STRUCT NETDEV_STAT_OPS(9)

🌸 struct netdev_stat_ops(9) 🌸

okay so bestie, welcome to the Kernel API Manual — this is the page for struct netdev_stat_ops, and it's literally the blueprint for getting fine-grained stats out of your netdev. let's get into it 💅

---

✨ NAME ✨

struct netdev_stat_ops — netdev ops for fine grained stats. that's it, that's the vibe 🎀

---

🦋 SYNOPSIS 🦋

struct netdev_stat_ops {
    void (*get_queue_stats_rx)(struct net_device *dev, int idx, struct
netdev_queue_stats_rx *stats);
    void (*get_queue_stats_tx)(struct net_device *dev, int idx, struct
netdev_queue_stats_tx *stats);
    void (*get_base_stats)(struct net_device *dev, struct
netdev_queue_stats_rx *rx, struct netdev_queue_stats_tx *tx);
 };

---

💖 Members 💖

• get_queue_stats_rx — grab the stats for one specific Rx queue, periodt 🍓
• get_queue_stats_tx — same deal but for a given Tx queue, no cap ✨
• get_base_stats — get the base stats, aka the ones that don't belong to any live instance 👑

---

🔥 Description 🔥

So like, the whole point is you're querying stats for a given object. and here's the crucial part bestie — the values of the statistics are undefined on entry, and yes, specifically they are not zero-initialized 😭 so don't go assuming. Drivers should assign values only to the statistics they actually collect. anything you don't collect? leave it undefined, don't touch it, hands off 🙈

Queue objects are not necessarily persistent, okay? only the currently active queues get queried by the per-queue callbacks. which means per-queue statistics will not generally add up to the total number of events for the device — it's giving "close but not the whole story". that's exactly where the get_base_stats callback comes in: it lets you fill in the delta between the events for currently live queues and the overall device history. get_base_stats can also be used to report any miscellaneous packets transferred outside of the main set of queues used by the networking stack. iconic multitasking honestly 🫶

Now, when the statistics for the entire device are queried, the order goes like this: first get_base_stats is issued to collect the delta, then a series of per-queue callbacks. Slay. And here's the twist — only statistics which are set in get_base_stats will be reported at the device level. Meaning that, unlike in queue callbacks, setting a statistic to zero in get_base_stats is a totally legitimate thing to do ⭐ this is because get_base_stats has a second function: designating which statistics are in fact correct for the entire device (think: when history for some of the events is not maintained, and a reliable "total" cannot be provided — yeah, in that case you genuinely can't give one, and that's fine).

Ops are called under the instance lock if netdev_need_ops_lock returns true, otherwise under rtnl_lock. Device drivers can assume that when collecting total device stats, the get_base_stats and subsequent per-queue calls are performed "atomically" (without releasing the relevant lock). 🔒 so like, don't worry about the lock slipping away mid-collection mid-character-energy.

💀 One thing drivers are encouraged to do: reset the per-queue statistics when the number of queues changes. Why? because the primary use case for per-queue statistics is currently to detect traffic imbalance — and stale numbers would absolutely ruin that vibe check.

---

🌈 SEE ALSO 🌈

• Kernel file ./include/net/netdev_queues.h
• struct netdev_config(9)
• struct netdev_queue_mgmt_ops(9)
• Lockless queue stopping / waking helpers.(9)
• netif_txq_maybe_stop(9)
• __netif_txq_completed_wake(9)

---

okay that's the whole page, you ate that 💅💖 this one came out of ./include/net — the Kernel API Manual source — dated October 2026. go forth and collect your fine-grained stats, bestie, ily 🧚✨👑