1fi_endpoint(3)                 Libfabric v1.12.1                fi_endpoint(3)
2
3
4

NAME

6       fi_endpoint - Fabric endpoint operations
7
8       fi_endpoint / fi_scalable_ep / fi_passive_ep / fi_close
9              Allocate or close an endpoint.
10
11       fi_ep_bind
12              Associate  an  endpoint  with  hardware resources, such as event
13              queues, completion queues, counters, address vectors, or  shared
14              transmit/receive contexts.
15
16       fi_scalable_ep_bind
17              Associate a scalable endpoint with an address vector
18
19       fi_pep_bind
20              Associate a passive endpoint with an event queue
21
22       fi_enable
23              Transitions an active endpoint into an enabled state.
24
25       fi_cancel
26              Cancel a pending asynchronous data transfer
27
28       fi_ep_alias
29              Create an alias to the endpoint
30
31       fi_control
32              Control endpoint operation.
33
34       fi_getopt / fi_setopt
35              Get or set endpoint options.
36
37       fi_rx_context / fi_tx_context / fi_srx_context / fi_stx_context
38              Open a transmit or receive context.
39
40       fi_tc_dscp_set / fi_tc_dscp_get
41              Convert between a DSCP value and a network traffic class
42
43       fi_rx_size_left / fi_tx_size_left (DEPRECATED)
44              Query the lower bound on how many RX/TX operations may be posted
45              without an operation returning -FI_EAGAIN.  This functions  have
46              been  deprecated  and will be removed in a future version of the
47              library.
48

SYNOPSIS

50              #include <rdma/fabric.h>
51
52              #include <rdma/fi_endpoint.h>
53
54              int fi_endpoint(struct fid_domain *domain, struct fi_info *info,
55                  struct fid_ep **ep, void *context);
56
57              int fi_scalable_ep(struct fid_domain *domain, struct fi_info *info,
58                  struct fid_ep **sep, void *context);
59
60              int fi_passive_ep(struct fi_fabric *fabric, struct fi_info *info,
61                  struct fid_pep **pep, void *context);
62
63              int fi_tx_context(struct fid_ep *sep, int index,
64                  struct fi_tx_attr *attr, struct fid_ep **tx_ep,
65                  void *context);
66
67              int fi_rx_context(struct fid_ep *sep, int index,
68                  struct fi_rx_attr *attr, struct fid_ep **rx_ep,
69                  void *context);
70
71              int fi_stx_context(struct fid_domain *domain,
72                  struct fi_tx_attr *attr, struct fid_stx **stx,
73                  void *context);
74
75              int fi_srx_context(struct fid_domain *domain,
76                  struct fi_rx_attr *attr, struct fid_ep **rx_ep,
77                  void *context);
78
79              int fi_close(struct fid *ep);
80
81              int fi_ep_bind(struct fid_ep *ep, struct fid *fid, uint64_t flags);
82
83              int fi_scalable_ep_bind(struct fid_ep *sep, struct fid *fid, uint64_t flags);
84
85              int fi_pep_bind(struct fid_pep *pep, struct fid *fid, uint64_t flags);
86
87              int fi_enable(struct fid_ep *ep);
88
89              int fi_cancel(struct fid_ep *ep, void *context);
90
91              int fi_ep_alias(struct fid_ep *ep, struct fid_ep **alias_ep, uint64_t flags);
92
93              int fi_control(struct fid *ep, int command, void *arg);
94
95              int fi_getopt(struct fid *ep, int level, int optname,
96                  void *optval, size_t *optlen);
97
98              int fi_setopt(struct fid *ep, int level, int optname,
99                  const void *optval, size_t optlen);
100
101              uint32_t fi_tc_dscp_set(uint8_t dscp);
102
103              uint8_t fi_tc_dscp_get(uint32_t tclass);
104
105              DEPRECATED ssize_t fi_rx_size_left(struct fid_ep *ep);
106
107              DEPRECATED ssize_t fi_tx_size_left(struct fid_ep *ep);
108

ARGUMENTS

110       fid    On creation, specifies a fabric  or  access  domain.   On  bind,
111              identifies  the  event  queue, completion queue, counter, or ad‐
112              dress vector to bind to the endpoint.  In other  cases,  it's  a
113              fabric identifier of an associated resource.
114
115       info   Details  about  the  fabric interface endpoint to be opened, ob‐
116              tained from fi_getinfo.
117
118       ep     A fabric endpoint.
119
120       sep    A scalable fabric endpoint.
121
122       pep    A passive fabric endpoint.
123
124       context
125              Context associated with the endpoint or asynchronous operation.
126
127       index  Index to retrieve a specific transmit/receive context.
128
129       attr   Transmit or receive context attributes.
130
131       flags  Additional flags to apply to the operation.
132
133       command
134              Command of control operation to perform on endpoint.
135
136       arg    Optional control argument.
137
138       level  Protocol level at which the desired option resides.
139
140       optname
141              The protocol option to read or set.
142
143       optval The option value that was read or to set.
144
145       optlen The size of the optval buffer.
146

DESCRIPTION

148       Endpoints are transport level communication  portals.   There  are  two
149       types  of endpoints: active and passive.  Passive endpoints belong to a
150       fabric domain and are most often used to listen for incoming connection
151       requests.   However, a passive endpoint may be used to reserve a fabric
152       address that can be granted to an active  endpoint.   Active  endpoints
153       belong to access domains and can perform data transfers.
154
155       Active  endpoints may be connection-oriented or connectionless, and may
156       provide data reliability.  The data  transfer  interfaces  --  messages
157       (fi_msg),  tagged  messages  (fi_tagged),  RMA  (fi_rma),  and  atomics
158       (fi_atomic) -- are associated with active endpoints.  In basic configu‐
159       rations, an active endpoint has transmit and receive queues.  In gener‐
160       al, operations that generate traffic on the fabric are  posted  to  the
161       transmit  queue.   This  includes  all RMA and atomic operations, along
162       with sent messages and sent tagged messages.  Operations that post buf‐
163       fers for receiving incoming data are submitted to the receive queue.
164
165       Active  endpoints are created in the disabled state.  They must transi‐
166       tion into an enabled state before accepting data  transfer  operations,
167       including  posting  of  receive buffers.  The fi_enable call is used to
168       transition an active endpoint into an enabled  state.   The  fi_connect
169       and  fi_accept  calls will also transition an endpoint into the enabled
170       state, if it is not already active.
171
172       In order to transition an endpoint into an enabled state,  it  must  be
173       bound  to one or more fabric resources.  An endpoint that will generate
174       asynchronous completions, either through data  transfer  operations  or
175       communication  establishment  events,  must be bound to the appropriate
176       completion queues or event queues, respectively, before being  enabled.
177       Additionally,  endpoints  that  use  manual progress must be associated
178       with relevant completion queues or  event  queues  in  order  to  drive
179       progress.   For  endpoints  that  are only used as the target of RMA or
180       atomic operations, this means binding  the  endpoint  to  a  completion
181       queue  associated  with  receive  processing.  Connectionless endpoints
182       must be bound to an address vector.
183
184       Once an endpoint has been activated, it may be associated with  an  ad‐
185       dress  vector.   Receive  buffers  may be posted to it and calls may be
186       made to connection establishment  routines.   Connectionless  endpoints
187       may also perform data transfers.
188
189       The behavior of an endpoint may be adjusted by setting its control data
190       and protocol options.  This allows the underlying provider to  redirect
191       function  calls to implementations optimized to meet the desired appli‐
192       cation behavior.
193
194       If an endpoint experiences a critical error, it  will  transition  back
195       into  a disabled state.  Critical errors are reported through the event
196       queue associated with the EP.  In certain cases,  a  disabled  endpoint
197       may  be  re-enabled.   The  ability  to transition back into an enabled
198       state is provider specific and depends on the type of  error  that  the
199       endpoint  experienced.   When  an endpoint is disabled as a result of a
200       critical error, all pending operations are discarded.
201
202   fi_endpoint / fi_passive_ep / fi_scalable_ep
203       fi_endpoint allocates a new active endpoint.  fi_passive_ep allocates a
204       new  passive  endpoint.   fi_scalable_ep allocates a scalable endpoint.
205       The properties and behavior of the endpoint are defined  based  on  the
206       provided  struct  fi_info.   See  fi_getinfo  for additional details on
207       fi_info.  fi_info flags that control the operation of an  endpoint  are
208       defined below.  See section SCALABLE ENDPOINTS.
209
210       If  an active endpoint is allocated in order to accept a connection re‐
211       quest, the fi_info parameter must be the same as the fi_info  structure
212       provided with the connection request (FI_CONNREQ) event.
213
214       An  active endpoint may acquire the properties of a passive endpoint by
215       setting the fi_info handle field to the  passive  endpoint  fabric  de‐
216       scriptor.   This  is  useful  for applications that need to reserve the
217       fabric address of an endpoint prior to knowing if the endpoint will  be
218       used  on the active or passive side of a connection.  For example, this
219       feature is useful for simulating socket semantics.  Once an active end‐
220       point  acquires  the properties of a passive endpoint, the passive end‐
221       point is no longer bound to any fabric resources and must no longer  be
222       used.  The user is expected to close the passive endpoint after opening
223       the active endpoint in order to free up any  lingering  resources  that
224       had been used.
225
226   fi_close
227       Closes an endpoint and release all resources associated with it.
228
229       When closing a scalable endpoint, there must be no opened transmit con‐
230       texts, or receive contexts associated with the scalable  endpoint.   If
231       resources are still associated with the scalable endpoint when attempt‐
232       ing to close, the call will return -FI_EBUSY.
233
234       Outstanding operations posted to the endpoint when fi_close  is  called
235       will be discarded.  Discarded operations will silently be dropped, with
236       no completions reported.  Additionally, a provider may  discard  previ‐
237       ously  completed  operations  from  the associated completion queue(s).
238       The behavior to discard completed operations is provider specific.
239
240   fi_ep_bind
241       fi_ep_bind is used to associate an endpoint with  other  allocated  re‐
242       sources,  such  as  completion queues, counters, address vectors, event
243       queues, shared contexts, and memory regions.  The type of objects  that
244       must be bound with an endpoint depend on the endpoint type and its con‐
245       figuration.
246
247       Passive endpoints must be bound with an  EQ  that  supports  connection
248       management  events.  Connectionless endpoints must be bound to a single
249       address vector.  If an endpoint is using a shared transmit  and/or  re‐
250       ceive context, the shared contexts must be bound to the endpoint.  CQs,
251       counters, AV, and shared contexts must be  bound  to  endpoints  before
252       they are enabled either explicitly or implicitly.
253
254       An endpoint must be bound with CQs capable of reporting completions for
255       any asynchronous operation initiated on the endpoint.  For example,  if
256       the  endpoint  supports  any  outbound  transfers (sends, RMA, atomics,
257       etc.), then it must be bound to a  completion  queue  that  can  report
258       transmit  completions.  This is true even if the endpoint is configured
259       to suppress successful completions, in order that operations that  com‐
260       plete in error may be reported to the user.
261
262       An  active  endpoint  may  direct asynchronous completions to different
263       CQs,  based  on  the  type  of  operation.   This  is  specified  using
264       fi_ep_bind flags.  The following flags may be OR'ed together when bind‐
265       ing an endpoint to a completion domain CQ.
266
267       FI_RECV
268              Directs the notification of inbound data transfers to the speci‐
269              fied  completion  queue.  This includes received messages.  This
270              binding automatically includes FI_REMOTE_WRITE, if applicable to
271              the endpoint.
272
273       FI_SELECTIVE_COMPLETION
274              By default, data transfer operations write CQ completion entries
275              into the associated completion queue after they have successful‐
276              ly completed.  Applications can use this bind flag to selective‐
277              ly enable when completions are generated.  If  FI_SELECTIVE_COM‐
278              PLETION is specified, data transfer operations will not generate
279              CQ entries for successful completions  unless  FI_COMPLETION  is
280              set  as an operational flag for the given operation.  Operations
281              that fail asynchronously will still generate  completions,  even
282              if  a completion is not requested.  FI_SELECTIVE_COMPLETION must
283              be OR'ed with FI_TRANSMIT and/or FI_RECV flags.
284
285       When FI_SELECTIVE_COMPLETION is set, the user must determine when a re‐
286       quest  that  does  NOT have FI_COMPLETION set has completed indirectly,
287       usually based on the completion of a subsequent operation or  by  using
288       completion  counters.   Use of this flag may improve performance by al‐
289       lowing the provider to avoid writing a CQ completion  entry  for  every
290       operation.
291
292       See Notes section below for additional information on how this flag in‐
293       teracts with the FI_CONTEXT and FI_CONTEXT2 mode bits.
294
295       FI_TRANSMIT
296              Directs the completion of outbound data transfer requests to the
297              specified  completion  queue.   This includes send message, RMA,
298              and atomic operations.
299
300       An endpoint may optionally be bound to a completion counter.  Associat‐
301       ing  an endpoint with a counter is in addition to binding the EP with a
302       CQ.  When binding an endpoint to a counter, the following flags may  be
303       specified.
304
305       FI_READ
306              Increments  the  specified  counter whenever an RMA read, atomic
307              fetch, or atomic compare operation initiated from  the  endpoint
308              has completed successfully or in error.
309
310       FI_RECV
311              Increments  the specified counter whenever a message is received
312              over the endpoint.  Received messages include  both  tagged  and
313              normal message operations.
314
315       FI_REMOTE_READ
316              Increments  the  specified  counter whenever an RMA read, atomic
317              fetch, or atomic compare operation is initiated  from  a  remote
318              endpoint  that targets the given endpoint.  Use of this flag re‐
319              quires that the endpoint be created using FI_RMA_EVENT.
320
321       FI_REMOTE_WRITE
322              Increments the specified counter whenever an RMA write  or  base
323              atomic  operation  is initiated from a remote endpoint that tar‐
324              gets the given endpoint.  Use of this  flag  requires  that  the
325              endpoint be created using FI_RMA_EVENT.
326
327       FI_SEND
328              Increments  the  specified  counter  whenever a message transfer
329              initiated over the endpoint has completed successfully or in er‐
330              ror.  Sent messages include both tagged and normal message oper‐
331              ations.
332
333       FI_WRITE
334              Increments the specified counter whenever an RMA write  or  base
335              atomic  operation initiated from the endpoint has completed suc‐
336              cessfully or in error.
337
338       An endpoint may only be bound to a single CQ or  counter  for  a  given
339       type of operation.  For example, a EP may not bind to two counters both
340       using FI_WRITE.  Furthermore, providers may limit CQ and counter  bind‐
341       ings to endpoints of the same endpoint type (DGRAM, MSG, RDM, etc.).
342
343   fi_scalable_ep_bind
344       fi_scalable_ep_bind  is  used  to associate a scalable endpoint with an
345       address vector.  See section on SCALABLE ENDPOINTS.   A  scalable  end‐
346       point  has  a  single  transport level address and can support multiple
347       transmit and receive contexts.  The transmit and receive contexts share
348       the  transport-level  address.  Address vectors that are bound to scal‐
349       able endpoints are implicitly bound to any transmit or receive contexts
350       created using the scalable endpoint.
351
352   fi_enable
353       This  call transitions the endpoint into an enabled state.  An endpoint
354       must be enabled before it may be used to perform data  transfers.   En‐
355       abling  an  endpoint  typically results in hardware resources being as‐
356       signed to it.  Endpoints making use  of  completion  queues,  counters,
357       event queues, and/or address vectors must be bound to them before being
358       enabled.
359
360       Calling connect or accept on an endpoint will implicitly enable an end‐
361       point if it has not already been enabled.
362
363       fi_enable  may also be used to re-enable an endpoint that has been dis‐
364       abled as a result  of  experiencing  a  critical  error.   Applications
365       should  check the return value from fi_enable to see if a disabled end‐
366       point has successfully be re-enabled.
367
368   fi_cancel
369       fi_cancel attempts to cancel  an  outstanding  asynchronous  operation.
370       Canceling an operation causes the fabric provider to search for the op‐
371       eration and, if it is still pending, complete it as  having  been  can‐
372       celed.   An error queue entry will be available in the associated error
373       queue with error code FI_ECANCELED.  On the other hand, if  the  opera‐
374       tion completed before the call to fi_cancel, then the completion status
375       of that operation will be available in the associated completion queue.
376       No specific entry related to fi_cancel itself will be posted.
377
378       Cancel uses the context parameter associated with an operation to iden‐
379       tify the request to cancel.  Operations posted without a valid  context
380       parameter  --  either  no context parameter is specified or the context
381       value was ignored by the provider -- cannot be canceled.   If  multiple
382       outstanding  operations  match  the context parameter, only one will be
383       canceled.  In this case, the operation which is  canceled  is  provider
384       specific.   The  cancel  operation  is  asynchronous, but will complete
385       within a bounded period of time.
386
387   fi_ep_alias
388       This call creates an alias to the specified endpoint.  Conceptually, an
389       endpoint alias provides an alternate software path from the application
390       to the underlying provider hardware.  An alias EP differs from its par‐
391       ent  endpoint only by its default data transfer flags.  For example, an
392       alias EP may be configured to use a different completion mode.  By  de‐
393       fault,  an alias EP inherits the same data transfer flags as the parent
394       endpoint.  An application can use fi_control to modify the alias EP op‐
395       erational flags.
396
397       When  allocating  an  alias,  an  application  may configure either the
398       transmit or receive operational flags.  This avoids needing a  separate
399       call to fi_control to set those flags.  The flags passed to fi_ep_alias
400       must include FI_TRANSMIT or FI_RECV (not both) with  other  operational
401       flags  OR'ed in.  This will override the transmit or receive flags, re‐
402       spectively, for operations posted through the alias endpoint.  All  al‐
403       located  aliases  must  be closed for the underlying endpoint to be re‐
404       leased.
405
406   fi_control
407       The control operation is used to adjust the default behavior of an end‐
408       point.  It allows the underlying provider to redirect function calls to
409       implementations optimized to meet the desired application behavior.  As
410       a  result,  calls to fi_ep_control must be serialized against all other
411       calls to an endpoint.
412
413       The base operation of an endpoint is  selected  during  creation  using
414       struct  fi_info.   The  following control commands and arguments may be
415       assigned to an endpoint.
416
417       **FI_BACKLOG - int *value**
418              This option only applies to passive endpoints.  It  is  used  to
419              set the connection request backlog for listening endpoints.
420
421       **FI_GETOPSFLAG -- uint64_t *flags**
422              Used  to retrieve the current value of flags associated with the
423              data transfer operations initiated on the endpoint.  The control
424              argument must include FI_TRANSMIT or FI_RECV (not both) flags to
425              indicate the type of data transfer flags to  be  returned.   See
426              below for a list of control flags.
427
428       FI_GETWAIT -- void **
429              This command allows the user to retrieve the file descriptor as‐
430              sociated with a socket endpoint.  The fi_control  arg  parameter
431              should  be  an  address where a pointer to the returned file de‐
432              scriptor will be written.  See fi_eq.3 for addition details  us‐
433              ing fi_control with FI_GETWAIT.  The file descriptor may be used
434              for notification that the endpoint is ready to send  or  receive
435              data.
436
437       **FI_SETOPSFLAG -- uint64_t *flags**
438              Used to change the data transfer operation flags associated with
439              an endpoint.  The control argument must include  FI_TRANSMIT  or
440              FI_RECV  (not  both)  to indicate the type of data transfer that
441              the flags should apply to, with other flags OR'ed in.  The given
442              flags will override the previous transmit and receive attributes
443              that were set when the  endpoint  was  created.   Valid  control
444              flags are defined below.
445
446   fi_getopt / fi_setopt
447       Endpoint  protocol  operations  may be retrieved using fi_getopt or set
448       using fi_setopt.  Applications specify the level that a desired  option
449       exists, identify the option, and provide input/output buffers to get or
450       set the option.  fi_setopt provides an  application  a  way  to  adjust
451       low-level protocol and implementation specific details of an endpoint.
452
453       The  following  option  levels  and option names and parameters are de‐
454       fined.
455
456       FI_OPT_ENDPOINT • .RS 2
457
458       FI_OPT_BUFFERED_LIMIT - size_t
459              Defines the maximum size of a buffered message that will be  re‐
460              ported  to  users  as  part  of  a  receive  completion when the
461              FI_BUFFERED_RECV mode is enabled on an endpoint.
462
463       fi_getopt() will return the  currently  configured  threshold,  or  the
464       provider's  default threshold if one has not be set by the application.
465       fi_setopt() allows an application to configure the threshold.   If  the
466       provider  cannot  support  the  requested  threshold,  it will fail the
467       fi_setopt()  call  with  FI_EMSGSIZE.   Calling  fi_setopt()  with  the
468       threshold  set  to  SIZE_MAX will set the threshold to the maximum sup‐
469       ported by the provider.  fi_getopt() can then be used to  retrieve  the
470       set size.
471
472       In  most  cases, the sending and receiving endpoints must be configured
473       to use the same threshold value, and the threshold must be set prior to
474       enabling the endpoint.  • .RS 2
475
476       FI_OPT_BUFFERED_MIN - size_t
477              Defines  the minimum size of a buffered message that will be re‐
478              ported.  Applications would set this to a size that's big enough
479              to decide whether to discard or claim a buffered receive or when
480              to claim a buffered receive on getting a buffered  receive  com‐
481              pletion.  The value is typically used by a provider when sending
482              a rendezvous protocol request  where  it  would  send  at  least
483              FI_OPT_BUFFERED_MIN  bytes of application data along with it.  A
484              smaller sized rendezvous protocol  message  usually  results  in
485              better latency for the overall transfer of a large message.
486       • .RS 2
487
488       FI_OPT_CM_DATA_SIZE - size_t
489              Defines  the size of available space in CM messages for user-de‐
490              fined data.  This value limits the amount of data that  applica‐
491              tions  can exchange between peer endpoints using the fi_connect,
492              fi_accept, and fi_reject operations.  The size returned  is  de‐
493              pendent  upon the properties of the endpoint, except in the case
494              of passive endpoints, in which the  size  reflects  the  maximum
495              size of the data that may be present as part of a connection re‐
496              quest event.  This option is read only.
497       • .RS 2
498
499       FI_OPT_MIN_MULTI_RECV - size_t
500              Defines the minimum receive buffer space available when the  re‐
501              ceive  buffer  is  released by the provider (see FI_MULTI_RECV).
502              Modifying this value is only guaranteed to set the minimum  buf‐
503              fer  space  needed  on  receives posted after the value has been
504              changed.  It is recommended that applications that want to over‐
505              ride the default MIN_MULTI_RECV value set this option before en‐
506              abling the corresponding endpoint.
507
508   fi_tc_dscp_set
509       This call converts a DSCP defined value into a libfabric traffic  class
510       value.   It should be used when assigning a DSCP value when setting the
511       tclass field in either domain or endpoint attributes
512
513   fi_tc_dscp_get
514       This call returns the DSCP value associated with the tclass  field  for
515       the domain or endpoint attributes.
516
517   fi_rx_size_left (DEPRECATED)
518       This  function has been deprecated and will be removed in a future ver‐
519       sion of the library.  It may not be supported by all providers.
520
521       The fi_rx_size_left call returns a lower bound on the number of receive
522       operations that may be posted to the given endpoint without that opera‐
523       tion returning -FI_EAGAIN.  Depending on the specific  details  of  the
524       subsequently  posted  receive  operations (e.g., number of iov entries,
525       which receive function is called, etc.), it may  be  possible  to  post
526       more receive operations than originally indicated by fi_rx_size_left.
527
528   fi_tx_size_left (DEPRECATED)
529       This  function has been deprecated and will be removed in a future ver‐
530       sion of the library.  It may not be supported by all providers.
531
532       The fi_tx_size_left call returns a lower bound on the number of  trans‐
533       mit  operations  that  may be posted to the given endpoint without that
534       operation returning -FI_EAGAIN.  Depending on the specific  details  of
535       the  subsequently  posted  transmit operations (e.g., number of iov en‐
536       tries, which transmit function is called, etc.), it may be possible  to
537       post   more   transmit   operations   than   originally   indicated  by
538       fi_tx_size_left.
539

ENDPOINT ATTRIBUTES

541       The fi_ep_attr structure defines the set of attributes associated  with
542       an  endpoint.   Endpoint  attributes  may  be further refined using the
543       transmit and receive context attributes as shown below.
544
545              struct fi_ep_attr {
546                  enum fi_ep_type type;
547                  uint32_t        protocol;
548                  uint32_t        protocol_version;
549                  size_t          max_msg_size;
550                  size_t          msg_prefix_size;
551                  size_t          max_order_raw_size;
552                  size_t          max_order_war_size;
553                  size_t          max_order_waw_size;
554                  uint64_t        mem_tag_format;
555                  size_t          tx_ctx_cnt;
556                  size_t          rx_ctx_cnt;
557                  size_t          auth_key_size;
558                  uint8_t         *auth_key;
559              };
560
561   type - Endpoint Type
562       If specified, indicates the type of fabric interface communication  de‐
563       sired.  Supported types are:
564
565       FI_EP_DGRAM
566              Supports  a  connectionless,  unreliable datagram communication.
567              Message boundaries are maintained, but the maximum message  size
568              may  be  limited to the fabric MTU.  Flow control is not guaran‐
569              teed.
570
571       FI_EP_MSG
572              Provides a reliable, connection-oriented data  transfer  service
573              with flow control that maintains message boundaries.
574
575       FI_EP_RDM
576              Reliable  datagram message.  Provides a reliable, connectionless
577              data transfer service with flow control that  maintains  message
578              boundaries.
579
580       FI_EP_SOCK_DGRAM
581              A  connectionless,  unreliable  datagram endpoint with UDP sock‐
582              et-like semantics.  FI_EP_SOCK_DGRAM is most useful for applica‐
583              tions  designed  around  using UDP sockets.  See the SOCKET END‐
584              POINT section for additional details and restrictions that apply
585              to datagram socket endpoints.
586
587       FI_EP_SOCK_STREAM
588              Data  streaming  endpoint  with TCP socket-like semantics.  Pro‐
589              vides a reliable, connection-oriented data transfer service that
590              does not maintain message boundaries.  FI_EP_SOCK_STREAM is most
591              useful for applications designed around using TCP sockets.   See
592              the  SOCKET ENDPOINT section for additional details and restric‐
593              tions that apply to stream endpoints.
594
595       FI_EP_UNSPEC
596              The type of endpoint is not specified.  This is usually provided
597              as  input, with other attributes of the endpoint or the provider
598              selecting the type.
599
600   Protocol
601       Specifies the low-level end to end protocol employed by  the  provider.
602       A  matching  protocol must be used by communicating endpoints to ensure
603       interoperability.  The following protocol values are defined.  Provider
604       specific  protocols are also allowed.  Provider specific protocols will
605       be indicated by having the upper bit of the protocol value set to one.
606
607       FI_PROTO_GNI
608              Protocol runs over Cray GNI low-level interface.
609
610       FI_PROTO_IB_RDM
611              Reliable-datagram protocol  implemented  over  InfiniBand  reli‐
612              able-connected queue pairs.
613
614       FI_PROTO_IB_UD
615              The  protocol  runs  over  Infiniband  unreliable datagram queue
616              pairs.
617
618       FI_PROTO_IWARP
619              The protocol runs over the  Internet  wide  area  RDMA  protocol
620              transport.
621
622       FI_PROTO_IWARP_RDM
623              Reliable-datagram  protocol implemented over iWarp reliable-con‐
624              nected queue pairs.
625
626       FI_PROTO_NETWORKDIRECT
627              Protocol runs over Microsoft NetworkDirect service provider  in‐
628              terface.   This  adds  reliable-datagram semantics over the Net‐
629              workDirect connection- oriented endpoint semantics.
630
631       FI_PROTO_PSMX
632              The protocol is based on an Intel proprietary protocol known  as
633              PSM,  performance scaled messaging.  PSMX is an extended version
634              of the PSM protocol to support the libfabric interfaces.
635
636       FI_PROTO_PSMX2
637              The protocol is based on an Intel proprietary protocol known  as
638              PSM2,  performance  scaled messaging version 2.  PSMX2 is an ex‐
639              tended version of the PSM2 protocol to support the libfabric in‐
640              terfaces.
641
642       FI_PROTO_PSMX3
643              The  protocol  is  Intel's  protocol  known as PSM3, performance
644              scaled messaging version 3.  PSMX3 is  implemented  over  RoCEv2
645              and verbs.
646
647       FI_PROTO_RDMA_CM_IB_RC
648              The  protocol  runs  over  Infiniband  reliable-connected  queue
649              pairs, using the RDMA CM protocol for connection establishment.
650
651       FI_PROTO_RXD
652              Reliable-datagram protocol implemented over datagram  endpoints.
653              RXD  is a libfabric utility component that adds RDM endpoint se‐
654              mantics over DGRAM endpoint semantics.
655
656       FI_PROTO_RXM
657              Reliable-datagram protocol implemented over  message  endpoints.
658              RXM  is a libfabric utility component that adds RDM endpoint se‐
659              mantics over MSG endpoint semantics.
660
661       FI_PROTO_SOCK_TCP
662              The protocol is layered over TCP packets.
663
664       FI_PROTO_UDP
665              The protocol sends and receives UDP datagrams.  For example,  an
666              endpoint  using  FI_PROTO_UDP will be able to communicate with a
667              remote peer that is using Berkeley SOCK_DGRAM sockets using  IP‐
668              PROTO_UDP.
669
670       FI_PROTO_UNSPEC
671              The  protocol is not specified.  This is usually provided as in‐
672              put, with other attributes of the socket or the provider select‐
673              ing the actual protocol.
674
675   protocol_version - Protocol Version
676       Identifies  which  version of the protocol is employed by the provider.
677       The protocol version allows providers to extend an  existing  protocol,
678       by adding support for additional features or functionality for example,
679       in a backward compatible manner.  Providers that support different ver‐
680       sions  of  the  same protocol should inter-operate, but only when using
681       the capabilities defined for the lesser version.
682
683   max_msg_size - Max Message Size
684       Defines the maximum size for an application data transfer as  a  single
685       operation.
686
687   msg_prefix_size - Message Prefix Size
688       Specifies  the  size of any required message prefix buffer space.  This
689       field will be 0 unless the FI_MSG_PREFIX mode is enabled.  If  msg_pre‐
690       fix_size is > 0 the specified value will be a multiple of 8-bytes.
691
692   Max RMA Ordered Size
693       The maximum ordered size specifies the delivery order of transport data
694       into target memory for RMA and atomic  operations.   Data  ordering  is
695       separate,  but dependent on message ordering (defined below).  Data or‐
696       dering is unspecified where message order is not defined.
697
698       Data ordering refers to the access of target memory by subsequent oper‐
699       ations.  When back to back RMA read or write operations access the same
700       registered memory location, data ordering indicates whether the  second
701       operation  reads  or writes the target memory after the first operation
702       has completed.  Because RMA ordering applies  between  two  operations,
703       and not within a single data transfer, ordering is defined per byte-ad‐
704       dressable memory location.  I.e.  ordering specifies whether location X
705       is accessed by the second operation after the first operation.  Nothing
706       is implied about the completion of the first operation before the  sec‐
707       ond operation is initiated.
708
709       In  order  to  support  large data transfers being broken into multiple
710       packets and sent using multiple paths through the fabric, data ordering
711       may  be  limited  to  transfers  of a specific size or less.  Providers
712       specify when data ordering is maintained through the following  values.
713       Note that even if data ordering is not maintained, message ordering may
714       be.
715
716       max_order_raw_size
717              Read after write size.  If set, an RMA or atomic read  operation
718              issued after an RMA or atomic write operation, both of which are
719              smaller than the size, will be ordered.  Where the target memory
720              locations overlap, the RMA or atomic read operation will see the
721              results of the previous RMA or atomic write.
722
723       max_order_war_size
724              Write after read size.  If set, an RMA or atomic write operation
725              issued  after an RMA or atomic read operation, both of which are
726              smaller than the size, will be ordered.  The RMA or atomic  read
727              operation  will see the initial value of the target memory loca‐
728              tion before a subsequent RMA or atomic write updates the value.
729
730       max_order_waw_size
731              Write after write size.  If set, an RMA or atomic  write  opera‐
732              tion  issued  after  an  RMA  or atomic write operation, both of
733              which are smaller than the size, will be  ordered.   The  target
734              memory  location  will  reflect the results of the second RMA or
735              atomic write.
736
737       An order size value of 0 indicates that ordering is not guaranteed.   A
738       value of -1 guarantees ordering for any data size.
739
740   mem_tag_format - Memory Tag Format
741       The  memory  tag  format  is  a  bit array used to convey the number of
742       tagged bits supported by a provider.  Additionally, it may be  used  to
743       divide  the bit array into separate fields.  The mem_tag_format option‐
744       ally begins with a series of bits set to 0, to signify bits  which  are
745       ignored by the provider.  Following the initial prefix of ignored bits,
746       the array will consist of alternating groups of bits set to all 1's  or
747       all 0's.  Each group of bits corresponds to a tagged field.  The impli‐
748       cation of defining a tagged field is that when a mask is applied to the
749       tagged  bit  array, all bits belonging to a single field will either be
750       set to 1 or 0, collectively.
751
752       For example, a mem_tag_format of 0x30FF indicates support for 14 tagged
753       bits, separated into 3 fields.  The first field consists of 2-bits, the
754       second field 4-bits, and the final field 8-bits.  Valid masks for  such
755       a tagged field would be a bitwise OR'ing of zero or more of the follow‐
756       ing values: 0x3000, 0x0F00, and 0x00FF.  The provider may not  validate
757       the mask provided by the application for performance reasons.
758
759       By  identifying fields within a tag, a provider may be able to optimize
760       their search routines.  An application which requests tag  fields  must
761       provide  tag  masks  that  either  set all mask bits corresponding to a
762       field to all 0 or all 1.  When negotiating tag fields,  an  application
763       can  request  a  specific number of fields of a given size.  A provider
764       must return a tag format that supports the requested number of  fields,
765       with each field being at least the size requested, or fail the request.
766       A provider may increase the size of the fields.  When reporting comple‐
767       tions (see FI_CQ_FORMAT_TAGGED), it is not guaranteed that the provider
768       would clear out any unsupported tag bits in the tag field of  the  com‐
769       pletion entry.
770
771       It is recommended that field sizes be ordered from smallest to largest.
772       A generic, unstructured tag and mask can be achieved  by  requesting  a
773       bit array consisting of alternating 1's and 0's.
774
775   tx_ctx_cnt - Transmit Context Count
776       Number  of  transmit  contexts  to associate with the endpoint.  If not
777       specified (0), 1 context will be assigned if the endpoint supports out‐
778       bound  transfers.   Transmit  contexts  are independent transmit queues
779       that may be separately configured.  Each transmit context may be  bound
780       to  a  separate CQ, and no ordering is defined between contexts.  Addi‐
781       tionally, no synchronization is needed when accessing contexts in  par‐
782       allel.
783
784       If  the  count is set to the value FI_SHARED_CONTEXT, the endpoint will
785       be configured to use a shared transmit context,  if  supported  by  the
786       provider.   Providers that do not support shared transmit contexts will
787       fail the request.
788
789       See the scalable endpoint and shared contexts sections  for  additional
790       details.
791
792   rx_ctx_cnt - Receive Context Count
793       Number  of  receive  contexts  to  associate with the endpoint.  If not
794       specified, 1 context will be assigned if the endpoint supports  inbound
795       transfers.  Receive contexts are independent processing queues that may
796       be separately configured.  Each receive context may be bound to a sepa‐
797       rate CQ, and no ordering is defined between contexts.  Additionally, no
798       synchronization is needed when accessing contexts in parallel.
799
800       If the count is set to the value FI_SHARED_CONTEXT, the  endpoint  will
801       be  configured  to  use  a  shared receive context, if supported by the
802       provider.  Providers that do not support shared receive  contexts  will
803       fail the request.
804
805       See  the  scalable endpoint and shared contexts sections for additional
806       details.
807
808   auth_key_size - Authorization Key Length
809       The length of the authorization key in bytes.  This field will be 0  if
810       authorization  keys  are  not available or used.  This field is ignored
811       unless the fabric is opened with API version 1.5 or greater.
812
813   auth_key - Authorization Key
814       If supported by the fabric, an authorization key (a.k.a.  job  key)  to
815       associate  with  the  endpoint.   An authorization key is used to limit
816       communication between endpoints.  Only peer  endpoints  that  are  pro‐
817       grammed  to use the same authorization key may communicate.  Authoriza‐
818       tion keys are often used to implement job keys, to ensure that process‐
819       es  running  in  different jobs do not accidentally cross traffic.  The
820       domain authorization key will be used if auth_key_size  is  set  to  0.
821       This  field is ignored unless the fabric is opened with API version 1.5
822       or greater.
823

TRANSMIT CONTEXT ATTRIBUTES

825       Attributes specific to the transmit capabilities  of  an  endpoint  are
826       specified using struct fi_tx_attr.
827
828              struct fi_tx_attr {
829                  uint64_t  caps;
830                  uint64_t  mode;
831                  uint64_t  op_flags;
832                  uint64_t  msg_order;
833                  uint64_t  comp_order;
834                  size_t    inject_size;
835                  size_t    size;
836                  size_t    iov_limit;
837                  size_t    rma_iov_limit;
838                  uint32_t  tclass;
839              };
840
841   caps - Capabilities
842       The  requested capabilities of the context.  The capabilities must be a
843       subset of those requested of the associated endpoint.  See the CAPABIL‐
844       ITIES  section  of  fi_getinfo(3)  for capability details.  If the caps
845       field is 0 on input to fi_getinfo(3), the  applicable  capability  bits
846       from the fi_info structure will be used.
847
848       The  following  capabilities  apply to the transmit attributes: FI_MSG,
849       FI_RMA, FI_TAGGED,  FI_ATOMIC,  FI_READ,  FI_WRITE,  FI_SEND,  FI_HMEM,
850       FI_TRIGGER,  FI_FENCE,  FI_MULTICAST, FI_RMA_PMEM, FI_NAMED_RX_CTX, and
851       FI_COLLECTIVE.
852
853       Many applications will be able to ignore this field and rely solely  on
854       the  fi_info::caps field.  Use of this field provides fine grained con‐
855       trol over the transmit capabilities associated with an endpoint.  It is
856       useful  when  handling  scalable endpoints, with multiple transmit con‐
857       texts, for example, and allows configuring a specific transmit  context
858       with  fewer  capabilities  than that supported by the endpoint or other
859       transmit contexts.
860
861   mode
862       The operational mode bits of the context.  The mode bits will be a sub‐
863       set  of  those  associated  with the endpoint.  See the MODE section of
864       fi_getinfo(3) for details.  A mode value of 0 will be ignored on  input
865       to fi_getinfo(3), with the mode value of the fi_info structure used in‐
866       stead.  On return from fi_getinfo(3), the mode  will  be  set  only  to
867       those constraints specific to transmit operations.
868
869   op_flags - Default transmit operation flags
870       Flags  that  control  the operation of operations submitted against the
871       context.  Applicable flags are listed in the Operation Flags section.
872
873   msg_order - Message Ordering
874       Message ordering refers to the order in which transport  layer  headers
875       (as  viewed  by the application) are identified and processed.  Relaxed
876       message order enables data transfers to be sent and received out of or‐
877       der,  which may improve performance by utilizing multiple paths through
878       the fabric from the initiating endpoint to a target endpoint.   Message
879       order  applies  only  between  a single source and destination endpoint
880       pair.  Ordering between different target endpoints is not defined.
881
882       Message order is determined using a set of ordering bits.  Each set bit
883       indicates  that  ordering  is  maintained between data transfers of the
884       specified type.  Message order is defined for [read | write | send] op‐
885       erations submitted by an application after [read | write | send] opera‐
886       tions.
887
888       Message ordering only applies to the end to end transmission of  trans‐
889       port  headers.   Message ordering is necessary, but does not guarantee,
890       the order in which message data is sent or received  by  the  transport
891       layer.   Message  ordering  requires matching ordering semantics on the
892       receiving side of a data transfer operation in order to guarantee  that
893       ordering is met.
894
895       FI_ORDER_ATOMIC_RAR
896              Atomic  read  after  read.   If set, atomic fetch operations are
897              transmitted in the order  submitted  relative  to  other  atomic
898              fetch operations.  If not set, atomic fetches may be transmitted
899              out of order from their submission.
900
901       FI_ORDER_ATOMIC_RAW
902              Atomic read after write.  If set, atomic  fetch  operations  are
903              transmitted in the order submitted relative to atomic update op‐
904              erations.  If not set, atomic fetches may be  transmitted  ahead
905              of atomic updates.
906
907       FI_ORDER_ATOMIC_WAR
908              RMA  write  after  read.   If  set, atomic update operations are
909              transmitted in the order submitted relative to atomic fetch  op‐
910              erations.   If  not set, atomic updates may be transmitted ahead
911              of atomic fetches.
912
913       FI_ORDER_ATOMIC_WAW
914              RMA write after write.  If set,  atomic  update  operations  are
915              transmitted  in the order submitted relative to other atomic up‐
916              date operations.  If not atomic updates may be  transmitted  out
917              of order from their submission.
918
919       FI_ORDER_NONE
920              No  ordering  is  specified.  This value may be used as input in
921              order to obtain the  default  message  order  supported  by  the
922              provider.  FI_ORDER_NONE is an alias for the value 0.
923
924       FI_ORDER_RAR
925              Read  after  read.   If  set, RMA and atomic read operations are
926              transmitted in the order submitted relative  to  other  RMA  and
927              atomic read operations.  If not set, RMA and atomic reads may be
928              transmitted out of order from their submission.
929
930       FI_ORDER_RAS
931              Read after send.  If set, RMA and  atomic  read  operations  are
932              transmitted  in the order submitted relative to message send op‐
933              erations, including tagged sends.  If not set,  RMA  and  atomic
934              reads may be transmitted ahead of sends.
935
936       FI_ORDER_RAW
937              Read  after  write.   If set, RMA and atomic read operations are
938              transmitted in the order submitted relative to  RMA  and  atomic
939              write  operations.   If  not  set,  RMA  and atomic reads may be
940              transmitted ahead of RMA and atomic writes.
941
942       FI_ORDER_RMA_RAR
943              RMA read after read.  If set, RMA read operations are  transmit‐
944              ted  in  the  order  submitted relative to other RMA read opera‐
945              tions.  If not set, RMA reads may be transmitted  out  of  order
946              from their submission.
947
948       FI_ORDER_RMA_RAW
949              RMA read after write.  If set, RMA read operations are transmit‐
950              ted in the order submitted relative to RMA write operations.  If
951              not set, RMA reads may be transmitted ahead of RMA writes.
952
953       FI_ORDER_RMA_WAR
954              RMA  write  after read.  If set, RMA write operations are trans‐
955              mitted in the order submitted relative to RMA  read  operations.
956              If not set, RMA writes may be transmitted ahead of RMA reads.
957
958       FI_ORDER_RMA_WAW
959              RMA  write after write.  If set, RMA write operations are trans‐
960              mitted in the order submitted relative to other RMA write opera‐
961              tions.   If  not set, RMA writes may be transmitted out of order
962              from their submission.
963
964       FI_ORDER_SAR
965              Send after read.  If set,  message  send  operations,  including
966              tagged sends, are transmitted in order submitted relative to RMA
967              and atomic read operations.  If not set, message  sends  may  be
968              transmitted ahead of RMA and atomic reads.
969
970       FI_ORDER_SAS
971              Send  after  send.   If  set, message send operations, including
972              tagged sends, are transmitted in the order submitted relative to
973              other  message send.  If not set, message sends may be transmit‐
974              ted out of order from their submission.
975
976       FI_ORDER_SAW
977              Send after write.  If set, message  send  operations,  including
978              tagged sends, are transmitted in order submitted relative to RMA
979              and atomic write operations.  If not set, message sends  may  be
980              transmitted ahead of RMA and atomic writes.
981
982       FI_ORDER_WAR
983              Write  after  read.  If set, RMA and atomic write operations are
984              transmitted in the order submitted relative to  RMA  and  atomic
985              read  operations.   If  not  set,  RMA  and atomic writes may be
986              transmitted ahead of RMA and atomic reads.
987
988       FI_ORDER_WAS
989              Write after send.  If set, RMA and atomic write  operations  are
990              transmitted  in the order submitted relative to message send op‐
991              erations, including tagged sends.  If not set,  RMA  and  atomic
992              writes may be transmitted ahead of sends.
993
994       FI_ORDER_WAW
995              Write  after write.  If set, RMA and atomic write operations are
996              transmitted in the order submitted relative  to  other  RMA  and
997              atomic  write operations.  If not set, RMA and atomic writes may
998              be transmitted out of order from their submission.
999
1000   comp_order - Completion Ordering
1001       Completion ordering refers to the order in which completed requests are
1002       written  into  the completion queue.  Completion ordering is similar to
1003       message order.  Relaxed completion order may enable faster reporting of
1004       completed  transfers,  allow  acknowledgments to be sent over different
1005       fabric paths, and support more sophisticated  retry  mechanisms.   This
1006       can  result  in lower-latency completions, particularly when using con‐
1007       nectionless endpoints.  Strict completion  ordering  may  require  that
1008       providers queue completed operations or limit available optimizations.
1009
1010       For transmit requests, completion ordering depends on the endpoint com‐
1011       munication type.  For unreliable communication, completion ordering ap‐
1012       plies  to all data transfer requests submitted to an endpoint.  For re‐
1013       liable communication, completion ordering only applies to requests that
1014       target  a single destination endpoint.  Completion ordering of requests
1015       that target different endpoints over a reliable transport  is  not  de‐
1016       fined.
1017
1018       Applications  should  specify the completion ordering that they support
1019       or require.  Providers should return the completion order that they ac‐
1020       tually  provide,  with  the  constraint  that  the returned ordering is
1021       stricter than that specified by the application.  Supported  completion
1022       order values are:
1023
1024       FI_ORDER_NONE
1025              No  ordering is defined for completed operations.  Requests sub‐
1026              mitted to the transmit context may complete in any order.
1027
1028       FI_ORDER_STRICT
1029              Requests complete in the order in which they  are  submitted  to
1030              the transmit context.
1031
1032   inject_size
1033       The  requested  inject operation size (see the FI_INJECT flag) that the
1034       context will support.  This is the maximum size data transfer that  can
1035       be  associated  with  an inject operation (such as fi_inject) or may be
1036       used with the FI_INJECT data transfer flag.
1037
1038   size
1039       The size of the transmit context.  The mapping of the size value to re‐
1040       sources  is provider specific, but it is directly related to the number
1041       of command entries allocated for the endpoint.  A  smaller  size  value
1042       consumes fewer hardware and software resources, while a larger size al‐
1043       lows queuing more transmit requests.
1044
1045       While the size attribute guides the size of underlying endpoint  trans‐
1046       mit  queue,  there  is  not  necessarily a one-to-one mapping between a
1047       transmit operation and a queue entry.  A single transmit operation  may
1048       consume multiple queue entries; for example, one per scatter-gather en‐
1049       try.  Additionally, the size field is intended to guide the  allocation
1050       of  the  endpoint's transmit context.  Specifically, for connectionless
1051       endpoints, there may be lower-level queues use to  track  communication
1052       on  a  per peer basis.  The sizes of any lower-level queues may only be
1053       significantly smaller than the endpoint's transmit size,  in  order  to
1054       reduce resource utilization.
1055
1056   iov_limit
1057       This is the maximum number of IO vectors (scatter-gather elements) that
1058       a single posted operation may reference.
1059
1060   rma_iov_limit
1061       This is the maximum number of RMA IO vectors (scatter-gather  elements)
1062       that  an RMA or atomic operation may reference.  The rma_iov_limit cor‐
1063       responds to the rma_iov_count values in RMA and atomic operations.  See
1064       struct fi_msg_rma and struct fi_msg_atomic in fi_rma.3 and fi_atomic.3,
1065       for additional details.  This limit applies to both the number  of  RMA
1066       IO  vectors that may be specified when initiating an operation from the
1067       local endpoint, as well as the maximum number of IO vectors that may be
1068       carried in a single request from a remote endpoint.
1069
1070   Traffic Class (tclass)
1071       Traffic classes can be a differentiated services code point (DSCP) val‐
1072       ue, one of the following defined labels, or a provider-specific defini‐
1073       tion.  If tclass is unset or set to FI_TC_UNSPEC, the endpoint will use
1074       the default traffic class associated with the domain.
1075
1076       FI_TC_BEST_EFFORT
1077              This is the default in the absence of any other local or  fabric
1078              configuration.   This  class carries the traffic for a number of
1079              applications executing concurrently over the same network infra‐
1080              structure.   Even  though it is shared, network capacity and re‐
1081              source allocation are distributed  fairly  across  the  applica‐
1082              tions.
1083
1084       FI_TC_BULK_DATA
1085              This  class is intended for large data transfers associated with
1086              I/O and is present to separate sustained I/O transfers from oth‐
1087              er application inter-process communications.
1088
1089       FI_TC_DEDICATED_ACCESS
1090              This  class operates at the highest priority, except the manage‐
1091              ment class.  It carries a high bandwidth allocation, minimum la‐
1092              tency targets, and the highest scheduling and arbitration prior‐
1093              ity.
1094
1095       FI_TC_LOW_LATENCY
1096              This class supports low latency, low jitter data patterns  typi‐
1097              cally  caused  by transactional data exchanges, barrier synchro‐
1098              nizations, and collective operations that are typical of HPC ap‐
1099              plications.   This class often requires maximum tolerable laten‐
1100              cies that data transfers must achieve for correct or performance
1101              operations.   Fulfillment  of  such  requests in this class will
1102              typically require accompanying bandwidth and message size  limi‐
1103              tations so as not to consume excessive bandwidth at high priori‐
1104              ty.
1105
1106       FI_TC_NETWORK_CTRL
1107              This class is intended for traffic directly  related  to  fabric
1108              (network) management, which is critical to the correct operation
1109              of the network.  Its use is typically restricted  to  privileged
1110              network management applications.
1111
1112       FI_TC_SCAVENGER
1113              This  class  is  used for data that is desired but does not have
1114              strict delivery requirements, such as in-band network or  appli‐
1115              cation  level monitoring data.  Use of this class indicates that
1116              the traffic is considered lower priority and should  not  inter‐
1117              fere with higher priority workflows.
1118
1119       fi_tc_dscp_set / fi_tc_dscp_get
1120              DSCP  values  are  supported via the DSCP get and set functions.
1121              The definitions for DSCP values are outside the scope of libfab‐
1122              ric.  See the fi_tc_dscp_set and fi_tc_dscp_get function defini‐
1123              tions for details on their use.
1124

RECEIVE CONTEXT ATTRIBUTES

1126       Attributes specific to the receive  capabilities  of  an  endpoint  are
1127       specified using struct fi_rx_attr.
1128
1129              struct fi_rx_attr {
1130                  uint64_t  caps;
1131                  uint64_t  mode;
1132                  uint64_t  op_flags;
1133                  uint64_t  msg_order;
1134                  uint64_t  comp_order;
1135                  size_t    total_buffered_recv;
1136                  size_t    size;
1137                  size_t    iov_limit;
1138              };
1139
1140   caps - Capabilities
1141       The  requested capabilities of the context.  The capabilities must be a
1142       subset of those requested of the associated endpoint.  See the CAPABIL‐
1143       ITIES  section  if  fi_getinfo(3)  for capability details.  If the caps
1144       field is 0 on input to fi_getinfo(3), the  applicable  capability  bits
1145       from the fi_info structure will be used.
1146
1147       The  following  capabilities  apply  to the receive attributes: FI_MSG,
1148       FI_RMA, FI_TAGGED, FI_ATOMIC, FI_REMOTE_READ, FI_REMOTE_WRITE, FI_RECV,
1149       FI_HMEM,  FI_TRIGGER,  FI_RMA_PMEM,  FI_DIRECTED_RECV, FI_VARIABLE_MSG,
1150       FI_MULTI_RECV, FI_SOURCE, FI_RMA_EVENT, FI_SOURCE_ERR,  and  FI_COLLEC‐
1151       TIVE.
1152
1153       Many  applications will be able to ignore this field and rely solely on
1154       the fi_info::caps field.  Use of this field provides fine grained  con‐
1155       trol  over the receive capabilities associated with an endpoint.  It is
1156       useful when handling scalable endpoints,  with  multiple  receive  con‐
1157       texts,  for  example, and allows configuring a specific receive context
1158       with fewer capabilities than that supported by the  endpoint  or  other
1159       receive contexts.
1160
1161   mode
1162       The operational mode bits of the context.  The mode bits will be a sub‐
1163       set of those associated with the endpoint.  See  the  MODE  section  of
1164       fi_getinfo(3)  for details.  A mode value of 0 will be ignored on input
1165       to fi_getinfo(3), with the mode value of the fi_info structure used in‐
1166       stead.   On  return  from  fi_getinfo(3),  the mode will be set only to
1167       those constraints specific to receive operations.
1168
1169   op_flags - Default receive operation flags
1170       Flags that control the operation of operations  submitted  against  the
1171       context.  Applicable flags are listed in the Operation Flags section.
1172
1173   msg_order - Message Ordering
1174       For  a  description of message ordering, see the msg_order field in the
1175       Transmit Context Attribute section.  Receive context  message  ordering
1176       defines  the order in which received transport message headers are pro‐
1177       cessed when received by an endpoint.  When ordering is  set,  it  indi‐
1178       cates that message headers will be processed in order, based on how the
1179       transmit side has identified the messages.  Typically, this means  that
1180       messages  will  be  handled  in order based on a message level sequence
1181       number.
1182
1183       The following ordering flags, as defined for  transmit  ordering,  also
1184       apply  to  the processing of received operations: FI_ORDER_NONE, FI_OR‐
1185       DER_RAR, FI_ORDER_RAW, FI_ORDER_RAS, FI_ORDER_WAR, FI_ORDER_WAW, FI_OR‐
1186       DER_WAS,  FI_ORDER_SAR,  FI_ORDER_SAW,  FI_ORDER_SAS, FI_ORDER_RMA_RAR,
1187       FI_ORDER_RMA_RAW,  FI_ORDER_RMA_WAR,  FI_ORDER_RMA_WAW,  FI_ORDER_ATOM‐
1188       IC_RAR,  FI_ORDER_ATOMIC_RAW,  FI_ORDER_ATOMIC_WAR,  and FI_ORDER_ATOM‐
1189       IC_WAW.
1190
1191   comp_order - Completion Ordering
1192       For a description of completion ordering, see the comp_order  field  in
1193       the Transmit Context Attribute section.
1194
1195       FI_ORDER_DATA
1196              When  set, this bit indicates that received data is written into
1197              memory in order.  Data ordering applies to  memory  accessed  as
1198              part of a single operation and between operations if message or‐
1199              dering is guaranteed.
1200
1201       FI_ORDER_NONE
1202              No ordering is defined for completed operations.  Receive opera‐
1203              tions  may complete in any order, regardless of their submission
1204              order.
1205
1206       FI_ORDER_STRICT
1207              Receive operations complete in the order in which they are  pro‐
1208              cessed by the receive context, based on the receive side msg_or‐
1209              der attribute.
1210
1211   total_buffered_recv
1212       This field is supported for backwards compatibility purposes.  It is  a
1213       hint to the provider of the total available space that may be needed to
1214       buffer messages that are received for which there is  no  matching  re‐
1215       ceive  operation.   The  provider may adjust or ignore this value.  The
1216       allocation of internal network  buffering  among  received  message  is
1217       provider specific.  For instance, a provider may limit the size of mes‐
1218       sages which can be buffered or the amount of buffering allocated  to  a
1219       single message.
1220
1221       If  receive  side buffering is disabled (total_buffered_recv = 0) and a
1222       message is received by an endpoint, then the behavior is  dependent  on
1223       whether  resource management has been enabled (FI_RM_ENABLED has be set
1224       or not).  See the Resource Management section of fi_domain.3  for  fur‐
1225       ther  clarification.   It  is  recommended that applications enable re‐
1226       source management if they  anticipate  receiving  unexpected  messages,
1227       rather than modifying this value.
1228
1229   size
1230       The  size of the receive context.  The mapping of the size value to re‐
1231       sources is provider specific, but it is directly related to the  number
1232       of  command  entries  allocated for the endpoint.  A smaller size value
1233       consumes fewer hardware and software resources, while a larger size al‐
1234       lows queuing more transmit requests.
1235
1236       While the size attribute guides the size of underlying endpoint receive
1237       queue, there is not necessarily a one-to-one mapping between a  receive
1238       operation  and  a  queue entry.  A single receive operation may consume
1239       multiple queue entries; for example, one per scatter-gather entry.  Ad‐
1240       ditionally,  the  size field is intended to guide the allocation of the
1241       endpoint's receive  context.   Specifically,  for  connectionless  end‐
1242       points, there may be lower-level queues use to track communication on a
1243       per peer basis.  The sizes of any lower-level queues may only  be  sig‐
1244       nificantly smaller than the endpoint's receive size, in order to reduce
1245       resource utilization.
1246
1247   iov_limit
1248       This is the maximum number of IO vectors (scatter-gather elements) that
1249       a single posted operating may reference.
1250

SCALABLE ENDPOINTS

1252       A  scalable  endpoint  is a communication portal that supports multiple
1253       transmit and receive contexts.  Scalable endpoints are loosely  modeled
1254       after  the  networking  concept  of transmit/receive side scaling, also
1255       known as multi-queue.  Support for scalable endpoints is domain specif‐
1256       ic.   Scalable  endpoints may improve the performance of multi-threaded
1257       and parallel applications, by allowing threads  to  access  independent
1258       transmit  and  receive queues.  A scalable endpoint has a single trans‐
1259       port level address, which can reduce the memory requirements needed  to
1260       store  remote  addressing data, versus using standard endpoints.  Scal‐
1261       able endpoints cannot be used directly  for  communication  operations,
1262       and  require  the application to explicitly create transmit and receive
1263       contexts as described below.
1264
1265   fi_tx_context
1266       Transmit contexts are independent transmit queues.  Ordering  and  syn‐
1267       chronization between contexts are not defined.  Conceptually a transmit
1268       context behaves similar to a send-only endpoint.   A  transmit  context
1269       may  be  configured  with fewer capabilities than the base endpoint and
1270       with different attributes (such as  ordering  requirements  and  inject
1271       size)  than  other contexts associated with the same scalable endpoint.
1272       Each transmit context has its own  completion  queue.   The  number  of
1273       transmit  contexts associated with an endpoint is specified during end‐
1274       point creation.
1275
1276       The fi_tx_context call is used to retrieve a specific context,  identi‐
1277       fied  by  an  index  (see  above  for  details  on transmit context at‐
1278       tributes).  Providers may dynamically allocate contexts when fi_tx_con‐
1279       text  is called, or may statically create all contexts when fi_endpoint
1280       is invoked.  By default, a transmit context inherits the properties  of
1281       its  associated  endpoint.   However,  applications may request context
1282       specific attributes through the attr parameter.  Support for per trans‐
1283       mit  context  attributes  is  provider  specific  and  not  guaranteed.
1284       Providers will return the actual attributes  assigned  to  the  context
1285       through the attr parameter, if provided.
1286
1287   fi_rx_context
1288       Receive  contexts are independent receive queues for receiving incoming
1289       data.  Ordering and synchronization between contexts  are  not  guaran‐
1290       teed.  Conceptually a receive context behaves similar to a receive-only
1291       endpoint.  A receive context may be configured with fewer  capabilities
1292       than  the base endpoint and with different attributes (such as ordering
1293       requirements and inject size) than other contexts associated  with  the
1294       same  scalable  endpoint.   Each receive context has its own completion
1295       queue.  The number of receive contexts associated with an  endpoint  is
1296       specified during endpoint creation.
1297
1298       Receive contexts are often associated with steering flows, that specify
1299       which incoming packets targeting a scalable endpoint to process.   How‐
1300       ever,  receive  contexts  may be targeted directly by the initiator, if
1301       supported by the underlying protocol.  Such contexts are referred to as
1302       'named'.   Support  for named contexts must be indicated by setting the
1303       caps FI_NAMED_RX_CTX capability when the corresponding endpoint is cre‐
1304       ated.   Support  for named receive contexts is coordinated with address
1305       vectors.  See fi_av(3) and fi_rx_addr(3).
1306
1307       The fi_rx_context call is used to retrieve a specific context,  identi‐
1308       fied by an index (see above for details on receive context attributes).
1309       Providers may  dynamically  allocate  contexts  when  fi_rx_context  is
1310       called,  or  may statically create all contexts when fi_endpoint is in‐
1311       voked.  By default, a receive context inherits the  properties  of  its
1312       associated endpoint.  However, applications may request context specif‐
1313       ic attributes through the attr parameter.  Support for per receive con‐
1314       text  attributes  is  provider  specific and not guaranteed.  Providers
1315       will return the actual attributes assigned to the context  through  the
1316       attr parameter, if provided.
1317

SHARED CONTEXTS

1319       Shared  contexts  are  transmit  and receive contexts explicitly shared
1320       among one or more endpoints.  A shareable context allows an application
1321       to  use  a  single dedicated provider resource among multiple transport
1322       addressable endpoints.  This can greatly reduce the resources needed to
1323       manage  communication  over multiple endpoints by multiplexing transmit
1324       and/or receive processing, with the potential cost of  serializing  ac‐
1325       cess  across multiple endpoints.  Support for shareable contexts is do‐
1326       main specific.
1327
1328       Conceptually, shareable transmit contexts are transmit queues that  may
1329       be accessed by many endpoints.  The use of a shared transmit context is
1330       mostly opaque to an application.  Applications must allocate  and  bind
1331       shared  transmit  contexts  to endpoints, but operations are posted di‐
1332       rectly to the endpoint.  Shared transmit contexts  are  not  associated
1333       with completion queues or counters.  Completed operations are posted to
1334       the CQs bound to the endpoint.  An endpoint may only be associated with
1335       a single shared transmit context.
1336
1337       Unlike  shared  transmit  contexts, applications interact directly with
1338       shared receive contexts.  Users post  receive  buffers  directly  to  a
1339       shared  receive  context, with the buffers usable by any endpoint bound
1340       to the shared receive context.  Shared receive contexts are not associ‐
1341       ated  with completion queues or counters.  Completed receive operations
1342       are posted to the CQs bound to the endpoint.  An endpoint may  only  be
1343       associated  with  a single receive context, and all connectionless end‐
1344       points associated with a shared receive context  must  also  share  the
1345       same address vector.
1346
1347       Endpoints  associated  with a shared transmit context may use dedicated
1348       receive contexts, and vice-versa.  Or an endpoint may use shared trans‐
1349       mit  and  receive  contexts.  And there is no requirement that the same
1350       group of endpoints sharing a context of one type also share the context
1351       of  an  alternate type.  Furthermore, an endpoint may use a shared con‐
1352       text of one type, but a scalable set of contexts of the alternate type.
1353
1354   fi_stx_context
1355       This call is used to open a shareable transmit context (see  above  for
1356       details on the transmit context attributes).  Endpoints associated with
1357       a shared transmit context must use a subset of the  transmit  context's
1358       attributes.   Note  that  this  is  the  reverse of the requirement for
1359       transmit contexts for scalable endpoints.
1360
1361   fi_srx_context
1362       This allocates a shareable receive context (see above  for  details  on
1363       the  receive  context  attributes).  Endpoints associated with a shared
1364       receive context must use a subset of the receive context's  attributes.
1365       Note  that  this is the reverse of the requirement for receive contexts
1366       for scalable endpoints.
1367

SOCKET ENDPOINTS

1369       The following feature and description should be  considered  experimen‐
1370       tal.  Until the experimental tag is removed, the interfaces, semantics,
1371       and data structures associated with socket endpoints may change between
1372       library versions.
1373
1374       This  section  applies  to  endpoints  of  type  FI_EP_SOCK_STREAM  and
1375       FI_EP_SOCK_DGRAM, commonly referred to as socket endpoints.
1376
1377       Socket endpoints are defined with semantics that  allow  them  to  more
1378       easily  be  adopted by developers familiar with the UNIX socket API, or
1379       by middleware that exposes the socket API, while still taking advantage
1380       of high-performance hardware features.
1381
1382       The  key difference between socket endpoints and other active endpoints
1383       are socket endpoints use synchronous data  transfers.   Buffers  passed
1384       into  send and receive operations revert to the control of the applica‐
1385       tion upon returning from the function  call.   As  a  result,  no  data
1386       transfer  completions  are reported to the application, and socket end‐
1387       points are not associated with completion queues or counters.
1388
1389       Socket endpoints support  a  subset  of  message  operations:  fi_send,
1390       fi_sendv,  fi_sendmsg,  fi_recv,  fi_recvv,  fi_recvmsg, and fi_inject.
1391       Because data transfers are synchronous, the return value from send  and
1392       receive operations indicate the number of bytes transferred on success,
1393       or a negative value on error, including -FI_EAGAIN if the endpoint can‐
1394       not  send  or receive any data because of full or empty queues, respec‐
1395       tively.
1396
1397       Socket endpoints are associated with event queues and address  vectors,
1398       and  process  connection  management  events asynchronously, similar to
1399       other endpoints.  Unlike UNIX sockets, socket endpoint  must  still  be
1400       declared as either active or passive.
1401
1402       Socket endpoints behave like non-blocking sockets.  In order to support
1403       select and poll semantics, active socket endpoints are associated  with
1404       a  file  descriptor  that is signaled whenever the endpoint is ready to
1405       send and/or receive data.  The file descriptor may be  retrieved  using
1406       fi_control.
1407

OPERATION FLAGS

1409       Operation  flags  are  obtained by OR-ing the following flags together.
1410       Operation flags define the default flags applied to an endpoint's  data
1411       transfer  operations,  where  a flags parameter is not available.  Data
1412       transfer operations that take flags as input override the op_flags val‐
1413       ue of transmit or receive context attributes of an endpoint.
1414
1415       FI_COMMIT_COMPLETE
1416              Indicates  that a completion should not be generated (locally or
1417              at the peer) until the result of an  operation  have  been  made
1418              persistent.   See  fi_cq(3) for additional details on completion
1419              semantics.
1420
1421       FI_COMPLETION
1422              Indicates that a completion queue entry should  be  written  for
1423              data  transfer operations.  This flag only applies to operations
1424              issued on an endpoint that was bound to a completion queue  with
1425              the  FI_SELECTIVE_COMPLETION flag set, otherwise, it is ignored.
1426              See the fi_ep_bind section above for more detail.
1427
1428       FI_DELIVERY_COMPLETE
1429              Indicates that a completion should be generated when the  opera‐
1430              tion  has  been  processed  by the destination endpoint(s).  See
1431              fi_cq(3) for additional details on completion semantics.
1432
1433       FI_INJECT
1434              Indicates that all outbound data buffers should be  returned  to
1435              the  user's  control  immediately after a data transfer call re‐
1436              turns, even if the operation is  handled  asynchronously.   This
1437              may  require that the provider copy the data into a local buffer
1438              and transfer out of that buffer.  A provider can limit the total
1439              amount  of  send  data that may be buffered and/or the size of a
1440              single send that can use this flag.  This limit is indicated us‐
1441              ing inject_size (see inject_size above).
1442
1443       FI_INJECT_COMPLETE
1444              Indicates  that a completion should be generated when the source
1445              buffer(s) may be reused.  See fi_cq(3) for additional details on
1446              completion semantics.
1447
1448       FI_MULTICAST
1449              Indicates that data transfers will target multicast addresses by
1450              default.  Any fi_addr_t passed into a  data  transfer  operation
1451              will be treated as a multicast address.
1452
1453       FI_MULTI_RECV
1454              Applies to posted receive operations.  This flag allows the user
1455              to post a single buffer that will receive multiple incoming mes‐
1456              sages.  Received messages will be packed into the receive buffer
1457              until the buffer has been consumed.  Use of this flag may  cause
1458              a  single  posted receive operation to generate multiple comple‐
1459              tions as messages are placed into the buffer.  The placement  of
1460              received  data into the buffer may be subjected to provider spe‐
1461              cific alignment restrictions.  The buffer will  be  released  by
1462              the  provider  when  the  available buffer space falls below the
1463              specified minimum (see FI_OPT_MIN_MULTI_RECV).
1464
1465       FI_TRANSMIT_COMPLETE
1466              Indicates that a completion should be generated when the  trans‐
1467              mit operation has completed relative to the local provider.  See
1468              fi_cq(3) for additional details on completion semantics.
1469

NOTES

1471       Users should call fi_close to release all resources  allocated  to  the
1472       fabric endpoint.
1473
1474       Endpoints  allocated  with  the FI_CONTEXT or FI_CONTEXT2 mode bits set
1475       must typically provide struct fi_context(2) as their per operation con‐
1476       text  parameter.   (See fi_getinfo.3 for details.) However, when FI_SE‐
1477       LECTIVE_COMPLETION is enabled to suppress CQ completion entries, and an
1478       operation  is  initiated  without  the FI_COMPLETION flag set, then the
1479       context parameter is ignored.  An application does not need to pass  in
1480       a valid struct fi_context(2) into such data transfers.
1481
1482       Operations  that  complete  in error that are not associated with valid
1483       operational context will use the endpoint context in any error  report‐
1484       ing structures.
1485
1486       Although  applications  typically associate individual completions with
1487       either completion queues or counters, an endpoint can  be  attached  to
1488       both  a  counter and completion queue.  When combined with using selec‐
1489       tive completions, this allows an application to use counters  to  track
1490       successful  completions,  with  a CQ used to report errors.  Operations
1491       that complete with an error increment the error counter and generate  a
1492       CQ completion event.
1493
1494       As  mentioned  in  fi_getinfo(3),  the ep_attr structure can be used to
1495       query providers that support various endpoint  attributes.   fi_getinfo
1496       can return provider info structures that can support the minimal set of
1497       requirements (such that the application maintains correctness).  Howev‐
1498       er, it can also return provider info structures that exceed application
1499       requirements.   As  an  example,  consider  an  application  requesting
1500       msg_order  as  FI_ORDER_NONE.  The resulting output from fi_getinfo may
1501       have all the ordering bits set.  The application can reset the ordering
1502       bits it does not require before creating the endpoint.  The provider is
1503       free to implement a stricter ordering than is required by the  applica‐
1504       tion.
1505

RETURN VALUES

1507       Returns 0 on success.  On error, a negative value corresponding to fab‐
1508       ric errno is returned.  For fi_cancel, a return value  of  0  indicates
1509       that the cancel request was submitted for processing.
1510
1511       Fabric errno values are defined in rdma/fi_errno.h.
1512

ERRORS

1514       -FI_EDOMAIN
1515              A  resource  domain  was not bound to the endpoint or an attempt
1516              was made to bind multiple domains.
1517
1518       -FI_ENOCQ
1519              The endpoint has not been configured with necessary event queue.
1520
1521       -FI_EOPBADSTATE
1522              The endpoint's state does not permit the requested operation.
1523

SEE ALSO

1525       fi_getinfo(3),   fi_domain(3),   fi_cq(3)   fi_msg(3),    fi_tagged(3),
1526       fi_rma(3)
1527

AUTHORS

1529       OpenFabrics.
1530
1531
1532
1533Libfabric Programmer's Manual     2021-02-10                    fi_endpoint(3)
Impressum