Memory layout of Julia Objects

    Each jl_value_t struct is contained in a jl_typetag_t struct that contains metadata information about the Julia object, such as its type and garbage collector (gc) reachability:

    1. typedef struct {
    2.     opaque metadata;
    3.     jl_value_t value;
    4. } jl_typetag_t;

    The type of any Julia object is an instance of a leaf jl_datatype_t object. The jl_typeof() function can be used to query for it:

    1. jl_value_t *jl_typeof(jl_value_t *v);

    The layout of the object depends on its type. Reflection methods can be used to inspect that layout. A field can be accessed by calling one of the get-field methods:

    1. jl_value_t *jl_get_nth_field_checked(jl_value_t *v, size_t i);
    2. jl_value_t *jl_get_field(jl_value_t *o, char *fld);

    If the field types are known, a priori, to be all pointers, the values can also be extracted directly as an array access:

    1. jl_value_t *v = value->fieldptr[n];

    As an example, a “boxed” uint16_t is stored as follows:

    1. struct {
    2.     struct {
    3.         uint16_t data;        // -- 2 bytes
    4.     } jl_value_t;

    This object is created by jl_box_uint16(). Note that the jl_value_t pointer references the data portion, not the metadata at the top of the struct.

    A value may be stored “unboxed” in many circumstances (just the data, without the metadata, and possibly not even stored but just kept in registers), so it is unsafe to assume that the address of a box is a unique identifier. The “egal” test (corresponding to the === function in Julia), should instead be used to compare two unknown objects for equivalence:

    This optimization should be relatively transparent to the API, since the object will be “boxed” on-demand, whenever a jl_value_t pointer is needed.

    Note that modification of a jl_value_t pointer in memory is permitted only if the object is mutable. Otherwise, modification of the value may corrupt the program and the result will be undefined. The mutability property of a value can be queried for with:

    1. int jl_is_mutable(jl_value_t *v);
    1. void jl_gc_wb(jl_value_t *parent, jl_value_t *ptr);

    However, the Embedding Julia section of the manual is also required reading at this point, for covering other details of boxing and unboxing various types, and understanding the gc interactions.

    Mirror structs for some of the built-in types are . The corresponding global jl_datatype_t objects are created by jl_init_types in jltypes.c.

    The garbage collector uses several bits from the metadata portion of the jl_typetag_t to track each object in the system. Further details about this algorithm can be found in the comments of the garbage collector implementation in gc.c.

    Most new objects are allocated by jl_new_structv():

    1. jl_value_t *jl_new_struct(jl_datatype_t *type, ...);
    2. jl_value_t *jl_new_structv(jl_datatype_t *type, jl_value_t **args, uint32_t na);

    Although, isbits objects can be also constructed directly from memory:

    1. jl_value_t *jl_new_bits(jl_value_t *bt, void *data)

    And some objects have special constructors that must be used instead of the above functions:

    Types:

    1. jl_datatype_t *jl_apply_type(jl_datatype_t *tc, jl_tuple_t *params);
    2. jl_datatype_t *jl_apply_array_type(jl_datatype_t *type, size_t dim);

    While these are the most commonly used options, there are more low-level constructors too, which you can find declared in . These are used in jl_init_types() to create the initial types needed to bootstrap the creation of the Julia system image.

    Tuples:

    1. typedef struct {
    2.     jl_value_t *data[length];
    3. } jl_tuple_t;

    However, in other cases, the tuple may be converted to an anonymous isbits type and stored unboxed, or it may not stored at all (if it is not being used in a generic context as a jl_value_t*).

    Symbols:

    1. jl_sym_t *jl_symbol(const char *str);

    Functions and MethodInstance:

    1. jl_function_t *jl_new_generic_function(jl_sym_t *name);
    2. jl_method_instance_t *jl_new_method_instance(jl_value_t *ast, jl_tuple_t *sparams);

    Arrays:

    1. jl_array_t *jl_new_array(jl_value_t *atype, jl_tuple_t *dims);
    2. jl_array_t *jl_new_arrayv(jl_value_t *atype, ...);
    3. jl_array_t *jl_alloc_array_1d(jl_value_t *atype, size_t nr);
    4. jl_array_t *jl_alloc_array_2d(jl_value_t *atype, size_t nr, size_t nc);
    5. jl_array_t *jl_alloc_array_3d(jl_value_t *atype, size_t nr, size_t nc, size_t z);
    6. jl_array_t *jl_alloc_vec_any(size_t n);

    Note that many of these have alternative allocation functions for various special-purposes. The list here reflects the more common usages, but a more complete list can be found by reading the .

    Internal to Julia, storage is typically allocated by newstruct() (or newobj() for the special types):

    1. jl_value_t *newstruct(jl_value_t *type);

    And at the lowest level, memory is getting allocated by a call to the garbage collector (in gc.c), then tagged with its type:

    Note that all objects are allocated in multiples of 4 bytes and aligned to the platform pointer size. Memory is allocated from a pool for smaller objects, or directly with malloc() for large objects.

    Singleton Types

    See Singleton Types and