ptrace_init_task(9)

💫 ptrace_init_task — like, the Kernel API Manual page, bestie ✨

Okay so this is ptrace_init_task(9) from the Kernel API Manual, and honestly? It's giving "small function, big responsibilities" 🎀 Stick with me and you'll have it down fr fr.

---

🌸 NAME 🌸

ptrace_init_task — sets up the ptrace state for a brand-new child task, no cap 💅

---

✨ SYNOPSIS ✨

void ptrace_init_task (struct task_struct *child , bool ptrace );

---

🦋 ARGUMENTS 🦋

• child — the new child task in question, the little main character 👶
• ptrace — a bool, and it's true if the child should be ptrace'd by the parent's tracer 🔍

---

💜 DESCRIPTION 💜

So here's the vibe: this gets called immediately after child has been added to its parent's children list. Like, right after. Not before, not later. Timing is everything, bestie ⏰

About that ptrace flag — false is the normal, everyday case, and true means "yep, ptrace this child." That's the whole deal 🤷‍♀️

🚨 Locking lore, pay attention: this is called with current's siglock and write_lock_irq(tasklist_lock) held. Both of them. Simultaneously. Do not come in here unlocked, that's not the move 😭🔒

---

🔥 SEE ALSO 🔥

Kernel file ./include/linux/ptrace.h, plus the whole friend group:

• ptrace_may_access(9)
• ptrace_parent(9)
• ptrace_event_enabled(9)
• ptrace_event(9)
• ptrace_event_pid(9)
• ptrace_release_task(9)
• arch_has_single_step(9)
• user_enable_single_step(9)
• user_disable_single_step(9)
• arch_has_block_step(9)
• user_enable_block_step(9)
• arch_ptrace_stop_needed(9)
• arch_ptrace_stop(9)
• ptrace_report_syscall_permit_entry(9)
• ptrace_report_syscall_exit(9)

---

And that's the whole page, periodt 🫶 This little gem lives in ./include/linux, dated October 2026 — go tell your friends, iykyk 💖✨