Updated: April 19, 2023

I/O mount structure


#include <sys/iofunc.h>

typedef struct _iofunc_funcs		iofunc_funcs_t;

typedef struct _iofunc_mount {
    uint32_t            flags;
    uint32_t            conf;
    dev_t               dev;
    int32_t             blocksize;
    iofunc_funcs_t      *funcs;
    void                *power;

    size_t              size;
    uint32_t            ext_flags;
    uint32_t            timeres;
    uint64_t            reserved[4];    
} iofunc_mount_t;

struct _iofunc_funcs {
   unsigned     nfuncs;
   IOFUNC_OCB_T *(*ocb_calloc) (resmgr_context_t *ctp,
                                IOFUNC_ATTR_T *attr);
   void         (*ocb_free) (IOFUNC_OCB_T *ocb);
   int          (*attr_lock)(IOFUNC_ATTR_T *attr);
   int          (*attr_unlock)(IOFUNC_ATTR_T *attr);
   int          (*attr_trylock)(IOFUNC_ATTR_T *attr);


The iofunc_mount_t structure contains per-mountpoint data items that are global to the entire mount device. Unlike the attribute (iofunc_attr_t) and OCB (iofunc_ocb_t) structures, which are mandatory, use of this structure is optional, and it need be used only if a resource manager needs to override some of the default settings it controls. For example, you need to provide this structure for the following:

(QNX Neutrino 7.0 or later) If you need to modify any members of the mount structure, use iofunc_mount_init() to initialize it. For example, if your program saves files to a filesystem, you can use iofunc_mount_init() and iofunc_mount_set_time() to set a timestamp resolution for open files that matches the resolution the filesystem uses (a POSIX requirement).

The members of the mount structure, specifically the conf and flags members, modify the behavior of some of the iofunc layer functions. This structure contains at least the following members:

A bitwise OR of mount flags, which include the following:
The offsets used by this resource manager are 32-bit (as opposed to the extended 64-bit offsets). This flag indicates that offset in the OCB and nbytes and inode in the attributes structure are 32-bit values.

In addition, you can define your own mount flags using the bits specified in IOFUNC_MOUNT_FLAGS_PRIVATE (see <sys/iofunc.h>).

A bitwise OR of configuration flags, which include the following:
Causes the default handler for the _IO_CHOWN message to behave in a manner defined by POSIX as chown-restricted”.
Has no effect on the iofunc layer libraries, but is returned by the iofunc layer's default _IO_PATHCONF handler.
Indicates that the resource manager supports synchronous I/O operations. If this bit isn't set, the following may occur:
Controls whether or not it's possible to create hard links to directories. If this bit isn't set, calls to link() and unlink() for directories fail; if this bit is set, only root is allowed to link and unlink directories.
Indicates whether or not the resource manager supports access control lists. Setting this bit enables ACL support for your resource manager in the resource manager library. For more information about ACLs, see Working with Access Control Lists (ACLs) in the QNX Neutrino Programmer's Guide.
(QNX Neutrino 7.0 or later) The iofunc_mount_t structure includes the size, ext_flags, and timeres members. This bit is set automatically when you call iofunc_mount_init() to set the iofunc_mount_t structure.

Note that the options mentioned above for the conf member are returned by the iofunc layer _IO_PATHCONF default handler.

The device number for the filesystem. This number is returned to the client's stat() function in the struct stat st_dev member.
The block size of the device. On resource managers that implement filesystems, this indicates the native blocksize of the disk, e.g., 512 bytes.
A pointer to an iofunc_funcs_t that specifies mount-specific handlers. The members include the following:
The number of functions present in the structure; you should fill it with the manifest constant _IOFUNC_NFUNCS.
ocb_calloc() and ocb_free()
Allows you to override the OCBs on a per-mountpoint basis (see Extending the OCB and attribute structures in the “Extending the POSIX-Layer Data Structures” chapter of Writing a Resource Manager). If these members are NULL, then the default library versions (iofunc_ocb_calloc() and iofunc_ocb_free()) are used. You must specify either both or neither of these functions; they operate as a matched pair.
attr_lock(), attr_unlock(), and attr_trylock()
Allows you to create customized versions of the functions that lock, unlock, and attempt to lock the iofunc_attr_t attribute structure. These are the preferred way to modify the attribute locking behavior for a resource manager, and are called by iofunc_lock_ocb_default() and iofunc_unlock_ocb_default() if they're supplied. The lock and unlock operations are used as a pair, and if you supply either, you must supply both. If you do override these, you may no longer use iofunc_attr_lock() and iofunc_attr_unlock().
Note: The iofunc_attr_lock() and iofunc_attr_unlock() functions use the lock member of the iofunc_attr_t structure. If you want to use the lockobj member to provide your own locking mechanism, you need to provide your own locking and unlocking routines.
Reserved for future use.
(QNX Neutrino 7.0 or later) Contains the size of the iofunc_mount_t structure. It's set by iofunc_mount_init().
(QNX Neutrino 7.0 or later) Extends the range of possible flags values.
(QNX Neutrino 7.0 or later) Contains the timestamp resolution in nanoseconds for iofunc functions and structures. It's set by iofunc_mount_set_time().
Reserved for future use.