__hw_addr_list_reconcile(9)

📦 version 7.3.1, in nixpkgs unstable

🦋 __hw_addr_list_reconcile — Kernel API Manual 🦋

okay bestie, so this page is all about __hw_addr_list_reconcile(9), and it lives in the Kernel API Manual, which is honestly giving very "the docs the docs don't want you to read" energy. sit down, hydrate, let's get into it 💅

✨ NAME ✨

__hw_addr_list_reconcile — literally the function that syncs your snapshot changes back and frees the snapshots afterwards. no cap, that's the whole job 🫶

🎀 SYNOPSIS 🎀

void __hw_addr_list_reconcile (struct netdev_hw_addr_list *real_list ,
struct netdev_hw_addr_list *work , struct netdev_hw_addr_list *ref , int
addr_len , struct netdev_hw_addr_list *cache );

five args, one void return, zero drama (well, okay, the lock drama is real — more on that below 😭). ✨

🌸 ARGUMENTS 🌸

  • real_list — the actual, real-deal address list that's getting updated. this is the main character of the whole operation 👑
  • work — the working snapshot ✨. heads up: your driver modifies this one via __hw_addr_sync_dev, so it's the "we've been busy" copy 🫶
  • ref — the reference snapshot, aka the untouched copy of the original state. it's the "before" photo, it does not get edited, it's just here for comparison 🎀
  • addr_len — the length of the addresses. that's it, that's the arg 🙂
  • cache — the entry cache where snapshot entries go back to get reused, because we recycle in this house ♻️💖

🔥 DESCRIPTION 🔥

so what does it actually do, bestie? it walks the reference snapshot (ref) and compares every single entry against the work snapshot (work) to compute the sync_cnt deltas. the way it just... does the math. obsessed 🧠✨

then it applies those deltas to real_list, because that's the list that's actually real (name checks out, fr fr).

after all that, it returns the snapshot entries to cache for reuse, and frees both snapshots — work and ref are gone, done, no takebacksies 💀. if you were planning on using them afterwards, don't 🙈.

⚠️ real talk though: the caller must hold netif_addr_lock_bh. like, this is not a suggestion, this is not a "eh, probably fine" — hold the lock, bestie. do not test me on this 🔒😤

🌈 SEE ALSO 🌈

kernel file ./net/core/dev_addr_lists.c (go read it, iykyk 👀) plus this whole friend group:

and that's the tea ☕✨ this lil page came from ./net/core, dated October 2026, and it's part of the Kernel API Manual — __hw_addr_list_reconcile(9), the one and only. stay locked, stay synced, ily, bye 🎀💜🦋

📦 where this page comes from