1fi_cq(3) Libfabric v1.7.0 fi_cq(3)
2
3
4
6 fi_cq - Completion queue operations
7
8 fi_cq_open / fi_close
9 Open/close a completion queue
10
11 fi_control
12 Control CQ operation or attributes.
13
14 fi_cq_read / fi_cq_readfrom / fi_cq_readerr
15 Read a completion from a completion queue
16
17 fi_cq_sread / fi_cq_sreadfrom
18 A synchronous (blocking) read that waits until a specified con‐
19 dition has been met before reading a completion from a comple‐
20 tion queue.
21
22 fi_cq_signal
23 Unblock any thread waiting in fi_cq_sread or fi_cq_sreadfrom.
24
25 fi_cq_strerror
26 Converts provider specific error information into a printable
27 string
28
30 #include <rdma/fi_domain.h>
31
32 int fi_cq_open(struct fid_domain *domain, struct fi_cq_attr *attr,
33 struct fid_cq **cq, void *context);
34
35 int fi_close(struct fid *cq);
36
37 int fi_control(struct fid *cq, int command, void *arg);
38
39 ssize_t fi_cq_read(struct fid_cq *cq, void *buf, size_t count);
40
41 ssize_t fi_cq_readfrom(struct fid_cq *cq, void *buf, size_t count,
42 fi_addr_t *src_addr);
43
44 ssize_t fi_cq_readerr(struct fid_cq *cq, struct fi_cq_err_entry *buf,
45 uint64_t flags);
46
47 ssize_t fi_cq_sread(struct fid_cq *cq, void *buf, size_t count,
48 const void *cond, int timeout);
49
50 ssize_t fi_cq_sreadfrom(struct fid_cq *cq, void *buf, size_t count,
51 fi_addr_t *src_addr, const void *cond, int timeout);
52
53 int fi_cq_signal(struct fid_cq *cq);
54
55 const char * fi_cq_strerror(struct fid_cq *cq, int prov_errno,
56 const void *err_data, char *buf, size_t len);
57
59 domain Open resource domain
60
61 cq Completion queue
62
63 attr Completion queue attributes
64
65 context
66 User specified context associated with the completion queue.
67
68 buf For read calls, the data buffer to write completions into. For
69 write calls, a completion to insert into the completion queue.
70 For fi_cq_strerror, an optional buffer that receives printable
71 error information.
72
73 count Number of CQ entries.
74
75 len Length of data buffer
76
77 src_addr
78 Source address of a completed receive operation
79
80 flags Additional flags to apply to the operation
81
82 command
83 Command of control operation to perform on CQ.
84
85 arg Optional control argument
86
87 cond Condition that must be met before a completion is generated
88
89 timeout
90 Time in milliseconds to wait. A negative value indicates infi‐
91 nite timeout.
92
93 prov_errno
94 Provider specific error value
95
96 err_data
97 Provider specific error data related to a completion
98
100 Completion queues are used to report events associated with data trans‐
101 fers. They are associated with message sends and receives, RMA, atom‐
102 ic, tagged messages, and triggered events. Reported events are usually
103 associated with a fabric endpoint, but may also refer to memory regions
104 used as the target of an RMA or atomic operation.
105
106 fi_cq_open
107 fi_cq_open allocates a new completion queue. Unlike event queues, com‐
108 pletion queues are associated with a resource domain and may be off‐
109 loaded entirely in provider hardware.
110
111 The properties and behavior of a completion queue are defined by
112 struct fi_cq_attr.
113
114 struct fi_cq_attr {
115 size_t size; /* # entries for CQ */
116 uint64_t flags; /* operation flags */
117 enum fi_cq_format format; /* completion format */
118 enum fi_wait_obj wait_obj; /* requested wait object */
119 int signaling_vector; /* interrupt affinity */
120 enum fi_cq_wait_cond wait_cond; /* wait condition format */
121 struct fid_wait *wait_set; /* optional wait set */
122 };
123
124 size Specifies the minimum size of a completion queue. A value of 0
125 indicates that the provider may choose a default value.
126
127 flags Flags that control the configuration of the CQ.
128
129 - FI_AFFINITY
130 Indicates that the signaling_vector field (see below) is valid.
131
132 format Completion queues allow the application to select the amount of
133 detail that it must store and report. The format attribute al‐
134 lows the application to select one of several completion for‐
135 mats, indicating the structure of the data that the completion
136 queue should return when read. Supported formats and the struc‐
137 tures that correspond to each are listed below. The meaning of
138 the CQ entry fields are defined in the Completion Fields sec‐
139 tion.
140
141 - FI_CQ_FORMAT_UNSPEC
142 If an unspecified format is requested, then the CQ will use a
143 provider selected default format.
144
145 - FI_CQ_FORMAT_CONTEXT
146 Provides only user specified context that was associated with
147 the completion.
148
149 struct fi_cq_entry {
150 void *op_context; /* operation context */
151 };
152 · .RS 2
153
154 FI_CQ_FORMAT_MSG
155 Provides minimal data for processing completions, with expanded
156 support for reporting information about received messages.
157
158 struct fi_cq_msg_entry {
159 void *op_context; /* operation context */
160 uint64_t flags; /* completion flags */
161 size_t len; /* size of received data */
162 };
163 · .RS 2
164
165 FI_CQ_FORMAT_DATA
166 Provides data associated with a completion. Includes support
167 for received message length, remote CQ data, and multi-receive
168 buffers.
169
170 struct fi_cq_data_entry {
171 void *op_context; /* operation context */
172 uint64_t flags; /* completion flags */
173 size_t len; /* size of received data */
174 void *buf; /* receive data buffer */
175 uint64_t data; /* completion data */
176 };
177 · .RS 2
178
179 FI_CQ_FORMAT_TAGGED
180 Expands completion data to include support for the tagged mes‐
181 sage interfaces.
182
183 struct fi_cq_tagged_entry {
184 void *op_context; /* operation context */
185 uint64_t flags; /* completion flags */
186 size_t len; /* size of received data */
187 void *buf; /* receive data buffer */
188 uint64_t data; /* completion data */
189 uint64_t tag; /* received tag */
190 };
191
192 wait_obj
193 CQ's may be associated with a specific wait object. Wait ob‐
194 jects allow applications to block until the wait object is sig‐
195 naled, indicating that a completion is available to be read.
196 Users may use fi_control to retrieve the underlying wait object
197 associated with a CQ, in order to use it in other system calls.
198 The following values may be used to specify the type of wait ob‐
199 ject associated with a CQ: FI_WAIT_NONE, FI_WAIT_UNSPEC,
200 FI_WAIT_SET, FI_WAIT_FD, and FI_WAIT_MUTEX_COND. The default is
201 FI_WAIT_NONE.
202
203 - FI_WAIT_NONE
204 Used to indicate that the user will not block (wait) for comple‐
205 tions on the CQ. When FI_WAIT_NONE is specified, the applica‐
206 tion may not call fi_cq_sread or fi_cq_sreadfrom.
207
208 - FI_WAIT_UNSPEC
209 Specifies that the user will only wait on the CQ using fabric
210 interface calls, such as fi_cq_sread or fi_cq_sreadfrom. In
211 this case, the underlying provider may select the most appropri‐
212 ate or highest performing wait object available, including cus‐
213 tom wait mechanisms. Applications that select FI_WAIT_UNSPEC
214 are not guaranteed to retrieve the underlying wait object.
215
216 - FI_WAIT_SET
217 Indicates that the completion queue should use a wait set object
218 to wait for completions. If specified, the wait_set field must
219 reference an existing wait set object.
220
221 - FI_WAIT_FD
222 Indicates that the CQ should use a file descriptor as its wait
223 mechanism. A file descriptor wait object must be usable in se‐
224 lect, poll, and epoll routines. However, a provider may signal
225 an FD wait object by marking it as readable, writable, or with
226 an error.
227
228 - FI_WAIT_MUTEX_COND
229 Specifies that the CQ should use a pthread mutex and cond vari‐
230 able as a wait object.
231
232 - FI_WAIT_CRITSEC_COND
233 Windows specific. Specifies that the CQ should use a critical
234 section and condition variable as a wait object.
235
236 signaling_vector
237 If the FI_AFFINITY flag is set, this indicates the logical cpu
238 number (0..max cpu - 1) that interrupts associated with the CQ
239 should target. This field should be treated as a hint to the
240 provider and may be ignored if the provider does not support in‐
241 terrupt affinity.
242
243 wait_cond
244 By default, when a completion is inserted into a CQ that sup‐
245 ports blocking reads (fi_cq_sread/fi_cq_sreadfrom), the corre‐
246 sponding wait object is signaled. Users may specify a condition
247 that must first be met before the wait is satisfied. This field
248 indicates how the provider should interpret the cond field,
249 which describes the condition needed to signal the wait object.
250
251 A wait condition should be treated as an optimization. Providers are
252 not required to meet the requirements of the condition before signaling
253 the wait object. Applications should not rely on the condition neces‐
254 sarily being true when a blocking read call returns.
255
256 If wait_cond is set to FI_CQ_COND_NONE, then no additional conditions
257 are applied to the signaling of the CQ wait object, and the insertion
258 of any new entry will trigger the wait condition. If wait_cond is set
259 to FI_CQ_COND_THRESHOLD, then the cond field is interpreted as a size_t
260 threshold value. The threshold indicates the number of entries that
261 are to be queued before at the CQ before the wait is satisfied.
262
263 This field is ignored if wait_obj is set to FI_WAIT_NONE.
264
265 wait_set
266 If wait_obj is FI_WAIT_SET, this field references a wait object
267 to which the completion queue should attach. When an event is
268 inserted into the completion queue, the corresponding wait set
269 will be signaled if all necessary conditions are met. The use
270 of a wait_set enables an optimized method of waiting for events
271 across multiple event and completion queues. This field is ig‐
272 nored if wait_obj is not FI_WAIT_SET.
273
274 fi_close
275 The fi_close call releases all resources associated with a completion
276 queue. Any completions which remain on the CQ when it is closed are
277 lost.
278
279 When closing the CQ, there must be no opened endpoints, transmit con‐
280 texts, or receive contexts associated with the CQ. If resources are
281 still associated with the CQ when attempting to close, the call will
282 return -FI_EBUSY.
283
284 fi_control
285 The fi_control call is used to access provider or implementation spe‐
286 cific details of the completion queue. Access to the CQ should be se‐
287 rialized across all calls when fi_control is invoked, as it may redi‐
288 rect the implementation of CQ operations. The following control com‐
289 mands are usable with a CQ.
290
291 FI_GETWAIT (void **)
292 This command allows the user to retrieve the low-level wait ob‐
293 ject associated with the CQ. The format of the wait-object is
294 specified during CQ creation, through the CQ attributes. The
295 fi_control arg parameter should be an address where a pointer to
296 the returned wait object will be written. See fi_eq.3 for addi‐
297 tion details using fi_control with FI_GETWAIT.
298
299 fi_cq_read
300 The fi_cq_read operation performs a non-blocking read of completion da‐
301 ta from the CQ. The format of the completion event is determined using
302 the fi_cq_format option that was specified when the CQ was opened.
303 Multiple completions may be retrieved from a CQ in a single call. The
304 maximum number of entries to return is limited to the specified count
305 parameter, with the number of entries successfully read from the CQ re‐
306 turned by the call. (See return values section below.)
307
308 CQs are optimized to report operations which have completed successful‐
309 ly. Operations which fail are reported 'out of band'. Such operations
310 are retrieved using the fi_cq_readerr function. When an operation that
311 has completed with an unexpected error is encountered, it is placed in‐
312 to a temporary error queue. Attempting to read from a CQ while an item
313 is in the error queue results in fi_cq_read failing with a return code
314 of -FI_EAVAIL. Applications may use this return code to determine when
315 to call fi_cq_readerr.
316
317 fi_cq_readfrom
318 The fi_cq_readfrom call behaves identical to fi_cq_read, with the ex‐
319 ception that it allows the CQ to return source address information to
320 the user for any received data. Source address data is only available
321 for those endpoints configured with FI_SOURCE capability. If
322 fi_cq_readfrom is called on an endpoint for which source addressing da‐
323 ta is not available, the source address will be set to FI_ADDR_NO‐
324 TAVAIL. The number of input src_addr entries must be the same as the
325 count parameter.
326
327 Returned source addressing data is converted from the native address
328 used by the underlying fabric into an fi_addr_t, which may be used in
329 transmit operations. Under most circumstances, returning fi_addr_t re‐
330 quires that the source address already have been inserted into the ad‐
331 dress vector associated with the receiving endpoint. This is true for
332 address vectors of type FI_AV_TABLE. In select providers when
333 FI_AV_MAP is used, source addresses may be converted algorithmically
334 into a usable fi_addr_t, even though the source address has not been
335 inserted into the address vector. This is permitted by the API, as it
336 allows the provider to avoid address look-up as part of receive message
337 processing. In no case do providers insert addresses into an AV sepa‐
338 rate from an application calling fi_av_insert or similar call.
339
340 For endpoints allocated using the FI_SOURCE_ERR capability, if the
341 source address cannot be converted into a valid fi_addr_t value,
342 fi_cq_readfrom will return -FI_EAVAIL, even if the data were received
343 successfully. The completion will then be reported through fi_cq_read‐
344 err with error code -FI_EADDRNOTAVAIL. See fi_cq_readerr for details.
345
346 If FI_SOURCE is specified without FI_SOURCE_ERR, source addresses which
347 cannot be mapped to a usable fi_addr_t will be reported as FI_ADDR_NO‐
348 TAVAIL.
349
350 fi_cq_sread / fi_cq_sreadfrom
351 The fi_cq_sread and fi_cq_sreadfrom calls are the blocking equivalent
352 operations to fi_cq_read and fi_cq_readfrom. Their behavior is similar
353 to the non-blocking calls, with the exception that the calls will not
354 return until either a completion has been read from the CQ or an error
355 or timeout occurs.
356
357 It is invalid for applications to call these functions if the CQ has
358 been configured with a wait object of FI_WAIT_NONE or FI_WAIT_SET.
359
360 fi_cq_readerr
361 The read error function, fi_cq_readerr, retrieves information regarding
362 any asynchronous operation which has completed with an unexpected er‐
363 ror. fi_cq_readerr is a non-blocking call, returning immediately
364 whether an error completion was found or not.
365
366 Error information is reported to the user through struct fi_cq_err_en‐
367 try. The format of this structure is defined below.
368
369 struct fi_cq_err_entry {
370 void *op_context; /* operation context */
371 uint64_t flags; /* completion flags */
372 size_t len; /* size of received data */
373 void *buf; /* receive data buffer */
374 uint64_t data; /* completion data */
375 uint64_t tag; /* message tag */
376 size_t olen; /* overflow length */
377 int err; /* positive error code */
378 int prov_errno; /* provider error code */
379 void *err_data; /* error data */
380 size_t err_data_size; /* size of err_data */
381 };
382
383 The general reason for the error is provided through the err field.
384 Provider specific error information may also be available through the
385 prov_errno and err_data fields. Users may call fi_cq_strerror to con‐
386 vert provider specific error information into a printable string for
387 debugging purposes. See field details below for more information on
388 the use of err_data and err_data_size.
389
390 Note that error completions are generated for all operations, including
391 those for which a completion was not requested (e.g. an endpoint is
392 configured with FI_SELECTIVE_COMPLETION, but the request did not have
393 the FI_COMPLETION flag set). In such cases, providers will return as
394 much information as made available by the underlying software and hard‐
395 ware about the failure, other fields will be set to NULL or 0. This
396 includes the op_context value, which may not have been provided or was
397 ignored on input as part of the transfer.
398
399 Notable completion error codes are given below.
400
401 FI_EADDRNOTAVAIL
402 This error code is used by CQs configured with FI_SOURCE_ERR to
403 report completions for which a usable fi_addr_t source address
404 could not be found. An error code of FI_EADDRNOTAVAIL indicates
405 that the data transfer was successfully received and processed,
406 with the fi_cq_err_entry fields containing information about the
407 completion. The err_data field will be set to the source ad‐
408 dress data. The source address will be in the same format as
409 specified through the fi_info addr_format field for the opened
410 domain. This may be passed directly into an fi_av_insert call
411 to add the source address to the address vector.
412
413 fi_cq_signal
414 The fi_cq_signal call will unblock any thread waiting in fi_cq_sread or
415 fi_cq_sreadfrom. This may be used to wake-up a thread that is blocked
416 waiting to read a completion operation. The fi_cq_signal operation is
417 only available if the CQ was configured with a wait object.
418
420 The CQ entry data structures share many of the same fields. The mean‐
421 ings of these fields are the same for all CQ entry structure formats.
422
423 op_context
424 The operation context is the application specified context value
425 that was provided with an asynchronous operation. The op_con‐
426 text field is valid for all completions that are associated with
427 an asynchronous operation.
428
429 For completion events that are not associated with a posted operation,
430 this field will be set to NULL. This includes completions generated at
431 the target in response to RMA write operations that carry CQ data
432 (FI_REMOTE_WRITE | FI_REMOTE_CQ_DATA flags set), when the FI_RX_CQ_DATA
433 mode bit is not required.
434
435 flags This specifies flags associated with the completed operation.
436 The Completion Flags section below lists valid flag values.
437 Flags are set for all relevant completions.
438
439 len This len field only applies to completed receive operations
440 (e.g. fi_recv, fi_trecv, etc.). It indicates the size of re‐
441 ceived message data -- i.e. how many data bytes were placed in‐
442 to the associated receive buffer by a corresponding
443 fi_send/fi_tsend/et al call. If an endpoint has been configured
444 with the FI_MSG_PREFIX mode, the len also reflects the size of
445 the prefix buffer.
446
447 buf The buf field is only valid for completed receive operations,
448 and only applies when the receive buffer was posted with the
449 FI_MULTI_RECV flag. In this case, buf points to the starting
450 location where the receive data was placed.
451
452 data The data field is only valid if the FI_REMOTE_CQ_DATA completion
453 flag is set, and only applies to receive completions. If FI_RE‐
454 MOTE_CQ_DATA is set, this field will contain the completion data
455 provided by the peer as part of their transmit request. The
456 completion data will be given in host byte order.
457
458 tag A tag applies only to received messages that occur using the
459 tagged interfaces. This field contains the tag that was includ‐
460 ed with the received message. The tag will be in host byte or‐
461 der.
462
463 olen The olen field applies to received messages. It is used to in‐
464 dicate that a received message has overrun the available buffer
465 space and has been truncated. The olen specifies the amount of
466 data that did not fit into the available receive buffer and was
467 discarded.
468
469 err This err code is a positive fabric errno associated with a com‐
470 pletion. The err value indicates the general reason for an er‐
471 ror, if one occurred. See fi_errno.3 for a list of possible er‐
472 ror codes.
473
474 prov_errno
475 On an error, prov_errno may contain a provider specific error
476 code. The use of this field and its meaning is provider specif‐
477 ic. It is intended to be used as a debugging aid. See
478 fi_cq_strerror for additional details on converting this error
479 value into a human readable string.
480
481 err_data
482 On an error, err_data may reference a provider specific amount
483 of data associated with an error. The use of this field and its
484 meaning is provider specific. It is intended to be used as a
485 debugging aid. See fi_cq_strerror for additional details on
486 converting this error data into a human readable string.
487
488 err_data_size
489 On input, err_data_size indicates the size of the err_data buf‐
490 fer in bytes. On output, err_data_size will be set to the num‐
491 ber of bytes copied to the err_data buffer. The err_data infor‐
492 mation is typically used with fi_cq_strerror to provide details
493 about the type of error that occurred.
494
495 For compatibility purposes, if err_data_size is 0 on input, or the fab‐
496 ric was opened with release < 1.5, err_data will be set to a data buf‐
497 fer owned by the provider. The contents of the buffer will remain
498 valid until a subsequent read call against the CQ. Applications must
499 serialize access to the CQ when processing errors to ensure that the
500 buffer referenced by err_data does not change.
501
503 Completion flags provide additional details regarding the completed op‐
504 eration. The following completion flags are defined.
505
506 FI_SEND
507 Indicates that the completion was for a send operation. This
508 flag may be combined with an FI_MSG or FI_TAGGED flag.
509
510 FI_RECV
511 Indicates that the completion was for a receive operation. This
512 flag may be combined with an FI_MSG or FI_TAGGED flag.
513
514 FI_RMA Indicates that an RMA operation completed. This flag may be
515 combined with an FI_READ, FI_WRITE, FI_REMOTE_READ, or FI_RE‐
516 MOTE_WRITE flag.
517
518 FI_ATOMIC
519 Indicates that an atomic operation completed. This flag may be
520 combined with an FI_READ, FI_WRITE, FI_REMOTE_READ, or FI_RE‐
521 MOTE_WRITE flag.
522
523 FI_MSG Indicates that a message-based operation completed. This flag
524 may be combined with an FI_SEND or FI_RECV flag.
525
526 FI_TAGGED
527 Indicates that a tagged message operation completed. This flag
528 may be combined with an FI_SEND or FI_RECV flag.
529
530 FI_MULTICAST
531 Indicates that a multicast operation completed. This flag may
532 be combined with FI_MSG and relevant flags. This flag is only
533 guaranteed to be valid for received messages if the endpoint has
534 been configured with FI_SOURCE.
535
536 FI_READ
537 Indicates that a locally initiated RMA or atomic read operation
538 has completed. This flag may be combined with an FI_RMA or
539 FI_ATOMIC flag.
540
541 FI_WRITE
542 Indicates that a locally initiated RMA or atomic write operation
543 has completed. This flag may be combined with an FI_RMA or
544 FI_ATOMIC flag.
545
546 FI_REMOTE_READ
547 Indicates that a remotely initiated RMA or atomic read operation
548 has completed. This flag may be combined with an FI_RMA or
549 FI_ATOMIC flag.
550
551 FI_REMOTE_WRITE
552 Indicates that a remotely initiated RMA or atomic write opera‐
553 tion has completed. This flag may be combined with an FI_RMA or
554 FI_ATOMIC flag.
555
556 FI_REMOTE_CQ_DATA
557 This indicates that remote CQ data is available as part of the
558 completion.
559
560 FI_MULTI_RECV
561 This flag applies to receive buffers that were posted with the
562 FI_MULTI_RECV flag set. This completion flag indicates that the
563 original receive buffer referenced by the completion has been
564 consumed and was released by the provider. Providers may set
565 this flag on the last message that is received into the multi-
566 recv buffer, or may generate a separate completion that indi‐
567 cates that the buffer has been released.
568
569 Applications can distinguish between these two cases by examining the
570 completion entry flags field. If additional flags, such as FI_RECV,
571 are set, the completion is associated with a received message. In this
572 case, the buf field will reference the location where the received mes‐
573 sage was placed into the multi-recv buffer. Other fields in the com‐
574 pletion entry will be determined based on the received message. If
575 other flag bits are zero, the provider is reporting that the multi-recv
576 buffer has been released, and the completion entry is not associated
577 with a received message.
578
579 FI_MORE
580 See the 'Buffered Receives' section in fi_msg(3) for more de‐
581 tails. This flag is associated with receive completions on end‐
582 points that have FI_BUFFERED_RECV mode enabled. When set to
583 one, it indicates that the buffer referenced by the completion
584 is limited by the FI_OPT_BUFFERED_LIMIT threshold, and addition‐
585 al message data must be retrieved by the application using an
586 FI_CLAIM operation.
587
588 FI_CLAIM
589 See the 'Buffered Receives' section in fi_msg(3) for more de‐
590 tails. This flag is set on completions associated with receive
591 operations that claim buffered receive data. Note that this
592 flag only applies to endpoints configured with the
593 FI_BUFFERED_RECV mode bit.
594
596 Libfabric defines several completion 'levels', identified using opera‐
597 tional flags. Each flag indicates the soonest that a completion event
598 may be generated by a provider, and the assumptions that an application
599 may make upon processing a completion. The operational flags are de‐
600 fined below, along with an example of how a provider might implement
601 the semantic. Note that only meeting the semantic is required of the
602 provider and not the implementation. Providers may implement stronger
603 completion semantics than necessary for a given operation, but only the
604 behavior defined by the completion level is guaranteed.
605
606 To help understand the conceptual differences in completion levels,
607 consider mailing a letter. Placing the letter into the local mailbox
608 for pick-up is similar to 'inject complete'. Having the letter picked
609 up and dropped off at the destination mailbox is equivalent to 'trans‐
610 mit complete'. The 'delivery complete' semantic is a stronger guaran‐
611 tee, with a person at the destination signing for the letter. However,
612 the person who signed for the letter is not necessarily the intended
613 recipient. The 'match complete' option is similar to delivery com‐
614 plete, but requires the intended recipient to sign for the letter.
615
616 The 'commit complete' level has different semantics than the previously
617 mentioned levels. Commit complete would be closer to the the letter
618 arriving at the destination and being placed into a fire proof safe.
619
620 The operational flags for the described completion levels are defined
621 below.
622
623 FI_INJECT_COMPLETE
624 Indicates that a completion should be generated when the source
625 buffer(s) may be reused. A completion guarantees that the buf‐
626 fers will not be read from again and the application may reclaim
627 them. No other guarantees are made with respect to the state of
628 the operation.
629
630 Example: A provider may generate this completion event after copying
631 the source buffer into a network buffer, either in host memory or on
632 the NIC. An inject completion does not indicate that the data has been
633 transmitted onto the network, and a local error could occur after the
634 completion event has been generated that could prevent it from being
635 transmitted.
636
637 Inject complete allows for the fastest completion reporting (and,
638 hence, buffer reuse), but provides the weakest guarantees against net‐
639 work errors.
640
641 Note: This flag is used to control when a completion entry is inserted
642 into a completion queue. It does not apply to operations that do not
643 generate a completion queue entry, such as the fi_inject operation, and
644 is not subject to the inject_size message limit restriction.
645
646 FI_TRANSMIT_COMPLETE
647 Indicates that a completion should be generated when the trans‐
648 mit operation has completed relative to the local provider. The
649 exact behavior is dependent on the endpoint type.
650
651 For reliable endpoints:
652
653 Indicates that a completion should be generated when the operation has
654 been delivered to the peer endpoint. A completion guarantees that the
655 operation is no longer dependent on the fabric or local resources. The
656 state of the operation at the peer endpoint is not defined.
657
658 Example: A provider may generate a transmit complete event upon receiv‐
659 ing an ack from the peer endpoint. The state of the message at the
660 peer is unknown and may be buffered in the target NIC at the time the
661 ack has been generated.
662
663 For unreliable endpoints:
664
665 Indicates that a completion should be generated when the operation has
666 been delivered to the fabric. A completion guarantees that the opera‐
667 tion is no longer dependent on local resources. The state of the oper‐
668 ation within the fabric is not defined.
669
670 FI_DELIVERY_COMPLETE
671 Indicates that a completion should not be generated until an op‐
672 eration has been processed by the destination endpoint(s). A
673 completion guarantees that the result of the operation is avail‐
674 able; however, additional steps may need to be taken at the des‐
675 tination to retrieve the results. For example, an application
676 may need to provide a receive buffers in order to retrieve mes‐
677 sages that were buffered by the provider.
678
679 Delivery complete indicates that the message has been processed by the
680 peer. If an application buffer was ready to receive the results of the
681 message when it arrived, then delivery complete indicates that the data
682 was placed into the application's buffer.
683
684 This completion mode applies only to reliable endpoints. For opera‐
685 tions that return data to the initiator, such as RMA read or atom‐
686 ic-fetch, the source endpoint is also considered a destination end‐
687 point. This is the default completion mode for such operations.
688
689 FI_MATCH_COMPLETE
690 Indicates that a completion should be generated only after the
691 operation has been matched with an application specified buffer.
692 Operations using this completion semantic are dependent on the
693 application at the target claiming the message or results. As a
694 result, match complete may involve additional provider level ac‐
695 knowledgements or lengthy delays. However, this completion mod‐
696 el enables peer applications to synchronize their execution.
697
698 FI_COMMIT_COMPLETE
699 Indicates that a completion should not be generated (locally or
700 at the peer) until the result of an operation have been made
701 persistent. A completion guarantees that the result is both
702 available and durable, in the case of power failure.
703
704 This completion mode applies only to operations that target persistent
705 memory regions over reliable endpoints. This completion mode is exper‐
706 imental.
707
709 A completion queue must be bound to at least one enabled endpoint be‐
710 fore any operation such as fi_cq_read, fi_cq_readfrom, fi_cq_sread,
711 fi_cq_sreadfrom etc. can be called on it.
712
713 Completion flags may be suppressed if the FI_NOTIFY_FLAGS_ONLY mode bit
714 has been set. When enabled, only the following flags are guaranteed to
715 be set in completion data when they are valid: FI_REMOTE_READ and
716 FI_REMOTE_WRITE (when FI_RMA_EVENT capability bit has been set), FI_RE‐
717 MOTE_CQ_DATA, and FI_MULTI_RECV.
718
719 If a completion queue has been overrun, it will be placed into an
720 'overrun' state. Read operations will continue to return any valid,
721 non-corrupted completions, if available. After all valid completions
722 have been retrieved, any attempt to read the CQ will result in it re‐
723 turning an FI_EOVERRUN error event. Overrun completion queues are con‐
724 sidered fatal and may not be used to report additional completions once
725 the overrun occurs.
726
728 fi_cq_open / fi_cq_signal
729 Returns 0 on success. On error, a negative value corresponding
730 to fabric errno is returned.
731
732 fi_cq_read / fi_cq_readfrom / fi_cq_readerr fi_cq_sread / fi_cq_sread‐
733 from : On success, returns the number of completion events retrieved
734 from the completion queue. On error, a negative value corresponding to
735 fabric errno is returned. If no completions are available to return
736 from the CQ, -FI_EAGAIN will be returned.
737
738 fi_cq_strerror
739 Returns a character string interpretation of the provider spe‐
740 cific error returned with a completion.
741
742 Fabric errno values are defined in rdma/fi_errno.h.
743
745 fi_getinfo(3), fi_endpoint(3), fi_domain(3), fi_eq(3), fi_cntr(3),
746 fi_poll(3)
747
749 OpenFabrics.
750
751
752
753Libfabric Programmer's Manual 2018-11-28 fi_cq(3)