CVE-2026-98147

A flaw was found in the Linux kernel printk subsystem. A local user could trigger a Denial of Service (DoS) by causing an unexpected system crash if the kernel is configured to panic on warnings. This issue occurs because the kernel inappropriately triggers a warning check when background print thread creation fails due to an interruption signal. By terminating the thread creation process, an attacker can provoke this warning condition and disrupt system availability.

A flaw was found in the Linux kernel printk subsystem where the creation of the background printing thread was guarded by WARN_ON(), so a kthread_run() failure triggered a kernel warning. Because __kthread_create_on_node() legitimately returns -EINTR when the creating task receives SIGKILL, a local user who terminates the thread-creation process can provoke this warning. The impact on Red Hat products is rated LOW: under the default kernel configuration this only produces a warning message and stack trace in the system log, and a system crash (kernel panic) causing denial of service is possible only when the non-default panic_on_warn setting is enabled, which also makes the crash scenario dependent on precise timing of signal delivery. The kernel in Red Hat Enterprise Linux 10 and Red Hat In-Vehicle OS 2 is affected, and the fix has been deferred for these streams given the low security impact and the warning-only nature of the flaw under default configurations; the residual risk is a log-only kernel warning on default installations and a potential panic only on systems configured with panic_on_warn. Earlier Red Hat Enterprise Linux versions and the OpenShift Container Platform RHEL CoreOS kernels are not affected because the vulnerable code is not present. The Red Hat impact rating of LOW is lower than the CVSS 3.1 base score of 5.5 (MODERATE) recorded for this flaw, reflecting that a denial of service requires a non-default kernel configuration.

The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.

The exposure to a system crash can be mitigated by disabling the panic-on-warning behavior: run 'sysctl -w kernel.panic_on_warn=0' and make the setting persistent by adding 'kernel.panic_on_warn = 0' to /etc/sysctl.conf or another sysctl configuration file. With panic_on_warn disabled, triggering this flaw produces only a kernel warning message and stack trace in the system log rather than a panic. panic_on_warn is not enabled by default, so default installations are exposed only to the log-only warning. Updating to a kernel release that contains the upstream fix also resolves the issue.