glib/glib/gvariant-core.c
Simon McVittie e9337a9c1d gvariant-core: Don't pass NULL second argument to memcpy
Similar to 3837b83f, glibc memcpy is declared with the first two
arguments annotated as non-null via an attribute, which results in the
undefined behaviour sanitizer considering it to be UB to pass a null
pointer there (even if we are copying no bytes, and hence not actually
dereferencing the pointer).

Signed-off-by: Simon McVittie <smcv@collabora.com>
2020-01-07 15:06:51 +00:00

1209 lines
38 KiB
C
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

/*
* Copyright © 2007, 2008 Ryan Lortie
* Copyright © 2010 Codethink Limited
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library 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
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/
#include "config.h"
#include <glib/gvariant-core.h>
#include <glib/gvariant-internal.h>
#include <glib/gvariant-serialiser.h>
#include <glib/gtestutils.h>
#include <glib/gbitlock.h>
#include <glib/gatomic.h>
#include <glib/gbytes.h>
#include <glib/gslice.h>
#include <glib/gmem.h>
#include <glib/grefcount.h>
#include <string.h>
/*
* This file includes the structure definition for GVariant and a small
* set of functions that are allowed to access the structure directly.
*
* This minimises the amount of code that can possibly touch a GVariant
* structure directly to a few simple fundamental operations. These few
* operations are written to be completely threadsafe with respect to
* all possible outside access. This means that we only need to be
* concerned about thread safety issues in this one small file.
*
* Most GVariant API functions are in gvariant.c.
*/
/**
* GVariant:
*
* #GVariant is an opaque data structure and can only be accessed
* using the following functions.
*
* Since: 2.24
**/
struct _GVariant
/* see below for field member documentation */
{
GVariantTypeInfo *type_info;
gsize size;
union
{
struct
{
GBytes *bytes;
gconstpointer data;
} serialised;
struct
{
GVariant **children;
gsize n_children;
} tree;
} contents;
gint state;
gatomicrefcount ref_count;
gsize depth;
};
/* struct GVariant:
*
* There are two primary forms of GVariant instances: "serialised form"
* and "tree form".
*
* "serialised form": A serialised GVariant instance stores its value in
* the GVariant serialisation format. All
* basic-typed instances (ie: non-containers) are in
* serialised format, as are some containers.
*
* "tree form": Some containers are in "tree form". In this case,
* instead of containing the serialised data for the
* container, the instance contains an array of pointers to
* the child values of the container (thus forming a tree).
*
* It is possible for an instance to transition from tree form to
* serialised form. This happens, implicitly, if the serialised data is
* requested (eg: via g_variant_get_data()). Serialised form instances
* never transition into tree form.
*
*
* The fields of the structure are documented here:
*
* type_info: this is a reference to a GVariantTypeInfo describing the
* type of the instance. When the instance is freed, this
* reference must be released with g_variant_type_info_unref().
*
* The type_info field never changes during the life of the
* instance, so it can be accessed without a lock.
*
* size: this is the size of the serialised form for the instance, if it
* is known. If the instance is in serialised form then it is, by
* definition, known. If the instance is in tree form then it may
* be unknown (in which case it is -1). It is possible for the
* size to be known when in tree form if, for example, the user
* has called g_variant_get_size() without calling
* g_variant_get_data(). Additionally, even when the user calls
* g_variant_get_data() the size of the data must first be
* determined so that a large enough buffer can be allocated for
* the data.
*
* Once the size is known, it can never become unknown again.
* g_variant_ensure_size() is used to ensure that the size is in
* the known state -- it calculates the size if needed. After
* that, the size field can be accessed without a lock.
*
* contents: a union containing either the information associated with
* holding a value in serialised form or holding a value in
* tree form.
*
* .serialised: Only valid when the instance is in serialised form.
*
* Since an instance can never transition away from
* serialised form, once these fields are set, they will
* never be changed. It is therefore valid to access
* them without holding a lock.
*
* .bytes: the #GBytes that contains the memory pointed to by
* .data, or %NULL if .data is %NULL. In the event that
* the instance was deserialised from another instance,
* then the bytes will be shared by both of them. When
* the instance is freed, this reference must be released
* with g_bytes_unref().
*
* .data: the serialised data (of size 'size') of the instance.
* This pointer should not be freed or modified in any way.
* #GBytes is responsible for memory management.
*
* This pointer may be %NULL in two cases:
*
* - if the serialised size of the instance is 0
*
* - if the instance is of a fixed-sized type and was
* deserialised out of a corrupted container such that
* the container contains too few bytes to point to the
* entire proper fixed-size of this instance. In this
* case, 'size' will still be equal to the proper fixed
* size, but this pointer will be %NULL. This is exactly
* the reason that g_variant_get_data() sometimes returns
* %NULL. For all other calls, the effect should be as
* if .data pointed to the appropriate number of nul
* bytes.
*
* .tree: Only valid when the instance is in tree form.
*
* Note that accesses from other threads could result in
* conversion of the instance from tree form to serialised form
* at any time. For this reason, the instance lock must always
* be held while performing any operations on 'contents.tree'.
*
* .children: the array of the child instances of this instance.
* When the instance is freed (or converted to serialised
* form) then each child must have g_variant_unref()
* called on it and the array must be freed using
* g_free().
*
* .n_children: the number of items in the .children array.
*
* state: a bitfield describing the state of the instance. It is a
* bitwise-or of the following STATE_* constants:
*
* STATE_LOCKED: the instance lock is held. This is the bit used by
* g_bit_lock().
*
* STATE_SERIALISED: the instance is in serialised form. If this
* flag is not set then the instance is in tree
* form.
*
* STATE_TRUSTED: for serialised form instances, this means that the
* serialised data is known to be in normal form (ie:
* not corrupted).
*
* For tree form instances, this means that all of the
* child instances in the contents.tree.children array
* are trusted. This means that if the container is
* serialised then the resulting data will be in
* normal form.
*
* If this flag is unset it does not imply that the
* data is corrupted. It merely means that we're not
* sure that it's valid. See g_variant_is_trusted().
*
* STATE_FLOATING: if this flag is set then the object has a floating
* reference. See g_variant_ref_sink().
*
* ref_count: the reference count of the instance
*
* depth: the depth of the GVariant in a hierarchy of nested containers,
* increasing with the level of nesting. The top-most GVariant has depth
* zero. This is used to avoid recursing too deeply and overflowing the
* stack when handling deeply nested untrusted serialised GVariants.
*/
#define STATE_LOCKED 1
#define STATE_SERIALISED 2
#define STATE_TRUSTED 4
#define STATE_FLOATING 8
/* -- private -- */
/* < private >
* g_variant_lock:
* @value: a #GVariant
*
* Locks @value for performing sensitive operations.
*/
static void
g_variant_lock (GVariant *value)
{
g_bit_lock (&value->state, 0);
}
/* < private >
* g_variant_unlock:
* @value: a #GVariant
*
* Unlocks @value after performing sensitive operations.
*/
static void
g_variant_unlock (GVariant *value)
{
g_bit_unlock (&value->state, 0);
}
/* < private >
* g_variant_release_children:
* @value: a #GVariant
*
* Releases the reference held on each child in the 'children' array of
* @value and frees the array itself. @value must be in tree form.
*
* This is done when freeing a tree-form instance or converting it to
* serialised form.
*
* The current thread must hold the lock on @value.
*/
static void
g_variant_release_children (GVariant *value)
{
gsize i;
g_assert (value->state & STATE_LOCKED);
g_assert (~value->state & STATE_SERIALISED);
for (i = 0; i < value->contents.tree.n_children; i++)
g_variant_unref (value->contents.tree.children[i]);
g_free (value->contents.tree.children);
}
/* This begins the main body of the recursive serialiser.
*
* There are 3 functions here that work as a team with the serialiser to
* get things done. g_variant_store() has a trivial role, but as a
* public API function, it has its definition elsewhere.
*
* Note that "serialisation" of an instance does not mean that the
* instance is converted to serialised form -- it means that the
* serialised form of an instance is written to an external buffer.
* g_variant_ensure_serialised() (which is not part of this set of
* functions) is the function that is responsible for converting an
* instance to serialised form.
*
* We are only concerned here with container types since non-container
* instances are always in serialised form. For these instances,
* storing their serialised form merely involves a memcpy().
*
* Serialisation is a two-step process. First, the size of the
* serialised data must be calculated so that an appropriately-sized
* buffer can be allocated. Second, the data is written into the
* buffer.
*
* Determining the size:
* The process of determining the size is triggered by a call to
* g_variant_ensure_size() on a container. This invokes the
* serialiser code to determine the size. The serialiser is passed
* g_variant_fill_gvs() as a callback.
*
* g_variant_fill_gvs() is called by the serialiser on each child of
* the container which, in turn, calls g_variant_ensure_size() on
* itself and fills in the result of its own size calculation.
*
* The serialiser uses the size information from the children to
* calculate the size needed for the entire container.
*
* Writing the data:
* After the buffer has been allocated, g_variant_serialise() is
* called on the container. This invokes the serialiser code to write
* the bytes to the container. The serialiser is, again, passed
* g_variant_fill_gvs() as a callback.
*
* This time, when g_variant_fill_gvs() is called for each child, the
* child is given a pointer to a sub-region of the allocated buffer
* where it should write its data. This is done by calling
* g_variant_store(). In the event that the instance is in serialised
* form this means a memcpy() of the serialised data into the
* allocated buffer. In the event that the instance is in tree form
* this means a recursive call back into g_variant_serialise().
*
*
* The forward declaration here allows corecursion via callback:
*/
static void g_variant_fill_gvs (GVariantSerialised *, gpointer);
/* < private >
* g_variant_ensure_size:
* @value: a #GVariant
*
* Ensures that the ->size field of @value is filled in properly. This
* must be done as a precursor to any serialisation of the value in
* order to know how large of a buffer is needed to store the data.
*
* The current thread must hold the lock on @value.
*/
static void
g_variant_ensure_size (GVariant *value)
{
g_assert (value->state & STATE_LOCKED);
if (value->size == (gsize) -1)
{
gpointer *children;
gsize n_children;
children = (gpointer *) value->contents.tree.children;
n_children = value->contents.tree.n_children;
value->size = g_variant_serialiser_needed_size (value->type_info,
g_variant_fill_gvs,
children, n_children);
}
}
/* < private >
* g_variant_serialise:
* @value: a #GVariant
* @data: an appropriately-sized buffer
*
* Serialises @value into @data. @value must be in tree form.
*
* No change is made to @value.
*
* The current thread must hold the lock on @value.
*/
static void
g_variant_serialise (GVariant *value,
gpointer data)
{
GVariantSerialised serialised = { 0, };
gpointer *children;
gsize n_children;
g_assert (~value->state & STATE_SERIALISED);
g_assert (value->state & STATE_LOCKED);
serialised.type_info = value->type_info;
serialised.size = value->size;
serialised.data = data;
serialised.depth = value->depth;
children = (gpointer *) value->contents.tree.children;
n_children = value->contents.tree.n_children;
g_variant_serialiser_serialise (serialised, g_variant_fill_gvs,
children, n_children);
}
/* < private >
* g_variant_fill_gvs:
* @serialised: a pointer to a #GVariantSerialised
* @data: a #GVariant instance
*
* This is the callback that is passed by a tree-form container instance
* to the serialiser. This callback gets called on each child of the
* container. Each child is responsible for performing the following
* actions:
*
* - reporting its type
*
* - reporting its serialised size (requires knowing the size first)
*
* - possibly storing its serialised form into the provided buffer
*/
static void
g_variant_fill_gvs (GVariantSerialised *serialised,
gpointer data)
{
GVariant *value = data;
g_variant_lock (value);
g_variant_ensure_size (value);
g_variant_unlock (value);
if (serialised->type_info == NULL)
serialised->type_info = value->type_info;
g_assert (serialised->type_info == value->type_info);
if (serialised->size == 0)
serialised->size = value->size;
g_assert (serialised->size == value->size);
serialised->depth = value->depth;
if (serialised->data)
/* g_variant_store() is a public API, so it
* it will reacquire the lock if it needs to.
*/
g_variant_store (value, serialised->data);
}
/* this ends the main body of the recursive serialiser */
/* < private >
* g_variant_ensure_serialised:
* @value: a #GVariant
*
* Ensures that @value is in serialised form.
*
* If @value is in tree form then this function ensures that the
* serialised size is known and then allocates a buffer of that size and
* serialises the instance into the buffer. The 'children' array is
* then released and the instance is set to serialised form based on the
* contents of the buffer.
*
* The current thread must hold the lock on @value.
*/
static void
g_variant_ensure_serialised (GVariant *value)
{
g_assert (value->state & STATE_LOCKED);
if (~value->state & STATE_SERIALISED)
{
GBytes *bytes;
gpointer data;
g_variant_ensure_size (value);
data = g_malloc (value->size);
g_variant_serialise (value, data);
g_variant_release_children (value);
bytes = g_bytes_new_take (data, value->size);
value->contents.serialised.data = g_bytes_get_data (bytes, NULL);
value->contents.serialised.bytes = bytes;
value->state |= STATE_SERIALISED;
}
}
/* < private >
* g_variant_alloc:
* @type: the type of the new instance
* @serialised: if the instance will be in serialised form
* @trusted: if the instance will be trusted
*
* Allocates a #GVariant instance and does some common work (such as
* looking up and filling in the type info), setting the state field,
* and setting the ref_count to 1.
*
* Returns: a new #GVariant with a floating reference
*/
static GVariant *
g_variant_alloc (const GVariantType *type,
gboolean serialised,
gboolean trusted)
{
GVariant *value;
value = g_slice_new (GVariant);
value->type_info = g_variant_type_info_get (type);
value->state = (serialised ? STATE_SERIALISED : 0) |
(trusted ? STATE_TRUSTED : 0) |
STATE_FLOATING;
value->size = (gssize) -1;
g_atomic_ref_count_init (&value->ref_count);
value->depth = 0;
return value;
}
/**
* g_variant_new_from_bytes:
* @type: a #GVariantType
* @bytes: a #GBytes
* @trusted: if the contents of @bytes are trusted
*
* Constructs a new serialised-mode #GVariant instance. This is the
* inner interface for creation of new serialised values that gets
* called from various functions in gvariant.c.
*
* A reference is taken on @bytes.
*
* The data in @bytes must be aligned appropriately for the @type being loaded.
* Otherwise this function will internally create a copy of the memory (since
* GLib 2.60) or (in older versions) fail and exit the process.
*
* Returns: (transfer none): a new #GVariant with a floating reference
*
* Since: 2.36
*/
GVariant *
g_variant_new_from_bytes (const GVariantType *type,
GBytes *bytes,
gboolean trusted)
{
GVariant *value;
guint alignment;
gsize size;
GBytes *owned_bytes = NULL;
GVariantSerialised serialised;
value = g_variant_alloc (type, TRUE, trusted);
g_variant_type_info_query (value->type_info,
&alignment, &size);
/* Ensure the alignment is correct. This is a huge performance hit if its
* not correct, but thats better than aborting if a caller provides data
* with the wrong alignment (which is likely to happen very occasionally, and
* only cause an abort on some architectures — so is unlikely to be caught
* in testing). Callers can always actively ensure they use the correct
* alignment to avoid the performance hit. */
serialised.type_info = value->type_info;
serialised.data = (guchar *) g_bytes_get_data (bytes, &serialised.size);
serialised.depth = 0;
if (!g_variant_serialised_check (serialised))
{
#ifdef HAVE_POSIX_MEMALIGN
gpointer aligned_data = NULL;
gsize aligned_size = g_bytes_get_size (bytes);
/* posix_memalign() requires the alignment to be a multiple of
* sizeof(void*), and a power of 2. See g_variant_type_info_query() for
* details on the alignment format. */
if (posix_memalign (&aligned_data, MAX (sizeof (void *), alignment + 1),
aligned_size) != 0)
g_error ("posix_memalign failed");
if (aligned_size != 0)
memcpy (aligned_data, g_bytes_get_data (bytes, NULL), aligned_size);
bytes = owned_bytes = g_bytes_new_with_free_func (aligned_data,
aligned_size,
free, aligned_data);
aligned_data = NULL;
#else
/* NOTE: there may be platforms that lack posix_memalign() and also
* have malloc() that returns non-8-aligned. if so, we need to try
* harder here.
*/
bytes = owned_bytes = g_bytes_new (g_bytes_get_data (bytes, NULL),
g_bytes_get_size (bytes));
#endif
}
value->contents.serialised.bytes = g_bytes_ref (bytes);
if (size && g_bytes_get_size (bytes) != size)
{
/* Creating a fixed-sized GVariant with a bytes of the wrong
* size.
*
* We should do the equivalent of pulling a fixed-sized child out
* of a brozen container (ie: data is NULL size is equal to the correct
* fixed size).
*/
value->contents.serialised.data = NULL;
value->size = size;
}
else
{
value->contents.serialised.data = g_bytes_get_data (bytes, &value->size);
}
g_clear_pointer (&owned_bytes, g_bytes_unref);
return value;
}
/* -- internal -- */
/* < internal >
* g_variant_new_from_children:
* @type: a #GVariantType
* @children: an array of #GVariant pointers. Consumed.
* @n_children: the length of @children
* @trusted: %TRUE if every child in @children in trusted
*
* Constructs a new tree-mode #GVariant instance. This is the inner
* interface for creation of new serialised values that gets called from
* various functions in gvariant.c.
*
* @children is consumed by this function. g_free() will be called on
* it some time later.
*
* Returns: a new #GVariant with a floating reference
*/
GVariant *
g_variant_new_from_children (const GVariantType *type,
GVariant **children,
gsize n_children,
gboolean trusted)
{
GVariant *value;
value = g_variant_alloc (type, FALSE, trusted);
value->contents.tree.children = children;
value->contents.tree.n_children = n_children;
return value;
}
/* < internal >
* g_variant_get_type_info:
* @value: a #GVariant
*
* Returns the #GVariantTypeInfo corresponding to the type of @value. A
* reference is not added, so the return value is only good for the
* duration of the life of @value.
*
* Returns: the #GVariantTypeInfo for @value
*/
GVariantTypeInfo *
g_variant_get_type_info (GVariant *value)
{
return value->type_info;
}
/* < internal >
* g_variant_is_trusted:
* @value: a #GVariant
*
* Determines if @value is trusted by #GVariant to contain only
* fully-valid data. All values constructed solely via #GVariant APIs
* are trusted, but values containing data read in from other sources
* are usually not trusted.
*
* The main advantage of trusted data is that certain checks can be
* skipped. For example, we don't need to check that a string is
* properly nul-terminated or that an object path is actually a
* properly-formatted object path.
*
* Returns: if @value is trusted
*/
gboolean
g_variant_is_trusted (GVariant *value)
{
return (value->state & STATE_TRUSTED) != 0;
}
/* -- public -- */
/**
* g_variant_unref:
* @value: a #GVariant
*
* Decreases the reference count of @value. When its reference count
* drops to 0, the memory used by the variant is freed.
*
* Since: 2.24
**/
void
g_variant_unref (GVariant *value)
{
g_return_if_fail (value != NULL);
if (g_atomic_ref_count_dec (&value->ref_count))
{
if G_UNLIKELY (value->state & STATE_LOCKED)
g_critical ("attempting to free a locked GVariant instance. "
"This should never happen.");
value->state |= STATE_LOCKED;
g_variant_type_info_unref (value->type_info);
if (value->state & STATE_SERIALISED)
g_bytes_unref (value->contents.serialised.bytes);
else
g_variant_release_children (value);
memset (value, 0, sizeof (GVariant));
g_slice_free (GVariant, value);
}
}
/**
* g_variant_ref:
* @value: a #GVariant
*
* Increases the reference count of @value.
*
* Returns: the same @value
*
* Since: 2.24
**/
GVariant *
g_variant_ref (GVariant *value)
{
g_return_val_if_fail (value != NULL, NULL);
g_atomic_ref_count_inc (&value->ref_count);
return value;
}
/**
* g_variant_ref_sink:
* @value: a #GVariant
*
* #GVariant uses a floating reference count system. All functions with
* names starting with `g_variant_new_` return floating
* references.
*
* Calling g_variant_ref_sink() on a #GVariant with a floating reference
* will convert the floating reference into a full reference. Calling
* g_variant_ref_sink() on a non-floating #GVariant results in an
* additional normal reference being added.
*
* In other words, if the @value is floating, then this call "assumes
* ownership" of the floating reference, converting it to a normal
* reference. If the @value is not floating, then this call adds a
* new normal reference increasing the reference count by one.
*
* All calls that result in a #GVariant instance being inserted into a
* container will call g_variant_ref_sink() on the instance. This means
* that if the value was just created (and has only its floating
* reference) then the container will assume sole ownership of the value
* at that point and the caller will not need to unreference it. This
* makes certain common styles of programming much easier while still
* maintaining normal refcounting semantics in situations where values
* are not floating.
*
* Returns: the same @value
*
* Since: 2.24
**/
GVariant *
g_variant_ref_sink (GVariant *value)
{
g_return_val_if_fail (value != NULL, NULL);
g_return_val_if_fail (!g_atomic_ref_count_compare (&value->ref_count, 0), NULL);
g_variant_lock (value);
if (~value->state & STATE_FLOATING)
g_variant_ref (value);
else
value->state &= ~STATE_FLOATING;
g_variant_unlock (value);
return value;
}
/**
* g_variant_take_ref:
* @value: a #GVariant
*
* If @value is floating, sink it. Otherwise, do nothing.
*
* Typically you want to use g_variant_ref_sink() in order to
* automatically do the correct thing with respect to floating or
* non-floating references, but there is one specific scenario where
* this function is helpful.
*
* The situation where this function is helpful is when creating an API
* that allows the user to provide a callback function that returns a
* #GVariant. We certainly want to allow the user the flexibility to
* return a non-floating reference from this callback (for the case
* where the value that is being returned already exists).
*
* At the same time, the style of the #GVariant API makes it likely that
* for newly-created #GVariant instances, the user can be saved some
* typing if they are allowed to return a #GVariant with a floating
* reference.
*
* Using this function on the return value of the user's callback allows
* the user to do whichever is more convenient for them. The caller
* will alway receives exactly one full reference to the value: either
* the one that was returned in the first place, or a floating reference
* that has been converted to a full reference.
*
* This function has an odd interaction when combined with
* g_variant_ref_sink() running at the same time in another thread on
* the same #GVariant instance. If g_variant_ref_sink() runs first then
* the result will be that the floating reference is converted to a hard
* reference. If g_variant_take_ref() runs first then the result will
* be that the floating reference is converted to a hard reference and
* an additional reference on top of that one is added. It is best to
* avoid this situation.
*
* Returns: the same @value
**/
GVariant *
g_variant_take_ref (GVariant *value)
{
g_return_val_if_fail (value != NULL, NULL);
g_return_val_if_fail (!g_atomic_ref_count_compare (&value->ref_count, 0), NULL);
g_atomic_int_and (&value->state, ~STATE_FLOATING);
return value;
}
/**
* g_variant_is_floating:
* @value: a #GVariant
*
* Checks whether @value has a floating reference count.
*
* This function should only ever be used to assert that a given variant
* is or is not floating, or for debug purposes. To acquire a reference
* to a variant that might be floating, always use g_variant_ref_sink()
* or g_variant_take_ref().
*
* See g_variant_ref_sink() for more information about floating reference
* counts.
*
* Returns: whether @value is floating
*
* Since: 2.26
**/
gboolean
g_variant_is_floating (GVariant *value)
{
g_return_val_if_fail (value != NULL, FALSE);
return (value->state & STATE_FLOATING) != 0;
}
/**
* g_variant_get_size:
* @value: a #GVariant instance
*
* Determines the number of bytes that would be required to store @value
* with g_variant_store().
*
* If @value has a fixed-sized type then this function always returned
* that fixed size.
*
* In the case that @value is already in serialised form or the size has
* already been calculated (ie: this function has been called before)
* then this function is O(1). Otherwise, the size is calculated, an
* operation which is approximately O(n) in the number of values
* involved.
*
* Returns: the serialised size of @value
*
* Since: 2.24
**/
gsize
g_variant_get_size (GVariant *value)
{
g_variant_lock (value);
g_variant_ensure_size (value);
g_variant_unlock (value);
return value->size;
}
/**
* g_variant_get_data:
* @value: a #GVariant instance
*
* Returns a pointer to the serialised form of a #GVariant instance.
* The returned data may not be in fully-normalised form if read from an
* untrusted source. The returned data must not be freed; it remains
* valid for as long as @value exists.
*
* If @value is a fixed-sized value that was deserialised from a
* corrupted serialised container then %NULL may be returned. In this
* case, the proper thing to do is typically to use the appropriate
* number of nul bytes in place of @value. If @value is not fixed-sized
* then %NULL is never returned.
*
* In the case that @value is already in serialised form, this function
* is O(1). If the value is not already in serialised form,
* serialisation occurs implicitly and is approximately O(n) in the size
* of the result.
*
* To deserialise the data returned by this function, in addition to the
* serialised data, you must know the type of the #GVariant, and (if the
* machine might be different) the endianness of the machine that stored
* it. As a result, file formats or network messages that incorporate
* serialised #GVariants must include this information either
* implicitly (for instance "the file always contains a
* %G_VARIANT_TYPE_VARIANT and it is always in little-endian order") or
* explicitly (by storing the type and/or endianness in addition to the
* serialised data).
*
* Returns: (transfer none): the serialised form of @value, or %NULL
*
* Since: 2.24
**/
gconstpointer
g_variant_get_data (GVariant *value)
{
g_variant_lock (value);
g_variant_ensure_serialised (value);
g_variant_unlock (value);
return value->contents.serialised.data;
}
/**
* g_variant_get_data_as_bytes:
* @value: a #GVariant
*
* Returns a pointer to the serialised form of a #GVariant instance.
* The semantics of this function are exactly the same as
* g_variant_get_data(), except that the returned #GBytes holds
* a reference to the variant data.
*
* Returns: (transfer full): A new #GBytes representing the variant data
*
* Since: 2.36
*/
GBytes *
g_variant_get_data_as_bytes (GVariant *value)
{
const gchar *bytes_data;
const gchar *data;
gsize bytes_size;
gsize size;
g_variant_lock (value);
g_variant_ensure_serialised (value);
g_variant_unlock (value);
bytes_data = g_bytes_get_data (value->contents.serialised.bytes, &bytes_size);
data = value->contents.serialised.data;
size = value->size;
if (data == NULL)
{
g_assert (size == 0);
data = bytes_data;
}
if (data == bytes_data && size == bytes_size)
return g_bytes_ref (value->contents.serialised.bytes);
else
return g_bytes_new_from_bytes (value->contents.serialised.bytes,
data - bytes_data, size);
}
/**
* g_variant_n_children:
* @value: a container #GVariant
*
* Determines the number of children in a container #GVariant instance.
* This includes variants, maybes, arrays, tuples and dictionary
* entries. It is an error to call this function on any other type of
* #GVariant.
*
* For variants, the return value is always 1. For values with maybe
* types, it is always zero or one. For arrays, it is the length of the
* array. For tuples it is the number of tuple items (which depends
* only on the type). For dictionary entries, it is always 2
*
* This function is O(1).
*
* Returns: the number of children in the container
*
* Since: 2.24
**/
gsize
g_variant_n_children (GVariant *value)
{
gsize n_children;
g_variant_lock (value);
if (value->state & STATE_SERIALISED)
{
GVariantSerialised serialised = {
value->type_info,
(gpointer) value->contents.serialised.data,
value->size,
value->depth,
};
n_children = g_variant_serialised_n_children (serialised);
}
else
n_children = value->contents.tree.n_children;
g_variant_unlock (value);
return n_children;
}
/**
* g_variant_get_child_value:
* @value: a container #GVariant
* @index_: the index of the child to fetch
*
* Reads a child item out of a container #GVariant instance. This
* includes variants, maybes, arrays, tuples and dictionary
* entries. It is an error to call this function on any other type of
* #GVariant.
*
* It is an error if @index_ is greater than the number of child items
* in the container. See g_variant_n_children().
*
* The returned value is never floating. You should free it with
* g_variant_unref() when you're done with it.
*
* There may be implementation specific restrictions on deeply nested values,
* which would result in the unit tuple being returned as the child value,
* instead of further nested children. #GVariant is guaranteed to handle
* nesting up to at least 64 levels.
*
* This function is O(1).
*
* Returns: (transfer full): the child at the specified index
*
* Since: 2.24
**/
GVariant *
g_variant_get_child_value (GVariant *value,
gsize index_)
{
g_return_val_if_fail (index_ < g_variant_n_children (value), NULL);
g_return_val_if_fail (value->depth < G_MAXSIZE, NULL);
if (~g_atomic_int_get (&value->state) & STATE_SERIALISED)
{
g_variant_lock (value);
if (~value->state & STATE_SERIALISED)
{
GVariant *child;
child = g_variant_ref (value->contents.tree.children[index_]);
g_variant_unlock (value);
return child;
}
g_variant_unlock (value);
}
{
GVariantSerialised serialised = {
value->type_info,
(gpointer) value->contents.serialised.data,
value->size,
value->depth,
};
GVariantSerialised s_child;
GVariant *child;
/* get the serialiser to extract the serialised data for the child
* from the serialised data for the container
*/
s_child = g_variant_serialised_get_child (serialised, index_);
/* Check whether this would cause nesting too deep. If so, return a fake
* child. The only situation we expect this to happen in is with a variant,
* as all other deeply-nested types have a static type, and hence should
* have been rejected earlier. In the case of a variant whose nesting plus
* the depth of its child is too great, return a unit variant () instead of
* the real child. */
if (!(value->state & STATE_TRUSTED) &&
g_variant_type_info_query_depth (s_child.type_info) >=
G_VARIANT_MAX_RECURSION_DEPTH - value->depth)
{
g_assert (g_variant_is_of_type (value, G_VARIANT_TYPE_VARIANT));
return g_variant_new_tuple (NULL, 0);
}
/* create a new serialised instance out of it */
child = g_slice_new (GVariant);
child->type_info = s_child.type_info;
child->state = (value->state & STATE_TRUSTED) |
STATE_SERIALISED;
child->size = s_child.size;
g_atomic_ref_count_init (&child->ref_count);
child->depth = value->depth + 1;
child->contents.serialised.bytes =
g_bytes_ref (value->contents.serialised.bytes);
child->contents.serialised.data = s_child.data;
return child;
}
}
/**
* g_variant_store:
* @value: the #GVariant to store
* @data: (not nullable): the location to store the serialised data at
*
* Stores the serialised form of @value at @data. @data should be
* large enough. See g_variant_get_size().
*
* The stored data is in machine native byte order but may not be in
* fully-normalised form if read from an untrusted source. See
* g_variant_get_normal_form() for a solution.
*
* As with g_variant_get_data(), to be able to deserialise the
* serialised variant successfully, its type and (if the destination
* machine might be different) its endianness must also be available.
*
* This function is approximately O(n) in the size of @data.
*
* Since: 2.24
**/
void
g_variant_store (GVariant *value,
gpointer data)
{
g_variant_lock (value);
if (value->state & STATE_SERIALISED)
{
if (value->contents.serialised.data != NULL)
memcpy (data, value->contents.serialised.data, value->size);
else
memset (data, 0, value->size);
}
else
g_variant_serialise (value, data);
g_variant_unlock (value);
}
/**
* g_variant_is_normal_form:
* @value: a #GVariant instance
*
* Checks if @value is in normal form.
*
* The main reason to do this is to detect if a given chunk of
* serialised data is in normal form: load the data into a #GVariant
* using g_variant_new_from_data() and then use this function to
* check.
*
* If @value is found to be in normal form then it will be marked as
* being trusted. If the value was already marked as being trusted then
* this function will immediately return %TRUE.
*
* There may be implementation specific restrictions on deeply nested values.
* GVariant is guaranteed to handle nesting up to at least 64 levels.
*
* Returns: %TRUE if @value is in normal form
*
* Since: 2.24
**/
gboolean
g_variant_is_normal_form (GVariant *value)
{
if (value->state & STATE_TRUSTED)
return TRUE;
g_variant_lock (value);
if (value->depth >= G_VARIANT_MAX_RECURSION_DEPTH)
return FALSE;
if (value->state & STATE_SERIALISED)
{
GVariantSerialised serialised = {
value->type_info,
(gpointer) value->contents.serialised.data,
value->size,
value->depth
};
if (g_variant_serialised_is_normal (serialised))
value->state |= STATE_TRUSTED;
}
else
{
gboolean normal = TRUE;
gsize i;
for (i = 0; i < value->contents.tree.n_children; i++)
normal &= g_variant_is_normal_form (value->contents.tree.children[i]);
if (normal)
value->state |= STATE_TRUSTED;
}
g_variant_unlock (value);
return (value->state & STATE_TRUSTED) != 0;
}