/*
* Copyright (c) 2019, 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. Oracle designates this
* particular file as subject to the "Classpath" exception as provided
* by Oracle in the LICENSE file that accompanied this code.
*
* 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.
*
*/
package jdk.incubator.foreign;
import jdk.internal.access.JavaLangInvokeAccess;
import jdk.internal.access.SharedSecrets;
import jdk.internal.foreign.Utils;
import sun.invoke.util.Wrapper;
import java.lang.invoke.MethodHandle;
import java.lang.invoke.MethodHandles;
import java.lang.invoke.MethodType;
import java.lang.invoke.VarHandle;
import java.nio.ByteOrder;
import java.util.List;
import java.util.Objects;
This class defines several factory methods for constructing and combining memory access var handles.
To obtain a memory access var handle, clients must start from one of the leaf methods (see varHandle(Class<?>, ByteOrder)
, varHandle(Class<?>, long, ByteOrder)
). This determines the variable type (all primitive types but void
and boolean
are supported), as well as the alignment constraint and the byte order associated to a memory access var handle. The resulting memory access var handle can then be combined in various ways to emulate different addressing modes. The var handles created by this class feature a mandatory coordinate type (of type MemorySegment
), and one long
coordinate type, which represents the offset, in bytes, relative to the segment, at which dereference should occur. As an example, consider the memory layout expressed by a GroupLayout
instance constructed as follows:
GroupLayout seq = MemoryLayout.ofStruct(
MemoryLayout.ofPaddingBits(32),
MemoryLayout.ofValueBits(32, ByteOrder.BIG_ENDIAN).withName("value")
);
To access the member layout named value
, we can construct a memory access var handle as follows:
VarHandle handle = MemoryHandles.varHandle(int.class, ByteOrder.BIG_ENDIAN); //(MemorySegment, long) -> int
handle = MemoryHandles.insertCoordinates(handle, 1, 4); //(MemorySegment) -> int
Unless otherwise specified, passing a null
argument, or an array argument containing one or more null
elements to a method in this class causes a NullPointerException
to be thrown.
Alignment and access modes
A memory access var handle is associated with an access size S
and an alignment constraint B
(both expressed in bytes). We say that a memory access operation is fully aligned if it occurs at a memory address A
which is compatible with both alignment constraints S
and B
. If access is fully aligned then following access modes are supported and are guaranteed to support atomic access:
- read write access modes for all
T
, with the exception of access modes get
and set
for long
and double
on 32-bit platforms. - atomic update access modes for
int
, long
, float
or double
. (Future major platform releases of the JDK may support additional types for certain currently unsupported access modes.) - numeric atomic update access modes for
int
and long
. (Future major platform releases of the JDK may support additional numeric types for certain currently unsupported access modes.) - bitwise atomic update access modes for
int
and long
. (Future major platform releases of the JDK may support additional numeric types for certain currently unsupported access modes.)
If T
is float
or double
then atomic update access modes compare values using their bitwise representation (see Float.floatToRawIntBits
and Double.doubleToRawLongBits
, respectively).
Alternatively, a memory access operation is partially aligned if it occurs at a memory address A
which is only compatible with the alignment constraint B
; in such cases, access for anything other than the get
and set
access modes will result in an IllegalStateException
. If access is partially aligned, atomic access is only guaranteed with respect to the largest power of two that divides the GCD of A
and S
.
Finally, in all other cases, we say that a memory access operation is misaligned; in such cases an IllegalStateException
is thrown, irrespective of the access mode being used.
/**
* This class defines several factory methods for constructing and combining memory access var handles.
* To obtain a memory access var handle, clients must start from one of the <em>leaf</em> methods
* (see {@link MemoryHandles#varHandle(Class, ByteOrder)},
* {@link MemoryHandles#varHandle(Class, long, ByteOrder)}). This determines the variable type
* (all primitive types but {@code void} and {@code boolean} are supported), as well as the alignment constraint and the
* byte order associated to a memory access var handle. The resulting memory access var handle can then be combined in various ways
* to emulate different addressing modes. The var handles created by this class feature a <em>mandatory</em> coordinate type
* (of type {@link MemorySegment}), and one {@code long} coordinate type, which represents the offset, in bytes, relative
* to the segment, at which dereference should occur.
* <p>
* As an example, consider the memory layout expressed by a {@link GroupLayout} instance constructed as follows:
* <blockquote><pre>{@code
GroupLayout seq = MemoryLayout.ofStruct(
MemoryLayout.ofPaddingBits(32),
MemoryLayout.ofValueBits(32, ByteOrder.BIG_ENDIAN).withName("value")
);
* }</pre></blockquote>
* To access the member layout named {@code value}, we can construct a memory access var handle as follows:
* <blockquote><pre>{@code
VarHandle handle = MemoryHandles.varHandle(int.class, ByteOrder.BIG_ENDIAN); //(MemorySegment, long) -> int
handle = MemoryHandles.insertCoordinates(handle, 1, 4); //(MemorySegment) -> int
* }</pre></blockquote>
*
* <p> Unless otherwise specified, passing a {@code null} argument, or an array argument containing one or more {@code null}
* elements to a method in this class causes a {@link NullPointerException NullPointerException} to be thrown. </p>
*
* <h2><a id="memaccess-mode"></a>Alignment and access modes</h2>
*
* A memory access var handle is associated with an access size {@code S} and an alignment constraint {@code B}
* (both expressed in bytes). We say that a memory access operation is <em>fully aligned</em> if it occurs
* at a memory address {@code A} which is compatible with both alignment constraints {@code S} and {@code B}.
* If access is fully aligned then following access modes are supported and are
* guaranteed to support atomic access:
* <ul>
* <li>read write access modes for all {@code T}, with the exception of
* access modes {@code get} and {@code set} for {@code long} and
* {@code double} on 32-bit platforms.
* <li>atomic update access modes for {@code int}, {@code long},
* {@code float} or {@code double}.
* (Future major platform releases of the JDK may support additional
* types for certain currently unsupported access modes.)
* <li>numeric atomic update access modes for {@code int} and {@code long}.
* (Future major platform releases of the JDK may support additional
* numeric types for certain currently unsupported access modes.)
* <li>bitwise atomic update access modes for {@code int} and {@code long}.
* (Future major platform releases of the JDK may support additional
* numeric types for certain currently unsupported access modes.)
* </ul>
*
* If {@code T} is {@code float} or {@code double} then atomic
* update access modes compare values using their bitwise representation
* (see {@link Float#floatToRawIntBits} and
* {@link Double#doubleToRawLongBits}, respectively).
* <p>
* Alternatively, a memory access operation is <em>partially aligned</em> if it occurs at a memory address {@code A}
* which is only compatible with the alignment constraint {@code B}; in such cases, access for anything other than the
* {@code get} and {@code set} access modes will result in an {@code IllegalStateException}. If access is partially aligned,
* atomic access is only guaranteed with respect to the largest power of two that divides the GCD of {@code A} and {@code S}.
* <p>
* Finally, in all other cases, we say that a memory access operation is <em>misaligned</em>; in such cases an
* {@code IllegalStateException} is thrown, irrespective of the access mode being used.
*/
public final class MemoryHandles {
private final static JavaLangInvokeAccess JLI = SharedSecrets.getJavaLangInvokeAccess();
private MemoryHandles() {
//sorry, just the one!
}
private static final MethodHandle LONG_TO_ADDRESS;
private static final MethodHandle ADDRESS_TO_LONG;
private static final MethodHandle INT_TO_BYTE;
private static final MethodHandle BYTE_TO_UNSIGNED_INT;
private static final MethodHandle INT_TO_SHORT;
private static final MethodHandle SHORT_TO_UNSIGNED_INT;
private static final MethodHandle LONG_TO_BYTE;
private static final MethodHandle BYTE_TO_UNSIGNED_LONG;
private static final MethodHandle LONG_TO_SHORT;
private static final MethodHandle SHORT_TO_UNSIGNED_LONG;
private static final MethodHandle LONG_TO_INT;
private static final MethodHandle INT_TO_UNSIGNED_LONG;
static {
try {
LONG_TO_ADDRESS = MethodHandles.lookup().findStatic(MemoryHandles.class, "longToAddress",
MethodType.methodType(MemoryAddress.class, long.class));
ADDRESS_TO_LONG = MethodHandles.lookup().findStatic(MemoryHandles.class, "addressToLong",
MethodType.methodType(long.class, MemoryAddress.class));
INT_TO_BYTE = MethodHandles.explicitCastArguments(MethodHandles.identity(byte.class),
MethodType.methodType(byte.class, int.class));
BYTE_TO_UNSIGNED_INT = MethodHandles.lookup().findStatic(Byte.class, "toUnsignedInt",
MethodType.methodType(int.class, byte.class));
INT_TO_SHORT = MethodHandles.explicitCastArguments(MethodHandles.identity(short.class),
MethodType.methodType(short.class, int.class));
SHORT_TO_UNSIGNED_INT = MethodHandles.lookup().findStatic(Short.class, "toUnsignedInt",
MethodType.methodType(int.class, short.class));
LONG_TO_BYTE = MethodHandles.explicitCastArguments(MethodHandles.identity(byte.class),
MethodType.methodType(byte.class, long.class));
BYTE_TO_UNSIGNED_LONG = MethodHandles.lookup().findStatic(Byte.class, "toUnsignedLong",
MethodType.methodType(long.class, byte.class));
LONG_TO_SHORT = MethodHandles.explicitCastArguments(MethodHandles.identity(short.class),
MethodType.methodType(short.class, long.class));
SHORT_TO_UNSIGNED_LONG = MethodHandles.lookup().findStatic(Short.class, "toUnsignedLong",
MethodType.methodType(long.class, short.class));
LONG_TO_INT = MethodHandles.explicitCastArguments(MethodHandles.identity(int.class),
MethodType.methodType(int.class, long.class));
INT_TO_UNSIGNED_LONG = MethodHandles.lookup().findStatic(Integer.class, "toUnsignedLong",
MethodType.methodType(long.class, int.class));
} catch (Throwable ex) {
throw new ExceptionInInitializerError(ex);
}
}
Creates a memory access var handle with the given carrier type and byte order. The returned var handle's type is carrier
and the list of coordinate types is (MemorySegment, long)
, where the long
coordinate type corresponds to byte offset into a given memory segment. The returned var handle accesses bytes at an offset in a given memory segment, composing bytes to or from a value of the type carrier
according to the given endianness; the alignment constraint (in bytes) for the resulting memory access var handle is the same as the size (in bytes) of the carrier type carrier
. Params: - carrier – the carrier type. Valid carriers are
byte
, short
, char
, int
, float
, long
, and double
. - byteOrder – the required byte order.
Throws: - IllegalArgumentException – when an illegal carrier type is used
API Note: the resulting var handle features certain access mode restrictions,
which are common to all memory access var handles. Returns: the new memory access var handle.
/**
* Creates a memory access var handle with the given carrier type and byte order.
*
* The returned var handle's type is {@code carrier} and the list of coordinate types is
* {@code (MemorySegment, long)}, where the {@code long} coordinate type corresponds to byte offset into
* a given memory segment. The returned var handle accesses bytes at an offset in a given
* memory segment, composing bytes to or from a value of the type {@code carrier} according to the given endianness;
* the alignment constraint (in bytes) for the resulting memory access var handle is the same as the size (in bytes) of the
* carrier type {@code carrier}.
*
* @apiNote the resulting var handle features certain <a href="#memaccess-mode">access mode restrictions</a>,
* which are common to all memory access var handles.
*
* @param carrier the carrier type. Valid carriers are {@code byte}, {@code short}, {@code char}, {@code int},
* {@code float}, {@code long}, and {@code double}.
* @param byteOrder the required byte order.
* @return the new memory access var handle.
* @throws IllegalArgumentException when an illegal carrier type is used
*/
public static VarHandle varHandle(Class<?> carrier, ByteOrder byteOrder) {
Objects.requireNonNull(carrier);
Objects.requireNonNull(byteOrder);
return varHandle(carrier,
carrierSize(carrier),
byteOrder);
}
Creates a memory access var handle with the given carrier type, alignment constraint, and byte order. The returned var handle's type is carrier
and the list of coordinate types is (MemorySegment, long)
, where the long
coordinate type corresponds to byte offset into a given memory segment. The returned var handle accesses bytes at an offset in a given memory segment, composing bytes to or from a value of the type carrier
according to the given endianness; the alignment constraint (in bytes) for the resulting memory access var handle is given by alignmentBytes
. Params: - carrier – the carrier type. Valid carriers are
byte
, short
, char
, int
, float
, long
, and double
. - alignmentBytes – the alignment constraint (in bytes). Must be a power of two.
- byteOrder – the required byte order.
Throws: - IllegalArgumentException – if an illegal carrier type is used, or if
alignmentBytes
is not a power of two.
API Note: the resulting var handle features certain access mode restrictions,
which are common to all memory access var handles. Returns: the new memory access var handle.
/**
* Creates a memory access var handle with the given carrier type, alignment constraint, and byte order.
*
* The returned var handle's type is {@code carrier} and the list of coordinate types is
* {@code (MemorySegment, long)}, where the {@code long} coordinate type corresponds to byte offset into
* a given memory segment. The returned var handle accesses bytes at an offset in a given
* memory segment, composing bytes to or from a value of the type {@code carrier} according to the given endianness;
* the alignment constraint (in bytes) for the resulting memory access var handle is given by {@code alignmentBytes}.
*
* @apiNote the resulting var handle features certain <a href="#memaccess-mode">access mode restrictions</a>,
* which are common to all memory access var handles.
*
* @param carrier the carrier type. Valid carriers are {@code byte}, {@code short}, {@code char}, {@code int},
* {@code float}, {@code long}, and {@code double}.
* @param alignmentBytes the alignment constraint (in bytes). Must be a power of two.
* @param byteOrder the required byte order.
* @return the new memory access var handle.
* @throws IllegalArgumentException if an illegal carrier type is used, or if {@code alignmentBytes} is not a power of two.
*/
public static VarHandle varHandle(Class<?> carrier, long alignmentBytes, ByteOrder byteOrder) {
Objects.requireNonNull(carrier);
Objects.requireNonNull(byteOrder);
checkCarrier(carrier);
if (alignmentBytes <= 0
|| (alignmentBytes & (alignmentBytes - 1)) != 0) { // is power of 2?
throw new IllegalArgumentException("Bad alignment: " + alignmentBytes);
}
return Utils.fixUpVarHandle(JLI.memoryAccessVarHandle(carrier, false, alignmentBytes - 1, byteOrder));
}
Adapt an existing var handle into a new var handle whose carrier type is MemorySegment
. That is, when calling VarHandle.get(Object...)
on the returned var handle, the read numeric value will be turned into a memory address (as if by calling MemoryAddress.ofLong(long)
); similarly, when calling VarHandle.set(Object...)
, the memory address to be set will be converted into a numeric value, and then written into memory. The amount of bytes read (resp. written) from (resp. to) memory depends on the carrier of the original memory access var handle. Params: - target – the memory access var handle to be adapted
Throws: - IllegalArgumentException – if the carrier type of
varHandle
is either boolean
, float
, or double
, or is not a primitive type.
Returns: the adapted var handle.
/**
* Adapt an existing var handle into a new var handle whose carrier type is {@link MemorySegment}.
* That is, when calling {@link VarHandle#get(Object...)} on the returned var handle,
* the read numeric value will be turned into a memory address (as if by calling {@link MemoryAddress#ofLong(long)});
* similarly, when calling {@link VarHandle#set(Object...)}, the memory address to be set will be converted
* into a numeric value, and then written into memory. The amount of bytes read (resp. written) from (resp. to)
* memory depends on the carrier of the original memory access var handle.
*
* @param target the memory access var handle to be adapted
* @return the adapted var handle.
* @throws IllegalArgumentException if the carrier type of {@code varHandle} is either {@code boolean},
* {@code float}, or {@code double}, or is not a primitive type.
*/
public static VarHandle asAddressVarHandle(VarHandle target) {
Objects.requireNonNull(target);
Class<?> carrier = target.varType();
if (!carrier.isPrimitive() || carrier == boolean.class ||
carrier == float.class || carrier == double.class) {
throw new IllegalArgumentException("Unsupported carrier type: " + carrier.getName());
}
if (carrier != long.class) {
// slow-path, we need to adapt
return filterValue(target,
MethodHandles.explicitCastArguments(ADDRESS_TO_LONG, MethodType.methodType(carrier, MemoryAddress.class)),
MethodHandles.explicitCastArguments(LONG_TO_ADDRESS, MethodType.methodType(MemoryAddress.class, carrier)));
} else {
// fast-path
return filterValue(target, ADDRESS_TO_LONG, LONG_TO_ADDRESS);
}
}
Adapts a target var handle by narrowing incoming values and widening
outgoing values, to and from the given type, respectively.
The returned var handle can be used to conveniently treat unsigned
primitive data types as if they were a wider signed primitive type. For
example, it is often convenient to model an unsigned short as a Java int
to avoid dealing with negative values, which would be the case if modeled as a Java short
. This is illustrated in the following example:
MemorySegment segment = MemorySegment.allocateNative(2);
VarHandle SHORT_VH = MemoryLayouts.JAVA_SHORT.varHandle(short.class);
VarHandle INT_VH = MemoryHandles.asUnsigned(SHORT_VH, int.class);
SHORT_VH.set(segment, (short)-1);
INT_VH.get(segment); // returns 65535
When calling e.g. VarHandle.set(Object...)
on the resulting var handle, the incoming value (of type adaptedType
) is converted by a narrowing primitive conversion and then passed to the
target
var handle. A narrowing primitive conversion may lose information about the overall magnitude of a numeric value. Conversely, when calling e.g. VarHandle.get(Object...)
on the resulting var handle, the returned value obtained from the target
var handle is converted by a unsigned widening conversion before being returned to the caller. In an unsigned widening conversion the high-order bits greater than that of the target
carrier type are zero, and the low-order bits (equal to the width of the target
carrier type) are equal to the bits of the value obtained from the target
var handle.
The returned var handle will feature the variable type adaptedType
, and the same access coordinates, the same access modes (see AccessMode
, and the same atomic access guarantees, as those featured by the target
var handle.
Params: - target – the memory access var handle to be adapted
- adaptedType – the adapted type
Throws: - IllegalArgumentException – if the carrier type of
target
is not one of byte
, short
, or int
; if
adaptedType
is not one of int
, or long
; if the bitwidth of the adaptedType
is not greater than that of the target
carrier type.
Returns: the adapted var handle. @jls 5.1.3 Narrowing Primitive Conversion
/**
* Adapts a target var handle by narrowing incoming values and widening
* outgoing values, to and from the given type, respectively.
* <p>
* The returned var handle can be used to conveniently treat unsigned
* primitive data types as if they were a wider signed primitive type. For
* example, it is often convenient to model an <i>unsigned short</i> as a
* Java {@code int} to avoid dealing with negative values, which would be
* the case if modeled as a Java {@code short}. This is illustrated in the following example:
* <blockquote><pre>{@code
MemorySegment segment = MemorySegment.allocateNative(2);
VarHandle SHORT_VH = MemoryLayouts.JAVA_SHORT.varHandle(short.class);
VarHandle INT_VH = MemoryHandles.asUnsigned(SHORT_VH, int.class);
SHORT_VH.set(segment, (short)-1);
INT_VH.get(segment); // returns 65535
* }</pre></blockquote>
* <p>
* When calling e.g. {@link VarHandle#set(Object...)} on the resulting var
* handle, the incoming value (of type {@code adaptedType}) is converted by a
* <i>narrowing primitive conversion</i> and then passed to the {@code
* target} var handle. A narrowing primitive conversion may lose information
* about the overall magnitude of a numeric value. Conversely, when calling
* e.g. {@link VarHandle#get(Object...)} on the resulting var handle, the
* returned value obtained from the {@code target} var handle is converted
* by a <i>unsigned widening conversion</i> before being returned to the
* caller. In an unsigned widening conversion the high-order bits greater
* than that of the {@code target} carrier type are zero, and the low-order
* bits (equal to the width of the {@code target} carrier type) are equal to
* the bits of the value obtained from the {@code target} var handle.
* <p>
* The returned var handle will feature the variable type {@code adaptedType},
* and the same access coordinates, the same access modes (see {@link
* java.lang.invoke.VarHandle.AccessMode}, and the same atomic access
* guarantees, as those featured by the {@code target} var handle.
*
* @param target the memory access var handle to be adapted
* @param adaptedType the adapted type
* @return the adapted var handle.
* @throws IllegalArgumentException if the carrier type of {@code target}
* is not one of {@code byte}, {@code short}, or {@code int}; if {@code
* adaptedType} is not one of {@code int}, or {@code long}; if the bitwidth
* of the {@code adaptedType} is not greater than that of the {@code target}
* carrier type.
*
* @jls 5.1.3 Narrowing Primitive Conversion
*/
public static VarHandle asUnsigned(VarHandle target, final Class<?> adaptedType) {
Objects.requireNonNull(target);
Objects.requireNonNull(adaptedType);
final Class<?> carrier = target.varType();
checkWidenable(carrier);
checkNarrowable(adaptedType);
checkTargetWiderThanCarrier(carrier, adaptedType);
if (adaptedType == int.class && carrier == byte.class) {
return filterValue(target, INT_TO_BYTE, BYTE_TO_UNSIGNED_INT);
} else if (adaptedType == int.class && carrier == short.class) {
return filterValue(target, INT_TO_SHORT, SHORT_TO_UNSIGNED_INT);
} else if (adaptedType == long.class && carrier == byte.class) {
return filterValue(target, LONG_TO_BYTE, BYTE_TO_UNSIGNED_LONG);
} else if (adaptedType == long.class && carrier == short.class) {
return filterValue(target, LONG_TO_SHORT, SHORT_TO_UNSIGNED_LONG);
} else if (adaptedType == long.class && carrier == int.class) {
return filterValue(target, LONG_TO_INT, INT_TO_UNSIGNED_LONG);
} else {
throw new InternalError("should not reach here");
}
}
Adapts a target var handle by pre-processing incoming and outgoing values using a pair of filter functions.
When calling e.g. VarHandle.set(Object...)
on the resulting var handle, the incoming value (of type T
, where T
is the last parameter type of the first filter function) is processed using the first filter and then passed to the target var handle. Conversely, when calling e.g. VarHandle.get(Object...)
on the resulting var handle, the return value obtained from the target var handle (of type T
, where T
is the last parameter type of the second filter function) is processed using the second filter and returned to the caller. More advanced access mode types, such as AccessMode.COMPARE_AND_EXCHANGE
might apply both filters at the same time.
For the boxing and unboxing filters to be well formed, their types must be of the form (A... , S) -> T
and (A... , T) -> S
, respectively, where T
is the type of the target var handle. If this is the case, the resulting var handle will have type S
and will feature the additional coordinates A...
(which will be appended to the coordinates of the target var handle).
The resulting var handle will feature the same access modes (see AccessMode
and atomic access guarantees as those featured by the target var handle.
Params: - target – the target var handle
- filterToTarget – a filter to convert some type
S
into the type of target
- filterFromTarget – a filter to convert the type of
target
to some type S
Throws: - IllegalArgumentException – if
filterFromTarget
and filterToTarget
are not well-formed, that is, they have types other than (A... , S) -> T
and (A... , T) -> S
, respectively, where T
is the type of the target var handle, or if either filterFromTarget
or filterToTarget
throws any checked exceptions.
Returns: an adapter var handle which accepts a new type, performing the provided boxing/unboxing conversions.
/**
* Adapts a target var handle by pre-processing incoming and outgoing values using a pair of filter functions.
* <p>
* When calling e.g. {@link VarHandle#set(Object...)} on the resulting var handle, the incoming value (of type {@code T}, where
* {@code T} is the <em>last</em> parameter type of the first filter function) is processed using the first filter and then passed
* to the target var handle.
* Conversely, when calling e.g. {@link VarHandle#get(Object...)} on the resulting var handle, the return value obtained from
* the target var handle (of type {@code T}, where {@code T} is the <em>last</em> parameter type of the second filter function)
* is processed using the second filter and returned to the caller. More advanced access mode types, such as
* {@link java.lang.invoke.VarHandle.AccessMode#COMPARE_AND_EXCHANGE} might apply both filters at the same time.
* <p>
* For the boxing and unboxing filters to be well formed, their types must be of the form {@code (A... , S) -> T} and
* {@code (A... , T) -> S}, respectively, where {@code T} is the type of the target var handle. If this is the case,
* the resulting var handle will have type {@code S} and will feature the additional coordinates {@code A...} (which
* will be appended to the coordinates of the target var handle).
* <p>
* The resulting var handle will feature the same access modes (see {@link java.lang.invoke.VarHandle.AccessMode} and
* atomic access guarantees as those featured by the target var handle.
*
* @param target the target var handle
* @param filterToTarget a filter to convert some type {@code S} into the type of {@code target}
* @param filterFromTarget a filter to convert the type of {@code target} to some type {@code S}
* @return an adapter var handle which accepts a new type, performing the provided boxing/unboxing conversions.
* @throws IllegalArgumentException if {@code filterFromTarget} and {@code filterToTarget} are not well-formed, that is, they have types
* other than {@code (A... , S) -> T} and {@code (A... , T) -> S}, respectively, where {@code T} is the type of the target var handle,
* or if either {@code filterFromTarget} or {@code filterToTarget} throws any checked exceptions.
*/
public static VarHandle filterValue(VarHandle target, MethodHandle filterToTarget, MethodHandle filterFromTarget) {
return JLI.filterValue(target, filterToTarget, filterFromTarget);
}
Adapts a target var handle by pre-processing incoming coordinate values using unary filter functions.
When calling e.g. VarHandle.get(Object...)
on the resulting var handle, the incoming coordinate values starting at position pos
(of type C1, C2 ... Cn
, where C1, C2 ... Cn
are the return type of the unary filter functions) are transformed into new values (of type S1, S2 ... Sn
, where S1, S2 ... Sn
are the parameter types of the unary filter functions), and then passed (along with any coordinate that was left unaltered by the adaptation) to the target var handle.
For the coordinate filters to be well formed, their types must be of the form S1 -> T1, S2 -> T1 ... Sn -> Tn
, where T1, T2 ... Tn
are the coordinate types starting at position pos
of the target var handle.
The resulting var handle will feature the same access modes (see AccessMode
) and atomic access guarantees as those featured by the target var handle.
Params: - target – the target var handle
- pos – the position of the first coordinate to be transformed
- filters – the unary functions which are used to transform coordinates starting at position
pos
Throws: - IllegalArgumentException – if the handles in
filters
are not well-formed, that is, they have types other than S1 -> T1, S2 -> T2, ... Sn -> Tn
where T1, T2 ... Tn
are the coordinate types starting at position pos
of the target var handle, if pos
is not between 0 and the target var handle coordinate arity, inclusive, or if more filters are provided than the actual number of coordinate types available starting at pos
, or if any of the filters throws any checked exceptions.
Returns: an adapter var handle which accepts new coordinate types, applying the provided transformation
to the new coordinate values.
/**
* Adapts a target var handle by pre-processing incoming coordinate values using unary filter functions.
* <p>
* When calling e.g. {@link VarHandle#get(Object...)} on the resulting var handle, the incoming coordinate values
* starting at position {@code pos} (of type {@code C1, C2 ... Cn}, where {@code C1, C2 ... Cn} are the return type
* of the unary filter functions) are transformed into new values (of type {@code S1, S2 ... Sn}, where {@code S1, S2 ... Sn} are the
* parameter types of the unary filter functions), and then passed (along with any coordinate that was left unaltered
* by the adaptation) to the target var handle.
* <p>
* For the coordinate filters to be well formed, their types must be of the form {@code S1 -> T1, S2 -> T1 ... Sn -> Tn},
* where {@code T1, T2 ... Tn} are the coordinate types starting at position {@code pos} of the target var handle.
* <p>
* The resulting var handle will feature the same access modes (see {@link java.lang.invoke.VarHandle.AccessMode}) and
* atomic access guarantees as those featured by the target var handle.
*
* @param target the target var handle
* @param pos the position of the first coordinate to be transformed
* @param filters the unary functions which are used to transform coordinates starting at position {@code pos}
* @return an adapter var handle which accepts new coordinate types, applying the provided transformation
* to the new coordinate values.
* @throws IllegalArgumentException if the handles in {@code filters} are not well-formed, that is, they have types
* other than {@code S1 -> T1, S2 -> T2, ... Sn -> Tn} where {@code T1, T2 ... Tn} are the coordinate types starting
* at position {@code pos} of the target var handle, if {@code pos} is not between 0 and the target var handle coordinate arity, inclusive,
* or if more filters are provided than the actual number of coordinate types available starting at {@code pos},
* or if any of the filters throws any checked exceptions.
*/
public static VarHandle filterCoordinates(VarHandle target, int pos, MethodHandle... filters) {
return JLI.filterCoordinates(target, pos, filters);
}
Provides a target var handle with one or more bound coordinates
in advance of the var handle's invocation. As a consequence, the resulting var handle will feature less
coordinate types than the target var handle.
When calling e.g. VarHandle.get(Object...)
on the resulting var handle, incoming coordinate values are joined with bound coordinate values, and then passed to the target var handle.
For the bound coordinates to be well formed, their types must be T1, T2 ... Tn
, where T1, T2 ... Tn
are the coordinate types starting at position pos
of the target var handle.
The resulting var handle will feature the same access modes (see AccessMode
) and atomic access guarantees as those featured by the target var handle.
Params: - target – the var handle to invoke after the bound coordinates are inserted
- pos – the position of the first coordinate to be inserted
- values – the series of bound coordinates to insert
Throws: - IllegalArgumentException – if
pos
is not between 0 and the target var handle coordinate arity, inclusive, or if more values are provided than the actual number of coordinate types available starting at pos
. - ClassCastException – if the bound coordinates in
values
are not well-formed, that is, they have types other than T1, T2 ... Tn
, where T1, T2 ... Tn
are the coordinate types starting at position pos
of the target var handle.
Returns: an adapter var handle which inserts an additional coordinates,
before calling the target var handle
/**
* Provides a target var handle with one or more <em>bound coordinates</em>
* in advance of the var handle's invocation. As a consequence, the resulting var handle will feature less
* coordinate types than the target var handle.
* <p>
* When calling e.g. {@link VarHandle#get(Object...)} on the resulting var handle, incoming coordinate values
* are joined with bound coordinate values, and then passed to the target var handle.
* <p>
* For the bound coordinates to be well formed, their types must be {@code T1, T2 ... Tn },
* where {@code T1, T2 ... Tn} are the coordinate types starting at position {@code pos} of the target var handle.
* <p>
* The resulting var handle will feature the same access modes (see {@link java.lang.invoke.VarHandle.AccessMode}) and
* atomic access guarantees as those featured by the target var handle.
*
* @param target the var handle to invoke after the bound coordinates are inserted
* @param pos the position of the first coordinate to be inserted
* @param values the series of bound coordinates to insert
* @return an adapter var handle which inserts an additional coordinates,
* before calling the target var handle
* @throws IllegalArgumentException if {@code pos} is not between 0 and the target var handle coordinate arity, inclusive,
* or if more values are provided than the actual number of coordinate types available starting at {@code pos}.
* @throws ClassCastException if the bound coordinates in {@code values} are not well-formed, that is, they have types
* other than {@code T1, T2 ... Tn }, where {@code T1, T2 ... Tn} are the coordinate types starting at position {@code pos}
* of the target var handle.
*/
public static VarHandle insertCoordinates(VarHandle target, int pos, Object... values) {
return JLI.insertCoordinates(target, pos, values);
}
Provides a var handle which adapts the coordinate values of the target var handle, by re-arranging them
so that the new coordinates match the provided ones.
The given array controls the reordering. Call #I
the number of incoming coordinates (the value newCoordinates.size()
, and call #O
the number of outgoing coordinates (the number of coordinates associated with the target var handle). Then the length of the reordering array must be #O
, and each element must be a non-negative number less than #I
. For every N
less than #O
, the N
-th outgoing coordinate will be taken from the I
-th incoming coordinate, where I
is reorder[N]
.
No coordinate value conversions are applied. The type of each incoming coordinate, as determined by newCoordinates
, must be identical to the type of the corresponding outgoing coordinate in the target var handle.
The reordering array need not specify an actual permutation.
An incoming coordinate will be duplicated if its index appears
more than once in the array, and an incoming coordinate will be dropped
if its index does not appear in the array.
The resulting var handle will feature the same access modes (see AccessMode
) and atomic access guarantees as those featured by the target var handle.
Params: - target – the var handle to invoke after the coordinates have been reordered
- newCoordinates – the new coordinate types
- reorder – an index array which controls the reordering
Throws: - IllegalArgumentException – if the index array length is not equal to the number of coordinates of the target var handle, or if any index array element is not a valid index for a coordinate of
newCoordinates
, or if two corresponding coordinate types in the target var handle and in newCoordinates
are not identical.
Returns: an adapter var handle which re-arranges the incoming coordinate values,
before calling the target var handle
/**
* Provides a var handle which adapts the coordinate values of the target var handle, by re-arranging them
* so that the new coordinates match the provided ones.
* <p>
* The given array controls the reordering.
* Call {@code #I} the number of incoming coordinates (the value
* {@code newCoordinates.size()}, and call {@code #O} the number
* of outgoing coordinates (the number of coordinates associated with the target var handle).
* Then the length of the reordering array must be {@code #O},
* and each element must be a non-negative number less than {@code #I}.
* For every {@code N} less than {@code #O}, the {@code N}-th
* outgoing coordinate will be taken from the {@code I}-th incoming
* coordinate, where {@code I} is {@code reorder[N]}.
* <p>
* No coordinate value conversions are applied.
* The type of each incoming coordinate, as determined by {@code newCoordinates},
* must be identical to the type of the corresponding outgoing coordinate
* in the target var handle.
* <p>
* The reordering array need not specify an actual permutation.
* An incoming coordinate will be duplicated if its index appears
* more than once in the array, and an incoming coordinate will be dropped
* if its index does not appear in the array.
* <p>
* The resulting var handle will feature the same access modes (see {@link java.lang.invoke.VarHandle.AccessMode}) and
* atomic access guarantees as those featured by the target var handle.
* @param target the var handle to invoke after the coordinates have been reordered
* @param newCoordinates the new coordinate types
* @param reorder an index array which controls the reordering
* @return an adapter var handle which re-arranges the incoming coordinate values,
* before calling the target var handle
* @throws IllegalArgumentException if the index array length is not equal to
* the number of coordinates of the target var handle, or if any index array element is not a valid index for
* a coordinate of {@code newCoordinates}, or if two corresponding coordinate types in
* the target var handle and in {@code newCoordinates} are not identical.
*/
public static VarHandle permuteCoordinates(VarHandle target, List<Class<?>> newCoordinates, int... reorder) {
return JLI.permuteCoordinates(target, newCoordinates, reorder);
}
Adapts a target var handle handle by pre-processing
a sub-sequence of its coordinate values with a filter (a method handle).
The pre-processed coordinates are replaced by the result (if any) of the
filter function and the target var handle is then called on the modified (usually shortened)
coordinate list.
If R
is the return type of the filter (which cannot be void), the target var handle must accept a value of type R
as its coordinate in position pos
, preceded and/or followed by any coordinate not passed to the filter. No coordinates are reordered, and the result returned from the filter replaces (in order) the whole subsequence of coordinates originally passed to the adapter.
The argument types (if any) of the filter replace zero or one coordinate types of the target var handle, at position pos
, in the resulting adapted var handle. The return type of the filter must be identical to the coordinate type of the target var handle at position pos
, and that target var handle coordinate is supplied by the return value of the filter.
The resulting var handle will feature the same access modes (see AccessMode
) and atomic access guarantees as those featured by the target var handle.
Params: - target – the var handle to invoke after the coordinates have been filtered
- pos – the position of the coordinate to be filtered
- filter – the filter method handle
Throws: - IllegalArgumentException – if the return type of
filter
is void, or it is not the same as the pos
coordinate of the target var handle, if pos
is not between 0 and the target var handle coordinate arity, inclusive, if the resulting var handle's type would have too many coordinates, or if filter
throws any checked exceptions.
Returns: an adapter var handle which filters the incoming coordinate values,
before calling the target var handle
/**
* Adapts a target var handle handle by pre-processing
* a sub-sequence of its coordinate values with a filter (a method handle).
* The pre-processed coordinates are replaced by the result (if any) of the
* filter function and the target var handle is then called on the modified (usually shortened)
* coordinate list.
* <p>
* If {@code R} is the return type of the filter (which cannot be void), the target var handle must accept a value of
* type {@code R} as its coordinate in position {@code pos}, preceded and/or followed by
* any coordinate not passed to the filter.
* No coordinates are reordered, and the result returned from the filter
* replaces (in order) the whole subsequence of coordinates originally
* passed to the adapter.
* <p>
* The argument types (if any) of the filter
* replace zero or one coordinate types of the target var handle, at position {@code pos},
* in the resulting adapted var handle.
* The return type of the filter must be identical to the
* coordinate type of the target var handle at position {@code pos}, and that target var handle
* coordinate is supplied by the return value of the filter.
* <p>
* The resulting var handle will feature the same access modes (see {@link java.lang.invoke.VarHandle.AccessMode}) and
* atomic access guarantees as those featured by the target var handle.
*
* @param target the var handle to invoke after the coordinates have been filtered
* @param pos the position of the coordinate to be filtered
* @param filter the filter method handle
* @return an adapter var handle which filters the incoming coordinate values,
* before calling the target var handle
* @throws IllegalArgumentException if the return type of {@code filter}
* is void, or it is not the same as the {@code pos} coordinate of the target var handle,
* if {@code pos} is not between 0 and the target var handle coordinate arity, inclusive,
* if the resulting var handle's type would have <a href="MethodHandle.html#maxarity">too many coordinates</a>,
* or if {@code filter} throws any checked exceptions.
*/
public static VarHandle collectCoordinates(VarHandle target, int pos, MethodHandle filter) {
return JLI.collectCoordinates(target, pos, filter);
}
Returns a var handle which will discard some dummy coordinates before delegating to the
target var handle. As a consequence, the resulting var handle will feature more
coordinate types than the target var handle.
The pos
argument may range between zero and N, where N is the arity of the target var handle's coordinate types. If pos
is zero, the dummy coordinates will precede the target's real arguments; if pos
is N they will come after.
The resulting var handle will feature the same access modes (see AccessMode
) and atomic access guarantees as those featured by the target var handle.
Params: - target – the var handle to invoke after the dummy coordinates are dropped
- pos – position of first coordinate to drop (zero for the leftmost)
- valueTypes – the type(s) of the coordinate(s) to drop
Throws: - IllegalArgumentException – if
pos
is not between 0 and the target var handle coordinate arity, inclusive.
Returns: an adapter var handle which drops some dummy coordinates,
before calling the target var handle
/**
* Returns a var handle which will discard some dummy coordinates before delegating to the
* target var handle. As a consequence, the resulting var handle will feature more
* coordinate types than the target var handle.
* <p>
* The {@code pos} argument may range between zero and <i>N</i>, where <i>N</i> is the arity of the
* target var handle's coordinate types. If {@code pos} is zero, the dummy coordinates will precede
* the target's real arguments; if {@code pos} is <i>N</i> they will come after.
* <p>
* The resulting var handle will feature the same access modes (see {@link java.lang.invoke.VarHandle.AccessMode}) and
* atomic access guarantees as those featured by the target var handle.
*
* @param target the var handle to invoke after the dummy coordinates are dropped
* @param pos position of first coordinate to drop (zero for the leftmost)
* @param valueTypes the type(s) of the coordinate(s) to drop
* @return an adapter var handle which drops some dummy coordinates,
* before calling the target var handle
* @throws IllegalArgumentException if {@code pos} is not between 0 and the target var handle coordinate arity, inclusive.
*/
public static VarHandle dropCoordinates(VarHandle target, int pos, Class<?>... valueTypes) {
return JLI.dropCoordinates(target, pos, valueTypes);
}
private static void checkAddressFirstCoordinate(VarHandle handle) {
if (handle.coordinateTypes().size() < 1 ||
handle.coordinateTypes().get(0) != MemorySegment.class) {
throw new IllegalArgumentException("Expected var handle with leading coordinate of type MemorySegment");
}
}
private static void checkCarrier(Class<?> carrier) {
if (!carrier.isPrimitive() || carrier == void.class || carrier == boolean.class) {
throw new IllegalArgumentException("Illegal carrier: " + carrier.getSimpleName());
}
}
private static long carrierSize(Class<?> carrier) {
long bitsAlignment = Math.max(8, Wrapper.forPrimitiveType(carrier).bitWidth());
return Utils.bitsToBytesOrThrow(bitsAlignment, IllegalStateException::new);
}
private static void checkWidenable(Class<?> carrier) {
if (!(carrier == byte.class || carrier == short.class || carrier == int.class)) {
throw new IllegalArgumentException("illegal carrier:" + carrier.getSimpleName());
}
}
private static void checkNarrowable(Class<?> type) {
if (!(type == int.class || type == long.class)) {
throw new IllegalArgumentException("illegal adapter type: " + type.getSimpleName());
}
}
private static void checkTargetWiderThanCarrier(Class<?> carrier, Class<?> target) {
if (Wrapper.forPrimitiveType(target).bitWidth() <= Wrapper.forPrimitiveType(carrier).bitWidth()) {
throw new IllegalArgumentException(
target.getSimpleName() + " is not wider than: " + carrier.getSimpleName());
}
}
private static MemoryAddress longToAddress(long value) {
return MemoryAddress.ofLong(value);
}
private static long addressToLong(MemoryAddress value) {
return value.toRawLongValue();
}
}