nft_pipapo_avx2_fill(9)

📦 version 7.3.1, in nixpkgs unstable

✨ nft_pipapo_avx2_fill(9) ✨

ok so this is a page from the Kernel API Manual and it's about one lil' function that's lowkey the backbone of a whole bitmap situation 💅 pulling up the whole gang: nft_pipapo_avx2_fill and friends, all documented in nft_pipapo_avx2_fill(9) — yes bestie, the function and the manpage have the same name, it's giving main character energy 👑

🎀 NAME 🎀

nft_pipapo_avx2_fill — fill a bitmap region with ones. that's it. that's the vibe. filling bits with 1s ✨

🌸 SYNOPSIS 🌸

void nft_pipapo_avx2_fill (unsigned long *data , int start , int len );

💖 ARGUMENTS 💖

  • data — the base memory area, aka where the bits live, aka the home address of your bitmap 🏠
  • start — the first bit to set. where the party begins 🎉
  • len — count of bits to fill. how many bits are getting invited 🫶

🦋 DESCRIPTION 🦋

ok so picture this: it's literally nothing else than a version of bitmap_set — the same one used, e.g., by pipapo_refill — except it's been tailored for the microarchitectures that use it and it's better suited for this specific usage 👑

and why is it built like that? because it's very likely that we'll set a small number of bits, and those bits are not crossing a word boundary — so your branch prediction needs to be correct, and honestly? correct branch prediction is critical here. no cap. the CPU is out here guessing and we cannot let it guess wrong 😭

plot twist: this function doesn't actually use any AVX2 instruction. the name is a little misleading, bestie. it's the vibe of AVX2, not the instructions 🙈

🔥 SEE ALSO 🔥

kernel file ./net/netfilter/nft_set_pipapo_avx2.c — go read the source, iykyk 📖

and the whole squad, no one left behind:

ok that's the whole page, slay 💜 this one came from ./net/netfilter, dated October 2026, and it's filed under nft_pipapo_avx2_fill(9) — go forth and fill those bits, bestie 🎀✨

📦 where this page comes from