ADC Home > Reference Library > Reference > Darwin > Kernel Framework Reference

IOPartitionScheme

Inherits from:	IOStorage
Declared In:	IOPartitionScheme.h

Overview

The common base class for all partition scheme objects.

Discussion

The IOPartitionScheme class is the common base class for all partition scheme objects. It extends the IOStorage class by implementing the appropriate open and close semantics for partition objects (standard semantics are to act as a multiplexor for incoming opens, producing one outgoing open with the correct access). It also implements the default read and write semantics, which pass all reads and writes through to the provider media unprocessed. For simple schemes, the default behavior is sufficient. More complex partition schemes such as RAID will want to do extra processing for reads and writes.

Functions

handleClose
handleIsOpen
handleOpen
read
synchronizeCache
write

handleClose

protected

virtual void handleClose(
    IOService *client,
    IOOptionBits options);

Parameters

client: Client requesting the close.
options: Options for the close. Set to zero.

Discussion

The handleClose method closes the client's access to this object.

This implementation replaces the IOService definition of handleClose().

handleIsOpen

protected

virtual bool handleIsOpen(
    const IOService * client) const;

Parameters

client: Client to check the open state of. Set to zero to check the open state of all clients.

Return Value

Returns true if the client was (or clients were) open, false otherwise.

Discussion

The handleIsOpen method determines whether the specified client, or any client if none is specified, presently has an open on this object.

This implementation replaces the IOService definition of handleIsOpen().

handleOpen

protected

virtual bool handleOpen(
    IOService *client, 
    IOOptionBits options, 
    void *access);

Parameters

client: Client requesting the open.
options: Options for the open. Set to zero.
access: Access level for the open. Set to kIOStorageAccessReader or kIOStorageAccessReaderWriter.

Return Value

Returns true if the open was successful, false otherwise.

Discussion

The handleOpen method grants or denies permission to access this object to an interested client. The argument is an IOStorageAccess value that specifies the level of access desired -- reader or reader-writer.

This method can be invoked to upgrade or downgrade the access level for an existing client as well. The previous access level will prevail for upgrades that fail, of course. A downgrade should never fail. If the new access level should be the same as the old for a given client, this method will do nothing and return success. In all cases, one, singular close-per-client is expected for all opens-per-client received.

This implementation replaces the IOService definition of handleOpen().

read

public

virtual void read(
    IOService *client, 
    UInt64 byteStart, 
    IOMemoryDescriptor *buffer, 
    IOStorageAttributes *attributes, 
    IOStorageCompletion *completion);

Parameters

client: Client requesting the read.
byteStart: Starting byte offset for the data transfer.
buffer: Buffer for the data transfer. The size of the buffer implies the size of the data transfer.
attributes: Attributes of the data transfer. See IOStorageAttributes. It is the responsibility of the callee to maintain the information for the duration of the data transfer, as necessary.
completion: Completion routine to call once the data transfer is complete. It is the responsibility of the callee to maintain the information for the duration of the data transfer, as necessary.

Discussion

Read data from the storage object at the specified byte offset into the specified buffer, asynchronously. When the read completes, the caller will be notified via the specified completion action.

The buffer will be retained for the duration of the read.

For simple partition schemes, the default behavior is to simply pass the read through to the provider media. More complex partition schemes such as RAID will need to do extra processing here.

synchronizeCache

public

virtual IOReturn synchronizeCache(
    IOService *client);

Parameters

client: Client requesting the cache synchronization.

Return Value

Returns the status of the cache synchronization.

Discussion

Flush the cached data in the storage object, if any, synchronously.

write

public

virtual void write(
    IOService *client, 
    UInt64 byteStart, 
    IOMemoryDescriptor *buffer, 
    IOStorageAttributes *attributes, 
    IOStorageCompletion *completion);

Parameters

client: Client requesting the write.
byteStart: Starting byte offset for the data transfer.
buffer: Buffer for the data transfer. The size of the buffer implies the size of the data transfer.
attributes: Attributes of the data transfer. See IOStorageAttributes. It is the responsibility of the callee to maintain the information for the duration of the data transfer, as necessary.
completion: Completion routine to call once the data transfer is complete. It is the responsibility of the callee to maintain the information for the duration of the data transfer, as necessary.

Discussion

Write data into the storage object at the specified byte offset from the specified buffer, asynchronously. When the write completes, the caller will be notified via the specified completion action.

The buffer will be retained for the duration of the write.

For simple partition schemes, the default behavior is to simply pass the write through to the provider media. More complex partition schemes such as RAID will need to do extra processing here.

Did this document help you?

Yes: Tell us what works for you.

It’s good, but: Report typos, inaccuracies, and so forth.

It wasn’t helpful: Tell us what would have helped.

Last Updated: 2008-12-19