[Problem] Unlocking Alienware Aurora R11 (i9-10900F) BIOS

Alienware Aurora R11 (i9-10900F) - XMP and CPU-OC entirely disabled in BIOS, ThrottleStop confirms CFG-Lock-class MSR lock; VBS/HVCI ruled out; extracted firmware + specific NVRAM variable identified, requesting help pinpointing the exact lock bit.

SYSTEM

Alienware Aurora R11 desktop, motherboard IPCML-SH (Z490, LGA1200)
CPU: Intel i9-10900F (non-K) - unchanged since factory purchase
BIOS: 1.0.24, released 2024-04-01/02 - confirmed AMI Aptio core via direct binary evidence (AFUWIN/$AFU flash-utility markers, AMIPFAT/BIOS Guard headers present in both the update package and the firmware image itself)
RAM: factory-original 2x16GB HyperX Fury DDR4 (Dell part N205T, “Extreme Memory Profile,” native XMP 3200MHz), later upgraded to 4x16GB (64GB) - XMP was confirmed working on this exact system when it was younger, on this same non-K CPU

PROBLEM

BIOS Setup’s XMP profile selector and Overclocking menu are both entirely greyed out and unselectable - not present-but-unstable, just absent. This is a regression: XMP worked previously on this exact hardware. No BIOS update changelog documents this removal.

WHAT WE RULED OUT

Disabling Windows VBS (Hyper-V, Virtual Machine Platform, Hypervisor Platform, HVCI/Memory Integrity) does NOT restore the BIOS XMP/OC menu. However, it did have a real, confirmed effect: it unlocked GPU overclocking access in Alienware Command Center, and it unlocked ThrottleStop’s control over the CPU’s Package Power Limit MSR (0x610, PL1/PL2/Turbo Time Limit), which had been forced read-only (“Disable Controls” checked, all fields padlocked) until VBS was off. This matches documented behavior: VBS/HVCI’s hypervisor filters MSR access from kernel-mode drivers via an internal allowlist (Microsoft Learn: “Virtualization Based Security System Resource Protections”), and the ThrottleStop developer has confirmed the same VBS-blocks-FIVR-access pattern elsewhere.

Critically, the CPU’s Turbo Ratio Limit MSR (0x194, the classic “CFG Lock” target) remained locked even with VBS fully off (confirmed via Win32_DeviceGuard showing VirtualizationBasedSecurityStatus=0). This proves the BIOS-level XMP/OC greyout is a separate, firmware-side lock, unrelated to VBS.

WHAT WE FOUND

Extracted the real firmware image without touching the live BIOS chip, via Windows’ own UEFI capsule rollback backup at C:\Windows\Firmware{77023273-e7d0-440e-8539-4d88381da792}\firmware.bin (16,121,856 bytes). This is the actual flashed image, not Dell’s wrapped installer EXE, and avoids the installer’s own compression layer.

Full UEFIExtract dump (1990 files) plus Ghidra analysis, using al3xtjames’ ghidra-firmware-utils extension for proper Firmware Volume / TE loading.

Located the CpuSetup NVRAM variable: GUID B08F97FF-E6E8-4193-A997-5E9E9B0ADB32, 738 bytes, at firmware offset 0xA82629. This is the packed struct holding the CPU Setup page’s checkbox/dropdown states, almost certainly including whatever bit gates Overclocking/XMP.

Located and traced a PEI-phase driver named PeiOverClock: GUID BD87C542-9CFF-4D4A-A890-02B6AF986F34. Its entry point calls InstallPpi with a PPI descriptor (Flags=0x80000010, TERMINATE_LIST|PPI, confirmed correct against the PI spec) whose GUID resolves to D80544AF-E8B5-4BCC-A80A-C72145A96A80, with a NULL interface pointer (a marker/flag-style PPI, not a callable interface).

Binary-searched all 664 extracted modules for that GUID: zero consumers found anywhere in the image. This PPI install appears orphaned in this particular retail build, plausibly leftover Intel reference-platform code Dell never wired a consumer to for this SKU, though the real consumer using a different signal (HOB handoff, etc.) can’t be ruled out.

CLOSEST KNOWN PRECEDENT

An Aurora R9 thread here on win-raid (“Is there Other way available to remove bios lock (dell Alienware)”) describes a user successfully flipping a similar BIOS Lock NVRAM variable to 0 via modGRUBShell/RU.efi setup_var, but hitting FPT error 368 from Intel’s flash write-protection (FPRR) underneath, blocking the change from actually persisting to SPI flash. That thread concluded an external SPI hardware programmer would be needed to fully defeat it. It does not appear to have been resolved.

WHAT WE’RE ASKING

  1. Does anyone recognize the exact CFG-Lock/Overclocking-Lock NVRAM variable name and byte offset for this specific board and BIOS build (IPCML-SH, 1.0.24)? We have CpuSetup’s exact GUID, offset, and size above as a starting point but haven’t pinpointed the specific bit within it.

  2. Is the FPRR/SPI write-protection issue from the R9 thread a hard blocker on this platform too, or is there a software-only path we’re missing?

  3. Does a modified/unlocked BIOS already exist for the Aurora R11, or is anyone willing to help build one?

  4. Any insight on the orphaned PeiOverClock/PPI path above - dead end, or are we missing where it’s actually consumed?

I’m a new member here so I can’t attach files to this post yet - happy to send the extracted firmware.bin and the full UEFIExtract dump directly to anyone who wants to dig in, just send a DM and I’ll get it over to you.

Happy to run anything and report back - this system isn’t in daily production use in a way that makes experimentation risky, and I’ve got Ghidra, UEFIExtract, and IFRExtractor-RS already set up against this exact image.


Edit by Fernando: Thread moved into the “BIOS Modding Problems” Category and title customized/shortened