GPU VulnDB

Database/Control plane, storage & DevOps

Linux kernel SUNRPC: use-after-free when tracing a newly accepted NFS server transport

UnscoredCVE-2026-99352Control plane, storage & DevOpscurated

Impact

On a kernel NFS server, svc_handle_xprt() puts a newly accepted transport on the temp-socket list before it emits the svc_xprt_accept tracepoint. If the client has already closed the connection, another nfsd thread can pick up the transport and free it. The tracepoint then reads the transport's addresses and network namespace from freed memory. The record describes this as a read of freed memory in a tracepoint. Tracepoint arguments are normally only evaluated while that tracepoint is enabled, so the realistic exposure is most likely limited to NFS servers where SUNRPC tracing is on. In that case the likely result is a kernel crash on the storage node serving shared datasets or home directories. The record does not mention privilege escalation or data exposure.

Who can reach it

Any network peer that can open a TCP connection to the nfsd service and close it straight away, racing the accepting thread. No authentication is needed, since the race happens at connection accept, before any NFS request. Only kernels running the in-kernel NFS server (nfsd) are affected, and most likely only while the svc_xprt_accept tracepoint is enabled.

What to do

Install a kernel that includes the stable fix (the commits listed below), which moves the tracepoint so it fires before the transport is published, then reboot each NFS server. That means a storage outage or a failover for clients. Until then, keeping SUNRPC/nfsd tracepoints disabled on production NFS servers should avoid the faulty read. The record does not list fixed kernel versions.

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.