CVE-2024-26720

What it actually does

Incorrect use of div_u64() on a 64-bit unsigned long thresh value in wb_dirty_limits() causes implicit truncation on 64-bit architectures, enabling a divide-by-zero condition when global writeback is configured with BDI_CAP_STRICTLIMIT and vm.dirty_bytes set to trigger thresh == (1<<32).

Why it matters to the business

Systems may experience kernel instability or crashes during memory writeback under specific dirty page configurations, increasing downtime risk on Debian 11 Linux deployments until the kernel package is updated; high downstream prevalence of the affected kernel amplifies operational exposure.

How an attack unfolds

An attacker with local access configures vm.dirty_bytes to a value that makes dtc->thresh equal (1<<32) on a BDI_CAP_STRICTLIMIT device, triggering the truncated divisor and divide-by-zero during writeback; the flaw is noted as easier to trigger in the global domain than memcg but remains difficult to weaponize reliably.

How to fix it

Upgrade the linux package on Debian 11 to version 5.10.216-1 or later; apply the corrected kernel that replaces div_u64() with div64_u64() and removes the redundant cast.

How to confirm the fix worked

Confirm the running kernel version reports 5.10.216-1 or newer via uname -r; review /proc/sys/vm/dirty_bytes after reboot to ensure configuration does not match the triggering value; re-test writeback paths under the previously problematic dirty_bytes setting.

Typical attack path

1. Obtain local shell or code execution on target system
2. Set vm.dirty_bytes via sysctl to (1<<32)*PAGE_SIZE on a strict-limit backing device
3. Induce writeback activity that calls wb_dirty_limits()
4. Trigger the truncated divisor leading to divide-by-zero

How to detect it

Search kernel logs (dmesg, /var/log/kern.log) for divide-by-zero faults or stack traces referencing wb_dirty_limits and dirty_throttle_control; inspect sysctl vm.dirty_bytes values for the triggering magnitude on strict-limit devices.

What your security team should do

• Immediately schedule upgrade of all Debian 11 systems to linux 5.10.216-1
• Audit sysctl settings for vm.dirty_bytes on production hosts and reset any values near (1<<32)*PAGE_SIZE
• Monitor kernel logs for divide-by-zero or wb_dirty_limits related oops entries until patches are deployed