CVE-2024-53136

What it actually does

The reverted change introduced locking that triggers deadlocks during concurrent shmem_getattr operations when tmpfs is exported over NFS, turning a theoretical data-race report into a practical availability issue.

Why it matters to the business

Systems serving tmpfs filesystems via NFS may experience sudden hangs or require reboots, disrupting file services and any dependent workloads until the kernel is updated.

How an attack unfolds

An attacker with local access or the ability to trigger NFS operations against a tmpfs mount repeatedly performs getattr and related file operations from multiple threads or clients, forcing the deadlock condition introduced by the earlier data-race patch.

How to fix it

Upgrade the Linux kernel to commit d5226fa6dbae0569ee43ecfc08bdcd6770fc4755 or later; apply the revert patch if building from the affected tree; test NFS+tmpfs workloads after deployment.

How to confirm the fix worked

Confirm the running kernel contains commit d5226fa6dbae or verify that shmem_getattr no longer holds the lock that caused the NFS deadlock; monitor for absence of hung tasks in NFS server threads under load.

Typical attack path

1. Mount a tmpfs filesystem and export it over NFS
2. From one or more clients, issue repeated stat/getattr calls while other processes modify the same inodes
3. Trigger the locking sequence added in the reverted commit, causing the NFS server thread to deadlock
4. Repeat until the system becomes unresponsive or requires manual intervention

How to detect it

Inspect kernel logs for hung-task warnings involving shmem or NFS server threads; look for repeated getattr traffic patterns on tmpfs mounts preceding system unresponsiveness.

What your security team should do

• Prioritize kernel updates on all systems exporting tmpfs over NFS
• Add the fixed commit to internal kernel build pipelines and CI regression tests
• Review NFS export policies to limit exposure of tmpfs mounts until patched
• Monitor vendor advisories for backported fixes in distribution kernels