xe_migrate_job_unlock(9)

📦 version 7.3.1, in nixpkgs unstable

💅 xe_migrate_job_unlock — Kernel API Manual ✨

okay so bestie, welcome to the xe_migrate_job_unlock(9) page, like the actual kernel API manual energy, and honestly? it's a whole vibe. buckle up 🎀

✨ NAME ✨

xe_migrate_job_unlock — Unlock migrate job lock 🔓💖

🌸 SYNOPSIS 🌸

void xe_migrate_job_unlock (struct xe_migrate *m , struct xe_exec_queue *q );

like, that's the whole signature, no cap. two args, void return, slay. ✨

🦋 ARGUMENTS 🦋

  • m — The migration context. that's your main character struct fr fr 💫
  • q — Queue associated with the operation which requires a lock. it's giving "this queue needs the lock too bestie" 🔐

💜 DESCRIPTION 💜

okay so here's the tea ☕ — this function unlocks the migrate job lock if the queue is a migration queue. that's the whole first branch, iconic.

but if it's not a migration queue? then it asserts that the VM's dma-resv is held. why? because user queues have their own locking, iykyk 🙈 — so no migrate-job-lock-unlocking happens there, and the assert is just the vibe check confirming the dma-resv is held.

bestie please note: that assert is not decorative. if the VM's dma-resv isn't held on the non-migration-queue path, that's a problem. the manpage says "assert," and we do not argue with asserts. 😭

🌈 SEE ALSO 🌈

Kernel file ./drivers/gpu/drm/xe/xe_migrate.c struct xe_migrate(9), plus the whole friend group 🫶:

and yes, xe_migrate_job_lock(9) is literally the sibling of this one — the lock/unlock duo ate and left no crumbs 👑✨

🎀 cute lil sign-off: this page lives at source/version ./drivers/gpu/drm/xe and it's dated October 2026, and the whole thing is filed under xe_migrate_job_unlock(9) — periodt. go unlock that lock bestie, you got this 💖🌸🧚
📦 where this page comes from