finish_task_switch(9)

💅 finish_task_switch(9) — Kernel API Manual ✨

🎀 NAME 🎀

finish_task_switch — clean up after a task-switch. That's the whole name, periodt.

⭐ SYNOPSIS ⭐

struct rq * finish_task_switch (struct task_struct *prev );

💖 ARGUMENTS 💖

• prev — the thread we just switched away from. Hiii, previous task, we still see you. 👑

🦋 DESCRIPTION 🦋

Okay so bestie, here's the situation: finish_task_switch must be called after the context switch, and it's paired with a prepare_task_switch call before the context switch. They're a matching set, no cap.

finish_task_switch will reconcile locking set up by prepare_task_switch, and do any other architecture-specific cleanup actions. It's giving cleanup era, and honestly? Slay.

Not me remembering that we may have delayed dropping an mm in context_switch. If so, we finish that here outside of the runqueue lock.

🚨 Doing it with the lock held can cause deadlocks; see schedule for details. So please don't. 💀

The context switch has flipped the stack from under us and restored the local variables which were saved when this task called schedule in the past. 'prev == current' is still correct, but we need to recalculate this_rq because prev may have moved to another CPU. Literally recalculating the vibe check. 🌈

🧚 SEE ALSO 🧚

Kernel file: ./kernel/sched/core.c

• wake_q_add(9)
• wake_q_add_safe(9)
• task_curr(9)
• try_to_wake_up(9)
• task_call_func(9)
• cpu_curr_snapshot(9)
• wake_up_process(9)
• preempt_notifier_register(9)
• preempt_notifier_unregister(9)
• prepare_task_switch(9)
• schedule_tail(9)
• schedule_preempt_disabled(9)
• preempt_schedule_notrace(9)
• init_idle(9)
• curr_task(9)

Cute little sign-off: this page comes from the ./kernel/sched source/version, stamped October 2026, and it's all about finish_task_switch(9). Bye bestie, go forth and finish that task switch correctly. 🫶💅