inet_proto_csum_replace16(9)

💖 inet_proto_csum_replace16(9) — the Kernel API Manual glow-up 💖

okay so bestie, this is inet_proto_csum_replace16(9) from the Kernel API Manual, and honestly? she's a little niche but she carries. we're talking layer 4 header checksum maintenance when an IPv6 address gets swapped. iykyk. let's get into it 🫶

---

✨ NAME ✨

inet_proto_csum_replace16 — update layer 4 header checksum field

like literally that's her whole job. one field. checksum. updated. slay 💅

---

🦋 SYNOPSIS 🦋

void inet_proto_csum_replace16 (__sum16 *sum , struct sk_buff *skb ,
const __be32 *from , const __be32 *to , bool pseudohdr );

no return value, no drama, just vibes and five arguments. periodt ✨

---

🎀 ARGUMENTS 🎀

• sum — the layer 4 header checksum field 💖 this is the thing getting updated, the main character of the whole operation honestly
• skb — the sk_buff for the packet, aka the packet's whole little suitcase 🧳
• from — the old IPv6 address, the one we're saying bye to 👋
• to — the new IPv6 address, the one sliding in 🌸
• pseudohdr — a bool; true if the layer 4 header checksum includes a pseudoheader. yes ma'am, that matters 👑

---

🌈 DESCRIPTION 🌈

so here's the tea 🍓: this function updates the layer 4 header as per the update in the IPv6 src/dst address. address moved, checksum follows. simple, iconic, done.

now the interesting part — and this is the bit you actually gotta remember for the vibe check 😭 — there is no need to update skb->csum in this function. why? because the update happens in two fields:

a.) the IPv6 src/dst address, and b.) the L4 header checksum

and those two cancel each other out for skb->csum calculation. like not me shipping a self-cancelling pair of updates 💀 flawless.

meanwhile inet_proto_csum_replace4 does need to update skb->csum, because that one updates three fields:

a.) the IPv4 src/dst address, b.) the IPv4 Header checksum, and c.) the L4 header checksum

and those result in the same diff as the L4 Header checksum for skb->csum calculation. so yeah, different math, different needs, both valid. we don't pit two icons against each other 💜

fr fr: the whole point here is that v6 gets to skip the skb->csum update and v4 doesn't. don't come for me, that's just what the math said 💅

---

🔥 SEE ALSO 🔥

• Kernel file ./net/core/utils.c
• in4_pton(9)
• in6_pton(9)
• inet_pton_with_scope(9)

go read the whole friend group, they're all in ./net/core/utils.c together and they ate 🍓✨

---

okay that's the whole page bestie, no cap, nothing skipped, nothing invented 🌸 this little gem is brought to you by the source at ./net/core, dated October 2026, and it lives at inet_proto_csum_replace16(9) forever and ever. mwah 💋🫶👑