Linux NFS localio: reference leak on nfs_uuid_add_file failure pins an nfsd_file and blocks netns teardown

Impact

When nfs_uuid_add_file() races with nfs_uuid_put() tearing down uuid->net, it returns -ENXIO without publishing nfl->nfs_uuid, so the later close path is a no-op. Meanwhile nfsd_open_local_fh() has already handed back a caller-owned nfsd_file reference and an embedded nfsd_net reference, and both are leaked on that error branch. Each occurrence pins one nfsd_file and its underlying struct file, dentry and inode, plus one nfsd_net_ref, which blocks nfsd_net and network-namespace teardown. On a node that repeatedly creates and destroys netns - anything running containers against a co-located NFS server with localio enabled - leaked namespaces accumulate and the node eventually needs a reboot to recover. The trigger is a race, so exposure depends on churn rather than on an attacker.

Who can reach it

Local, on a node that both serves and mounts NFS with the localio fast path active. Triggered by the race between opening a localio file handle and namespace teardown; no authentication step and no remote reach.

What to do

Patch the kernel on nodes using NFS localio so the error path releases the caller-owned file ref and its net ref, and reboot to reclaim already-pinned namespaces. If a reboot is not available, disabling the localio fast path removes the code path, at a performance cost. Nodes without localio are unaffected.

References

Related entries

• KVM arm64 vGICv2: guest write of an out-of-range GICV_DIR INTID triggers a host WARN and panics panic_on_warn hostsCVE-2026-89727 · Linux kernel KVM arm64 vGICv2 (GICV_DIR INTID handling)Unscored
• Linux x86/tdx: off-by-one GENMASK makes the port I/O value mask one bit too wide in a TDX guestCVE-2026-89751 · Linux kernel x86/tdx (handle_in/handle_out port I/O mask)Unscored
• Linux memcg: concurrent memory.max writers reclaim toward a stale target and can loop indefinitelyCVE-2026-89752 · Linux kernel memcg (memory.high / memory.max writer reclaim loop)Unscored
• Linux mm/vmscan: direct reclaim never reports an RCU-tasks quiescent state, stalling rcu_tasks under memory pressureCVE-2026-89753 · Linux kernel mm/vmscan (shrink_lruvec RCU-tasks quiescent state)Unscored
• Microsoft Hyper-V: vmswitch fails to validate guest OID requestsCVE-2021-28476 · Microsoft Hyper-VCritical
• Incus: instance snapshots bypass restricted.containers.lowlevel, giving command execution on the hostCVE-2026-48751 · Incus (instance snapshots ignore restricted.containers.lowlevel)Critical