resctrl_exit(9)

💅 resctrl_exit(9) — a Kernel API Manual moment ✨

okay bestie, buckle up, we're in the kernel API manual today (yes, the manual, we're fancy now 🎀) and this little function is lowkey one of the most dramatic exits in the whole filesystem game. no cap.

---

🌸 NAME 🌸

resctrl_exit — the one that removes the resctrl filesystem and frees the resources. it's giving cleanup era, it's giving "we're so done here" 💅

---

✨ SYNOPSIS ✨

void resctrl_exit (void );

literally nothing to pass in, she takes no notes, she takes no prisoners 🫶

---

🦋 ARGUMENTS 🦋

• void — no arguments. none. zero. she needs nothing from you, main character energy fr fr 👑

---

💖 DESCRIPTION 💖

so here's the tea ☕ — the architecture code calls this one in response to a fatal error. yeah, we're in our crisis-management era 😭. what she does is remove the resctrl files and structures from kernfs so nobody can be out here doing further configuration. configuration is cancelled. the vibe has been shut down.

⚠️ okay this part matters, bestie, don't skip it: when the architecture code calls resctrl_exit, all CPUs and resctrl domains must be offline. this is not a suggestion, this is the whole reason the thing is safe.

why tho? because when everything's offline, that ensures the limbo and overflow handlers are not scheduled to run 🌈 — which means the data structures they access can actually be freed by resctrl_l3_mon_resource_exit. like, they're not gonna reach back in and grab something that's already been yeeted 💀.

after resctrl_exit returns, the architecture code should return an error from all resctrl_arch_ functions that can do this. all of them. periodt.

and here's the iconic little twist 👑: resctrl_arch_get_resource must continue to return struct rdt_resources with the correct rid field — because that's what ensures the filesystem can be unmounted. she's still doing her job even after the party's over, we stan a reliable queen 🫶

---

🧚 SEE ALSO 🧚

the whole friend group got invited to this section, so let's give them their moment 💜:

• Kernel file ./fs/resctrl/rdtgroup.c
• closid_allocated(9)
• rdtgroup_mode_by_closid(9)
• rdtgroup_mode_str(9)
• rdtgroup_remove(9)
• rdtgroup_tasks_assigned(9)
• __rdtgroup_cbm_overlaps(9)
• rdtgroup_cbm_overlaps(9)
• rdtgroup_mode_test_exclusive(9)
• rdtgroup_cbm_to_size(9)
• rdtgroup_kn_mode_restrict(9)
• rdtgroup_kn_mode_restore(9)
• mon_get_kn_priv(9)
• mon_put_kn_priv(9)
• cbm_ensure_valid(9)
• mongrp_reparent(9)
• domain_setup_l3_mon_state(9)

like, that's the whole gc in one list 🔥

---

and that's the page!! 🎀 this little icon came straight out of ./fs/resctrl, dated October 2026, in her full resctrl_exit(9) era. bye bestie, go free those resources responsibly 💖✨