HIGH 7.8

CVE-2026-53356: Linux i915 Kernel Graphics Driver Buffer Offset Memory Access Vulnerability

A bug in the Linux kernel's graphics driver (i915) causes incorrect memory access when reading or writing data from physical memory-backed graphics objects with non-zero offsets. The flaw stems from a pointer type mismatch in the sg_page() function, which returns a struct page pointer but was being treated as a void pointer, causing wrong scaling calculations. This affects older Intel graphics hardware (Gen3/945G/Lakeport) that used physical memory mapping for overlay or cursor planes. An attacker with local access could exploit this to read or modify sensitive kernel memory.

Source data · NVD / CISA · public domain

CVSS
3.1 · 7.8 HIGH · CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Weaknesses (CWE)
Affected products
8 configuration(s)
Published / Modified
2026-07-01 / 2026-07-22

NVD description (verbatim)

In the Linux kernel, the following vulnerability has been resolved: drm/i915/gem: Fix phys BO pread/pwrite with offset sg_page() returns struct page pointer not (void *) so the scaling of pread/pwrite is wrong for phys BO and wrong parts of BO would be accessed if non-zero offset is used. Last impacted platform with overlay or cursor planes using phys mapping was Gen3/945G/Lakeport. (cherry picked from commit 3e49a2f85070b2fb672c1e0fdba281a4ea3aebe6)

8 reference(s) · View on NVD →

SEC.co analysis · AI-assisted, reviewed against source

Technical summary

The vulnerability exists in the i915 DRM driver's handling of physically-backed buffer objects (phys BO). When performing pread/pwrite operations with a non-zero offset, sg_page() returns a struct page pointer rather than a void pointer. The code incorrectly scales the offset value, treating it as if it were already a void pointer, resulting in accessing unintended memory locations. The type confusion causes the offset calculation to be off by a factor of the page structure size, leading to reads or writes at wrong addresses. This is particularly critical for systems using physical memory mapping for graphics planes, a configuration found on older Intel integrated graphics from the Gen3 era through Lakeport.

Business impact

For most modern systems, the practical business impact is limited because physical memory mapping for graphics planes has not been used since the Gen3/Lakeport era of Intel hardware. However, organizations still operating legacy systems with affected Intel integrated graphics may face confidentiality and integrity risks. A local attacker could read kernel memory or corrupt graphics-related kernel structures, potentially leading to privilege escalation or denial of service. The severity escalates in environments where legacy hardware is still in critical use, such as industrial control systems or specialized appliances.

Affected systems

Linux kernel versions containing the vulnerable i915 driver code are affected. The practical impact is restricted to systems with Intel Gen3 integrated graphics (945G/Lakeport era) that use overlay or cursor planes with physical memory mapping. Modern Intel graphics hardware uses different memory management schemes and is not vulnerable. Organizations should inventory systems running affected kernel versions, particularly those with legacy Intel integrated graphics hardware.

Exploitability

The vulnerability requires local code execution privileges to exploit, limiting the threat model to authenticated users or processes on the affected system. It is not remotely exploitable. The bug is triggered when an application calls pread/pwrite on a graphics buffer object with a non-zero offset, and requires understanding of the i915 driver's internal structures. Exploitability is moderate because the preconditions are specific, but local privilege escalation is feasible.

Remediation

Update the Linux kernel to a version containing the fix from commit 3e49a2f85070b2fb672c1e0fdba281a4ea3aebe6 or later. Verify the specific patch version for your distribution and kernel branch. As an interim mitigation, restricting local user access and disabling unnecessary graphics hardware acceleration on legacy systems reduces attack surface. For most organizations with modern hardware, no action is required.

Patch guidance

Patch availability depends on your Linux distribution and kernel branch. Consult your vendor's security advisory and kernel update channels for the specific fixed version. The upstream fix is commit 3e49a2f85070b2fb672c1e0fdba281a4ea3aebe6; verify that your patch includes this cherry-picked commit. Test patches in non-production environments before rollout, especially on systems with legacy graphics hardware where the risk is highest.

Detection guidance

Monitor kernel logs for unusual memory access patterns or graphics driver errors on systems with Gen3/Lakeport Intel graphics. Audit local user and process permissions to identify which accounts have graphics device access. In environments with affected hardware, monitor for unexpected privilege escalation or graphics driver crashes. Kernel Address Sanitizer (KASAN) builds may flag the incorrect offset calculation at runtime. File integrity monitoring on kernel memory regions can detect exploitation attempts.

Why prioritize this

Prioritize patching based on hardware inventory. Systems with modern Intel graphics or non-Intel GPUs require no immediate action. Legacy systems with Gen3/945G/Lakeport hardware should be patched before allowing untrusted local users to access graphics APIs. The CVSS score of 7.8 reflects the high local impact, but the practical risk is proportional to both the presence of vulnerable hardware and the trust model for local access. Organizations should defer this to a scheduled update cycle unless legacy graphics systems are exposed to hostile local users.

Risk score, explained

The CVSS 3.1 score of 7.8 (HIGH) reflects a local attack vector (AV:L), no special configuration required (AC:L), and a requirement for local user privileges (PR:L). The impact is severe: High confidence (C:H), integrity compromise (I:H), and availability impact (A:H) are all achievable through kernel memory manipulation. The score assumes a single impacted user session (S:U). In practice, risk is significantly lower for organizations without Gen3-era hardware, but remains high for those that do and that permit local user access to graphics subsystems.

Frequently asked questions

Does this vulnerability affect modern Intel graphics hardware?

No. This vulnerability is specific to Intel Gen3 integrated graphics (945G/Lakeport era) that used physical memory mapping for cursor and overlay planes. Modern Intel, AMD, and NVIDIA graphics use different memory management architectures and are not affected by this bug.

What is the required privilege level to exploit this?

An attacker must have local user-level access to the system and be able to execute code that calls pread/pwrite operations on i915 graphics buffer objects. Remote exploitation is not possible.

Will my Linux distribution automatically patch this through updates?

Yes, most distributions include upstream kernel fixes in their regular security updates. Check your distribution's advisory for the specific fixed kernel version and apply standard security updates to your systems.

What should I do if my organization still uses legacy Intel graphics hardware?

Prioritize applying the kernel patch, restrict local user access to graphics subsystems, and consider air-gapping or isolating legacy systems. For new deployments, migrate to modern hardware with current security maintenance.

This analysis is based on publicly available vulnerability data and vendor advisories. The CVSS score, published date, and affected product list reflect the authoritative source record. Patch versions and availability timelines should be verified against your specific Linux distribution and vendor advisory. This vulnerability analysis is for informational purposes and does not constitute professional security advice. Organizations should conduct their own risk assessment based on hardware inventory and local access policies. No exploit code or weaponized proof-of-concept is provided. References to commit hashes and fix details are accurate as of the source data; verify against upstream kernel repositories for the most current status. Source: NVD (public-domain), retrieved 2026-08-10. Analysis generated by SEC.co (claude-haiku-4-5).