Security researcher: ASUS ROG laptop UEFI violates PCIe spec, triggers crashes and black screens

There’s another bug in ASUS ROG laptop firmware

Discovered by the same researcher that revealed stutter issues in ASUS ROG laptops.

Security researcher Mohamed Maatallah, better known online as Zephkek , says he has found another firmware-level problem in ASUS ROG laptops. His new write-up argues this is baked into the UEFI and shows up regardless of OS, so it is not a Windows issue and not a driver issue.

His focus this time is a PCIe spec requirement around L1.2 power management timing. The rule is that both ends of the link should be programmed with matching values, but he says ASUS sets them differently on every boot . In his example, the CPU root port ends up at 765 microseconds, while the NVIDIA GPU side is set to 0 nanoseconds.

However, the ASUS UEFI programs the L1.2 Timing Thresholds incorrectly on every boot:

  • CPU Root Port: LTR_L1.2_THRESHOLD = 765us
  • NVIDIA GPU: LTR_L1.2_THRESHOLD = 0ns

The Consequence:

The GPU and CPU disagree on sleep exit timing, causing the PCIe link to desynchronize during power transitions.

In the forensic report for an ASUS ROG Strix G533ZW, the executive summary calls this a UEFI firmware defect caused by the L1.2 threshold mismatch, and ties it to instability during power transitions. The symptoms he lists line up with what users often describe as “random” instability: WHEA 0x124 crashes, black screens, hangs, and driver weirdness that can look different from one system to another.

We reposted the video from the Zephkek cloud storage:

VideoCardz : ASUS ROG Laptops are Broken by Design (ZephKeks) 8 views

He lists the 2022 Strix Scar 15 and 2025 Strix Scar 16 as confirmed affected, and suspects other ROG models could share the same firmware codebase. He also says he reported it to ASUS engineering 24 days ago with full register dumps, but the mismatch still shows up in the latest firmware.

This is also not his first ASUS deep dive. About three months ago he published a long technical investigation into an ASUS gaming laptop ACPI firmware bug on GitHub, and now he is back with another one. Twenty four days is not a realistic turnaround for this sort of BIOS work, especially at the end of the year, but it is still wild that one person keeps turning up issues at this pace.

ASUS, just hire him already?

Source: Reddit , Google Drive (materials)