Windows 11 May Silently Remove Discrete GPU Drivers Due to Eco Mode on Gaming Laptops
Windows 11 has introduced an unexpected behavior that can delete drivers for discrete graphics cards on gaming laptops when the device remains powered off for an extended period in Eco Mode.
The issue was first reported by the owner of an Asus ROG Zephyrus G14 equipped with a mobile GeForce RTX 5070 Ti. In this power-saving mode the discrete GPU is physically disconnected from the PCIe bus. After several weeks of operation, Windows concluded that the hardware had disappeared permanently and initiated cleanup.
Analysis of the setupapi.dev.log showed that the system component pnpclean.dll removed the video card after only 15 days of absence and then deleted the now supposedly unnecessary Nvidia driver package. Although the standard timeout is 30 days, certain devices can trigger earlier removal.
Upon returning to Standard Mode the GPU appeared in Device Manager as a Microsoft Basic Display Adapter with error code 10. The Nvidia application reported driver incompatibility, and the only reliable fix proved to be a complete driver reinstallation.
While the problem has been confirmed primarily on Asus ROG laptops, it may affect other models that fully disconnect discrete graphics. One owner of an older G14 with AMD Radeon graphics reported identical behavior. The developer of the G-Helper utility has already added a warning to the application.
The simplest preventive measure is to switch to Standard Mode every couple of weeks to reset the timer. Automatic cleanup can also be disabled with the following PowerShell command run as administrator:
Set-ItemProperty 'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Explorer\VolumeCaches\Device Driver Packages' -Name Autorun -Value 0
Disabling the feature stops Windows from automatically removing old drivers and unused devices. No official confirmation has been received from Microsoft so far.
Related articles
Critical PHP Vulnerabilities CVE-2026-109618 and CVE-2026-109619 Enable Unauthenticated Remote Code Execution
Two critical vulnerabilities in PHP, tracked as CVE-2026-109618 and CVE-2026-109619, both scoring CVSS 9.2, were disclosed on October 21. The flaws allow attackers to execute arbitrary code without authentication by triggering out-of-memory conditions that lead to use-after-free memory corruption. Exploitation involves specially crafted requests that manipulate uninitialized arrays and function pointers, with full chains demonstrated via PHP_SESSION_UPLOAD_PROGRESS and fatal error backtraces. The issues affect PHP versions starting from 5.4 and were discovered using AI tools Gemini 4 Argon and GPT-5.6 Sol, resulting in a patch exceeding 10,000 lines of code. Amazon detected attacks as early as October 17, with Cloudflare later reporting millions of exploitation attempts. Mitigation requires updating to patched versions such as PHP 8.2.35, 8.3.36, 8.4.27 or 8.5.12, while temporary workarounds include disabling session upload progress and fatal error backtraces.
Trend Micro ZDI Reports on Critical Linux KSMBD Vulnerabilities Leak to Public Mailing Lists Before Patches
Trend Micro Zero Day Initiative inadvertently published detailed vulnerability reports for the Linux kernel's KSMBD component into public mailing lists. Two flaws received CVSS scores of 9.0, enabling remote code execution under specific conditions. One issue involves a race condition during parallel SMB request handling, while the second stems from an incorrect buffer size calculation in the SMB Direct mechanism that requires RDMA and a writable share. Technical details including root causes and reproduction steps became publicly archived before fixes reached end users. The KSMBD maintainer has already merged a patch for ZDI-CAN-33774 into their repository, yet distribution kernels may still lack the update. Additional privilege-escalation and information-disclosure risks in other kernel components were also described in the leaked messages.
Node.js Issues Security Releases for 26.x, 24.x and 22.x Fixing High-Severity TLS and Brotli Flaws
Node.js published security updates on October 21, 2026 for the 26.x, 24.x and 22.x release lines. The patches address multiple high-severity vulnerabilities affecting TLS hostname validation and the Brotli decoder, along with updates to critical dependencies including OpenSSL and Undici. Several CVEs cover bypasses in certificate hostname checks when IP addresses use full-width digits or when hosts resolve loosely to IP addresses. A particularly serious out-of-bounds write in the Brotli decoder (CVE-2026-111497) could enable remote code execution on services processing compressed responses. Additional fixes resolve keep-alive socket timing issues, DNS resolution truncation with NUL characters, and entropy caching problems in crypto.randomUUID and randomInt functions. The releases also update c-ares, Undici, OpenSSL, SQLite and ca-cert bundles following the eMudhra intermediate CA compromise.
Atlassian October Security Bulletin Flags 176 Third-Party Library Vulnerabilities Across Jira, Confluence and Bitbucket
Atlassian released its October security bulletin detailing 176 vulnerability entries distributed across seven products including Jira, Confluence, Bamboo, Bitbucket, Crowd, Fisheye and Crucible. Of these, 162 entries carry high severity and 14 are rated critical, though the total does not represent 176 unique flaws because the same identifiers recur when multiple products share the same third-party library. All critical issues originate in external components rather than Atlassian code, with the most severe being CVE-2026-65637 in the Coyote component of Apache Tomcat used by Confluence, which allows interception of server-client connections. Affected versions and corresponding fixes were published for Confluence Data Center, Bamboo, Bitbucket Data Center, Jira Software and Jira Service Management, with the company stressing that administrators must compare exact build numbers against the full corrected-version table. No exploitation of the listed issues had been reported as of 22 October, and Atlassian classifies the inherited exposure from third-party libraries as lower priority, issuing separate critical alerts only for immediate-risk cases.