kvm_pgtable_...estroy_range(9)

📦 version 7.3.1, in nixpkgs unstable

💅 kvm_pgtable_stage2_destroy_range(9) 💅

okay bestie, welcome to the Kernel API Manual, the vibe check you didn't know you needed 💖 This one's about yeeting an unlinked range of addresses out of existence, and yes the page title itself is doing that dramatic thing where it's abbreviated as kvm_pgtable_...estroy_range(9) — the ... is literally the manpage being mysterious on purpose iykyk 🎀

🌸 NAME 🌸

kvm_pgtable_stage2_destroy_range — Destroy the unlinked range of addresses. that's the whole job. iconic, focused, no notes 👑

✨ SYNOPSIS ✨

void kvm_pgtable_stage2_destroy_range (struct kvm_pgtable *pgt , u64 addr , u64 size );

it returns void because it gives nothing back and owes you nothing 💅

🦋 ARGUMENTS 🦋

  • pgt — the page-table structure, and bestie it has to be one that's already been initialised by kvm_pgtable_stage2_init*(). like, don't hand it a raw struct and expect magic 🌈
  • addr — the intermediate physical address where the mapping should be placed 🗺️
  • size — size of the mapping, fr fr, this is how much you're working with 🍓

💜 DESCRIPTION 💜

So here's the tea 🫖 — the page-table is assumed to be unreachable by any hardware walkers before all this freeing goes down. And because of that assumption, there is no TLB invalidation performed. none. zero. not a single flush. 🙈

⚠️ lowkey important: that's an assumption, not a promise the kernel makes for you. If some hardware walker out there can still see this page-table, you've got a whole situation on your hands. the manual is just telling you what it presumes, not what it checks 💀

🧚 SEE ALSO 🧚

a whole friend group of references, all here, all accounted for, all of them from the Kernel file ./arch/arm64/include/asm/kvm_pgtable.h:

and that's the whole page, bestie 🫶 signed off from ./arch/arm64/include/asm, dated October 2026, and filed under the very dramatic kvm_pgtable_...estroy_range(9) — go free some unlinked ranges, love u, drink water 💖✨🌸

📦 where this page comes from