]> git.saurik.com Git - apple/xnu.git/blob - osfmk/man/mach_msg.html
xnu-2422.1.72.tar.gz
[apple/xnu.git] / osfmk / man / mach_msg.html
1 <h2>mach_msg</h2>
2 <hr>
3 <p>
4 <strong>System Trap</strong> / <strong>Function</strong> - Send and/or receive a message from the target port.
5 <h3>SYNOPSIS</h3>
6 <pre>
7 <strong>mach_msg_return_t mach_msg</strong>
8 <strong>(mach_msg_header_t</strong> <var>msg</var>,
9 <strong>mach_msg_option_t</strong> <var>option</var>,
10 <strong>mach_msg_size_t</strong> <var>send_size</var>,
11 <strong>mach_msg_size_t</strong> <var>receive_limit</var>,
12 <strong>mach_port_t</strong> <var>receive_name</var>,
13 <strong>mach_msg_timeout_t</strong> <var>timeout</var>,
14 <strong>mach_port_t</strong> <var>notify</var><strong>);</strong>
15
16 <strong>mach_msg_return_t mach_msg_overwrite</strong>
17 <strong>(mach_msg_header_t*</strong> <var>send_msg</var>,
18 <strong>mach_msg_option_t</strong> <var>option</var>,
19 <strong>mach_msg_size_t</strong> <var>send_size</var>,
20 <strong>mach_msg_size_t</strong> <var>receive_limit</var>,
21 <strong>mach_port_t</strong> <var>receive_name</var>,
22 <strong>mach_msg_timeout_t</strong> <var>timeout</var>,
23 <strong>mach_port_t</strong> <var>notify</var>,
24 <strong>mach_msg_header_t</strong> <var>*receive_msg</var>,
25 <strong>mach_msg_size_t</strong> <var>receive_msg_size</var><strong>);</strong>
26 </pre>
27 <h3>PARAMETERS</h3>
28 <dl>
29 <p>
30 <dt> <var>msg</var>
31 <dd>
32 [pointer to in/out structure containing random and reply rights] A
33 message buffer used by <strong>mach_msg</strong> both for send and receive. This must
34 be naturally aligned.
35 <p>
36 <dt> <var>send_msg</var>
37 <dd>
38 [pointer to in structure containing random and reply rights] The mes-
39 sage buffer to be sent. This must be naturally aligned.
40 <p>
41 <dt> <var>option</var>
42 <dd>
43 [in scalar] Message options are bit values, combined with bitwise-or.
44 One or both of MACH_SEND_MSG and MACH_RCV_MSG should be used. Other
45 options act as modifiers.
46 <p>
47 <dt> <var>send_size</var>
48 <dd>
49 [in scalar] When sending a message, specifies the size of the message
50 buffer to be sent (the size of the header and body) in
51 bytes. Otherwise zero should be supplied.
52 <p>
53 <dt> <var>receive_limit</var>
54 <dd>
55 [in scalar] When receiving a message, specifies the maximum size of
56 the msg or receive_msg buffer in bytes. Otherwise zero should be sup-
57 plied.
58 <p>
59 <dt> <var>receive_name</var>
60 <dd>
61 [in random right] When receiving a message, specifies the port or port
62 set. Otherwise MACH_PORT_NULL should be supplied.
63 <p>
64 <dt> <var>timeout</var>
65 <dd>
66 [in scalar] When using the MACH_SEND_TIMEOUT and MACH_RCV_TIMEOUT
67 options, specifies the time in milliseconds to wait before giving
68 up. Otherwise MACH_MSG_TIMEOUT_NONE should be supplied.
69 <p>
70 <dt> <var>notify</var>
71 <dd>
72 [in notify receive right] When using the MACH_SEND_CANCEL and
73 MACH_RCV_NOTIFY options, specifies the port used for the
74 notification. Otherwise MACH_PORT_NULL should be supplied.
75 <p>
76 <dt> <var>receive_msg</var>
77 <dd>
78 [pointer to in/out structure] A message buffer into which a message
79 (header and body) will be received. This must be naturally aligned. By
80 default (<strong>mach_msg</strong>), any received message will overwrite the send
81 message buffer. This buffer is in/out only if the MACH_RCV_OVERWRITE
82 option is used; otherwise this buffer is out only.
83 <p>
84 <dt> <var>receive_msg_size</var>
85 <dd>
86 [in scalar] When using the MACH_RCV_OVERWRITE option, specifies the
87 size (in bytes) of the receive "message" that is to be used by
88 <strong>mach_msg</strong> to indicate the disposition of received out-of-line regions.
89 </dl>
90 <h3>DESCRIPTION</h3>
91 <p>
92 The <strong>mach_msg</strong> system call sends and receives Mach messages. Mach
93 messages contain data, which can include port rights and addresses of
94 large regions of memory. <strong>mach_msg</strong> uses the same buffer for sending and
95 receiving a message; the other calls permit separate send and receive
96 buffers (although they may be specified to be the same).
97 If the option argument contains MACH_SEND_MSG, the call sends a
98 message. The <var>send_size</var> argument specifies the size of the message
99 buffer (header and body) to send. The msgh_remote_port field of the
100 message header specifies the destination of the message.
101 If the option argument contains MACH_RCV_MSG, it receives a
102 message. The receive_limit argument specifies the size of a buffer
103 that will receive the message; messages that are larger are not
104 received. The receive_name argument specifies the port or port set
105 from which to receive.
106 <p>
107 If the option argument contains both MACH_SEND_MSG and MACH_RCV_MSG,
108 then <strong>mach_msg</strong> does both send and receive operations (in that
109 order). If the send operation encounters an error (any return code
110 other than MACH_MSG_SUCCESS), the call returns immediately
111 without attempting the receive operation. Semantically the combined
112 call is equivalent to separate send and receive calls, but it saves
113 a system call and enables other internal optimizations.
114
115 If the option argument specifies neither MACH_SEND_MSG nor
116 MACH_RCV_MSG, <strong>mach_msg</strong> does nothing.
117 Some options, like MACH_SEND_TIMEOUT and MACH_RCV_TIMEOUT, share a
118 supporting argument. If these options are used together, they make
119 independent use of the supporting argument's value.
120 <h3>NOTES</h3>
121 <p>
122 The Mach kernel provides message-oriented, capability-based
123 inter-process communication. The inter-process communication (IPC)
124 primitives efficiently support many different styles of interaction,
125 including remote procedure calls, object-oriented distributed
126 programming, streaming of data, and sending very large amounts of
127 data.
128 <h4>Major Concepts</h4>
129 <p>
130 The IPC primitives operate on three abstractions: messages, ports, and
131 port sets. User tasks access all other kernel services and
132 abstractions via the IPC primitives.
133 <p>
134 The message primitives let tasks send and receive messages. Tasks send
135 messages to ports. Messages sent to a port are delivered reliably
136 (messages may not be lost) and are received in the order in which they
137 were sent via send rights by a given sending task (or a given
138 kernel). (Messages sent to send-once rights are unordered.)
139 <p>
140 Messages
141 contain a fixed-size header and a variable-sized message body
142 containing kernel and user data, and a variable-size trailer of kernel
143 appended message attributes. The header describes the destination
144 and the size of the message (header plus body). The message body
145 contains descriptions of additional port rights to be transmitted,
146 descriptions of "out-of-line" memory regions to be sent and a
147 variable amount of user data, which typically includes type conversion
148 information. The out-of-line memory regions (including out-of-line
149 port arrays) are (typically) disjoint from the message body.
150 The IPC implementation makes use of the VM system to efficiently
151 transfer large amounts of data. The message can contain the addresses
152 of regions of the sender's address space which should be transferred
153 as part of the message.
154 <p>
155 When a task receives a message containing
156 such out-of-line regions of data, the data can appear in unused
157 portions or overwrite an existing portion of the receiver's address
158 space (depending on the requested receive options). Under favorable
159 circumstances, the transmission of out-of-line data is optimized so
160 that sender and receiver share the physical pages of data
161 copy-on-write, and no actual data copy occurs unless the pages are
162 written. Regions of memory up to 4 gigabytes may be sent in this
163 manner.
164 <p>
165 Ports hold a queue of messages. Tasks operate on a port to send and
166 receive messages by exercising capabilities (rights) for the
167 port. Multiple tasks can hold send rights for a port.
168 Tasks can also
169 hold send-once rights, which grant the ability to send a single
170 message. Only one task can hold the receive capability (receive
171 right) for a port.
172 <p>
173 Port rights can be transferred between tasks via
174 messages. The sender of a message can specify in the message that the
175 message contains a port right. If a message contains a receive right
176 for a port, the receive right is removed from the sender of the
177 message and transferred to the receiver of the
178 message. While the receive right is in transit, tasks holding send
179 rights can still send messages to the port, and they are queued until
180 a task acquires the receive right and uses it to receive the messages.
181 <p>
182 Tasks can receive messages from ports and port sets. The port set
183 abstraction allows a single thread to wait for a message from any of
184 several ports. Tasks manipulate port sets with a port set name,
185 which is taken from the same name space as are the port rights. The
186 port-set name may not be transferred in a message. A port set holds
187 receive rights, and a receive operation on a port set blocks waiting
188 for a message sent to any of the constituent ports. A port may not be-
189 long to more than one port set, and if a port is a member of a port
190 set, the holder of the receive right can't receive directly from the
191 port.
192 <p>
193 Port rights are a secure, location-independent way of naming
194 ports. The port queue is a protected data structure, only accessible
195 via the kernel's exported message primitives. Rights are also
196 protected by the kernel; there is no way for a malicious user task to
197 guess a port's internal name and send a message to a port to which it
198 shouldn't have access. Port rights do not carry any location in-
199 formation. When a receive right for a port moves from task to task,
200 and even between tasks on different machines, the send rights for
201 the port remain unchanged and continue to function.
202 <h4>Port Rights</h4>
203 <p>
204 Each task has its own space of port rights. Port rights are named with
205 positive (unsigned) integers. For all architectures, sizeof
206 (mach_port_t) = sizeof (mach_port_name_t) = sizeof (void*) and so user
207 space addresses may be used as port names, except for the reserved
208 values MACH_PORT_NULL (0) and MACH_PORT_DEAD (all 1 bits). When the
209 kernel chooses a name for a new right, however, it is free to pick any
210 unused name (one which denotes no right) in the space.
211 <p>
212 There are three basic kinds of rights: receive rights, send rights and
213 send-once rights. A port name can name any of these types of rights,
214 or name a port-set, be a dead name, or name nothing. Dead names are
215 not capabilities. They act as place-holders to prevent a name from
216 being otherwise used.
217 <p>
218 A port is destroyed, or dies, when its receive right is
219 de-allocated. When a port dies, send and send-once rights for the port
220 turn into dead names. Any messages queued at the port are destroyed,
221 which de-allocates the port rights and out-of-line memory in the
222 messages.
223 <p>
224 Each send-once right held by a task has a different name. In contrast,
225 when a task holds send rights or a receive right for a port, the
226 rights share a single name.
227 <p>
228 Tasks may hold multiple user-references for send rights. When a task
229 receives a send right which it already holds, the kernel increments
230 the right's user-reference count. When a task de-allocates a send
231 right, the kernel decrements its user-reference count, and the task
232 only loses the send right when the count goes to zero.
233 <p>
234 Send-once rights always have a user reference count of one. Tasks may
235 hold multiple user references for dead names.
236 Each send-once right generated guarantees the receipt of a single
237 message, either a message sent to that send-once right or, if the
238 send-once right is in any way destroyed, a send-once notification.
239 <p>
240 A message can carry port rights; the msgh_remote or msgh_local fields
241 in the message header or the disposition field in a message body
242 descriptor specify the type of port right and how the port right is to
243 be extracted from the caller. The values MACH_PORT_NULL and
244 MACH_PORT_DEAD are valid in place of a port right in a message body.
245 <p>
246 In a sent message, the following mach_msg_type_name_t values denote
247 port rights:
248 <dl>
249 <dt> MACH_MSG_TYPE_MAKE_SEND
250 <dd>
251 The message will carry a send right, but the caller must supply a
252 receive right. The send right is created from the receive right, and the
253 receive right's make-send count is incremented.
254 <dt> MACH_MSG_TYPE_COPY_SEND
255 <dd>
256 The message will carry a send right, and the caller must supply a send
257 right. The user reference count for the supplied send right is not
258 changed. The caller may also supply a dead name and the receiving
259 task will get MACH_PORT_DEAD.
260 <dt> MACH_MSG_TYPE_MOVE_SEND
261 <dd>
262 The message will carry a send right, and the caller must supply a send
263 right. The user reference count for the supplied send right is
264 decremented, and the right is destroyed if the count becomes
265 zero. Unless a receive right remains, the name becomes available for
266 recycling. The caller may also supply a dead name, which loses a user
267 reference, and the receiving task will get MACH_PORT_DEAD.
268 <dt> MACH_MSG_TYPE_MAKE_SEND_ONCE
269 <dd>
270 The message will carry a send-once right, but the caller must supply a
271 receive right. The send-once right is created from the receive right.
272 Note that send once rights can only be created from the receive right.
273 <dt> MACH_MSG_TYPE_MOVE_SEND_ONCE
274 <dd>
275 The message will carry a send-once right, and the caller must supply a
276 send-once right. The caller loses the supplied send-once right. The
277 caller may also supply a dead name, which loses a user reference,
278 and the receiving task will get MACH_PORT_DEAD.
279 <dt> MACH_MSG_TYPE_MOVE_RECEIVE
280 <dd>
281 The message will carry a receive right, and the caller must supply a
282 receive right. The caller loses the supplied receive right, but
283 retains any send rights with the same name. The make-send count and
284 sequence number of the receive right are reset to zero and
285 no-more-senders notification requests are cancelled (with a
286 send-once notification being sent to the no-more-senders notification
287 right), but the port retains other attributes like queued messages
288 and extant send and send-once rights.
289 If a message carries a send or send-once right, and the port dies
290 while the message is in transit, then the receiving task will get
291 MACH_PORT_DEAD instead of a right.
292 </dl>
293 <p>
294 The following mach_msg_type_name_t values in a received message
295 indicate that it carries port rights:
296 <dl>
297 <dt> MACH_MSG_TYPE_PORT_SEND
298 <dd>
299 This value is an alias for MACH_MSG_TYPE_MOVE_SEND. The
300 message carried a send right. If the receiving task already has send and/
301 or receive rights for the port, then that name for the port will be reused.
302 Otherwise, the right will have a new, previously unused, name. If the
303 task already has send rights, it gains a user reference for the right (un-
304 less this would cause the user-reference count to overflow). Otherwise,
305 it acquires send rights, with a user-reference count of one.
306 <dt> MACH_MSG_TYPE_PORT_SEND_ONCE
307 <dd>
308 This value is an alias for MACH_MSG_TYPE_MOVE_SEND_ONCE. The message
309 carried a send-once right. The right will have a new, previously
310 unused, name.
311 <dt> MACH_MSG_TYPE_PORT_RECEIVE
312 <dd>
313 This value is an alias for MACH_MSG_TYPE_MOVE_RECEIVE. The message
314 carried a receive right. If the receiving task already has send rights
315 for the port, then that name for the port will be reused; otherwise,
316 the right will have a new, previously unused name.
317 </dl>
318 <p>
319 It is also possible to send a (nearly unbounded) array of port rights
320 "out-of-line". All of the rights named by the array must be of the
321 same type. The array is physically copied with the message body
322 proper. The array of port right (names) can be received by the
323 receiver using the same options available for out-of-line data
324 reception described below.
325 <h4>Memory</h4>
326 <p>
327 A message can contain one or more regions of the sender's address
328 space which are to be transferred as part of the message. The message
329 carries a logical copy of the memory. For this "out-of-line" memory,
330 the kernel can copy the data or use virtual memory techniques to defer
331 any actual page copies unless the sender or the receiver modifies
332 the data, the physical pages remain shared.
333 <p>
334 The sender of the message must explicitly request an out-of-line
335 transfer. Such a region is described as an arbitrary region of the
336 sender's address space. The sender always sees this memory as being
337 copied to the receiver.
338 <p>
339 For each region, the sender has a de-allocate option. If the option is
340 set and the out-of-line memory region is not null, then the region is
341 implicitly de-allocated from the sender, as if by vm_deallocate. In
342 particular, the start address is truncated down and the end address
343 rounded up so that every page overlapped by the memory region is
344 de-allocated (thereby possibly de-allocating more memory than is
345 effectively transmitted). The use of this option effectively changes
346 the memory copy to a memory movement. Aside from possibly optimizing
347 the sender's use of memory, the de-allocation option allows the kernel
348 to more efficiently handle the transfer of memory.
349 <p>
350 For each region, the sender has the choice of permitting the kernel to
351 choose a transmission strategy or the choice of requiring physical
352 copy:
353 <dl>
354 <dt> MACH_MSG_VIRTUAL_COPY
355 <dd>
356 In a sent message, this flag allows the kernel to choose any mechanism
357 to transmit the data. For large regions, this involves constructing a
358 virtual copy of the pages containing the region. The portion of the
359 first page preceding the data and the portion of the last page
360 following the data are not copied (and will appear as zero if the
361 virtual copy is dynamically allocated in the receiver).
362 <p>
363 In a received message, this flag indicates that the kernel transmitted
364 a virtual copy. Access to the received memory may involve interactions
365 with the memory manager managing the sender's original data. Integri-
366 ty-conscious receivers should exercise caution when dealing with out-
367 of-line memory from un-trustworthy sources. Receivers concerned about
368 deterministic access time should also exercise caution. The dynamic
369 allocation option guarantees that the virtual copy will not be di-
370 rectly referenced during the act of receiving the message.
371 <dt> MACH_MSG_PHYSICAL_COPY
372 <dd>
373 In a sent message, this flag requires that the kernel construct an
374 actual copy of the memory (either into wired kernel memory or default
375 memory managed space). There is a (fairly large) limit on the amount
376 of data that can be physically copied in a message. Port arrays always
377 assume this option when sent.
378 <p>
379 In a received message, this flag indicates that the kernel did
380 transmit a physical copy.
381 </dl>
382 <p>
383 The receiver has two options for the reception of out-of-line memory
384 (or "out-of-line" port arrays): allocation and overwrite.
385 In the absence of the MACH_RCV_OVERWRITE option, all out-of-line re-
386 gions are dynamically allocated. Allocated out-of-line memory arrives
387 somewhere in the receiver's address space as new memory. It has the
388 same inheritance and protection attributes as newly vm_allocate'ed
389 memory. The receiver has the responsibility of de-allocating (with
390 vm_deallocate) the memory when it is no longer needed. If the message
391 contains more than one region, each will be allocated its own region,
392 not necessarily contiguously. If the sender's data was transmitted as
393 a virtual copy the allocated region will have the same data alignment
394 within the page; otherwise, the received data will appear starting at
395 the beginning of a page.
396 <p>
397 If the MACH_RCV_OVERWRITE option is set, the receiver can specify how
398 each received region is to be processed (dynamically allocated as
399 described above, or written over existing memory). With this option,
400 the contents of the receive buffer (receive_msg) are examined by the
401 kernel. The kernel scans the descriptors in the receive buffer
402 "message" to determine how to handle each out-of-line region. (Note:
403 whereas receive_limit is the maximum size of the receive buffer,
404 receive_msg_size is the amount filled in with this "message".) The
405 kernel uses each out-of-line data descriptor (in order) to specify
406 the processing for each received data region in turn, each out-of-line
407 port array descriptor is used correspondingly. (Intermingled port
408 descriptors are ignored when matching descriptors between the
409 incoming message and the receive buffer list.)
410 <p>
411 The copy option in the
412 matching descriptor specifies the processing:
413 <dl>
414 <dt> MACH_MSG_OVERWRITE
415 <dd>
416 This flag indicates that the region should write over a specified
417 region of the receiver's address space, as indicated by the address
418 and size/ count fields of the descriptor. The full range overwritten
419 must already exist (be allocated or mapped) in the receiver's address
420 space. Depending on the nature of the data transmission this
421 overwrite may involve virtual memory manipulations or it may involve
422 actual data copy.
423 <dt> MACH_MSG_ALLOCATE
424 <dd>
425 This flag indicates that the region is to be dynamically allocated. No
426 other descriptor values are relevant.
427 </dl>
428 <p>
429 If not enough descriptors appear in the receive buffer to describe all
430 received regions, additional regions are dynamically allocated. If
431 the receiver specifies more descriptors than there are regions in the
432 received message, the additional descriptors are ignored (and do not
433 appear in the final received message).
434 <p>
435 Note that the receive buffer descriptors will be overwritten:
436 The size fields in descriptors will be updated (when scanned, they
437 specified the maximum sizes of regions, when received, they specify
438 the actual sizes of received regions).
439 The copy fields in descriptors will be updated (when scanned, they
440 specified allocate versus overwrite, when received, they indicate
441 whether the region was physically or virtually copied).
442 The descriptors may appear in different positions (given intermingled
443 port descriptors).
444 Descriptors that were not used (because there were not that many
445 received regions) will be discarded.
446 <p>
447 Null out-of-line memory is legal. If the out-of-line region size is
448 zero, then the region's specified address is ignored. A receive
449 allocated null out-of-line memory region always has a zero address.
450 Unaligned addresses and region sizes that are not page multiples are
451 legal. A received message can also contain regions with unaligned
452 addresses and sizes which are not multiples of the page size.
453 <h4>Message Send</h4>
454 <p>
455 The send operation queues a message to a port. The message carries a
456 copy of the caller's data. After the send, the caller can freely
457 modify the message buffer or the out-of-line memory regions and the
458 message contents will remain unchanged.
459 <p>
460 The message carries with it the security ID of the sender, which the
461 receiver can request in the message trailer.
462 <p>
463 Message delivery is reliable and sequenced. Reception of a message
464 guarantees that all messages previously sent to the port by a single
465 task (or a single kernel) via send rights have been received and that
466 they are received in the order in which they were sent. Messages sent
467 to send-once rights are unordered.
468 <p>
469 If the destination port's queue is full, several things can happen. If
470 the message is sent to a send-once right (msgh_remote_port carries a
471 send-once right), then the kernel ignores the queue limit and delivers
472 the message. Otherwise the caller blocks until there is room in the
473 queue, unless the MACH_SEND_TIMEOUT option is used. If a port has
474 several blocked senders, then any of them may queue the next message
475 when space in the queue becomes available, with the proviso that a
476 blocked sender will not be indefinitely starved.
477 These options modify MACH_SEND_MSG. If MACH_SEND_MSG is not also
478 specified, they are ignored.
479 <dl>
480 <dt> MACH_SEND_TIMEOUT
481 <dd>
482 The timeout argument should specify a maximum time (in milliseconds)
483 for the call to block before giving up. If the message can't be queued
484 before the timeout interval elapses, then the call returns
485 MACH_SEND_TIMED_OUT. A zero timeout is legitimate.
486 <dt> MACH_SEND_INTERRUPT
487 <dd>
488 If specified, the <strong>mach_msg</strong> call will return
489 MACH_SEND_INTERRUPTED if a software interrupt aborts the call.
490 Otherwise, the send operation will be retried.
491 <dt> MACH_SEND_TRAILER
492 <dd>
493 If set, the kernel, instead of determining the message attributes
494 itself, will accept a formatted message trailer from the sender. The
495 supplied trailer must be of the latest version supported by the
496 kernel, and must contain all message attributes defined by the
497 kernel. Only tasks with a security ID of KERNEL_SECURITY_ID can use
498 this option; the intended use of this option is in support of the
499 Net Message server. The trailer must follow the message in memory as
500 it would appear in a received message. (The send_size argument to
501 <strong>mach_msg</strong> still indicates the size of the message proper, not including
502 this trailer.)
503 </dl>
504 <p>
505 The queueing of a message carrying receive rights may create a
506 circular loop of receive rights and messages, which can never be
507 received. For example, a message carrying a receive right can be
508 sent to that receive right. This situation is not an error, but the
509 kernel will garbage-collect such loops, destroying the messages.
510 Some return codes, like MACH_SEND_TIMED_OUT, imply that the message
511 was almost sent, but could not be queued. In these situations, the
512 kernel tries to return the message contents to the caller with a
513 pseudo-receive operation. This prevents the loss of port rights or
514 memory which only exist in the message, for example, a receive right
515 which was moved into the message, or out-of-line memory sent with
516 the de-allocate option.
517 <p>
518 The intent of the pseudo-receive operation is to restore, as best as
519 possible, the state prior to attempting the send. This involves
520 restoring the port rights and out-of-line memory regions contained in
521 the message. The port right names and out-of-line addresses in the
522 message send buffer are updated to reflect the new values resulting
523 from their effective reception. The pseudo-receive handles the des-
524 tination and reply rights as any other rights; they are not reversed
525 as is the appearance in a normal received message. Also, no trailer is
526 appended to the message. After the pseudo-receive, the message is
527 ready to be resent. If the message is not resent, note that
528 out-of-line memory regions may have moved and some port rights may
529 have changed names.
530 <p>
531 Although unlikely, the pseudo-receive operation may encounter resource
532 shortages. This is similar to a MACH_RCV_BODY_ERROR return code from
533 a receive operation. When this happens, the normal send return codes
534 are augmented with the MACH_MSG_IPC_SPACE, MACH_MSG_VM_SPACE,
535 MACH_MSG_IPC_KERNEL and MACH_MSG_VM_KERNEL bits to indicate the
536 nature of the resource shortage.
537 <h4>Message Receive</h4>
538 <p>
539 The receive operation de-queues a message from a port. The receiving
540 task acquires the port rights and out-of-line memory regions carried
541 in the message.
542 The receive_name argument specifies a port or port set from which to
543 receive. If a port is specified, the caller must possess the receive
544 right for the port and the port must not be a member of a port set. If
545 no message is present, the call blocks, subject to the
546 MACH_RCV_TIMEOUT option.
547 <p>
548 If a port set is specified, the call will receive a message sent to
549 any of the member ports. It is permissible for the port set to have
550 no member ports, and ports may be added and removed while a receive
551 from the port set is in progress. The received message can come from
552 any of the member ports which have messages, with the proviso that a
553 member port with messages will not be indefinitely starved. The
554 msgh_local_port field in the received message header specifies from
555 which port in the port set the message came.
556 <p>
557 The receive_limit argument specifies the size of the caller's message
558 buffer (which must be big enough for the message header, body and
559 trailer); the msgh_size field of the received message indicates the
560 actual size of the received message header and body. The <strong>mach_msg</strong> call
561 will not receive a message larger than receive_limit. Messages that
562 are too large are destroyed, unless the MACH_RCV_LARGE option is used.
563 Following the received data, at the next natural boundary, is a
564 message trailer. The msgh_size field of the received message does not
565 include the length of this trailer; the trailer's length is given by
566 the msgh_trailer_size field within the trailer. The receiver of a
567 message is given a choice as to what trailer format is desired, and,
568 within that format, which of the leading trailer attributes are
569 desired (that is, to get trailer element three, the receiver must also
570 accept elements one and two). For any given trailer format (of which
571 there is currently only one), the trailer is compatibly extended by
572 adding additional elements to the end.
573 <p>
574 Received messages are stamped (in the trailer) with a sequence number,
575 taken from the port from which the message was received. (Messages
576 received from a port set are stamped with a sequence number from the
577 appropriate member port.) Newly created ports start with a zero
578 sequence number, and the sequence number is reset to zero whenever the
579 port's receive right moves between tasks. When a message is de-queued
580 from the port, it is stamped with the port's sequence number and the
581 port's sequence number is then incremented. (Note that this occurs
582 whether or not the receiver requests the sequence number in the trail-
583 er.) The de-queue and increment operations are atomic, so that
584 multiple threads receiving messages from a port can use the msgh_seqno
585 field to reconstruct the original order of the messages.
586 <p>
587 The destination and reply ports are reversed in a received message
588 header. The msgh_local_port field carries the name of the destination
589 port, from which the message was received, and the msgh_remote_port
590 field carries the reply port right. The bits in msgh_bits are also
591 reversed. The MACH_MSGH_BITS_LOCAL bits have a value of
592 MACH_MSG_TYPE_PORT_SEND_ONCE or MACH_MSG_TYPE_PORT_SEND depending on
593 the type of right to which the message was sent. The
594 MACH_MSGH_BITS_REMOTE bits describe the reply port right.
595 <p>
596 A received message can contain port rights and out-of-line memory. The
597 msgh_local_port field does not carry a port right; the act of
598 receiving the message consumes the send or send-once right for the
599 destination port. The msgh_remote_port field does carry a port right,
600 and the message can carry additional port rights and memory if the
601 MACH_MSGH_BITS_COMPLEX bit is set. Received port rights and memory
602 should be consumed or de-allocated in some fashion.
603 In almost all cases, msgh_local_port will specify the name of a
604 receive right, either receive_name, or, if receive_name is a port
605 set, a member of receive_name.
606 <p>
607 If other threads are concurrently
608 manipulating the receive right, the situation is more complicated. If
609 the receive right is renamed during the call, then msgh_local_port
610 specifies the right's new name. If the caller loses the receive right
611 after the message was de-queued from it, then <strong>mach_msg</strong> will proceed
612 instead of returning MACH_RCV_PORT_DIED. If the receive right was
613 destroyed, then msgh_local_port specifies MACH_PORT_DEAD. If the
614 receive right still exists, but isn't held by the caller, then
615 msgh_local_port specifies MACH_PORT_NULL.
616 <p>
617 The following options modify MACH_RCV_MSG. If MACH_RCV_MSG is not also
618 specified, they are ignored.
619 <dl>
620 <dt> MACH_RCV_TIMEOUT
621 <dd>
622 The timeout argument should specify a maximum time (in milliseconds)
623 for the call to block before giving up. If no message arrives before
624 the timeout interval elapses, then the call returns
625 MACH_RCV_TIMED_OUT. A zero timeout is legitimate.
626 <dt> MACH_RCV_NOTIFY
627 <dd>
628 The notify argument should specify a receive right for a notify
629 port. If receiving the reply port creates a new port right in the
630 caller, then the notify port is used to request a dead-name
631 notification for the new port right.
632 <dt> MACH_RCV_INTERRUPT
633 <dd>
634 If specified, the <strong>mach_msg</strong> call will return MACH_RCV_INTERRUPTED if a
635 software interrupt aborts the call. Otherwise, the receive operation
636 will be retried.
637 <dt> MACH_RCV_OVERWRITE
638 <dd>
639 If specified, the message buffer specified by receive_msg (or msg), of
640 length receive_msg_size, will be scanned for out-of-line descriptors to
641 specify the processing to be done when receiving out-of-line regions.
642 This option is only allowed for <strong>mach_msg_overwrite</strong>.
643 <dt> MACH_RCV_LARGE
644 <dd>
645 If the message is larger than receive_limit or an out-of-line region
646 is larger than the size allowed by a corresponding receive descriptor
647 (MACH_RCV_OVERWRITE), the message remains queued instead of being
648 destroyed. If the header, trailer and body would not fit into
649 receive_limit, only the message header (mach_msg_header) and trailer
650 header (mach_msg_trailer) are returned with the actual size of the
651 message returned in the msgh_size field, the actual size of the
652 trailer returned in the msgh_trailer_size field and an error return
653 value of MACH_RCV_TOO_LARGE. If receive_limit is sufficient but an
654 out-of-line descriptor is not, the message header, trailer and body
655 are received, with out-of-line descriptors set to indicate the
656 nature and size of the out-of-line regions, with an error return of
657 MACH_RCV_SCATTER_SMALL. No out-of-line regions or port rights
658 (including the reply right) will be received. If this option is not
659 specified, messages too large will be de-queued and then destroyed;
660 the caller receives the message header, with all fields correct,
661 including the destination port but excepting the reply port, which is
662 MACH_PORT_NULL and an empty (no additional element) message trailer.
663 <dt> MACH_RCV_TRAILER_TYPE(value)
664 <dd>
665 This macro encodes the type of trailer the kernel must return with the
666 message. If the kernel does not recognize this type, it returns
667 MACH_RCV_INVALID_TRAILER. Currently, only MACH_MSG_TRAILER_FORMAT_0 is
668 supported.
669 <dt> MACH_RCV_TRAILER_ELEMENTS(value)
670 <dd>
671 This macro encodes the number of trailer elements desired. If the ker-
672 nel does not support this number for the requested trailer type, the
673 kernel returns MACH_RCV_INVALID_TRAILER. Zero is a legal value.
674 </dl>
675 <p>
676 The following trailer elements are supported:
677 <dl>
678 <dt> MACH_RCV_TRAILER_SEQNO
679 <dd>
680 Returns the sequence number of the message relative to its port. This
681 value is of type mach_port_seqno_t.
682 <dt> MACH_RCV_TRAILER_SENDER
683 <dd>
684 Returns the security ID of the task that sent the message. This value
685 is of type security_id_t.
686 </dl>
687 <p>
688 If a resource shortage prevents the reception of a port right, the
689 port right is destroyed and the caller sees the name
690 MACH_PORT_NULL. If a resource shortage prevents the reception of an
691 out-of-line memory region, the region is destroyed and the caller sees
692 a zero address. In addition, the corresponding element in the size
693 array is set to zero. A task never receives port rights or memory for
694 which it is not told.
695 <p>
696 The MACH_RCV_HEADER_ERROR return code indicates a resource shortage
697 in the reception of the message header. The reply port and all port
698 rights and memory in the message are destroyed. The caller receives
699 the message header with all fields correct except for the reply
700 port.
701 <p>
702 The MACH_RCV_BODY_ERROR return code indicates a resource shortage in
703 the reception of the message body. The message header, including the
704 reply port, is correct. The kernel attempts to transfer all port
705 rights and memory regions in the body, and only destroys those that
706 can't be transferred.
707 <h4>Atomicity</h4>
708 <p>
709 The <strong>mach_msg</strong> call handles port rights in the message header
710 atomically. Out-of-line memory and port rights in the message body do
711 not enjoy this atomicity guarantee. These elements may be processed
712 front-to-back, back-to-front, in some random order, or even
713 atomically.
714 <p>
715 For example, consider sending a message with the destination port
716 specified as MACH_MSG_TYPE_MOVE_SEND and the reply port specified as
717 MACH_MSG_TYPE_COPY_SEND. The same send right, with one user-refer-
718 ence, is supplied for both the msgh_remote_port and msgh_local_port
719 fields. Because <strong>mach_msg</strong> processes the port rights atomically, this
720 succeeds. If msgh_remote_port were processed before msgh_local_port,
721 then <strong>mach_msg</strong> would return MACH_SEND_INVALID_REPLY in this situation.
722 <p>
723 On the other hand, suppose the destination and reply port are both
724 specified as MACH_MSG_TYPE_MOVE_SEND, and again the same send right
725 with one user-reference is supplied for both. Now the send operation
726 fails, but because it processes the rights atomically, <strong>mach_msg</strong> can
727 return either MACH_SEND_INVALID_DEST or MACH_SEND_INVALID_REPLY.
728 <p>
729 For example, consider receiving a message at the same time another
730 thread is deallocating the destination receive right. Suppose the
731 reply port field carries a send right for the destination port. If the
732 de-allocation happens before the dequeuing, the receiver gets
733 MACH_RCV_PORT_DIED. If the de-allocation happens after the receive,
734 the msgh_local_port and the msgh_remote_port fields both specify
735 the same right, which becomes a dead name when the receive right is
736 de-allocated. If the de-allocation happens between the de-queue and
737 the receive, the msgh_local_port and msgh_remote_port fields both
738 specify MACH_PORT_DEAD. Because the rights are processed atomically,
739 it is not possible for just one of the two fields to hold
740 MACH_PORT_DEAD.
741 <p>
742 The MACH_RCV_NOTIFY option provides a more likely example. Suppose a
743 message carrying a send-once right reply port is received with
744 MACH_RCV_NOTIFY at the same time the reply port is destroyed. If the
745 reply port is destroyed first, then msgh_remote_port specifies
746 MACH_PORT_DEAD and the kernel does not generate a dead-name
747 notification. If the reply port is destroyed after it is received,
748 then msgh_remote_port specifies a dead name for which the kernel
749 generates a dead-name notification. Either the reply port is dead on
750 arrival or notification is requested.
751 <h4>Implementation</h4>
752 <p>
753 <strong>mach_msg</strong> and <strong>mach_msg_overwrite</strong> are wrappers for a system call. They
754 have the responsibility for repeating the interrupted system call.
755 <h3>CAUTIONS</h3>
756 <p>
757 If MACH_RCV_TIMEOUT is used without MACH_RCV_INTERRUPT, then the
758 timeout duration might not be accurate. When the call is interrupted
759 and automatically retried, the original timeout is used. If
760 interrupts occur frequently enough, the timeout interval might never
761 expire. MACH_SEND_TIMEOUT without MACH_SEND_INTERRUPT suffers from the
762 same problem.
763 <h3>RETURN VALUES</h3>
764 <p>
765 The send operation can generate the following return codes. These
766 return codes imply that the call did nothing:
767 <dl>
768 <p>
769 <dt> MACH_SEND_MSG_TOO_SMALL
770 <dd>
771 The specified send_size was smaller than the minimum size for a
772 message.
773 <p>
774 <dt> MACH_SEND_NO_BUFFER
775 <dd>
776 A resource shortage prevented the kernel from allocating a message
777 buffer.
778 <p>
779 <dt> MACH_SEND_INVALID_DATA
780 <dd>
781 The supplied message buffer was not readable.
782 <p>
783 <dt> MACH_SEND_INVALID_HEADER
784 <dd>
785 The msgh_bits value was invalid.
786 <p>
787 <dt> MACH_SEND_INVALID_DEST
788 <dd>
789 The msgh_remote_port value was invalid.
790 <p>
791 <dt> MACH_SEND_INVALID_NOTIFY
792 <dd>
793 When using MACH_SEND_CANCEL, the notify argument did not
794 denote a valid receive right.
795 <p>
796 <dt> MACH_SEND_INVALID_REPLY
797 <dd>
798 The msgh_local_port value was invalid.
799 <p>
800 <dt> MACH_SEND_INVALID_TRAILER
801 <dd>
802 The trailer to be sent does not correspond to the current kernel format,
803 or the sending task does not have the privilege to supply the message
804 attributes.
805 </dl>
806 <p>
807 These return codes imply that some or all of the message was destroyed:
808 <dl>
809 <p>
810 <dt> MACH_SEND_INVALID_MEMORY
811 <dd>
812 The message body specified out-of-line data that was not readable.
813 <p>
814 <dt> MACH_SEND_INVALID_RIGHT
815 <dd>
816 The message body specified a port right which the caller didn't possess.
817 <p>
818 <dt> MACH_SEND_INVALID_TYPE
819 <dd>
820 A kernel processed descriptor was invalid.
821 <p>
822 <dt> MACH_SEND_MSG_TOO_SMALL
823 <dd>
824 The last data item in the message ran over the end of the message.
825 </dl>
826 <p>
827 These return codes imply that the message was returned to the caller with a
828 pseudo-receive operation:
829 <dl>
830 <p>
831 <dt> MACH_SEND_TIMED_OUT
832 <dd>
833 The timeout interval expired.
834 <p>
835 <dt> MACH_SEND_INTERRUPTED
836 <dd>
837 A software interrupt occurred.
838 </dl>
839 <p>
840 This return code implies that the message was queued:
841 <dl>
842 <p>
843 <dt> MACH_MSG_SUCCESS
844 <dd>
845 The message was queued.
846 </dl>
847 <p>
848 The receive operation can generate the following return codes. These return
849 codes imply that the call did not de-queue a message:
850 <dl>
851 <p>
852 <dt> MACH_RCV_INVALID_NAME
853 <dd>
854 The specified receive_name was invalid.
855 <p>
856 <dt> MACH_RCV_IN_SET
857 <dd>
858 The specified port was a member of a port set.
859 <p>
860 <dt> MACH_RCV_TIMED_OUT
861 <dd>
862 The timeout interval expired.
863 <p>
864 <dt> MACH_RCV_INTERRUPTED
865 <dd>
866 A software interrupt occurred.
867 <p>
868 <dt> MACH_RCV_PORT_DIED
869 <dd>
870 The caller lost the rights specified by receive_name.
871 <p>
872 <dt> MACH_RCV_PORT_CHANGED
873 <dd>
874 receive_name specified a receive right which was moved into a port set
875 during the call.
876 <p>
877 <dt> MACH_RCV_TOO_LARGE
878 <dd>
879 When using MACH_RCV_LARGE, the message was larger than
880 receive_limit. The message is left queued, and its actual size is
881 returned in the message header/message body.
882 <p>
883 <dt> MACH_RCV_SCATTER_SMALL
884 <dd>
885 When using MACH_RCV_LARGE with MACH_RCV_OVERWRITE, one or more scatter
886 list descriptors specified an overwrite region smaller than the
887 corresponding incoming region. The message is left queued, and the
888 proper descriptors are returned in the message header/message body.
889 <p>
890 <dt> MACH_RCV_INVALID_TRAILER
891 <dd>
892 The trailer type desired, or the number of trailer elements desired, is
893 not supported by the kernel.
894 </dl>
895 <p>
896 These return codes imply that a message was de-queued and destroyed:
897 <dl>
898 <p>
899 <dt> MACH_RCV_HEADER_ERROR
900 <dd>
901 A resource shortage prevented the reception of the port rights in the
902 message header.
903 <p>
904 <dt> MACH_RCV_INVALID_NOTIFY
905 <dd>
906 When using MACH_RCV_NOTIFY, the notify argument did not denote a
907 valid receive right.
908 <p>
909 <dt> MACH_RCV_INVALID_DATA
910 <dd>
911 The specified message buffer was not writable.
912 <p>
913 <dt> MACH_RCV_TOO_LARGE
914 <dd>
915 When not using MACH_RCV_LARGE, a message larger than
916 receive_limit was de-queued and destroyed.
917 <p>
918 <dt> MACH_RCV_SCATTER_SMALL
919 <dd>
920 When not using MACH_RCV_LARGE with MACH_RCV_OVERWRITE, one or more
921 scatter list descriptors specified an overwrite region smaller than
922 the corresponding incoming region. The message was de-queued and
923 destroyed.
924 <p>
925 <dt> MACH_RCV_OVERWRITE_ERROR
926 <dd>
927 A region specified by a receive overwrite descriptor
928 (MACH_RCV_OVERWRITE) was not allocated or could not be written.
929 <p>
930 <dt> MACH_RCV_INVALID_TYPE
931 <dd>
932 When using MACH_RCV_OVERWRITE, one or more scatter list descriptors
933 did not have the type matching the corresponding incoming message
934 descriptor or had an invalid copy (disposition) field.
935 <p>
936 <dt> MACH_RCV_LIMITS
937 <dd>
938 The combined size of all out-of-line memory regions or the total num-
939 ber of port rights in the message exceeds the limit set for the port.
940 These return codes imply that a message was received:
941 <p>
942 <dt> MACH_RCV_BODY_ERROR
943 <dd>
944 A resource shortage prevented the reception of a port right or out-of-
945 line memory region in the message body.
946 <p>
947 <dt> MACH_MSG_SUCCESS
948 <dd>
949 A message was received.
950 </dl>
951 <p>
952 Resource shortages can occur after a message is de-queued, while
953 transferring port rights and out-of-line memory regions to the
954 receiving task. The <strong>mach_msg</strong> call returns MACH_RCV_HEADER_ERROR or
955 MACH_RCV_BODY_ERROR in this situation. These return codes always carry
956 extra bits (bitwise-or'ed) that indicate the nature of the resource
957 shortage:
958 <dl>
959 <p>
960 <dt> MACH_MSG_IPC_SPACE
961 <dd>
962 There was no room in the task's IPC name space for another port name.
963 <p>
964 <dt> MACH_MSG_VM_SPACE
965 <dd>
966 There was no room in the task's VM address space for an out-of-line
967 memory region.
968 <p>
969 <dt> MACH_MSG_IPC_KERNEL
970 <dd>
971 A kernel resource shortage prevented the reception of a port right.
972 <p>
973 <dt> MACH_MSG_VM_KERNEL
974 <dd>
975 A kernel resource shortage prevented the reception of an out-of-line
976 memory region.
977 </dl>
978 <h3>RELATED INFORMATION</h3>
979 <p>
980 Functions:
981 <a href="vm_allocate.html"><strong>vm_allocate</strong></a>,
982 <a href="vm_deallocate.html"><strong>vm_deallocate</strong></a>,
983 <a href="vm_write.html"><strong>vm_write</strong></a>,
984 <a href="MP_request_notification.html"><strong>mach_port_request_notification</strong></a>,
985 <p>
986 Data Structures:
987 mach_msg_header.