plucky (3) rdma_join_multicast_ex.3.gz

NAME

       rdma_join_multicast_ex - Joins a multicast group with extended options.

SYNOPSIS

       #include <rdma/rdma_cma.h>

       int  rdma_join_multicast_ex  (struct  rdma_cm_id  *id, struct rdma_cm_join_mc_attr_ex *mc_join_attr, void
       *context);

ARGUMENTS

       id                  Communication identifier associated with the request.

       mc_join_attr        Is an rdma_cm_join_mc_attr_ex struct, as defined in <rdma/rdma_cma.h>.

       context             User-defined context associated with the join request.

DESCRIPTION

       Joins a multicast group (MCG) with extended options.  Currently supporting MC join with a specified  join
       flag.

       struct rdma_cm_join_mc_attr_ex {
               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 */
       };

       The supported join flags are:

       RDMA_MC_JOIN_FLAG_FULLMEMBER  - Create multicast group, Send multicast messages to MCG, Receive multicast
       messages from MCG.

       RDMA_MC_JOIN_FLAG_SENDONLY_FULLMEMBER - Create multicast group, Send multicast  messages  to  MCG,  Don't
       receive multicast messages from MCG (send-only).

       Initiating a MC join as "Send Only Full Member" on InfiniBand requires SM support, otherwise joining will
       fail.

       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.

RETURN VALUE

       Returns  0  on  success,  or  -1 on error.  If an error occurs, errno will be set to indicate the failure
       reason.

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.

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)

AUTHORS

       Alex Vesker <valex@mellanox.com>