GPU VulnDB

Database/Control plane, storage & DevOps

Linux nfsd: POSIX ACL refs leak on every malformed NFSv4 OPEN compound

UnscoredCVE-2026-89664Control plane, storage & DevOpscurated

Impact

nfsd4_decode_fattr4() allocates refcounted posix_acl objects into open->op_pacl and op_dpacl, but OP_OPEN had no .op_release, and the release pair lived inside nfsd4_open(). When nfsd4_decode_open_claim4() returns a non-seqid-mutating error the dispatcher short-circuits and nfsd4_open() never runs, so both ACL references leak on every malformed OPEN compound that carried valid POSIX ACL createhow4 attributes. Resources attached to op->u also leaked on the v4.0 replay encoding path. A client can repeat the malformed compound freely, so this is an unbounded kernel memory leak driven remotely against a shared storage server - the failure mode is memory pressure and eventual OOM on a node the whole fleet mounts.

Who can reach it

Any client that can mount the NFSv4 export and send OPEN compounds, under normal export authentication. No server-local access needed.

What to do

Upgrade to a stable kernel with the referenced commits and reboot the NFS server, draining or failing over clients first. Until then, watch slab growth on the server; there is no runtime knob that disables the path short of dropping NFSv4. The record names commits, not fixed release numbers.

References

Related entries

All Control plane, storage & DevOps entries

This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.