CVE-2026-90261

Name of the Vulnerable Software and Affected Versions Linux kernel (affected versions not specified)

Description An issue exists in the btrfs zoned filesystem where the btree writepages() function writes dirty metadata in ascending logical address order. On zoned filesystems, only one metadata and one system block group can be active for writing simultaneously. When the active block group is at a higher logical address than another group containing dirty metadata, the system must finish the active group to activate the lower one. However, if there is unsent IO during specific writeback conditions, the system refuses to wait under the fs info->zoned meta io lock to avoid a deadlock. This causes the btrfs check meta write pointer() function to return -EAGAIN, which the btrfs write and wait transaction() function treats as a fatal error, aborting the transaction and forcing the filesystem into read-only mode. This behavior occurs intermittently during metadata-heavy relocation.

Recommendations At the moment, there is no information about a newer version that contains a fix for this vulnerability.