CVE-2024-50032

What it actually does

Race condition in RCU NOCB handling allows synchronized IPI from softirq on an offline CPU, attempting scheduler calls after hrtimer migration and triggering warnings in smp_call_function_single.

Why it matters to the business

Linux systems may experience kernel warnings or instability during CPU offline events under RCU callback load, increasing operational risk until patched.

How an attack unfolds

A workload or timer enqueues RCU callbacks via softirq on a NOCB CPU transitioning offline; the deferred rcuog wake-up attempts a synchronized IPI to an online CPU, hitting the buggy scheduler path before rcutree_report_cpu_dead completes.

How to fix it

Upgrade the Linux kernel package to version 6.11.4-1 or later on all Debian 13 systems; no other mitigations are effective.

How to confirm the fix worked

Confirm installed kernel version is 6.11.4-1 or newer; run rcutorture or simulated CPU offline tests and verify absence of smp_call_function_single warnings in dmesg.

Typical attack path

1. CPU begins offline transition via multi_cpu_stop
2. Softirq enqueues RCU callbacks on NOCB CPU
3. Deferred wake-up issues smp_call_function_single IPI
4. Warning triggered in kernel/smp.c during irq_exit_rcu

How to detect it

Search kernel logs and dmesg for the smp_call_function_single warning triggered from swake_up_one_online during run_timer_softirq on offline CPUs; correlate with rcutorture or migration/ threads.

• WARNING: CPU: .* PID: .* at kernel/smp.c:633 smp_call_function_single
• Stopper: multi_cpu_stop+0x0/0x320
• swake_up_one_online
• __call_rcu_nocb_wake
• rcutorture

What your security team should do

• Apply the 6.11.4-1 kernel update across all affected systems without delay
• Monitor dmesg and kernel logs for the specific smp_call_function_single warning until patching completes
• Prioritize systems with frequent CPU hotplug or high RCU callback activity