irdma_complete_cqp_request(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… irdma_complete_cqp_request β€” the kernel API manual, bestie edition πŸŽ€

okay so real talk, this one lives in section 9 of the Kernel API Manual and it's called irdma_complete_cqp_request. yes the name is a mouthful, no we're not changing it, it's iconic. 🫢


✨ SYNOPSIS ✨

void irdma_complete_cqp_request (struct irdma_cqp *cqp , struct
       irdma_cqp_request *cqp_request );
it returns void btw β€” nothing comes back out, it just Does The Thing. very "no thoughts, just vibes, but the vibes are cleanup". πŸ’…

🌸 ARGUMENTS 🌸

  • cqp β€” the device CQP. that's the command queue pair, the main character of this whole operation πŸ‘‘
  • cqp_request β€” the CQP request itself. the little guy getting completed πŸ₯Ί

πŸ’– DESCRIPTION πŸ’–

okay so here's what happens, step by step, no cap:

this function marks the CQP request as done, then it wakes up the thread that's been waiting on it (or invokes the callback function instead β€” whichever applies, iykyk), and then it releases/frees the CQP request. that's it. that's the whole entire job. post-completion cleanup, slay. ✨

so it's literally: mark done β†’ wake the waiter or run the callback β†’ free the request. three little steps and it ate. πŸ”₯


πŸ¦‹ SEE ALSO πŸ¦‹

the kernel file lives right here:

./drivers/infiniband/hw/irdma/hw.c

and now, the friend group β€” literally the whole irdma extended universe, we're listing every single one because nobody gets left out πŸ’œ:

like, the whole roster. go read them, they're all right there iykyk πŸ“š


anyway bestie that's the whole page πŸ’– this one's brought to you from ./drivers/infiniband/hw/irdma, dated October 2026, and it's irdma_complete_cqp_request(9) all day every day. go clean up your requests, periodt. πŸ’…βœ¨πŸŽ€

πŸ“¦ where this page comes from