medium · CVSS v3 5.5

CVE-2026-98623

CVE-2026-98623 is a medium‑severity vulnerability in the Linux kernel’s Nouveau driver that allows an out‑of‑bounds read via the debugfs int

Overview

CVE-2026-98623 is a medium‑severity vulnerability in the Linux kernel’s Nouveau driver that allows an out‑of‑bounds read via the debugfs interface. The flaw occurs when an invalid pstate ID is supplied to nvkm_clk_ustate_update, causing a kernel memory read. It can potentially expose sensitive kernel data to privileged users.

Description

In the Linux kernel, the following vulnerability has been resolved: drm/nouveau/clk: fix list cursor use after loop in nvkm_clk_ustate_update If list_for_each_entry() exits without hitting a break then "pstate" is not a valid pstate pointer. Introduce a "found" variable instead. The check is reachable from userspace: nvkm_clk_ustate_update() takes the pstate id straight from the 'pstate' debugfs file, so requesting an id that is not in clk->states - or any id at all when the perf tables are broken and the list is empty - makes the pstate->pstate != req test dereference the list head cast to a struct nvkm_pstate, which is an out-of-bounds read. [Francesco: rebased on drm-misc-next, expanded the commit message]

Impact

The vulnerability permits a confidential read of kernel memory, potentially leaking sensitive data. It is limited to systems with the Nouveau driver and debugfs enabled. Attackers must have local or privileged access to read the /sys/kernel/debug/drm/nouveau/clk interface. The flaw does not directly modify data or crash the system.

Remediation

Apply the kernel patch that introduces a 'found' flag in nvkm_clk_ustate_update. Upgrade to a kernel version that includes the fix. Disable or restrict access to the Nouveau debugfs interface (/sys/kernel/debug/drm/nouveau/clk) to prevent unauthorized reads. Verify that performance tables are intact and that the list of pstates is not empty.

Risk context

The CVSS v3 score of 5.5 indicates medium risk. No EPSS data is available, so the urgency is moderate. Defenders should treat this as a medium‑severity issue and apply the patch promptly.

Affected products

  • Linux Kernel
  • Nouveau Driver
  • NVIDIA GPU

Scores

Severity
medium
CVSS v2
5
CVSS v3
5.5
CVSS v4
—
EPSS
—

kernel nouveau debugfs memory-read medium linux gpu

← All CVEs

Related CVEs

medium

CVE-2026-98723

In the Linux kernel, the following vulnerability has been resolved: nvmet: copy the hostid into the ctrl before creating PR pc_refs Commit 6202783184bf ("nvmet: Improve nvmet_alloc_ctrl() interface and implementation") added a second uuid_copy() of args->hostid near the end of nvmet_alloc_ctrl(), and commit 7b658153f1b8 ("nvmet: Remove duplicate uuid_copy") removed the original copy that sat before nvmet_ctrl_init_pr() instead of the new one. Since then nvmet_ctrl_init_pr() snapshots ctrl->hostid into the per-controller per-namespace reservation refs while the uuid_copy() from the connect data runs later, after the controller is published. The ctrl is allocated with kzalloc(), so every pc_ref created on this path stores the nil UUID. pc_ref->hostid has a single consumer: nvmet_pr_set_ctrl_to_abort() matches it against the preempted registrant's hostid to kill and drain the victim's in-flight I/O for Preempt and Abort. The match can never hit with the nil UUID, so whenever a namespace with reservations enabled exists before a host connects, which includes every reconnect, Preempt and Abort silently degrades into a plain Preempt: the preempting host sees success while the victim's in-flight I/O is still in the air. Copy the hostid where the rest of the connect data is consumed, before the controller is published and before nvmet_ctrl_init_pr() takes its snapshot.

View analysis

medium

CVE-2026-98619

A race condition in the Linux kernel’s ethtool ETHTOOL_TEST ioctl allows an offline self test to tear down and reconfigure a network interface without holding the required rtnl_lock. This can lead to device instability or crashes on ops‑locked drivers. The issue is limited to the ioctl path and affects specific drivers such as fbnic and bnxt.

View analysis

medium

CVE-2026-98622

This CVE involves a benign warning in the Linux kernel's traffic control conntrack action that can be triggered by an unprivileged user to cause a kernel panic when panic_on_warn is enabled. The issue arises when flow_offload_alloc() fails during a conntrack flush race. The patch removes the warning, preventing the panic.

View analysis

medium

CVE-2026-98630

In the Linux kernel, the following vulnerability has been resolved: eth: fbnic: Fix payload page pool error cleanup The payload page pool pointer contains an error pointer when its allocation fails. The cleanup path passes that error pointer to page_pool_destroy() instead of destroying the header page pool. This can dereference the error pointer and leave the header page pool allocated. Destroy the header page pool instead.

View analysis