fm10k_add_rx_frag(9)

🎀 fm10k_add_rx_frag(9) — the group chat edition 🎀

okay so real talk, bestie: we're in the fm10k section 9 kernel API manual, and today's main character is fm10k_add_rx_frag. she's a tiny little helper in drivers/net/ethernet/intel/fm10k/fm10k_main.c and honestly? she eats. let's get into it 💅

---

✨ NAME ✨

fm10k_add_rx_frag — Add contents of Rx buffer to sk_buff

like, literally what it says on the tin: the Rx buffer's contents get added into the sk_buff. no cap. 🦋

---

🌸 SYNOPSIS 🌸

bool fm10k_add_rx_frag (struct fm10k_rx_buffer *rx_buffer , unsigned int
size , union fm10k_rx_desc *rx_desc , struct sk_buff *skb );

she takes four things and gives you back a bool. iconic signature tbh. returns true/false, not a struct, not a pointer, just vibes and a boolean. 🫶

---

💖 ARGUMENTS 💖

• rx_buffer — the buffer containing the page to add. that's her whole moment, she's holding a page and she's ready to hand it over 🎀
• size — packet size from rx_desc. yes, the size comes from the descriptor, keep up bestie ✨
• rx_desc — the descriptor containing the length of the buffer written by hardware. hardware did the writing, rx_desc is the receipt 🔥
• skb — the sk_buff to place the data into. this is the destination, the final boss, the place where the data lives 💜

---

🦋 DESCRIPTION 🦋

okay so here's the tea. this function will add the data contained in rx_buffer->page to the skb. that's the whole assignment. 📚

how she does it is genuinely kind of clever — it depends:

• if the data in the buffer is less than the skb header size, she does a direct copy. simple, straightforward, no drama.
• otherwise, she just attaches the page as a frag to the skb. not me avoiding the copy… lowkey efficient 💅

then, the function will update the page offset if necessary — like, if things shifted, she handles it, don't worry — and return true if the buffer can be reused by the interface. so that bool wasn't decorative! it's a whole vibe check on whether your buffer lives to see another packet. ⭐

bestie psa: the return value is not a "did it work" — it's a "can the interface reuse this buffer now". read it right or you WILL be confused. 😭

---

👑 SEE ALSO 👑

Kernel file ./drivers/net/ethernet/intel/fm10k/fm10k_main.c

and the whole extended fam, in the order they were listed (yes, all of them, I'm not skipping anyone, everyone gets their moment 💖):

fm10k_init_module(9), fm10k_exit_module(9), fm10k_alloc_rx_buffers(9), fm10k_reuse_rx_page(9), fm10k_process_skb_fields(9), fm10k_is_non_eop(9), fm10k_cleanup_headers(9), fm10k_receive_skb(9), fm10k_get_tx_pending(9), fm10k_tx_timeout_reset(9), fm10k_clean_tx_irq(9), fm10k_update_itr(9), fm10k_set_qos_queues(9), fm10k_set_rss_queues(9), fm10k_set_num_queues(9), fm10k_reset_num_queues(9), fm10k_alloc_q_vector(9), fm10k_free_q_vector(9), fm10k_alloc_q_vectors(9), fm10k_free_q_vectors(9), fm10k_reset_msix_capability(9), fm10k_init_msix_capability(9), fm10k_cache_ring_qos(9), fm10k_cache_ring_rss(9), fm10k_assign_rings(9), fm10k_init_queueing_scheme(9), fm10k_clear_queueing_scheme(9)

go read them. all of them. you'll be smarter, i promise. 🌈

---

and that's the whole page, bestie!! 🍓 source is that lil abbreviated path ./drivers/net/ethernet/intel/f... and the date is October 2026 — so yes, this manual page is from the future, and she's still serving fm10k_add_rx_frag(9) realness. mwah 💋🧚‍♀️✨