments
// (TypeStackSlotEntries), a return type (ReturnTypeEntry) and a
// number of cells. Because the number of cells for the return type is
// smaller than the number of cells for the type of an arguments, the
// number of cells is used to tell how many arguments are profiled and
// whether a return value is profiled. See has_arguments() and
// has_return().
class TypeEntriesAtCall {
private:
static int stack_slot_local_offset(int i) {
return header_cell_count() + TypeStackSlotEntries::stack_slot_local_offset(i);
}
static int argument_type_local_offset(int i) {
return header_cell_count() + TypeStackSlotEntries::type_local_offset(i);
}
public:
static int header_cell_count() {
return 1;
}
static int cell_count_local_offset() {
return 0;
}
static int compute_cell_count(BytecodeStream* stream);
static void initialize(DataLayout* dl, int base, int cell_count) {
int off = base + cell_count_local_offset();
dl->set_cell_at(off, cell_count - base - header_cell_count());
}
static bool arguments_profiling_enabled();
static bool return_profiling_enabled();
// Code generation support
static ByteSize cell_count_offset() {
return in_ByteSize(cell_count_local_offset() * DataLayout::cell_size);
}
static ByteSize args_data_offset() {
return in_ByteSize(header_cell_count() * DataLayout::cell_size);
}
static ByteSize stack_slot_offset(int i) {
return in_ByteSize(stack_slot_local_offset(i) * DataLayout::cell_size);
}
static ByteSize argument_type_offset(int i) {
return in_ByteSize(argument_type_local_offset(i) * DataLayout::cell_size);
}
static ByteSize return_only_size() {
return ReturnTypeEntry::size() + in_ByteSize(header_cell_count() * DataLayout::cell_size);
}
};
// CallTypeData
//
// A CallTypeData is used to access profiling information about a non
// virtual call for which we collect type information about arguments
// and return value.
class CallTypeData : public CounterData {
private:
// entries for arguments if any
TypeStackSlotEntries _args;
// entry for return type if any
ReturnTypeEntry _ret;
int cell_count_global_offset() const {
return CounterData::static_cell_count() + TypeEntriesAtCall::cell_count_local_offset();
}
// number of cells not counting the header
int cell_count_no_header() const {
return uint_at(cell_count_global_offset());
}
void check_number_of_arguments(int total) {
assert(number_of_arguments() == total, "should be set in DataLayout::initialize");
}
public:
CallTypeData(DataLayout* layout) :
CounterData(layout),
_args(CounterData::static_cell_count()+TypeEntriesAtCall::header_cell_count(), number_of_arguments()),
_ret(cell_count() - ReturnTypeEntry::static_cell_count())
{
assert(layout->tag() == DataLayout::call_type_data_tag, "wrong type");
// Some compilers (VC++) don't want this passed in member initialization list
_args.set_profile_data(this);
_ret.set_profile_data(this);
}
const TypeStackSlotEntries* args() const {
assert(has_arguments(), "no profiling of arguments");
return &_args;
}
const ReturnTypeEntry* ret() const {
assert(has_return(), "no profiling of return value");
return &_ret;
}
virtual bool is_CallTypeData() const { return true; }
static int static_cell_count() {
return -1;
}
static int compute_cell_count(BytecodeStream* stream) {
return CounterData::static_cell_count() + TypeEntriesAtCall::compute_cell_count(stream);
}
static void initialize(DataLayout* dl, int cell_count) {
TypeEntriesAtCall::initialize(dl, CounterData::static_cell_count(), cell_count);
}
virtual void post_initialize(BytecodeStream* stream, MethodData* mdo);
virtual int cell_count() const {
return CounterData::static_cell_count() +
TypeEntriesAtCall::header_cell_count() +
int_at_unchecked(cell_count_global_offset());
}
int number_of_arguments() const {
return cell_count_no_header() / TypeStackSlotEntries::per_arg_count();
}
void set_argument_type(int i, Klass* k) {
assert(has_arguments(), "no arguments!");
intptr_t current = _args.type(i);
_args.set_type(i, TypeEntries::with_status(k, current));
}
void set_return_type(Klass* k) {
assert(has_return(), "no return!");
intptr_t current = _ret.type();
_ret.set_type(TypeEntries::with_status(k, current));
}
// An entry for a return value takes less space than an entry for an
// argument so if the number of cells exceeds the number of cells
// needed for an argument, this object contains type information for
// at least one argument.
bool has_arguments() const {
bool res = cell_count_no_header() >= TypeStackSlotEntries::per_arg_count();
assert (!res || TypeEntriesAtCall::arguments_profiling_enabled(), "no profiling of arguments");
return res;
}
// An entry for a return value takes less space than an entry for an
// argument, so if the remainder of the number of cells divided by
// the number of cells for an argument is not null, a return value
// is profiled in this object.
bool has_return() const {
bool res = (cell_count_no_header() % TypeStackSlotEntries::per_arg_count()) != 0;
assert (!res || TypeEntriesAtCall::return_profiling_enabled(), "no profiling of return values");
return res;
}
// Code generation support
static ByteSize args_data_offset() {
return cell_offset(CounterData::static_cell_count()) + TypeEntriesAtCall::args_data_offset();
}
ByteSize argument_type_offset(int i) {
return _args.type_offset(i);
}
ByteSize return_type_offset() {
return _ret.type_offset();
}
// GC support
virtual void clean_weak_klass_links(bool always_clean) {
if (has_arguments()) {
_args.clean_weak_klass_links(always_clean);
}
if (has_return()) {
_ret.clean_weak_klass_links(always_clean);
}
}
virtual void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// ReceiverTypeData
//
// A ReceiverTypeData is used to access profiling information about a
// dynamic type check. It consists of a counter which counts the total times
// that the check is reached, and a series of (Klass*, count) pairs
// which are used to store a type profile for the receiver of the check.
class ReceiverTypeData : public CounterData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
enum {
#if INCLUDE_JVMCI
// Description of the different counters
// ReceiverTypeData for instanceof/checkcast/aastore:
// count is decremented for failed type checks
// JVMCI only: nonprofiled_count is incremented on type overflow
// VirtualCallData for invokevirtual/invokeinterface:
// count is incremented on type overflow
// JVMCI only: nonprofiled_count is incremented on method overflow
// JVMCI is interested in knowing the percentage of type checks involving a type not explicitly in the profile
nonprofiled_count_off_set = counter_cell_count,
receiver0_offset,
#else
receiver0_offset = counter_cell_count,
#endif
count0_offset,
receiver_type_row_cell_count = (count0_offset + 1) - receiver0_offset
};
public:
ReceiverTypeData(DataLayout* layout) : CounterData(layout) {
assert(layout->tag() == DataLayout::receiver_type_data_tag ||
layout->tag() == DataLayout::virtual_call_data_tag ||
layout->tag() == DataLayout::virtual_call_type_data_tag, "wrong type");
}
virtual bool is_ReceiverTypeData() const { return true; }
static int static_cell_count() {
return counter_cell_count + (uint) TypeProfileWidth * receiver_type_row_cell_count JVMCI_ONLY(+ 1);
}
virtual int cell_count() const {
return static_cell_count();
}
// Direct accessors
static uint row_limit() {
return TypeProfileWidth;
}
static int receiver_cell_index(uint row) {
return receiver0_offset + row * receiver_type_row_cell_count;
}
static int receiver_count_cell_index(uint row) {
return count0_offset + row * receiver_type_row_cell_count;
}
Klass* receiver(uint row) const {
assert(row < row_limit(), "oob");
Klass* recv = (Klass*)intptr_at(receiver_cell_index(row));
assert(recv == NULL || recv->is_klass(), "wrong type");
return recv;
}
void set_receiver(uint row, Klass* k) {
assert((uint)row < row_limit(), "oob");
set_intptr_at(receiver_cell_index(row), (uintptr_t)k);
}
uint receiver_count(uint row) const {
assert(row < row_limit(), "oob");
return uint_at(receiver_count_cell_index(row));
}
void set_receiver_count(uint row, uint count) {
assert(row < row_limit(), "oob");
set_uint_at(receiver_count_cell_index(row), count);
}
void clear_row(uint row) {
assert(row < row_limit(), "oob");
// Clear total count - indicator of polymorphic call site.
// The site may look like as monomorphic after that but
// it allow to have more accurate profiling information because
// there was execution phase change since klasses were unloaded.
// If the site is still polymorphic then MDO will be updated
// to reflect it. But it could be the case that the site becomes
// only bimorphic. Then keeping total count not 0 will be wrong.
// Even if we use monomorphic (when it is not) for compilation
// we will only have trap, deoptimization and recompile again
// with updated MDO after executing method in Interpreter.
// An additional receiver will be recorded in the cleaned row
// during next call execution.
//
// Note: our profiling logic works with empty rows in any slot.
// We do sorting a profiling info (ciCallProfile) for compilation.
//
set_count(0);
set_receiver(row, NULL);
set_receiver_count(row, 0);
#if INCLUDE_JVMCI
if (!this->is_VirtualCallData()) {
// if this is a ReceiverTypeData for JVMCI, the nonprofiled_count
// must also be reset (see "Description of the different counters" above)
set_nonprofiled_count(0);
}
#endif
}
// Code generation support
static ByteSize receiver_offset(uint row) {
return cell_offset(receiver_cell_index(row));
}
static ByteSize receiver_count_offset(uint row) {
return cell_offset(receiver_count_cell_index(row));
}
#if INCLUDE_JVMCI
static ByteSize nonprofiled_receiver_count_offset() {
return cell_offset(nonprofiled_count_off_set);
}
uint nonprofiled_count() const {
return uint_at(nonprofiled_count_off_set);
}
void set_nonprofiled_count(uint count) {
set_uint_at(nonprofiled_count_off_set, count);
}
#endif // INCLUDE_JVMCI
static ByteSize receiver_type_data_size() {
return cell_offset(static_cell_count());
}
// GC support
virtual void clean_weak_klass_links(bool always_clean);
void print_receiver_data_on(outputStream* st) const;
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// VirtualCallData
//
// A VirtualCallData is used to access profiling information about a
// virtual call. For now, it has nothing more than a ReceiverTypeData.
class VirtualCallData : public ReceiverTypeData {
public:
VirtualCallData(DataLayout* layout) : ReceiverTypeData(layout) {
assert(layout->tag() == DataLayout::virtual_call_data_tag ||
layout->tag() == DataLayout::virtual_call_type_data_tag, "wrong type");
}
virtual bool is_VirtualCallData() const { return true; }
static int static_cell_count() {
// At this point we could add more profile state, e.g., for arguments.
// But for now it's the same size as the base record type.
return ReceiverTypeData::static_cell_count();
}
virtual int cell_count() const {
return static_cell_count();
}
// Direct accessors
static ByteSize virtual_call_data_size() {
return cell_offset(static_cell_count());
}
void print_method_data_on(outputStream* st) const NOT_JVMCI_RETURN;
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// VirtualCallTypeData
//
// A VirtualCallTypeData is used to access profiling information about
// a virtual call for which we collect type information about
// arguments and return value.
class VirtualCallTypeData : public VirtualCallData {
private:
// entries for arguments if any
TypeStackSlotEntries _args;
// entry for return type if any
ReturnTypeEntry _ret;
int cell_count_global_offset() const {
return VirtualCallData::static_cell_count() + TypeEntriesAtCall::cell_count_local_offset();
}
// number of cells not counting the header
int cell_count_no_header() const {
return uint_at(cell_count_global_offset());
}
void check_number_of_arguments(int total) {
assert(number_of_arguments() == total, "should be set in DataLayout::initialize");
}
public:
VirtualCallTypeData(DataLayout* layout) :
VirtualCallData(layout),
_args(VirtualCallData::static_cell_count()+TypeEntriesAtCall::header_cell_count(), number_of_arguments()),
_ret(cell_count() - ReturnTypeEntry::static_cell_count())
{
assert(layout->tag() == DataLayout::virtual_call_type_data_tag, "wrong type");
// Some compilers (VC++) don't want this passed in member initialization list
_args.set_profile_data(this);
_ret.set_profile_data(this);
}
const TypeStackSlotEntries* args() const {
assert(has_arguments(), "no profiling of arguments");
return &_args;
}
const ReturnTypeEntry* ret() const {
assert(has_return(), "no profiling of return value");
return &_ret;
}
virtual bool is_VirtualCallTypeData() const { return true; }
static int static_cell_count() {
return -1;
}
static int compute_cell_count(BytecodeStream* stream) {
return VirtualCallData::static_cell_count() + TypeEntriesAtCall::compute_cell_count(stream);
}
static void initialize(DataLayout* dl, int cell_count) {
TypeEntriesAtCall::initialize(dl, VirtualCallData::static_cell_count(), cell_count);
}
virtual void post_initialize(BytecodeStream* stream, MethodData* mdo);
virtual int cell_count() const {
return VirtualCallData::static_cell_count() +
TypeEntriesAtCall::header_cell_count() +
int_at_unchecked(cell_count_global_offset());
}
int number_of_arguments() const {
return cell_count_no_header() / TypeStackSlotEntries::per_arg_count();
}
void set_argument_type(int i, Klass* k) {
assert(has_arguments(), "no arguments!");
intptr_t current = _args.type(i);
_args.set_type(i, TypeEntries::with_status(k, current));
}
void set_return_type(Klass* k) {
assert(has_return(), "no return!");
intptr_t current = _ret.type();
_ret.set_type(TypeEntries::with_status(k, current));
}
// An entry for a return value takes less space than an entry for an
// argument, so if the remainder of the number of cells divided by
// the number of cells for an argument is not null, a return value
// is profiled in this object.
bool has_return() const {
bool res = (cell_count_no_header() % TypeStackSlotEntries::per_arg_count()) != 0;
assert (!res || TypeEntriesAtCall::return_profiling_enabled(), "no profiling of return values");
return res;
}
// An entry for a return value takes less space than an entry for an
// argument so if the number of cells exceeds the number of cells
// needed for an argument, this object contains type information for
// at least one argument.
bool has_arguments() const {
bool res = cell_count_no_header() >= TypeStackSlotEntries::per_arg_count();
assert (!res || TypeEntriesAtCall::arguments_profiling_enabled(), "no profiling of arguments");
return res;
}
// Code generation support
static ByteSize args_data_offset() {
return cell_offset(VirtualCallData::static_cell_count()) + TypeEntriesAtCall::args_data_offset();
}
ByteSize argument_type_offset(int i) {
return _args.type_offset(i);
}
ByteSize return_type_offset() {
return _ret.type_offset();
}
// GC support
virtual void clean_weak_klass_links(bool always_clean) {
ReceiverTypeData::clean_weak_klass_links(always_clean);
if (has_arguments()) {
_args.clean_weak_klass_links(always_clean);
}
if (has_return()) {
_ret.clean_weak_klass_links(always_clean);
}
}
virtual void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// RetData
//
// A RetData is used to access profiling information for a ret bytecode.
// It is composed of a count of the number of times that the ret has
// been executed, followed by a series of triples of the form
// (bci, count, di) which count the number of times that some bci was the
// target of the ret and cache a corresponding data displacement.
class RetData : public CounterData {
protected:
enum {
bci0_offset = counter_cell_count,
count0_offset,
displacement0_offset,
ret_row_cell_count = (displacement0_offset + 1) - bci0_offset
};
void set_bci(uint row, int bci) {
assert((uint)row < row_limit(), "oob");
set_int_at(bci0_offset + row * ret_row_cell_count, bci);
}
void release_set_bci(uint row, int bci);
void set_bci_count(uint row, uint count) {
assert((uint)row < row_limit(), "oob");
set_uint_at(count0_offset + row * ret_row_cell_count, count);
}
void set_bci_displacement(uint row, int disp) {
set_int_at(displacement0_offset + row * ret_row_cell_count, disp);
}
public:
RetData(DataLayout* layout) : CounterData(layout) {
assert(layout->tag() == DataLayout::ret_data_tag, "wrong type");
}
virtual bool is_RetData() const { return true; }
enum {
no_bci = -1 // value of bci when bci1/2 are not in use.
};
static int static_cell_count() {
return counter_cell_count + (uint) BciProfileWidth * ret_row_cell_count;
}
virtual int cell_count() const {
return static_cell_count();
}
static uint row_limit() {
return BciProfileWidth;
}
static int bci_cell_index(uint row) {
return bci0_offset + row * ret_row_cell_count;
}
static int bci_count_cell_index(uint row) {
return count0_offset + row * ret_row_cell_count;
}
static int bci_displacement_cell_index(uint row) {
return displacement0_offset + row * ret_row_cell_count;
}
// Direct accessors
int bci(uint row) const {
return int_at(bci_cell_index(row));
}
uint bci_count(uint row) const {
return uint_at(bci_count_cell_index(row));
}
int bci_displacement(uint row) const {
return int_at(bci_displacement_cell_index(row));
}
// Interpreter Runtime support
address fixup_ret(int return_bci, MethodData* mdo);
// Code generation support
static ByteSize bci_offset(uint row) {
return cell_offset(bci_cell_index(row));
}
static ByteSize bci_count_offset(uint row) {
return cell_offset(bci_count_cell_index(row));
}
static ByteSize bci_displacement_offset(uint row) {
return cell_offset(bci_displacement_cell_index(row));
}
// Specific initialization.
void post_initialize(BytecodeStream* stream, MethodData* mdo);
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// BranchData
//
// A BranchData is used to access profiling data for a two-way branch.
// It consists of taken and not_taken counts as well as a data displacement
// for the taken case.
class BranchData : public JumpData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
enum {
not_taken_off_set = jump_cell_count,
branch_cell_count
};
void set_displacement(int displacement) {
set_int_at(displacement_off_set, displacement);
}
public:
BranchData(DataLayout* layout) : JumpData(layout) {
assert(layout->tag() == DataLayout::branch_data_tag, "wrong type");
}
virtual bool is_BranchData() const { return true; }
static int static_cell_count() {
return branch_cell_count;
}
virtual int cell_count() const {
return static_cell_count();
}
// Direct accessor
uint not_taken() const {
return uint_at(not_taken_off_set);
}
void set_not_taken(uint cnt) {
set_uint_at(not_taken_off_set, cnt);
}
uint inc_not_taken() {
uint cnt = not_taken() + 1;
// Did we wrap? Will compiler screw us??
if (cnt == 0) cnt--;
set_uint_at(not_taken_off_set, cnt);
return cnt;
}
// Code generation support
static ByteSize not_taken_offset() {
return cell_offset(not_taken_off_set);
}
static ByteSize branch_data_size() {
return cell_offset(branch_cell_count);
}
// Specific initialization.
void post_initialize(BytecodeStream* stream, MethodData* mdo);
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// ArrayData
//
// A ArrayData is a base class for accessing profiling data which does
// not have a statically known size. It consists of an array length
// and an array start.
class ArrayData : public ProfileData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
friend class DataLayout;
enum {
array_len_off_set,
array_start_off_set
};
uint array_uint_at(int index) const {
int aindex = index + array_start_off_set;
return uint_at(aindex);
}
int array_int_at(int index) const {
int aindex = index + array_start_off_set;
return int_at(aindex);
}
oop array_oop_at(int index) const {
int aindex = index + array_start_off_set;
return oop_at(aindex);
}
void array_set_int_at(int index, int value) {
int aindex = index + array_start_off_set;
set_int_at(aindex, value);
}
// Code generation support for subclasses.
static ByteSize array_element_offset(int index) {
return cell_offset(array_start_off_set + index);
}
public:
ArrayData(DataLayout* layout) : ProfileData(layout) {}
virtual bool is_ArrayData() const { return true; }
static int static_cell_count() {
return -1;
}
int array_len() const {
return int_at_unchecked(array_len_off_set);
}
virtual int cell_count() const {
return array_len() + 1;
}
// Code generation support
static ByteSize array_len_offset() {
return cell_offset(array_len_off_set);
}
static ByteSize array_start_offset() {
return cell_offset(array_start_off_set);
}
};
// MultiBranchData
//
// A MultiBranchData is used to access profiling information for
// a multi-way branch (*switch bytecodes). It consists of a series
// of (count, displacement) pairs, which count the number of times each
// case was taken and specify the data displacement for each branch target.
class MultiBranchData : public ArrayData {
friend class VMStructs;
friend class JVMCIVMStructs;
protected:
enum {
default_count_off_set,
default_disaplacement_off_set,
case_array_start
};
enum {
relative_count_off_set,
relative_displacement_off_set,
per_case_cell_count
};
void set_default_displacement(int displacement) {
array_set_int_at(default_disaplacement_off_set, displacement);
}
void set_displacement_at(int index, int displacement) {
array_set_int_at(case_array_start +
index * per_case_cell_count +
relative_displacement_off_set,
displacement);
}
public:
MultiBranchData(DataLayout* layout) : ArrayData(layout) {
assert(layout->tag() == DataLayout::multi_branch_data_tag, "wrong type");
}
virtual bool is_MultiBranchData() const { return true; }
static int compute_cell_count(BytecodeStream* stream);
int number_of_cases() const {
int alen = array_len() - 2; // get rid of default case here.
assert(alen % per_case_cell_count == 0, "must be even");
return (alen / per_case_cell_count);
}
uint default_count() const {
return array_uint_at(default_count_off_set);
}
int default_displacement() const {
return array_int_at(default_disaplacement_off_set);
}
uint count_at(int index) const {
return array_uint_at(case_array_start +
index * per_case_cell_count +
relative_count_off_set);
}
int displacement_at(int index) const {
return array_int_at(case_array_start +
index * per_case_cell_count +
relative_displacement_off_set);
}
// Code generation support
static ByteSize default_count_offset() {
return array_element_offset(default_count_off_set);
}
static ByteSize default_displacement_offset() {
return array_element_offset(default_disaplacement_off_set);
}
static ByteSize case_count_offset(int index) {
return case_array_offset() +
(per_case_size() * index) +
relative_count_offset();
}
static ByteSize case_array_offset() {
return array_element_offset(case_array_start);
}
static ByteSize per_case_size() {
return in_ByteSize(per_case_cell_count) * cell_size;
}
static ByteSize relative_count_offset() {
return in_ByteSize(relative_count_off_set) * cell_size;
}
static ByteSize relative_displacement_offset() {
return in_ByteSize(relative_displacement_off_set) * cell_size;
}
// Specific initialization.
void post_initialize(BytecodeStream* stream, MethodData* mdo);
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
class ArgInfoData : public ArrayData {
public:
ArgInfoData(DataLayout* layout) : ArrayData(layout) {
assert(layout->tag() == DataLayout::arg_info_data_tag, "wrong type");
}
virtual bool is_ArgInfoData() const { return true; }
int number_of_args() const {
return array_len();
}
uint arg_modified(int arg) const {
return array_uint_at(arg);
}
void set_arg_modified(int arg, uint val) {
array_set_int_at(arg, val);
}
void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// ParametersTypeData
//
// A ParametersTypeData is used to access profiling information about
// types of parameters to a method
class ParametersTypeData : public ArrayData {
private:
TypeStackSlotEntries _parameters;
static int stack_slot_local_offset(int i) {
assert_profiling_enabled();
return array_start_off_set + TypeStackSlotEntries::stack_slot_local_offset(i);
}
static int type_local_offset(int i) {
assert_profiling_enabled();
return array_start_off_set + TypeStackSlotEntries::type_local_offset(i);
}
static bool profiling_enabled();
static void assert_profiling_enabled() {
assert(profiling_enabled(), "method parameters profiling should be on");
}
public:
ParametersTypeData(DataLayout* layout) : ArrayData(layout), _parameters(1, number_of_parameters()) {
assert(layout->tag() == DataLayout::parameters_type_data_tag, "wrong type");
// Some compilers (VC++) don't want this passed in member initialization list
_parameters.set_profile_data(this);
}
static int compute_cell_count(Method* m);
virtual bool is_ParametersTypeData() const { return true; }
virtual void post_initialize(BytecodeStream* stream, MethodData* mdo);
int number_of_parameters() const {
return array_len() / TypeStackSlotEntries::per_arg_count();
}
const TypeStackSlotEntries* parameters() const { return &_parameters; }
uint stack_slot(int i) const {
return _parameters.stack_slot(i);
}
void set_type(int i, Klass* k) {
intptr_t current = _parameters.type(i);
_parameters.set_type(i, TypeEntries::with_status((intptr_t)k, current));
}
virtual void clean_weak_klass_links(bool always_clean) {
_parameters.clean_weak_klass_links(always_clean);
}
virtual void print_data_on(outputStream* st, const char* extra = NULL) const;
static ByteSize stack_slot_offset(int i) {
return cell_offset(stack_slot_local_offset(i));
}
static ByteSize type_offset(int i) {
return cell_offset(type_local_offset(i));
}
};
// SpeculativeTrapData
//
// A SpeculativeTrapData is used to record traps due to type
// speculation. It records the root of the compilation: that type
// speculation is wrong in the context of one compilation (for
// method1) doesn't mean it's wrong in the context of another one (for
// method2). Type speculation could have more/different data in the
// context of the compilation of method2 and it's worthwhile to try an
// optimization that failed for compilation of method1 in the context
// of compilation of method2.
// Space for SpeculativeTrapData entries is allocated from the extra
// data space in the MDO. If we run out of space, the trap data for
// the ProfileData at that bci is updated.
class SpeculativeTrapData : public ProfileData {
protected:
enum {
speculative_trap_method,
#ifndef _LP64
// The size of the area for traps is a multiple of the header
// size, 2 cells on 32 bits. Packed at the end of this area are
// argument info entries (with tag
// DataLayout::arg_info_data_tag). The logic in
// MethodData::bci_to_extra_data() that guarantees traps don't
// overflow over argument info entries assumes the size of a
// SpeculativeTrapData is twice the header size. On 32 bits, a
// SpeculativeTrapData must be 4 cells.
padding,
#endif
speculative_trap_cell_count
};
public:
SpeculativeTrapData(DataLayout* layout) : ProfileData(layout) {
assert(layout->tag() == DataLayout::speculative_trap_data_tag, "wrong type");
}
virtual bool is_SpeculativeTrapData() const { return true; }
static int static_cell_count() {
return speculative_trap_cell_count;
}
virtual int cell_count() const {
return static_cell_count();
}
// Direct accessor
Method* method() const {
return (Method*)intptr_at(speculative_trap_method);
}
void set_method(Method* m) {
assert(!m->is_old(), "cannot add old methods");
set_intptr_at(speculative_trap_method, (intptr_t)m);
}
static ByteSize method_offset() {
return cell_offset(speculative_trap_method);
}
virtual void print_data_on(outputStream* st, const char* extra = NULL) const;
};
// MethodData*
//
// A MethodData* holds information which has been collected about
// a method. Its layout looks like this:
//
// -----------------------------
// | header |
// | klass |
// -----------------------------
// | method |
// | size of the MethodData* |
// -----------------------------
// | Data entries... |
// | (variable size) |
// | |
// . .
// . .
// . .
// | |
// -----------------------------
//
// The data entry area is a heterogeneous array of DataLayouts. Each
// DataLayout in the array corresponds to a specific bytecode in the
// method. The entries in the array are sorted by the corresponding
// bytecode. Access to the data is via resource-allocated ProfileData,
// which point to the underlying blocks of DataLayout structures.
//
// During interpretation, if profiling in enabled, the interpreter
// maintains a method data pointer (mdp), which points at the entry
// in the array corresponding to the current bci. In the course of
// interpretation, when a bytecode is encountered that has profile data
// associated with it, the entry pointed to by mdp is updated, then the
// mdp is adjusted to point to the next appropriate DataLayout. If mdp
// is NULL to begin with, the interpreter assumes that the current method
// is not (yet) being profiled.
//
// In MethodData* parlance, "dp" is a "data pointer", the actual address
// of a DataLayout element. A "di" is a "data index", the offset in bytes
// from the base of the data entry array. A "displacement" is the byte offset
// in certain ProfileData objects that indicate the amount the mdp must be
// adjusted in the event of a change in control flow.
//
class CleanExtraDataClosure : public StackObj {
public:
virtual bool is_live(Method* m) = 0;
};
#if INCLUDE_JVMCI
// Encapsulates an encoded speculation reason. These are linked together in
// a list that is atomically appended to during deoptimization. Entries are
// never removed from the list.
// @see jdk.vm.ci.hotspot.HotSpotSpeculationLog.HotSpotSpeculationEncoding
class FailedSpeculation: public CHeapObj<mtCompiler> {
private:
// The length of HotSpotSpeculationEncoding.toByteArray(). The data itself
// is an array embedded at the end of this object.
int _data_len;
// Next entry in a linked list.
FailedSpeculation* _next;
FailedSpeculation(address data, int data_len);
FailedSpeculation** next_adr() { return &_next; }
// Placement new operator for inlining the speculation data into
// the FailedSpeculation object.
void* operator new(size_t size, size_t fs_size) throw();
public:
char* data() { return (char*)(((address) this) + sizeof(FailedSpeculation)); }
int data_len() const { return _data_len; }
FailedSpeculation* next() const { return _next; }
// Atomically appends a speculation from nm to the list whose head is at (*failed_speculations_address).
// Returns false if the FailedSpeculation object could not be allocated.
static bool add_failed_speculation(nmethod* nm, FailedSpeculation** failed_speculations_address, address speculation, int speculation_len);
// Frees all entries in the linked list whose head is at (*failed_speculations_address).
static void free_failed_speculations(FailedSpeculation** failed_speculations_address);
};
#endif
class ciMethodData;
class MethodData : public Metadata {
friend class VMStructs;
friend class JVMCIVMStructs;
private:
friend class ProfileData;
friend class TypeEntriesAtCall;
friend class ciMethodData;
// If you add a new field that points to any metaspace object, you
// must add this field to MethodData::metaspace_pointers_do().
// Back pointer to the Method*
Method* _method;
// Size of this oop in bytes
int _size;
// Cached hint for bci_to_dp and bci_to_data
int _hint_di;
Mutex _extra_data_lock;
MethodData(const methodHandle& method);
public:
static MethodData* allocate(ClassLoaderData* loader_data, const methodHandle& method, TRAPS);
virtual bool is_methodData() const { return true; }
void initialize();
// Whole-method sticky bits and flags
enum {
_trap_hist_limit = Deoptimization::Reason_TRAP_HISTORY_LENGTH,
_trap_hist_mask = max_jubyte,
_extra_data_count = 4 // extra DataLayout headers, for trap history
}; // Public flag values
// Compiler-related counters.
class CompilerCounters {
friend class VMStructs;
friend class JVMCIVMStructs;
uint _nof_decompiles; // count of all nmethod removals
uint _nof_overflow_recompiles; // recompile count, excluding recomp. bits
uint _nof_overflow_traps; // trap count, excluding _trap_hist
union {
intptr_t _align;
// JVMCI separates trap history for OSR compilations from normal compilations
u1 _array[JVMCI_ONLY(2 *) MethodData::_trap_hist_limit];
} _trap_hist;
public:
CompilerCounters() : _nof_decompiles(0), _nof_overflow_recompiles(0), _nof_overflow_traps(0) {
#ifndef ZERO
// Some Zero platforms do not have expected alignment, and do not use
// this code. static_assert would still fire and fail for them.
static_assert(sizeof(_trap_hist) % HeapWordSize == 0, "align");
#endif
uint size_in_words = sizeof(_trap_hist) / HeapWordSize;
Copy::zero_to_words((HeapWord*) &_trap_hist, size_in_words);
}
// Return (uint)-1 for overflow.
uint trap_count(int reason) const {
assert((uint)reason < ARRAY_SIZE(_trap_hist._array), "oob");
return (int)((_trap_hist._array[reason]+1) & _trap_hist_mask) - 1;
}
uint inc_trap_count(int reason) {
// Count another trap, anywhere in this method.
assert(reason >= 0, "must be single trap");
assert((uint)reason < ARRAY_SIZE(_trap_hist._array), "oob");
uint cnt1 = 1 + _trap_hist._array[reason];
if ((cnt1 & _trap_hist_mask) != 0) { // if no counter overflow...
_trap_hist._array[reason] = cnt1;
return cnt1;
} else {
return _trap_hist_mask + (++_nof_overflow_traps);
}
}
uint overflow_trap_count() const {
return _nof_overflow_traps;
}
uint overflow_recompile_count() const {
return _nof_overflow_recompiles;
}
uint inc_overflow_recompile_count() {
return ++_nof_overflow_recompiles;
}
uint decompile_count() const {
return _nof_decompiles;
}
uint inc_decompile_count() {
return ++_nof_decompiles;
}
// Support for code generation
static ByteSize trap_history_offset() {
return byte_offset_of(CompilerCounters, _trap_hist._array);
}
};
private:
CompilerCounters _compiler_counters;
// Support for interprocedural escape analysis, from Thomas Kotzmann.
intx _eflags; // flags on escape information
intx _arg_local; // bit set of non-escaping arguments
intx _arg_stack; // bit set of stack-allocatable arguments
intx _arg_returned; // bit set of returned arguments
int _creation_mileage; // method mileage at MDO creation
// How many invocations has this MDO seen?
// These counters are used to determine the exact age of MDO.
// We need those because in tiered a method can be concurrently
// executed at different levels.
InvocationCounter _invocation_counter;
// Same for backedges.
InvocationCounter _backedge_counter;
// Counter values at the time profiling started.
int _invocation_counter_start;
int _backedge_counter_start;
uint _tenure_traps;
int _invoke_mask; // per-method Tier0InvokeNotifyFreqLog
int _backedge_mask; // per-method Tier0BackedgeNotifyFreqLog
#if INCLUDE_RTM_OPT
// State of RTM code generation during compilation of the method
int _rtm_state;
#endif
// Number of loops and blocks is computed when compiling the first
// time with C1. It is used to determine if method is trivial.
short _num_loops;
short _num_blocks;
// Does this method contain anything worth profiling?
enum WouldProfile {unknown, no_profile, profile};
WouldProfile _would_profile;
#if INCLUDE_JVMCI
// Support for HotSpotMethodData.setCompiledIRSize(int)
int _jvmci_ir_size;
FailedSpeculation* _failed_speculations;
#endif
// Size of _data array in bytes. (Excludes header and extra_data fields.)
int _data_size;
// data index for the area dedicated to parameters. -1 if no
// parameter profiling.
enum { no_parameters = -2, parameters_uninitialized = -1 };
int _parameters_type_data_di;
// Beginning of the data entries
intptr_t _data[1];
// Helper for size computation
static int compute_data_size(BytecodeStream* stream);
static int bytecode_cell_count(Bytecodes::Code code);
static bool is_speculative_trap_bytecode(Bytecodes::Code code);
enum { no_profile_data = -1, variable_cell_count = -2 };
// Helper for initialization
DataLayout* data_layout_at(int data_index) const {
assert(data_index % sizeof(intptr_t) == 0, "unaligned");
return (DataLayout*) (((address)_data) + data_index);
}
// Initialize an individual data segment. Returns the size of
// the segment in bytes.
int initialize_data(BytecodeStream* stream, int data_index);
// Helper for data_at
DataLayout* limit_data_position() const {
return data_layout_at(_data_size);
}
bool out_of_bounds(int data_index) const {
return data_index >= data_size();
}
// Give each of the data entries a chance to perform specific
// data initialization.
--> --------------------
--> maximum size reached
--> --------------------
¤ Dauer der Verarbeitung: 0.75 Sekunden
(vorverarbeitet)
¤
|
Haftungshinweis
Die Informationen auf dieser Webseite wurden
nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit,
noch Qualität der bereit gestellten Informationen zugesichert.
Bemerkung:
Die farbliche Syntaxdarstellung ist noch experimentell.
|
