CVE-2024-44971

What it actually does

Missing phy_device_free() after of_phy_find_device() in bcm_sf2_mdio_register() leaves an elevated refcount on struct device objects, causing a memory leak on each invocation of the registration routine.

Why it matters to the business

Long-running systems using the Broadcom SF2 DSA switch driver may experience gradual kernel memory exhaustion, leading to degraded network performance, instability, or eventual crashes that disrupt connectivity and require reboots.

How an attack unfolds

An attacker with local or privileged access triggers repeated calls to bcm_sf2_mdio_register(), for example by unloading and reloading the DSA driver or reconfiguring switch ports; each iteration leaks a small amount of memory until the system is starved of resources.

How to fix it

Update the Linux kernel to a version containing commit b7b8d9f5e679af60c94251fd6728dde34be69a71 or later; rebuild and reboot with the patched kernel. Where live patching is available, apply the vendor backport for the affected stable trees.

How to confirm the fix worked

Confirm the running kernel contains the fix by checking git log for b7b8d9f5e679af60c94251fd6728dde34be69a71 in drivers/net/dsa/bcm_sf2.c; monitor /proc/meminfo or slabtop for unexpected growth in device or PHY-related caches after repeated driver reloads.

Typical attack path

1. Obtain local or administrative access to a Linux system using the bcm_sf2 DSA driver
2. Trigger repeated MDIO device registration by cycling driver load/unload or port reconfiguration
3. Accumulate leaked kernel memory objects until resource exhaustion occurs
4. Cause denial of service through OOM killer activity or system crash

How to detect it

Monitor kernel logs for repeated device registration messages, track slab allocator growth for phy_device or device_private objects, and correlate with DSA switch configuration events; review audit logs for driver module load/unload activity on affected hosts.

What your security team should do

• Prioritize kernel upgrades on all systems exposing the bcm_sf2 driver within the next maintenance window
• Add kernel memory usage alerts for device and PHY slab caches in production monitoring
• Review switch configuration automation to avoid unnecessary driver reload cycles
• Track downstream distributions for backported packages and schedule updates accordingly