rdma_connect_locked(9)
πΈ rdma_connect_locked(9) β a Kernel API Manual moment πΈ
okay so bestie, welcome to the Kernel API Manual entry for rdma_connect_locked(9) β yes the (9) is part of the name, it's giving section-nine energy and we respect it π
β¨ NAME β¨
rdma_connect_locked β literally just Initiate an active connection request. no cap, that's the whole vibe π«Ά
π¦ SYNOPSIS π¦
int rdma_connect_locked (struct rdma_cm_id *id , struct rdma_conn_param
*conn_param );
look at her. she returns an int, she takes two pointers, she's iconic π
π ARGUMENTS π
idβ the Connection identifier to connect. that's your main character right there, the thing you're actually connecting πconn_paramβ Connection information used for connected QPs. all the deets your connected QPs need to know, packed up cute π
π DESCRIPTION π
So she's Same as rdma_connect β same energy, same job β BUT (and this is the important part, don't skip it π) she can only be called from the RDMA_CM_EVENT_ROUTE_RESOLVED handler callback.
fr fr, that's a hard requirement, not a suggestion. if you're not inside that handler callback, this is not your function, bestie. rdma_connect_locked is the locked one for a reason and we do not play with that π₯
π SEE ALSO π
the whole friend group, assembled π and also the Kernel file ./drivers/infiniband/core/cma.c:
rdma_is_consumer_reject(9)rdma_iw_cm_id(9)cma_acquire_dev_by_src_ip(9)cma_ib_acquire_dev(9)rdma_set_ack_timeout(9)rdma_set_min_rnr_timer(9)rdma_connect(9)rdma_connect_ece(9)rdma_accept(9)
and that's the page!! π§β¨
cute little sign-off from the source at ./drivers/infiniband/core, dated October 2026, filed under rdma_connect_locked(9) β mwah, go slay your connection request π
ππ¦
π¦ where this page comes from
- nixpkgs unstable: linux-manual 7.3.1
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash-1017 on : a language model's rewrite of the original, under its terms