page_pool_e...ct_recycling(9)

💅 page_pool_e...ct_recycling(9) — like, the whole page, bestie

okay so this is from the Kernel API Manual, and the function name got truncated in the header (it's page_pool_enable_direct_recycling, iykyk) — the real one is a little long for a header, tragic, we move.

---

✨ NAME ✨

page_pool_enable_direct_recycling — marks a page pool as owned by NAPI 🫶

---

🌸 SYNOPSIS 🌸

void page_pool_enable_direct_recycling (struct page_pool *pool , struct
napi_struct *napi );

---

🎀 ARGUMENTS 🎀

• pool — the page pool you're modifying, that's the one, she's the main character here 💅
• napi — the NAPI instance you're associating the page pool with 🦋

---

🔥 DESCRIPTION 🔥

So the vibe is: you're associating a page pool with a NAPI instance for lockless page recycling. Iconic. Basically this is useful when a new page pool has to be added to a NAPI instance without disabling that NAPI instance — you use this to mark the exact point at which the control path "hands over" the page pool to the NAPI instance. The handoff. The moment. ⭐

In most cases a driver can simply set the napi field in struct page_pool_params, and does not have to call this helper. fr fr, check before you commit 💀

Now the fine print, and bestie, listen:

• The function is idempotent — calling it again changes nothing, she's stable like that ✨
• But it does not implement any refcounting. 🚨 A single page_pool_disable_direct_recycling will disable recycling, no matter how many times enable was called. So don't go thinking you stacked up enable calls for safety, you didn't, that's not how this works 😭

---

🌈 SEE ALSO 🌈

• Kernel file ./net/core/page_pool.c
• page_pool_get_stats(9)
• page_pool_create_percpu(9)
• page_pool_create(9)
• page_pool_put_netmem_bulk(9)

---

and that's the whole page, no crumbs left 🍓. this lil beauty lives in the kernel source under ./net/core, dated October 2026 — go off, October 2026, you ate 👑💖