CVE-2024-57939

What it actually does

die() in the RISC-V exception path acquires a spinlock_t that can sleep under PREEMPT_RT, violating atomic context rules and triggering warnings or oops during illegal-instruction traps.

Why it matters to the business

RHEL 7 systems running real-time or PREEMPT_RT kernels on RISC-V hardware may experience unexpected warnings, kernel oopses, or instability during exception handling, increasing operational risk until patched; the large number of affected kernel packages amplifies exposure across server and client deployments.

How an attack unfolds

An attacker triggers an illegal instruction exception on a vulnerable RISC-V system with PREEMPT_RT enabled; die() is invoked from the handler, attempts to take the sleeping spinlock, and produces a BUG warning or further oops, potentially aiding denial-of-service or crash-based attacks.

How to fix it

Upgrade all listed Red Hat Enterprise Linux 7 kernel packages to the fixed versions 0:3.10.0-327.22.2.el7 or later; rebuild and reboot systems after package installation; apply the upstream raw_spinlock_t change if building custom kernels.

How to confirm the fix worked

Confirm package versions match or exceed 0:3.10.0-327.22.2.el7 on all affected RHEL 7 server and client systems; run kernel self-tests exercising die() paths and verify absence of sleeping-function warnings in dmesg under PREEMPT_RT configuration.

Typical attack path

1. Trigger illegal instruction or other synchronous exception on RISC-V target
2. Exception handler calls die()
3. die() acquires spinlock_t under atomic context with PREEMPT_RT
4. Kernel logs sleeping-function warning and may oops or panic

How to detect it

Search kernel logs for sleeping-function warnings originating from die() or rt_spin_lock calls during exception handling; monitor for oops messages referencing do_trap_insn_illegal or _new_vmalloc_restore_context on RISC-V hardware.

What your security team should do

• Prioritize patching all RHEL 7 RISC-V systems within the next maintenance window due to the CISA-style exposure of kernel exception paths
• Scan for and remove any custom kernels or modules still using the pre-fix spinlock_t pattern
• Update monitoring rules to alert on die() related warnings or oopses until patches are deployed fleet-wide