1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
|
.TH "RDMA_JOIN_MULTICAST_EX" 3 "2017-11-17" "librdmacm" "Librdmacm Programmer's Manual" librdmacm
.SH NAME
rdma_join_multicast_ex \- Joins a multicast group with extended options.
.SH SYNOPSIS
.B "#include <rdma/rdma_cma.h>"
.P
.B "int" rdma_join_multicast_ex
.BI "(struct rdma_cm_id *" id ","
.BI "struct rdma_cm_join_mc_attr_ex *" mc_join_attr ","
.BI "void *" context ");"
.SH ARGUMENTS
.IP "id" 20
Communication identifier associated with the request.
.IP "mc_join_attr" 20
Is an rdma_cm_join_mc_attr_ex struct, as defined in <rdma/rdma_cma.h>.
.IP "context" 20
User-defined context associated with the join request.
.SH "DESCRIPTION"
Joins a multicast group (MCG) with extended options.
Currently supporting MC join with a specified join flag.
.P
.nf
struct rdma_cm_join_mc_attr_ex {
.in +8
uint32_t comp_mask; /* Bitwise OR between "rdma_cm_join_mc_attr_mask" enum */
uint32_t join_flags; /* Use a single flag from "rdma_cm_mc_join_flags" enum */
struct sockaddr *addr; /* Multicast address identifying the group to join */
.in -8
};
.fi
.P
The supported join flags are:
.P
.B RDMA_MC_JOIN_FLAG_FULLMEMBER
- Create multicast group, Send multicast messages to MCG, Receive multicast messages from MCG.
.P
.B RDMA_MC_JOIN_FLAG_SENDONLY_FULLMEMBER
- Create multicast group, Send multicast messages to MCG, Don't receive multicast messages from MCG (send-only).
.P
Initiating a MC join as "Send Only Full Member" on InfiniBand requires SM support, otherwise joining will fail.
.P
Initiating a MC join as "Send Only Full Member" on RoCEv2/ETH will not send any IGMP messages unlike a Full Member MC join.
When "Send Only Full Member" is used the QP will not be attached to the MCG.
.P
.SH "RETURN VALUE"
Returns 0 on success, or -1 on error. If an error occurs, errno will be
set to indicate the failure reason.
.SH "NOTES"
Before joining a multicast group, the rdma_cm_id must be bound to
an RDMA device by calling rdma_bind_addr or rdma_resolve_addr. Use of
rdma_resolve_addr requires the local routing tables to resolve the
multicast address to an RDMA device, unless a specific source address
is provided. The user must call rdma_leave_multicast to leave the
multicast group and release any multicast resources. After the join
operation completes, if a QP is associated with the rdma_cm_id,
it is automatically attached to the multicast group when the multicast
event is retrieved by the user. Otherwise, the user is responsible
for calling ibv_attach_mcast to bind the QP to the multicast group.
The join context is returned to the user through the private_data
field in the rdma_cm_event.
.SH "SEE ALSO"
rdma_join_multicast(3), rdma_leave_multicast(3), rdma_bind_addr(3), rdma_resolve_addr(3), rdma_create_qp(3),
rdma_get_cm_event(3)
.SH "AUTHORS"
.TP
Alex Vesker <valex@mellanox.com>
|
