Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-98279 Overview
CVE-2026-98279 is a Linux kernel vulnerability in the Btrfs filesystem's fs-verity integration. When enable_verity() encounters a qgroup limit, the subsequent rollback_verity() call lacks its own metadata reservation. The failed rollback forces the entire filesystem read-only, even when the qgroup limit only applies to a single subvolume. Orphan cleanup at the next mount fails identically, so leftover verity items are never removed. With -EDQUOT, the affected subvolume remains unreachable. With -ENOSPC on a full filesystem, the next read-write mount fails outright.
Critical Impact
A per-subvolume quota exhaustion during verity enablement escalates into a filesystem-wide denial of service, blocking subsequent read-write mounts.
Affected Products
- Linux kernel Btrfs subsystem with fs-verity support
- Stable kernel branches referenced by commits 4b7ecab, 76bf149, and efa5780
- Systems using Btrfs qgroups alongside fs-verity
Discovery Timeline
- 2026-10-06 - CVE-2026-98279 published to NVD
- 2026-10-06 - Last updated in NVD database
Technical Details for CVE-2026-98279
Vulnerability Analysis
The flaw sits in the Btrfs verity enablement and rollback paths. When enable_verity() fails because a qgroup quota (-EDQUOT) or filesystem space limit (-ENOSPC) is reached, Btrfs invokes rollback_verity() to remove the partially written verity items. The rollback path begins a transaction without reserving its own metadata, so it fails under the same resource pressure that triggered the rollback.
Btrfs treats this failed cleanup as a serious inconsistency and forces the entire filesystem read-only. The impact extends beyond the affected subvolume because the global filesystem state is flipped, not just the quota-constrained subvolume.
The orphan cleanup path exhibits the same defect. On the next mount, btrfs_orphan_cleanup() attempts to remove leftover verity items but hits the same reservation failure. On full filesystems, this prevents read-write remount entirely, keeping the volume inaccessible for writes until space is freed through other means.
Root Cause
The root cause is missing use of the global metadata reserve in cleanup paths that only delete items and free space. btrfs_orphan_cleanup(), drop_verity_items(), and rollback_verity() each start transactions through standard paths that enforce qgroup accounting and ordinary reservations. Cleanup operations should be allowed to proceed against the global reserve because they ultimately return space rather than consume it.
Attack Vector
The vulnerability is triggered by local operations on a Btrfs volume with qgroups enabled. A user with write access to a subvolume subject to a qgroup limit can invoke fs-verity enablement on a file large enough to cross the quota threshold. The failed enablement cascades into a failed rollback, which flips the full filesystem to read-only. The patched commits replace the transaction start calls with btrfs_start_transaction_fallback_global_rsv(), allowing cleanup to draw from the global reserve and bypass the qgroup limit. See the Linux Kernel Commit 4b7ecab, Linux Kernel Commit 76bf149, and Linux Kernel Commit efa5780 for the fix details.
Detection Methods for CVE-2026-98279
Indicators of Compromise
- Kernel log entries showing Btrfs forced read-only transitions after fs-verity ioctls on quota-constrained subvolumes.
- -EDQUOT or -ENOSPC errors emitted by rollback_verity or btrfs_orphan_cleanup in dmesg.
- Failed read-write mounts on Btrfs volumes reporting leftover orphan verity items.
Detection Strategies
- Monitor dmesg and /var/log/kern.log for Btrfs transaction abort messages tied to verity operations.
- Audit systems running Btrfs with both qgroups and fs-verity enabled, as these are the required preconditions.
- Correlate filesystem read-only events with recent FS_IOC_ENABLE_VERITY ioctl activity by unprivileged users.
Monitoring Recommendations
- Alert on unexpected ro remounts of Btrfs volumes through host telemetry.
- Track btrfs qgroup show quota usage spikes immediately preceding filesystem errors.
- Collect kernel ring buffer events into a central log pipeline to detect repeat triggers across fleets.
How to Mitigate CVE-2026-98279
Immediate Actions Required
- Apply the stable kernel updates containing commits 4b7ecab, 76bf149, and efa5780.
- Reboot affected hosts after patching to load the fixed Btrfs module.
- Inventory Btrfs volumes that combine qgroups with fs-verity and prioritize those for remediation.
Patch Information
The fix starts transactions with btrfs_start_transaction_fallback_global_rsv() in btrfs_orphan_cleanup(), drop_verity_items(), and rollback_verity(). These paths only delete items and free space at completion, so drawing from the global reserve and skipping qgroup accounting is safe. Patched kernels are available through the stable tree references listed in the external advisories.
Workarounds
- Disable qgroups on Btrfs volumes where fs-verity is used until the kernel patch is deployed.
- Avoid enabling fs-verity on files within subvolumes approaching their qgroup limit.
- Maintain sufficient free space on Btrfs volumes to reduce the chance of -ENOSPC during cleanup.
# Check kernel version and Btrfs qgroup status before enabling fs-verity
uname -r
btrfs qgroup show -reF /mnt/subvolume
# Temporarily disable qgroups on the affected filesystem
btrfs quota disable /mnt/btrfs
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

