nfsd4_revoke_states(9)

🌸 NAME 🌸

bestie, welcome to the Kernel API Manual entry for nfsd4_revoke_states — it's the function that revokes all nfsv4 states associated with a given filesystem. periodt. 💅

✨ SYNOPSIS ✨

void nfsd4_revoke_states (struct nfsd_net *nn , struct super_block *sb
       );

🎀 ARGUMENTS 🎀

• nn — used to identify the instance of nfsd (there is one per net namespace), okay bestie?
• sb — the super_block used to identify the target filesystem.

💖 DESCRIPTION 💖

okay so here's the tea: all nfs4 states held by the server instance and associated with a file on the given filesystem — that's open, lock, delegation, and layout states — will be revoked. That results in any files being closed, and so all references from nfsd to the filesystem get released. The glow-up? nfsd will no longer prevent the filesystem from being unmounted. slay. ✨

The clients that own those states will subsequently be notified that the states have been "admin-revoked". not me being dramatic about a kernel notification, but it's iconic. 💀

🦋 CONTEXT 🦋

bestie, this is a must not a maybe: the caller must hold nfsd_mutex with NFSD_NET_UP set. Outside that window, nn->conf_id_hashtbl is unallocated or freed, so the walk would dereference a NULL or dangling pointer. no cap, that's the bad timeline. 😭

🌈 SEE ALSO 🌈

• Kernel file ./fs/nfsd/nfs4state.c
• nfs4_delegation_exists(9)
• hash_delegation_locked(9)
• revoke_delegation(9)
• nfs4_replay_free_cache(9)
• nfsd4_revoke_export_states(9)
• nfsd_slot_shrinker_count(9)
• nfsd_slot_shrinker_scan(9)
• nfsd4_put_client(9)
• nfsd_wait_for_delegreturn(9)
• nfsd_breaker_owns_lease(9)
• nfsd4_deleg_lm_open_conflict(9)
• nfsd4_process_open2(9)
• nfsd4_force_end_grace(9)
• nfs4_preprocess_stateid_op(9)
• nfs4_preprocess_seqid_op(9)
• nfsd4_release_lockowner(9)
• nfsd4_vet_deleg_time(9)
• nfsd4_deleg_getattr_conflict(9)
• nfsd_get_dir_deleg(9)
• nfsd_update_cmtime_attr(9)

signed off from the ./fs/nfsd source with the October 2026 date, for nfsd4_revoke_states(9) — byeee bestie 🫶