Interface MemoryLayout

All Known Subinterfaces:: AddressLayout, GroupLayout, PaddingLayout, SequenceLayout, StructLayout, UnionLayout, ValueLayout, ValueLayout.OfBoolean, ValueLayout.OfByte, ValueLayout.OfChar, ValueLayout.OfDouble, ValueLayout.OfFloat, ValueLayout.OfInt, ValueLayout.OfLong, ValueLayout.OfShort

public sealed interface MemoryLayout permits SequenceLayout, GroupLayout, PaddingLayout, ValueLayout

A memory layout describes the contents of a memory segment.

There are two leaves in the layout hierarchy, value layouts, which are used to represent values of given size and kind and padding layouts which are used, as the name suggests, to represent a portion of a memory segment whose contents should be ignored, and which are primarily present for alignment reasons. Some common value layout constants, such as ValueLayout.JAVA_INT and ValueLayout.JAVA_FLOAT_UNALIGNED are defined in the ValueLayout class. A special kind of value layout, namely an address layout, is used to model values that denote the address of a region of memory.

More complex layouts can be derived from simpler ones: a sequence layout denotes a homogeneous repetition of zero or more occurrences of an element layout; a group layout denotes a heterogeneous aggregation of zero or more member layouts. Group layouts come in two flavors: struct layouts, where member layouts are laid out one after the other, and union layouts where member layouts are laid out at the same starting offset.

Layouts can be optionally associated with a name. A layout name can be referred to when constructing layout paths.

Consider the following struct declaration in C:

typedef struct {
    char kind;
    int value;
} TaggedValues[5];

The above declaration can be modeled using a layout object, as follows:

SequenceLayout TAGGED_VALUES = MemoryLayout.sequenceLayout(5,
    MemoryLayout.structLayout(
        ValueLayout.JAVA_BYTE.withName("kind"),
        MemoryLayout.paddingLayout(3),
        ValueLayout.JAVA_INT.withName("value")
    )
).withName("TaggedValues");

Characteristics of memory layouts

All layouts have a size (expressed in bytes), which is defined as follows:

The size of a value layout is determined by the ValueLayout.carrier() associated with the value layout. That is, the constant ValueLayout.JAVA_INT has carrier int, and size of 4 bytes;
The size of an address layout is platform-dependent. That is, the constant ValueLayout.ADDRESS has a size of 8 bytes on a 64-bit platform;
The size of a padding layout is always provided explicitly, on construction;
The size of a sequence layout whose element layout is E and element count is L, is the size of E, multiplied by L;
The size of a struct layout with member layouts M1, M2, ... Mn whose sizes are S1, S2, ... Sn, respectively, is S1 + S2 + ... + Sn;
The size of a union layout U with member layouts M1, M2, ... Mn whose sizes are S1, S2, ... Sn, respectively, is max(S1, S2, ... Sn).

Furthermore, all layouts have a natural alignment (expressed in bytes) which is defined as follows:

The natural alignment of a padding layout is 1;
The natural alignment of a value layout whose size is N is N;
The natural alignment of a sequence layout whose element layout is E is the alignment of E;
The natural alignment of a group layout with member layouts M1, M2, ... Mn whose alignments are A1, A2, ... An, respectively, is max(A1, A2 ... An).

A layout's alignment can be overridden if needed (see withByteAlignment(long)), which can be useful to describe layouts with weaker or stronger alignment constraints.

Layout paths

A layout path is used to unambiguously select a layout that is nested in some other layout. Layout paths are typically expressed as a sequence of one or more path elements. (A more formal definition of layout paths is provided below).

Layout paths can be used to:

obtain offsets of arbitrarily nested layouts;
obtain a var handle that can be used to access the value corresponding to the selected layout;
select an arbitrarily nested layout.

For instance, given the taggedValues sequence layout constructed above, we can obtain the offset, in bytes, of the member layout named value in the first sequence element, as follows:

long valueOffset = TAGGED_VALUES.byteOffset(PathElement.sequenceElement(0),
                                          PathElement.groupElement("value")); // yields 4

Similarly, we can select the member layout named value, as follows:

MemoryLayout value = TAGGED_VALUES.select(PathElement.sequenceElement(),
                                         PathElement.groupElement("value"));

Open path elements

Some layout path elements, said open path elements, can select multiple layouts at once. For instance, the open path elements MemoryLayout.PathElement.sequenceElement(), MemoryLayout.PathElement.sequenceElement(long, long) select an unspecified element in a sequence layout. A var handle derived from a layout path containing one or more open path element features additional coordinates of type long, which can be used by clients to bind the open elements in the path:

VarHandle valueHandle = TAGGED_VALUES.varHandle(PathElement.sequenceElement(),
                                               PathElement.groupElement("value"));
MemorySegment taggedValues = ...
// reads the "value" field of the third struct in the array (taggedValues[2].value)
int val = (int) valueHandle.get(taggedValues,
        0L,  // base offset
        2L); // sequence index

Open path elements also affect the creation of offset-computing method handles. Each open path element becomes an additional long parameter in the obtained method handle. This parameter can be used to specify the index of the sequence element whose offset is to be computed:

MethodHandle offsetHandle = TAGGED_VALUES.byteOffsetHandle(PathElement.sequenceElement(),
                                                          PathElement.groupElement("kind"));
long offset1 = (long) offsetHandle.invokeExact(0L, 1L); // 0 + (1 * 8) = 8
long offset2 = (long) offsetHandle.invokeExact(0L, 2L); // 0 + (2 * 8) = 16

Dereference path elements

A special kind of path element, called dereference path element, allows var handles obtained from memory layouts to follow pointers. Consider the following layout:

StructLayout RECTANGLE = MemoryLayout.structLayout(
        ValueLayout.ADDRESS.withTargetLayout(
                MemoryLayout.sequenceLayout(4,
                        MemoryLayout.structLayout(
                                ValueLayout.JAVA_INT.withName("x"),
                                ValueLayout.JAVA_INT.withName("y")
                        ).withName("point")
                 )
         ).withName("points")
);

This layout is a struct layout describing a rectangle. It contains a single field, namely points, an address layout whose target layout is a sequence layout of four struct layouts. Each struct layout describes a two-dimensional point, and is defined as a pair or ValueLayout.JAVA_INT coordinates, with names x and y, respectively.

With dereference path elements, we can obtain a var handle that accesses the y coordinate of one of the point in the rectangle, as follows:

VarHandle rectPointYs = RECTANGLE.varHandle(
        PathElement.groupElement("points"),
        PathElement.dereferenceElement(),
        PathElement.sequenceElement(),
        PathElement.groupElement("y")
);

MemorySegment rect = ...
// dereferences the third point struct in the "points" array, and reads its "y" coordinate (rect.points[2]->y)
int rect_y_2 = (int) rectPointYs.get(rect,
    0L,  // base offset
    2L); // sequence index

Layout path well-formedness

A layout path is applied to a layout C_0, also called the initial layout. Each path element in a layout path can be thought of as a function that updates the current layout C_i-1 to some other layout C_i. That is, for each path element E1, E2, ... En, in a layout path P, we compute C_i = f_i(C_i-1), where f_i is the selection function associated with the path element under consideration, denoted as E_i. The final layout C_i is also called the selected layout.

A layout path P is considered well-formed for an initial layout C_0 if all its path elements E1, E2, ... En are well-formed for their corresponding input layouts C_0, C_1, ... C_n-1. A path element E is considered well-formed for a layout L if any of the following is true:

L is a sequence layout and E is a sequence path element (one of MemoryLayout.PathElement.sequenceElement(long), MemoryLayout.PathElement.sequenceElement(long, long) or MemoryLayout.PathElement.sequenceElement()). Moreover, if E contains one or more sequence indices, such indices have to be compatible with the sequence layout's element count;
L is a group layout and E is a group path element (one of MemoryLayout.PathElement.groupElement(String) or MemoryLayout.PathElement.groupElement(long)). Moreover, the group path element must refer to a valid member layout in L, either by name, or index;
L is an address layout and E is a dereference path element. Moreover, L must define some target layout.

Any attempt to provide a layout path P that is not well-formed for an initial layout C_0 will result in an IllegalArgumentException.

Access mode restrictions

A var handle returned by varHandle(PathElement...) or ValueLayout.varHandle() features certain access characteristics, which are derived from the selected layout L:

A carrier type T, derived from L.carrier()
An alignment constraint A, derived from L.byteAlignment()
An access size S, derived from L.byteSize()

Depending on the above characteristics, the returned var handle might feature certain access mode restrictions. We say that a var handle is aligned if its alignment constraint A is compatible with the access size S, that is if A >= S. An aligned var handle is guaranteed to support the following access modes:

read write access modes for all T. Access modes get and set for long, double and MemorySegment are supported but have no atomicity guarantee, as described in Section Â§17.7 of The Java Language Specification.
atomic update access modes for int, long, float, double and MemorySegment. (Future major platform releases of the JDK may support additional types for certain currently unsupported access modes.)
numeric atomic update access modes for int, long and MemorySegment. (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, long and MemorySegment. (Future major platform releases of the JDK may support additional numeric types for certain currently unsupported access modes.)

If T is float, double or MemorySegment then atomic update access modes compare values using their bitwise representation (see Float.floatToRawIntBits(float), Double.doubleToRawLongBits(double) and MemorySegment.address(), respectively).

Alternatively, a var handle is unaligned if its alignment constraint A is incompatible with the access size S, that is, if A < S. An unaligned var handle only supports the get and set access modes. All other access modes will result in UnsupportedOperationException being thrown. Moreover, while supported, access modes get and set might lead to word tearing.

Working with variable-length arrays

We have seen how sequence layouts are used to describe the contents of an array whose size is known statically. There are cases, however, where the array size is only known dynamically. We call such arrays variable-length arrays. There are two common kinds of variable-length arrays:

a toplevel variable-length array whose size depends on the value of some unrelated variable, or parameter;
an variable-length array nested in a struct, whose size depends on the value of some other field in the enclosing struct.

While variable-length arrays cannot be modeled directly using sequence layouts, clients can still enjoy structured access to elements of variable-length arrays using var handles as demonstrated in the following sections.

Toplevel variable-length arrays

Consider the following struct declaration in C:

typedef struct {
    int x;
    int y;
} Point;

In the above code, a point is modeled as two coordinates (x and y respectively). Now consider the following snippet of C code:

int size = ...
Point *points = (Point*)malloc(sizeof(Point) * size);
for (int i = 0 ; i < size ; i++) {
   ... points[i].x ...
}

Here, we allocate an array of points (points). Crucially, the size of the array is dynamically bound to the value of the size variable. Inside the loop, the x coordinate of all the points in the array is accessed.

To model this code in Java, let's start by defining a layout for the Point struct, as follows:

StructLayout POINT = MemoryLayout.structLayout(
            ValueLayout.JAVA_INT.withName("x"),
            ValueLayout.JAVA_INT.withName("y")
);

Since we know we need to create and access an array of points, it would be tempting to create a sequence layout modelling the variable-length array, and then derive the necessary access var handles from the sequence layout. But this approach is problematic, as the size of the variable-length array is not known. Instead, a var handle that provides structured access to the elements of a variable-length array can be obtained directly from the layout describing the array elements (e.g. the point layout), as demonstrated below:

VarHandle POINT_ARR_X = POINT.arrayElementVarHandle(PathElement.groupElement("x"));

int size = ...
MemorySegment points = ...
for (int i = 0 ; i < size ; i++) {
    ... POINT_ARR_X.get(points, 0L, (long)i) ...
}

Here, the coordinate x of subsequent point in the array is accessed using the POINT_ARR_X var handle, which is obtained using the arrayElementVarHandle(PathElement...) method. This var handle features two long coordinates: the first is a base offset (set to 0L), while the second is a logical index that can be used to stride over all the elements of the point array.

The base offset coordinate allows clients to express complex access operations, by injecting additional offset computation into the var handle (we will see an example of that below). In cases where the base offset is constant (as in the previous example) clients can, if desired, drop the base offset parameter and make the access expression simpler. This is achieved using the MethodHandles.insertCoordinates(VarHandle, int, Object...) var handle adapter.

Nested variable-length arrays

Consider the following struct declaration in C:

typedef struct {
    int size;
    Point points[];
} Polygon;

In the above code, a polygon is modeled as a size (the number of edges in the polygon) and an array of points (one for each vertex in the polygon). The number of vertices depends on the number of edges in the polygon. As such, the size of the points array is left unspecified in the C declaration, using a Flexible Array Member (a feature standardized in C99).

Again, clients can perform structured access to elements in the nested variable-length array using the arrayElementVarHandle(PathElement...) method, as demonstrated below:

StructLayout POLYGON = MemoryLayout.structLayout(
            ValueLayout.JAVA_INT.withName("size"),
            MemoryLayout.sequenceLayout(0, POINT).withName("points")
);

VarHandle POLYGON_SIZE = POLYGON.varHandle(0, PathElement.groupElement("size"));
long POINTS_OFFSET = POLYGON.byteOffset(PathElement.groupElement("points"));

The POLYGON layout contains a sequence layout of size zero. The element layout of the sequence layout is the POINT layout, shown previously. The polygon layout is used to obtain a var handle that provides access to the polygon size, as well as an offset (POINTS_OFFSET) to the start of the variable-length points array.

The x coordinates of all the points in a polygon can then be accessed as follows:

MemorySegment polygon = ...
int size = POLYGON_SIZE.get(polygon, 0L);
for (int i = 0 ; i < size ; i++) {
    ... POINT_ARR_X.get(polygon, POINTS_OFFSET, (long)i) ...
}

Here, we first obtain the polygon size, using the POLYGON_SIZE var handle. Then, in a loop, we read the x coordinates of all the points in the polygon. This is done by providing a custom offset (namely, POINTS_OFFSET) to the offset coordinate of the POINT_ARR_X var handle. As before, the loop induction variable i is passed as the index of the POINT_ARR_X var handle, to stride over all the elements of the variable-length array.

Implementation Requirements:: Implementations of this interface are immutable, thread-safe and value-based.
Sealed Class Hierarchy Graph:
Since:: 22

Nested Class Summary

Modifier and Type	Interface	Description
`static interface`	`MemoryLayout.PathElement`	An element in a layout path.

Method Summary

Modifier and Type	Method	Description
`VarHandle`	`arrayElementVarHandle(MemoryLayout.PathElement... elements)`	Creates a var handle that accesses adjacent elements in a memory segment at offsets selected by the given layout path, where the accessed elements have this layout, and where the initial layout in the path is this layout.
`long`	`byteAlignment()`	Returns the alignment constraint associated with this layout, expressed in bytes.
`long`	`byteOffset(MemoryLayout.PathElement... elements)`	Computes the offset, in bytes, of the layout selected by the given layout path, where the initial layout in the path is this layout.
`MethodHandle`	`byteOffsetHandle(MemoryLayout.PathElement... elements)`	Creates a method handle that computes the offset, in bytes, of the layout selected by the given layout path, where the initial layout in the path is this layout.
`long`	`byteSize()`	Returns the layout size, in bytes.
`boolean`	`equals(Object other)`	Compares the specified object with this layout for equality.
`int`	`hashCode()`	Returns the hash code value for this layout.
`Optional<String>`	`name()`	Returns the name (if any) associated with this layout.
`static PaddingLayout`	`paddingLayout(long byteSize)`	Creates a padding layout with the given byte size.
`long`	`scale(long offset, long index)`	Returns `offset + (byteSize() * index)`.
`MethodHandle`	`scaleHandle()`	Returns a method handle that can be used to invoke `scale(long, long)` on this layout.
`MemoryLayout`	`select(MemoryLayout.PathElement... elements)`	Returns the layout selected from the provided path, where the initial layout in the path is this layout.
`static SequenceLayout`	`sequenceLayout(long elementCount, MemoryLayout elementLayout)`	Creates a sequence layout with the given element layout and element count.
`MethodHandle`	`sliceHandle(MemoryLayout.PathElement... elements)`	Creates a method handle which, given a memory segment, returns a slice corresponding to the layout selected by the given layout path, where the initial layout in the path is this layout.
`static StructLayout`	`structLayout(MemoryLayout... elements)`	Creates a struct layout with the given member layouts.
`String`	`toString()`	Returns the string representation of this layout.
`static UnionLayout`	`unionLayout(MemoryLayout... elements)`	Creates a union layout with the given member layouts.
`VarHandle`	`varHandle(MemoryLayout.PathElement... elements)`	Creates a var handle that accesses a memory segment at the offset selected by the given layout path, where the initial layout in the path is this layout.
`MemoryLayout`	`withByteAlignment(long byteAlignment)`	Returns a memory layout with the same characteristics as this layout, but with the given alignment constraint (in bytes).
`MemoryLayout`	`withName(String name)`	Returns a memory layout with the same characteristics as this layout, but with the given name.
`MemoryLayout`	`withoutName()`	Returns a memory layout with the same characteristics as this layout, but with no name.

Method Details

byteSize

long byteSize()

Returns the layout size, in bytes.

Returns:: the layout size, in bytes

name

Optional<String> name()

Returns the name (if any) associated with this layout.

Returns:

the name (if any) associated with this layout

See Also:

withName

MemoryLayout withName(String name)

Returns a memory layout with the same characteristics as this layout, but with the given name.

Parameters:

name - the layout name

Returns:

a memory layout with the same characteristics as this layout, but with the given name

See Also:

withoutName

MemoryLayout withoutName()

Returns a memory layout with the same characteristics as this layout, but with no name.

API Note:

This can be useful to compare two layouts that have different names, but are otherwise equal.

Returns:

a memory layout with the same characteristics as this layout, but with no name

See Also:

byteAlignment

long byteAlignment()

Returns the alignment constraint associated with this layout, expressed in bytes.

Layout alignment defines a power of two A which is the byte-wise alignment of the layout, where A is the number of bytes that must be aligned for any pointer that correctly points to this layout. Thus:

A=1 means unaligned (in the usual sense), which is common in packets.
A=8 means word aligned (on LP64), A=4 int aligned, A=2 short aligned, etc.
A=64 is the most strict alignment required by the x86/SV ABI (for AVX-512 data).

If no explicit alignment constraint was set on this layout ( see withByteAlignment(long)), then this method returns the natural alignment constraint (in bytes) associated with this layout.

Returns:: the alignment constraint associated with this layout, expressed in bytes

withByteAlignment

MemoryLayout withByteAlignment(long byteAlignment)

Returns a memory layout with the same characteristics as this layout, but with the given alignment constraint (in bytes).

Parameters:: byteAlignment - the layout alignment constraint, expressed in bytes
Returns:: a memory layout with the same characteristics as this layout, but with the given alignment constraint (in bytes)
Throws:: IllegalArgumentException - if byteAlignment is not a power of two

scale

long scale(long offset, long index)

Returns offset + (byteSize() * index).

Parameters:: offset - the base offset; index - the index to be scaled by the byte size of this layout
Returns:: offset + (byteSize() * index)
Throws:: IllegalArgumentException - if offset or index is negative; ArithmeticException - if either the addition or multiplication overflows

scaleHandle

MethodHandle scaleHandle()

Returns a method handle that can be used to invoke scale(long, long) on this layout.

Returns:: a method handle that can be used to invoke scale(long, long) on this layout

byteOffset

long byteOffset(MemoryLayout.PathElement... elements)

Computes the offset, in bytes, of the layout selected by the given layout path, where the initial layout in the path is this layout.

Parameters:: elements - the layout path elements
Returns:: The offset, in bytes, of the layout selected by the layout path in elements
Throws:: IllegalArgumentException - if the layout path is not well-formed for this layout; IllegalArgumentException - if the layout path contains one or more open path elements; IllegalArgumentException - if the layout path contains one or more dereference path elements

byteOffsetHandle

MethodHandle byteOffsetHandle(MemoryLayout.PathElement... elements)

Creates a method handle that computes the offset, in bytes, of the layout selected by the given layout path, where the initial layout in the path is this layout.

The returned method handle has the following characteristics:

its return type is long;
it has one leading long parameter representing the base offset;
it has as zero or more trailing parameters of type long, one for each open path element in the provided layout path. The order of these parameters corresponds to the order in which the open path elements occur in the provided layout path.

The final offset returned by the method handle is computed as follows:

offset = b + c_1 + c_2 + ... + c_m + (x_1 * s_1) + (x_2 * s_2) + ... + (x_n * s_n)

where b represents the base offset provided as a dynamic long argument, x_1, x_2, ... x_n represent indices into sequences provided as dynamic long arguments, whereas s_1, s_2, ... s_n are static stride constants derived from the size of the element layout of a sequence, and c_1, c_2, ... c_m are other static offset constants (such as field offsets) which are derived from the layout path.

For any given dynamic argument x_i, it must be that 0 <= x_i < size_i, where size_i is the size of the open path element associated with x_i. Otherwise, the returned method handle throws IndexOutOfBoundsException. Moreover, the value of b must be such that the computation for offset does not overflow, or the returned method handle throws ArithmeticException.

API Note:: The returned method handle can be used to compute a layout offset, similarly to byteOffset(PathElement...), but more flexibly, as some indices can be specified when invoking the method handle.
Parameters:: elements - the layout path elements
Returns:: a method handle that computes the offset, in bytes, of the layout selected by the given layout path
Throws:: IllegalArgumentException - if the layout path is not well-formed for this layout; IllegalArgumentException - if the layout path contains one or more dereference path elements

varHandle

VarHandle varHandle(MemoryLayout.PathElement... elements)

Creates a var handle that accesses a memory segment at the offset selected by the given layout path, where the initial layout in the path is this layout.

The returned var handle has the following characteristics:

its type is derived from the carrier of the selected value layout;
it has a leading parameter of type MemorySegment representing the accessed segment
a following long parameter, corresponding to the base offset, denoted as B;
it has zero or more trailing access coordinates of type long, one for each open path element in the provided layout path, denoted as I1, I2, ... In, respectively. The order of these access coordinates corresponds to the order in which the open path elements occur in the provided layout path.

If the provided layout path P contains no dereference elements, then the offset O of the access operation is computed as follows:

O = this.byteOffsetHandle(P).invokeExact(B, I1, I2, ... In);

Accessing a memory segment using the var handle returned by this method is subject to the following checks:

The physical address of the accessed memory segment must be aligned according to the alignment constraint of the root layout (this layout), or an IllegalArgumentException is thrown. Note that the alignment constraint of the root layout can be more strict (but not less) than the alignment constraint of the selected value layout.
The access operation must fall inside the spatial bounds of the accessed memory segment, or an IndexOutOfBoundsException is thrown. This is the case when B + A <= S, where B is the base offset (defined above), A is the size of this layout and S is the size of the accessed memory segment. Note that the size of this layout might be bigger than the size of the accessed layout (e.g. when accessing a struct member).
If the provided layout path has an open path element whose size is S, its corresponding trailing long coordinate value I must be 0 <= I < S, or an IndexOutOfBoundsException is thrown.
The accessed memory segment must be accessible from the thread performing the access operation, or a WrongThreadException is thrown.
For write operations, the accessed memory segment must not be read only, or an IllegalArgumentException is thrown.
The scope associated with the accessed segment must be alive, or an IllegalStateException is thrown.

If the selected layout is an address layout, calling VarHandle.get(Object...) on the returned var handle will return a new memory segment. The segment is associated with the global scope. Moreover, the size of the segment depends on whether the address layout has a target layout. More specifically:

If the address layout has a target layout T, then the size of the returned segment is T.byteSize();
Otherwise, the address layout has no target layout and the size of the returned segment is zero.

Moreover, if the selected layout is an address layout, calling VarHandle.set(Object...) can throw IllegalArgumentException if the memory segment representing the address to be written is not a native memory segment.

If the provided layout path has size m and contains a dereference path element in position k (where k <= m) then two layout paths P and Q are derived, where P contains all the path elements from 0 to k - 1 and Q contains all the path elements from k + 1 to m (Q could be an empty layout path if k == m). Then, the returned var handle is computed as follows:

VarHandle baseHandle = this.varHandle(P);
MemoryLayout target = ((AddressLayout)this.select(P)).targetLayout().get();
VarHandle targetHandle = target.varHandle(Q);
targetHandle = MethodHandles.insertCoordinates(targetHandle, 1, 0L); // always access nested targets at offset 0
targetHandle = MethodHandles.collectCoordinates(targetHandle, 0,
        baseHandle.toMethodHandle(VarHandle.AccessMode.GET));

(The above can be trivially generalized to cases where the provided layout path contains more than one dereference path elements).

As an example, consider the memory layout expressed by a GroupLayout instance constructed as follows:

    GroupLayout grp = java.lang.foreign.MemoryLayout.structLayout(
            MemoryLayout.paddingLayout(4),
            ValueLayout.JAVA_INT.withOrder(ByteOrder.BIG_ENDIAN).withName("value")
    );

To access the member layout named value, we can construct a var handle as follows:

    VarHandle handle = grp.varHandle(PathElement.groupElement("value")); //(MemorySegment, long) -> int

API Note:: The resulting var handle features certain access mode restrictions, which are common to all var handles derived from memory layouts.
Parameters:: elements - the layout path elements
Returns:: a var handle that accesses a memory segment at the offset selected by the given layout path
Throws:: IllegalArgumentException - if the layout path is not well-formed for this layout; IllegalArgumentException - if the layout selected by the provided path is not a value layout

arrayElementVarHandle

VarHandle arrayElementVarHandle(MemoryLayout.PathElement... elements)

Creates a var handle that accesses adjacent elements in a memory segment at offsets selected by the given layout path, where the accessed elements have this layout, and where the initial layout in the path is this layout.

The returned var handle has the following characteristics:

its type is derived from the carrier of the selected value layout;
it has a leading parameter of type MemorySegment representing the accessed segment
a following long parameter, corresponding to the base offset, denoted as B;
a following long parameter, corresponding to the array index, denoted as I0. The array index is used to scale the accessed offset by this layout size;
it has zero or more trailing access coordinates of type long, one for each open path element in the provided layout path, denoted as I1, I2, ... In, respectively. The order of these access coordinates corresponds to the order in which the open path elements occur in the provided layout path.

If the provided layout path P contains no dereference elements, then the offset O of the access operation is computed as follows:

O = this.byteOffsetHandle(P).invokeExact(this.scale(B, I0), I1, I2, ... In);

More formally, the method handle returned by this method is obtained from varHandle(PathElement...), as follows:

MethodHandles.collectCoordinates(varHandle(elements), 1, scaleHandle())

Accessing a memory segment using the var handle returned by this method is subject to the following checks:

The physical address of the accessed memory segment must be aligned according to the alignment constraint of the root layout (this layout), or an IllegalArgumentException is thrown. Note that the alignment constraint of the root layout can be more strict (but not less) than the alignment constraint of the selected value layout.
The access operation must fall inside the spatial bounds of the accessed memory segment, or an IndexOutOfBoundsException is thrown. This is the case when B + A <= S, where B is the base offset (defined above), A is the size of this layout and S is the size of the accessed memory segment. Note that the size of this layout might be bigger than the size of the accessed layout (e.g. when accessing a struct member).
If the provided layout path has an open path element whose size is S, its corresponding trailing long coordinate value I must be 0 <= I < S, or an IndexOutOfBoundsException is thrown.
The accessed memory segment must be accessible from the thread performing the access operation, or a WrongThreadException is thrown.
For write operations, the accessed memory segment must not be read only, or an IllegalArgumentException is thrown.
The scope associated with the accessed segment must be alive, or an IllegalStateException is thrown.

API Note:: As the leading index coordinate I0 is not bound by any sequence layout, it can assume any non-negative value - provided that the resulting offset computation does not overflow, or that the computed offset does not fall outside the spatial bound of the accessed memory segment. As such, the var handles returned from this method can be especially useful when accessing variable-length arrays.
Parameters:: elements - the layout path elements
Returns:: a var handle that accesses adjacent elements in a memory segment at offsets selected by the given layout path
Throws:: IllegalArgumentException - if the layout path is not well-formed for this layout; IllegalArgumentException - if the layout selected by the provided path is not a value layout

sliceHandle

MethodHandle sliceHandle(MemoryLayout.PathElement... elements)

Creates a method handle which, given a memory segment, returns a slice corresponding to the layout selected by the given layout path, where the initial layout in the path is this layout.

The returned method handle has the following characteristics:

its return type is MemorySegment;
it has a leading parameter of type MemorySegment corresponding to the memory segment to be sliced
a following long parameter, corresponding to the base offset
it has as zero or more trailing parameters of type long, one for each open path element in the provided layout path. The order of these parameters corresponds to the order in which the open path elements occur in the provided layout path.

The offset O of the returned segment is computed as if by a call to a byte offset handle constructed using the given path elements.

Computing a slice of a memory segment using the method handle returned by this method is subject to the following checks:

The physical address of the accessed memory segment must be aligned according to the alignment constraint of the root layout (this layout), or an IllegalArgumentException will be issued. Note that the alignment constraint of the root layout can be more strict (but not less) than the alignment constraint of the selected layout.
The slicing operation must fall inside the spatial bounds of the accessed memory segment, or an IndexOutOfBoundsException is thrown. This is the case when B + A <= S, where B is the base offset (defined above), A is the size of this layout and S is the size of the accessed memory segment. Note that the size of this layout might be bigger than the size of the accessed layout (e.g. when accessing a struct member).
If the provided layout path has an open path element whose size is S, its corresponding trailing long coordinate value I must be 0 <= I < S, or an IndexOutOfBoundsException is thrown.

API Note:: The returned method handle can be used to obtain a memory segment slice, similarly to MemorySegment.asSlice(long, long), but more flexibly, as some indices can be specified when invoking the method handle.
Parameters:: elements - the layout path elements
Returns:: a method handle that is used to slice a memory segment at the offset selected by the given layout path
Throws:: IllegalArgumentException - if the layout path is not well-formed for this layout; IllegalArgumentException - if the layout path contains one or more dereference path elements

select

MemoryLayout select(MemoryLayout.PathElement... elements)

Returns the layout selected from the provided path, where the initial layout in the path is this layout.

Parameters:: elements - the layout path elements
Returns:: the layout selected by the layout path in elements
Throws:: IllegalArgumentException - if the layout path is not well-formed for this layout; IllegalArgumentException - if the layout path contains one or more dereference path elements; IllegalArgumentException - if the layout path contains one or more path elements that select one or more sequence element indices, such as MemoryLayout.PathElement.sequenceElement(long) and MemoryLayout.PathElement.sequenceElement(long, long))

equals

boolean equals(Object other)

Compares the specified object with this layout for equality. Returns true if and only if the specified object is also a layout, and it is equal to this layout. Two layouts are considered equal if they are of the same kind, have the same size, name and alignment constraint. Furthermore, depending on the layout kind, additional conditions must be satisfied:

two value layouts are considered equal if they have the same order, and carrier. Additionally, two address layouts are considered equal if they also have the same target layout;
two sequence layouts are considered equal if they have the same element count (see SequenceLayout.elementCount()), and if their element layouts (see SequenceLayout.elementLayout()) are also equal;
two group layouts are considered equal if they are of the same type (see StructLayout, UnionLayout) and if their member layouts (see GroupLayout.memberLayouts()) are also equal.

Overrides:

equals in class Object

Parameters:

other - the object to be compared for equality with this layout

Returns:

true if the specified object is equal to this layout

See Also:

hashCode

int hashCode()

Returns the hash code value for this layout.

Overrides:

hashCode in class Object

Returns:

the hash code value for this layout

See Also:

toString

String toString()

Returns the string representation of this layout.

Overrides:: toString in class Object
Returns:: the string representation of this layout

paddingLayout

static PaddingLayout paddingLayout(long byteSize)

Creates a padding layout with the given byte size. The alignment constraint of the returned layout is 1. As such, regardless of its size, in the absence of an explicit alignment constraint, a padding layout does not affect the natural alignment of the group or sequence layout it is nested into.

Parameters:: byteSize - the padding size (expressed in bytes)
Returns:: the new selector layout
Throws:: IllegalArgumentException - if byteSize <= 0

sequenceLayout

static SequenceLayout sequenceLayout(long elementCount, MemoryLayout elementLayout)

Creates a sequence layout with the given element layout and element count.

Parameters:: elementCount - the sequence element count; elementLayout - the sequence element layout
Returns:: the new sequence layout with the given element layout and size
Throws:: IllegalArgumentException - if elementCount is negative; IllegalArgumentException - if elementLayout.byteSize() * elementCount overflows; IllegalArgumentException - if elementLayout.byteSize() % elementLayout.byteAlignment() != 0

structLayout

static StructLayout structLayout(MemoryLayout... elements)

Creates a struct layout with the given member layouts.

API Note:

This factory does not automatically align element layouts, by inserting additional padding layout elements. As such, the following struct layout creation will fail with an exception:

structLayout(JAVA_SHORT, JAVA_INT);

To avoid the exception, clients can either insert additional padding layout elements:

structLayout(JAVA_SHORT, MemoryLayout.paddingLayout(2), JAVA_INT);

Or, alternatively, they can use a member layout that features a smaller alignment constraint. This will result in a packed struct layout:

structLayout(JAVA_SHORT, JAVA_INT.withByteAlignment(2));

Parameters:

elements - The member layouts of the struct layout

Returns:

a struct layout with the given member layouts

Throws:

IllegalArgumentException - if the sum of the byte sizes of the member layouts overflows

IllegalArgumentException - if a member layout in elements occurs at an offset (relative to the start of the struct layout) which is not compatible with its alignment constraint

unionLayout

static UnionLayout unionLayout(MemoryLayout... elements)

Creates a union layout with the given member layouts.

Parameters:: elements - The member layouts of the union layout
Returns:: a union layout with the given member layouts

© 1993, 2025, Oracle and/or its affiliates. All rights reserved.
Documentation extracted from Debian's OpenJDK Development Kit package.
Licensed under the GNU General Public License, version 2, with the Classpath Exception.
Various third party code in OpenJDK is licensed under different licenses (see Debian package).
Java and OpenJDK are trademarks or registered trademarks of Oracle and/or its affiliates.