GPU VulnDB

Database/Kernel, userspace & hypervisor

Linux kernel FUSE: setattr writeback failure leaks the invalidate lock on a DAX truncate

UnscoredCVE-2026-80856Kernel, userspace & hypervisorcurated

Impact

A second lock leak in the same FUSE DAX truncate machinery, on a different path from CVE-2026-80855. fuse_do_setattr() takes filemap_invalidate_lock() for a DAX truncate and normally drops it at the out:/error: labels, but when the request also needs a writeback flush and write_inode_now() fails - for example the FUSE daemon returns -EIO - the function returns directly and leaks the rwsem. The reachable trigger given in the record is truncate(2) on a setuid file, where privilege stripping adds ATTR_MODE alongside ATTR_SIZE. Afterwards every fault or truncate on that inode stalls forever, hanging the jobs that touch it and leaving uninterruptible tasks that make the node awkward to drain.

Who can reach it

Local unprivileged user on a host or guest with a DAX-enabled FUSE/virtiofs mount that also uses the writeback cache. The caller truncates a setuid file while the FUSE daemon returns a writeback error; no elevated privilege is needed. Nodes without FUSE DAX are unaffected.

What to do

Update to a stable kernel containing the fix (the linked commits add an unlock label covering the writeback-failure and fuse_dax_break_layouts() error returns) and reboot the node. Until then, mounts without the DAX window are not exposed, so inventory virtiofs mount options before deciding the window is urgent. The kernel commits are the only fix reference in this record.

References

Related entries

All Kernel, userspace & hypervisor 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.