rdma_restrack_begin_del(9)

💅 rdma_restrack_begin_del 💅

okay so we're in the Kernel API Manual, bestie — welcome to the resource-tracking corner of the kernel, it's giving very backstage-pass energy 🎀

---

✨ NAME ✨

rdma_restrack_begin_del — invalidates the object from the resource tracking database, but preserves its index in the array 💖 lowkey the whole personality of this function is "you're dead to me but you keep your seat".

And here's the part you actually need to know, no cap: since this keeps the index in the array until rdma_restrack_commit_del() gets called, if rdma_restrack_add() is called in between with an old QP ID, it can result in an untracked QP 😭🙈 like. that's the trap. that's the vibe check failing.

---

🌸 SYNOPSIS 🌸

void rdma_restrack_begin_del (struct rdma_restrack_entry *res );

---

🦋 ARGUMENTS 🦋

• res — resource entry 👑 (that's the struct of the thing you're starting to delete, mwah)

---

⭐ SEE ALSO ⭐

Kernel file ./drivers/infiniband/core/restrack.c — go read it, iykyk 📖✨ and while you're at it, the whole friend group:

• rdma_restrack_init(9)
• rdma_restrack_clean(9)
• rdma_restrack_count(9)
• rdma_restrack_attach_task(9)
• rdma_restrack_set_name(9)
• rdma_restrack_parent_name(9)
• rdma_restrack_new(9)
• rdma_restrack_add(9)
• rdma_restrack_abort_del(9)
• rdma_restrack_get_byid(9)
• rdma_restrack_sync(9)
• rdma_restrack_del(9)
• rdma_restrack_commit_del(9)

---

💀 reminder for the girlies: begin_del is not the whole delete. it's the beginning. don't be out here acting like the index is free when it literally is not 🫶

---

that's the whole page, bestie — brought to you straight outta ./drivers/infiniband/core, dated October 2026, in all its Kernel API Manual glory 🌈💜 slay, go track some resources responsibly 💅