drm_sched_fini(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… drm_sched_fini(9) β€” your fave scheduler teardown era πŸ’…

okay so this is the Kernel API Manual page for drm_sched_fini(9), and bestie, it is literally the Destroy a gpu scheduler function. No cap. 🫢

✨ NAME ✨

drm_sched_fini β€” Destroy a gpu scheduler πŸ’₯

πŸŽ€ SYNOPSIS πŸŽ€

void drm_sched_fini (struct drm_gpu_scheduler *sched );

πŸ¦‹ ARGUMENTS πŸ¦‹

  • sched β€” scheduler instance 🫢

🌈 DESCRIPTION 🌈

Okay so drm_sched_fini tears down and cleans up the scheduler. That's the whole assignment, and it ate. πŸ’…

It stops submission of new jobs to the hardware through struct drm_sched_backend_ops.run_job. Literally no new jobs sliding into the hardware queue. βœ‹βœ¨

If struct drm_sched_backend_ops.cancel_job is implemented, all jobs will be canceled through it and afterwards cleaned up through struct drm_sched_backend_ops.free_job. Orderly-ish? Iconic. πŸ’–

⚠️ real talk: if cancel_job is not implemented, memory could leak. 😭 That's not a slay, bestie β€” that's a leak. Keep it in mind.

πŸ‘‘ SEE ALSO πŸ‘‘

Kernel file ./drivers/gpu/drm/scheduler/sched_main.c is the main character here, and the whole linked crew is:

That's the whole page, bestie β€” from ./drivers/gpu/drm/scheduler, dated October 2026, all about drm_sched_fini(9). πŸ’–βœ¨

πŸ“¦ where this page comes from