nfp_nfd3_tx_tso(9)

🦋 nfp_nfd3_tx_tso(9) 🦋

okay so bestie, this lil page comes straight out of the Kernel API Manual and it's all about setting up a Tx descriptor for LSO. yes, that's the whole vibe. let's get into it 💅

---

✨ NAME ✨

nfp_nfd3_tx_tso — Set up Tx descriptor for LSO 🎀

---

🌸 SYNOPSIS 🌸

void nfp_nfd3_tx_tso (struct nfp_net_r_vector *r_vec , struct
       nfp_nfd3_tx_buf *txbuf , struct nfp_nfd3_tx_desc *txd , struct sk_buff
       *skb , u32 md_bytes );

it's giving void function, five args, zero drama ✨

---

💖 ARGUMENTS 💖

• r_vec — the per-ring structure, like the ring's whole personality 🫶
• txbuf — Pointer to driver soft TX descriptor (the soft one, the vibes one)
• txd — Pointer to HW TX descriptor (the hardware one, main character energy 👑)
• skb — Pointer to SKB 🌸
• md_bytes — Prepend length, lowkey important

---

🎀 DESCRIPTION 🎀

So what it does: Set up Tx descriptor for LSO, and for non-LSO skbs it does absolutely nothing — not a single thing, no cap, it just minds its business 💀

And it'll return error on packet header greater than maximum supported LSO header size. so if your header is too thicc for the max supported LSO header size, it's not gonna slay, it's gonna error 🙈

---

🔥 SEE ALSO 🔥

Kernel file ./drivers/net/ethernet/netronome/nfp/nfd3/dp.c — go read the source bestie, iykyk

• nfp_nfd3_tx_ring_stop(9)
• nfp_nfd3_tx_csum(9)
• nfp_nfd3_tx(9)
• nfp_nfd3_tx_complete(9)
• nfp_nfd3_rx_give_one(9)
• nfp_nfd3_rx_ring_fill_freelist(9)
• nfp_nfd3_rx_csum_has_errors(9)
• nfp_nfd3_rx_csum(9)
• nfp_nfd3_rx(9)
• nfp_nfd3_poll(9)

---

and that's the whole page, periodt 💜 it's sourced from ./drivers/net/ethernet/netrono..., dated October 2026, and it's nfp_nfd3_tx_tso(9) — stan the manpage, bye bestie 🎀✨🦋