Cloudanix VulnDB

CVE-2026-68185

None

CVSS

—

EPSS

0.2%

Published

Aug 10, 2026

CISA due date

—

Cloudanix analysis

Schedule a fix SSVC: Track

Certain Linux kernel settings on LoongArch hardware can make the computer refuse to start when an administrator tries to turn memory tracking off at boot.

Exploitation
Physical Access
Attacker needs
High Privilege
Complexity
High
Impact
Denial Of Service
Component
OS kernel
Class
Other
Remediation
Upgrade
Effort
Medium
Patch risk
Low

What it actually does

Kernel boot fails on LoongArch when CONFIG_MEM_ALLOC_PROFILING and CONFIG_MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT are both enabled and the sysctl.vm.mem_profiling=0 boot parameter is supplied, because early_param handlers can modify static keys before jump_label_init has run.

Why it matters to the business

Systems using these memory-profiling options cannot reliably boot after adding the documented disable parameter, causing unplanned downtime and requiring kernel rebuilds or parameter changes.

How an attack unfolds

An attacker with the ability to append or alter kernel boot parameters supplies sysctl.vm.mem_profiling=0 on a system built with the two profiling options; the resulting static-key modification before jump_label_init leaves the kernel unable to complete initialization and panics during early boot.

How to fix it

Update to a kernel that moves jump_label_init before parse_early_param on LoongArch; rebuild and deploy the corrected kernel image. As a workaround, avoid the combination of the two CONFIG options or do not pass the disabling boot parameter.

How to confirm the fix worked

Boot a test kernel with the two profiling options enabled and the sysctl.vm.mem_profiling=0 parameter; confirm the system reaches userspace without panic. Verify the ordering change is present in the source or in the built image via symbol inspection.

Typical attack path

  1. Identify target systems built with both memory-profiling Kconfig options enabled
  2. Obtain the ability to control or append kernel command-line parameters
  3. Append the parameter sysctl.vm.mem_profiling=0
  4. Trigger a reboot; the kernel panics before userspace starts

How to detect it

Examine kernel boot logs and dmesg for early panic messages referencing static keys or jump labels on LoongArch systems; search for the exact boot parameter string in grub.cfg, systemd-boot entries, or bootloader configuration.

  • sysctl.vm.mem_profiling=0
  • jump_label_init
  • parse_early_param

What your security team should do

  • Prioritize patching LoongArch kernels that enable memory allocation profiling
  • Audit boot-parameter management processes to prevent accidental introduction of the disabling parameter on affected builds
  • Add the corrected kernel to the standard update pipeline and test boot with profiling options

The record

In the Linux kernel, the following vulnerability has been resolved: LoongArch: Move jump_label_init() before parse_early_param() When enabling both CONFIG_MEM_ALLOC_PROFILING=y and CONFIG_MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT=y, then diabling memory profiling by adding the boot parameter 'sysctl.vm.mem_profiling=0' will cause the kernel failed to boot. After analysis, this is because jump_label_init() must be called before parse_early_param(), the early param handlers may modify static keys by static_branch_enable/disable(). Fix this by moving jump_label_init() to before parse_early_param(). The solution is similar to other architectures.

Affected packages

Ecosystem Package Introduced Fixed