Location:
e32std.h
Link against: euser.lib
class RChunk : public RHandleBase;
A handle to a chunk.
The chunk itself is a kernel side object.
RHandleBase
- A handle to an object
RChunk
- A handle to a chunk
Defined in RChunk
:
Adjust()
, AdjustDoubleEnded()
, Allocate()
, Base()
, Bottom()
, Commit()
, CreateDisconnectedGlobal()
, CreateDisconnectedLocal()
, CreateDoubleEndedGlobal()
, CreateDoubleEndedLocal()
, CreateGlobal()
, CreateLocal()
, CreateLocalCode()
, Decommit()
, EPreventAdjust
, IsReadable()
, IsWritable()
, MaxSize()
, Open()
, Open()
, Open()
, OpenGlobal()
, SetRestrictions()
, Size()
, TRestrictions
, Top()
Inherited from RHandleBase
:
Attributes()
,
Close()
,
Duplicate()
,
FullName()
,
Handle()
,
HandleInfo()
,
Name()
,
SetHandle()
,
SetHandleNC()
,
SetReturnedHandle()
,
iHandle
inline TInt Open(const TFindChunk &aFind, TOwnerType aType=EOwnerProcess);
Opens a handle to the global chunk found using a TFindChunk
object.
ATFindChunk
object is used to find all chunks whose full names match a specified pattern.
By default, ownership of this chunk handle is vested in the current process, but can be vested in the current thread by passing EOwnerThread as the second parameter to this function.
|
|
IMPORT_C TInt CreateLocal(TInt aSize, TInt aMaxSize, TOwnerType aType=EOwnerProcess);
Creates a local chunk.
The chunk is local to the process creating it; i.e. it is private to the process creating it and is not intended for access by other user processes.
aMaxSize specifies the maximum size of the chunk and aSize specifies the number of bytes to be committed on creation of the chunk. Both values are rounded up to the next nearest processor page boundary value if they are not already on a processor page boundary.
The committed region always starts at the bottom of the reserved region.
By default, ownership of this chunk handle is vested in the current process. Ownership of the chunk handle can be vested in the current thread by passing EOwnerThread as the third parameter to this function.
|
|
|
IMPORT_C TInt CreateLocalCode(TInt aSize, TInt aMaxSize, TOwnerType aType=EOwnerProcess);
Creates a user writable chunk that is marked by the kernel as containing code.
The chunk is local to the process creating it, i.e. it is private to the process creating it and is not intended for access by other user processes.
On systems using a Harvard cache, this type of chunk removes the need to flush the instruction cache (I-Cache) on a context switch. However, the instruction Translation Look-aside Buffer (ITLB) still needs to be flushed when switching to or from a process with one of these chunks in its address space. Systems with a dynamic branch predictor may also need to flush their branch target buffer when switching from one process using this type of chunk to another.
|
|
|
IMPORT_C TInt CreateGlobal(const TDesC &aName, TInt aSize, TInt aMaxSize, TOwnerType aType=EOwnerProcess);
Creates a global chunk.
The chunk is global; i.e. it is potentially visible to all processes and is intended for access by other user processes.
aMaxSize specifies the maximum size of the chunk and aSize specifies the number of bytes to be committed on creation of the chunk. Both values are rounded up to the next nearest processor page boundary value ,if they are not already on a processor page boundary value.
The committed region always starts at the bottom of the reserved region.
The descriptor aName contains the name to be assigned to this global chunk. If this name is empty, the chunk will be anonymous. Anonymous chunks cannot be accessed by other processes unless the creator explicitly passes them a handle to the chunk - this can be used to transfer large amounts of data between processes in a secure fashion.
By default, ownership of this chunk handle is vested in the current process. Ownership of the chunk handle can be vested in the current thread by passing EOwnerThread as the third parameter to this function.
|
|
|
IMPORT_C TInt CreateDoubleEndedLocal(TInt aInitialBottom, TInt aInitialTop, TInt aMaxSize, TOwnerType aType=EOwnerProcess);
Creates a local, double ended, chunk.
The chunk is local to the process creating it; i.e. it is private to the process creating it and is not intended for access by other user processes.
The committed region of a double ended chunk can be any contiguous subset of the reserved region.
aMaxSize specifies the maximum size of the chunk.
The difference between aInitialTop and aInitialBottom gives the number of bytes to be committed, on creation of the chunk; aInitialBottom gives the offset of the bottom of the committed region from the base of the chunk's reserved region; aInitialTop gives the offset of the top of the committed region from the base of the chunk's reserved region.
Both aInitialBottom and aInitialTop are rounded up to the next nearest processor page boundary value, if they are not already on a processor page boundary value.
By default, ownership of this chunk handle is vested in the current process. Ownership of the chunk handle can be vested in the current thread by passing EOwnerThread as the third parameter to this function.
Note that:
1. the lowest valid address in a double ended chunk is the sum of the base of the chunk's reserved region plus the adjusted value of aInitialBottom
2. the highest valid address in a double ended chunk is the the sum of the base of the chunk's reserved region plus the adjusted value of aInitialTop - 1.
|
|
|
IMPORT_C TInt CreateDoubleEndedGlobal(const TDesC &aName, TInt aInitialBottom, TInt aInitialTop, TInt aMaxSize, TOwnerType
aType=EOwnerProcess);
Creates a global, double ended, chunk.
The chunk is global; i.e. it is visible to all processes and is intended for access by other user processes.
The committed region of a double ended chunk can be any contiguous subset of the reserved region.
aMaxSize specifies the maximum size of the chunk.
The difference between aInitialTop and aInitialBottom gives the number of bytes to be committed, on creation of the chunk; aInitialBottom gives the offset of the bottom of the committed region from the base of the chunk's reserved region; aInitialTop gives the offset of the top of the committed region from the base of the chunk's reserved region.
Both aInitialBottom and aInitialTop are rounded up to the next nearest processor page boundary value, if they are not already on a processor page boundary value.
The descriptor aName contains the name to be assigned to this global chunk.
By default, ownership of this chunk handle is vested in the current process. Ownership of the chunk handle can be vested in the current thread by passing EOwnerThread as the third parameter to this function.
Note that:
1. the lowest valid address in a double ended chunk is the sum of the base of the chunk's reserved region plus the adjusted value of aInitialBottom
2. the highest valid address in a double ended chunk is the the sum of the base of the chunk's reserved region plus the adjusted value of aInitialTop - 1.
|
|
|
IMPORT_C TInt CreateDisconnectedLocal(TInt aInitialBottom, TInt aInitialTop, TInt aMaxSize, TOwnerType aType=EOwnerProcess);
Creates a local, disconnected chunk.
The chunk is local to the process creating it; i.e. it is private to the process creating it and is not intended for access by other user processes.
A disconnected chunk has a committed region consisting of an arbitrary set of MMU pages within the reserved region, i.e. each page-sized address range within the reserved region which begins on a page boundary may be committed independently.
aMaxSize specifies the maximum size of the chunk.
The difference between aInitialTop and aInitialBottom gives the number of bytes to be committed, on creation of the chunk; aInitialBottom gives the offset of the bottom of the committed region from the base of the chunk's reserved region; aInitialTop gives the offset of the top of the committed region from the base of the chunk's reserved region.
Both aInitialBottom and aInitialTop are rounded up to the next nearest processor page boundary value, if they are not already on a processor page boundary value.
By default, ownership of this chunk handle is vested in the current process. Ownership of the chunk handle can be vested in the current thread by passing EOwnerThread as the third parameter to this function.
|
|
|
IMPORT_C TInt CreateDisconnectedGlobal(const TDesC &aName, TInt aInitialBottom, TInt aInitialTop, TInt aMaxSize, TOwnerType
aType=EOwnerProcess);
Creates a global, disconnected, chunk.
The chunk is global; i.e. it is visible to all processes and is intended for access by other user processes.
A disconnected chunk has a committed region consisting of an arbitrary set of MMU pages within the reserved region, i.e. each page-sized address range within the reserved region which begins on a page boundary may be committed independently.
aMaxSize specifies the maximum size of the chunk.
The difference between aInitialTop and aInitialBottom gives the number of bytes to be committed, on creation of the chunk; aInitialBottom gives the offset of the bottom of the committed region from the base of the chunk's reserved region; aInitialTop gives the offset of the top of the committed region from the base of the chunk's reserved region.
Both aInitialBottom and aInitialTop are rounded up to the next nearest processor page boundary value, if they are not already on a processor page boundary value.
The descriptor aName contains the name to be assigned to this global chunk.
By default, ownership of this chunk handle is vested in the current process. Ownership of the chunk handle can be vested in the current thread by passing EOwnerThread as the third parameter to this function.
|
|
|
IMPORT_C TInt SetRestrictions(TUint aFlags);
Sets or removes restrictions on the ability of the chunk to change.
For example, to adjust, commit etc
|
|
IMPORT_C TInt OpenGlobal(const TDesC &aName, TBool isReadOnly, TOwnerType aType=EOwnerProcess);
Opens a handle to a specific named global chunk.
Full read/write access can be allowed or access can be limited to read only.
By default, ownership of this process handle is vested in the current process, but can be vested in the current thread by passing EOwnerThread as the second parameter to this function.
|
|
IMPORT_C TInt Open(RMessagePtr2 aMessage, TInt aParam, TBool isReadOnly, TOwnerType aType=EOwnerProcess);
Opens a handle to a chunk using a handle number sent by a client to a server.
This function is called by the server.
|
|
IMPORT_C TInt Open(TInt aArgumentIndex, TOwnerType aType=EOwnerProcess);
Opens a handle to a chunk using a handle number passed as an environment data item to the child process during the creation of that child process.
Note that this function can only be called successfully once.
|
|
IMPORT_C TInt Adjust(TInt aNewSize) const;
Changes the number of bytes committed to the chunk.
This value is always rounded up to the next nearest processor page boundary.
|
|
|
IMPORT_C TInt AdjustDoubleEnded(TInt aBottom, TInt aTop) const;
Changes the number of bytes and the position of this double ended chunk's committed region.
The difference between aTop and aBottom gives the new size of the committed region; aBottom gives the offset of the bottom of the committed region from the base of the chunk's reserved region.
Both aBottom and aTop are rounded up to the next nearest processor page boundary.
The function fails if this chunk is not a double ended chunk; for a standard chunk, use the Adjust()
function.
Note that if the initial and final committed regions intersect, the contents of the intersection are unchanged. Other parts of the committed region have undefined contents.
Note also that:
1. the lowest valid address in a double ended chunk is the sum of the base of the chunk's reserved region plus the adjusted value of aBottom
2. the highest valid address in a double ended chunk is the the sum of the base of the chunk's reserved region plus the adjusted value of aTop - 1.
|
|
|
IMPORT_C TInt Commit(TInt anOffset, TInt aSize) const;
Commits memory to a disconnected chunk.
Memory is committed in blocks of the MMU page size. E.g. Commit(pageSize-1,2) which asks for the last byte of the first page and the first byte of the second page and will result in the first 2 pages in the chunk being committed. For this reason it is best to only use values for aOffset and aSize which are multiples of the MMU page size. This size can be obtained with the following code.
TInt pageSize;
HAL::Get(HAL::EMemoryPageSize,pageSize)
|
|
|
IMPORT_C TInt Allocate(TInt aSize) const;
Allocates and commits to a disconnected chunk.
|
|
|
IMPORT_C TInt Decommit(TInt anOffset, TInt aSize) const;
Decommits memory from a disconnected chunk.
Memory is decommitted in blocks of the MMU page size. E.g. Decommit(pageSize-1,2) which asks for the last byte of the first page and the first byte of the second page and will result in the first 2 pages in the chunk being decommitted. For this reason it is best to only use values for aOffset and aSize which are multiples of the MMU page size. This size can be obtained with the following code.
TInt pageSize;
HAL::Get(HAL::EMemoryPageSize,pageSize)
|
|
|
IMPORT_C TUint8 *Base() const;
Gets a pointer to the base of the chunk's reserved region.
|
IMPORT_C TInt Size() const;
Gets the current size of this chunk's committed region.
|
IMPORT_C TInt Bottom() const;
Gets the offset of the bottom of the double ended chunk's committed region from the base of the chunk's reserved region.
Note that the lowest valid address in a double ended chunk is the sum of the base of the chunk's reserved region plus the
value of Bottom()
.
|
IMPORT_C TInt Top() const;
Gets the offset of the top of the double ended chunk's committed region from the base of the chunk's reserved region.
Note that the highest valid address in a double ended chunk is the the sum of the base of the chunk's reserved region plus
the value of Top()
- 1.
|
IMPORT_C TInt MaxSize() const;
Gets the maximum size of this chunk.
This maximum size of this chunk is set when the chunk is created.
|
inline TBool IsReadable() const;
Tests whether the chunk is mapped into its process address space.
|
inline TBool IsWritable() const;
Tests whether the chunk mapped into its process address space and is writable.
|