ib_process_cq_direct(9)

📦 version 7.3.1, in nixpkgs unstable

💅 ib_process_cq_direct(9) — Kernel API Manual ✨

Okay bestie, so this page is literally all about ib_process_cq_direct from the Kernel API Manual — it's giving "I do my own CQ processing, no middleman" energy and honestly? Iconic. 🎀

🌸 NAME 🌸

ib_process_cq_direct — process a CQ in caller context. That's it, that's the whole vibe. 💖

✨ SYNOPSIS ✨

int ib_process_cq_direct (struct ib_cq *cq , int budget );

🦋 ARGUMENTS 🦋

  • cq — the CQ to process, aka the main character of this whole situation 👑
  • budget — number of CQEs to poll for. That's how many completions we're willing to scoop up, fr fr. 🔥

💜 DESCRIPTION 💜

So this function is used to process all outstanding CQ entries — like, every single one that's sitting there waiting. 🫶

The way that it does not offload CQ processing to a different context? Honestly a serve. And it also does not ask for completion interrupts from the HCA — it's fully DIY, no interruptions needed. 💅

⚠️ But bestie, listen: using direct processing on a CQ with non IB_POLL_DIRECT type may trigger concurrent processing. So like… be aware of that, okay? 😭

💀 NOTE 💀

Do not pass -1 as budget unless it is guaranteed that the number of completions that will be processed is small. Like seriously, do not do it. That's the whole note. 🙈

🧚 SEE ALSO 🧚

Kernel file ./drivers/infiniband/core/cq.c — and then the whole friend group:

And that's the page! 🍓 Sourced from ./drivers/infiniband/core, dated October 2026 — catch you in the next manpage, bestie! 🌈💖

📦 where this page comes from