Skip to main content
Vulnerability Database/CVE-2026-98272

CVE-2026-98272: Linux Kernel mvpp2 Buffer Overflow Flaw

CVE-2026-98272 is a buffer overflow vulnerability in the Linux kernel mvpp2 network driver that can cause out-of-bounds memory access during MTU changes. This article covers technical details, impact, and mitigation.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-98272 Overview

CVE-2026-98272 is a buffer overflow vulnerability in the Linux kernel's Marvell PPv2 (mvpp2) network driver. The flaw resides in the mvpp2_change_mtu() code path, which calls mvpp2_bm_switch_buffers(priv, true) without validating pool count constraints. When the per-processor buffering scheme is activated, the number of pools (nrxqs * 2) can exceed MVPP2_BM_MAX_POOLS (8), causing an out-of-bounds write to the priv->page_pool array in mvpp2_bm_init(). The issue was identified by the Linux Verification Center (linuxtesting.org) using the SVACE static analyzer. The vulnerability affects systems using Marvell Armada networking hardware with the mvpp2 driver.

Critical Impact

An MTU change operation on a vulnerable mvpp2 interface can trigger an out-of-bounds array access, potentially leading to kernel memory corruption and denial of service.

Affected Products

  • Linux kernel versions containing the mvpp2 driver prior to the fix commits
  • Systems using Marvell Armada SoCs with Packet Processor v2 (MVPP22 and newer) networking
  • Embedded and networking appliances relying on the Marvell PPv2 Ethernet driver

Discovery Timeline

  • 2026-10-06 - CVE-2026-98272 published to NVD
  • 2026-10-06 - Last updated in NVD database

Technical Details for CVE-2026-98272

Vulnerability Analysis

The mvpp2 driver supports a per-CPU buffer management scheme that allocates a page pool for each receive queue. The scheme requires that the total pool count, computed as nrxqs * 2, remain within the compile-time bound MVPP2_BM_MAX_POOLS, which equals 8. The mvpp2_probe() function enforces this bound during initial activation of percpu_pools.

The mvpp2_change_mtu() handler can invoke mvpp2_bm_switch_buffers(priv, true) to re-enable per-CPU buffers after an MTU change. This call path omits the pool count validation present in mvpp2_probe(). When mvpp2_get_nrxqs() returns MVPP2_PORT_MAX_RXQ, doubling the value produces an index that exceeds the priv->page_pool array bounds inside mvpp2_bm_init().

Root Cause

The root cause is missing input validation on a code path reachable after device initialization. The priv->page_pool array is sized for MVPP2_PORT_MAX_RXQ entries, but the per-CPU allocation logic indexes it with nrxqs * 2. Without a guard check on the hardware version and pool count before switching modes, the driver writes past the end of the array.

Attack Vector

Exploitation requires the ability to change the MTU on an interface backed by the mvpp2 driver. On affected hardware, a privileged local operation such as ip link set dev <iface> mtu <value> can trigger mvpp2_change_mtu() and reach the vulnerable mvpp2_bm_switch_buffers() path. The resulting out-of-bounds access corrupts adjacent kernel memory. The vulnerability was discovered through static analysis rather than an in-the-wild exploit.

See the upstream fix in Kernel Git Commit 14cb1e77 for the exact code changes.

Detection Methods for CVE-2026-98272

Indicators of Compromise

  • Unexpected kernel oops, panic, or KASAN: slab-out-of-bounds messages referencing mvpp2_bm_init or mvpp2_bm_switch_buffers in dmesg
  • Interface-level failures or driver resets on Marvell Armada platforms following an MTU change
  • Repeated or scripted MTU reconfiguration events on mvpp2-backed network interfaces

Detection Strategies

  • Enable Kernel Address Sanitizer (KASAN) in test environments to catch the out-of-bounds write at the point of occurrence
  • Audit kernel versions deployed on Marvell Armada hardware and compare against the fixed stable releases referenced in the upstream commits
  • Correlate netlink or rtnetlink MTU change events with subsequent kernel log anomalies on affected hosts

Monitoring Recommendations

  • Centralize dmesg and /var/log/kern.log ingestion from embedded and networking appliances to flag mvpp2 fault signatures
  • Alert on unprivileged processes attempting to invoke privileged network configuration commands
  • Track MTU change frequency per interface as a baseline deviation metric

How to Mitigate CVE-2026-98272

Immediate Actions Required

  • Identify fleet systems running the mvpp2 driver, typically Marvell Armada 7K/8K and compatible networking SoCs
  • Apply the stable kernel updates containing the upstream commits that add the pool count and hardware version check
  • Restrict network configuration privileges to administrators to limit who can trigger mvpp2_change_mtu()

Patch Information

The fix adds a guard that verifies the hardware version is MVPP22 or newer and that nrxqs * 2 does not exceed MVPP2_BM_MAX_POOLS before switching to per-CPU mode. The patch is distributed across multiple stable branches via the following commits: 14cb1e77, 1734c3fc, 4ac1d5ad, 5cefdb7a, 6569fd85, dfd46b55, e7f30dcf, and f0e7d62c.

Workarounds

  • Avoid runtime MTU changes on mvpp2 interfaces until the patched kernel is deployed
  • Where feasible, disable the per-CPU buffering scheme or constrain the number of RX queues so that nrxqs * 2 stays well below MVPP2_BM_MAX_POOLS
  • Use kernel command-line hardening and least-privilege policies to prevent unauthorized MTU reconfiguration

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.