Handling ionotify(), poll(), and select()
A client uses ionotify(), poll(), and select() to ask a resource manager about the status of certain conditions (e.g., whether input data is available). The conditions may or may not have been met. The resource manager can be asked to:
- check the status of the conditions immediately, and return if any have been met
- deliver an event later on when a condition is met (this is referred to as arming the resource manager)
The poll() and select() functions differ from ionotify() in that most of the work is done in the library. For example, the client code would be unaware that any event is involved, nor would it be aware of the blocking function that waits for the event. This is all hidden in the library code for poll() and select().
However, from a resource manager's point of view, there's no difference between ionotify(), poll(), and select(); they're handled with the same code.
For more information on the ionotify(), poll(), and select() functions, see the QNX OS C Library Reference.
struct _io_notify {
uint16_t type;
uint16_t combine_len;
int32_t action;
int32_t flags;
struct __sigevent32 event;
/* The fields `mgr` to `timo` are only valid if (flags & _NOTIFY_COND_EXTEN)
* The full header must be present regardless of the flags. */
int32_t mgr[2]; /* For use by manager */
int32_t flags_extra_mask;
int32_t flags_exten;
int32_t nfds;
int32_t fd_first;
int32_t nfds_ready;
int64_t timo;
/* struct pollfd fds[nfds]; */
};
struct _io_notify64 {
uint16_t type;
uint16_t combine_len;
int32_t action;
int32_t flags;
struct __sigevent32 old_event;
/* The fields `mgr` to `timo` are only valid if (flags & _NOTIFY_COND_EXTEN)
* The full header must be present regardless of the flags. */
int32_t mgr[2]; /* For use by manager */
int32_t flags_extra_mask;
int32_t flags_exten;
int32_t nfds;
int32_t fd_first;
int32_t nfds_ready;
int64_t timo;
union {
struct __sigevent32 event32;
struct __sigevent64 event64;
};
/* struct pollfd fds[nfds]; */
};
struct _io_notify_reply {
uint32_t zero;
uint32_t flags; /* actions above */
int32_t flags2; /* flags above */
struct __sigevent32 event;
/* Following fields only updated by new managers (if valid) */
int32_t mgr[2]; /* For use by manager */
int32_t flags_extra_mask;
int32_t flags_exten;
int32_t nfds;
int32_t fd_first;
int32_t nfds_ready;
int64_t timo;
/* struct pollfd fds[nfds]; */
};
struct _io_notify_reply64 {
uint32_t zero;
uint32_t flags; /* actions */
int32_t flags2; /* flags above */
struct __sigevent32 old_event;
/* Following fields only updated by new managers (if valid) */
int32_t mgr[2]; /* For use by manager */
int32_t flags_extra_mask;
int32_t flags_exten;
int32_t nfds;
int32_t fd_first;
int32_t nfds_ready;
int64_t timo;
union {
struct __sigevent32 event32;
struct __sigevent64 event64;
};
/* struct pollfd fds[nfds]; */
};
typedef union {
struct _io_notify i;
struct _io_notify64 i64;
struct _io_notify_reply o;
struct _io_notify_reply64 o64;
} io_notify_t;
io_notify_t *msg
This is the pointer to the union containing the message. The main items in the input structure are:
- type
- combine_len
- action
- flags
- event32 or event64
The type member has the value _IO_NOTIFY or _IO_NOTIFY64.
The combine_len field has meaning for a combine message; see the Combine Messages chapter.
The action member is used by the iofunc_notify() helper function to tell it whether it should:
- just check for conditions now
- check for conditions now, and if none are met, arm them
- just arm for transitions
Since iofunc_notify() looks at this, you don't have to worry about it.
The flags member contains the conditions that the client is interested in and can be any mixture of the following:
- _NOTIFY_COND_INPUT
- This condition is met when there are one or more units of input data available (i.e., clients can now issue reads). The number of units defaults to 1, but you can change it. The definition of a unit is up to you: for a character device such as a serial port, it would be a character; for a POSIX message queue, it would be a message. Each resource manager selects an appropriate object.
- _NOTIFY_COND_OUTPUT
- This condition is met when there's room in the output buffer for one or more units of data (i.e., clients can now issue writes). The number of units defaults to 1, but you can change it. The definition of a unit is up to you—some resource managers may default to an empty output buffer, while others may choose some percentage of the buffer empty.
- _NOTIFY_COND_OBAND
- The condition is met when one or more units of out-of-band data are available. The number of units defaults to 1, but you can change it. The definition of out-of-band data is specific to the resource manager.
- _NOTIFY_COND_EXTEN
- The conditions are defined with some extended flags; used internally.
The event* member is what the resource manager delivers once a condition is met.
A resource manager needs to keep a list of clients that want to be notified as conditions are met, along with the events to use to do the notifying. When a condition is met, the resource manager must traverse the list to look for clients that are interested in that condition, and then deliver the appropriate event. As well, if a client closes its file descriptor, then any notification entries for that client must be removed from the list.
To make all this easier, the following structure and helper functions are provided for you to use in a resource manager:
- iofunc_notify_t structure
- Contains the three notification lists, one for each possible condition. Each is a list of the clients to be notified for that condition.
- iofunc_notify()
- Adds or removes notification entries; also polls for conditions. Call this function inside your io_notify handler function.
- iofunc_notify_trigger(), iofunc_notify_trigger_strict()
- Sends notifications to queued clients. Call either of these functions when one or more conditions have been met. When the client closes its file descriptor, call iofunc_notify_trigger_strict() for each condition.
- iofunc_notify_remove(), iofunc_notify_remove_strict()
- Removes notification entries from the list. Call iofunc_notify_remove_strict() when the client closes its file descriptor.
In a multi-threaded server, you must serialize access to each iofunc_notify_t structure. In many cases, this is done implicitly by the default locking on the extended attribute structure that holds device-specific data and notification data copied from that other structure. (For an example of this implicit locking, see the io_write handler in the coding example shown in the next section.) But if an iofunc_notify_t structure is accessed somewhere that this default locking has not occurred (e.g., in an interrupt-handling thread), then explicit locking must be done, either of the extended attribute structure or with an explicitly added lock for the iofunc_notify_t structure.
Failure to serialize access to these structures can lead to server-side race conditions, which can introduce bugs that are difficult to detect and fix.