migrate_vma_...uge_pmd_page(9)

📦 version 7.3.1, in nixpkgs unstable

💅 migrate_vma_...uge_pmd_page(9) — Kernel API Manual, but make it make sense

okay so bestie, welcome to the Kernel API Manual page for migrate_vma_insert_huge_pmd_page(9) — yes the title is truncated with ... and no we're not fixing it, that's just how the man page header do. 🎀

✨ NAME ✨

migrate_vma_insert_huge_pmd_page — Insert a huge folio into @migrate->vma->vm_mm at @addr. The folio is already allocated as a part of the migration process with large page. literally that's the whole vibe. 🦋

🌸 SYNOPSIS 🌸

int migrate_vma_insert_huge_pmd_page (struct migrate_vma *migrate ,
unsigned long addr , struct page *page , unsigned long *src , pmd_t
*pmdp );

like, five arguments, returns an int, she's serving signature realness ✨

💖 ARGUMENTS 💖

  • migrate — the migrate_vma arguments. the whole context package, the main character of the call fr fr.
  • addr — address where the folio will be inserted. this is where the magic happens 📍
  • page — page to be inserted at addr. she's the one being placed, no cap.
  • src — src pfn which is being migrated. the origin story 💫
  • pmdp — pointer to the pmd. handle with care bestie 🫶

🔥 DESCRIPTION 🔥

so page needs to be initialized and setup after it's allocated. not optional, not a suggestion — she has to be ready before she gets inserted. 👑

The code bits here follow closely the code in __do_huge_pmd_anonymous_page. like, closely closely. if you know that function, you already know the vibe iykyk 💅

⚠️ big important caveat: This API does not support THP zero pages. that's a hard no, not a "maybe later". don't come for her about it. 💀

🌈 SEE ALSO 🌈

Kernel file ./mm/migrate_device.c — and the whole friend group:

they ate tbh. go read them. 📚✨

and that's the page!! 🎀🦋💖 brought to you straight out of ./mm, dated October 2026, still titled migrate_vma_...uge_pmd_page(9) because the ellipsis is iconic and we respect the lore. periodt. 👑💅

📦 where this page comes from