Skip to content

Tool for partial deblobbing of Intel ME/TXE firmware images

License

Notifications You must be signed in to change notification settings

jwgarber/me_cleaner

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

89 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

me_cleaner Donation

me_cleaner is a Python script able to modify an Intel ME firmware image with the final purpose of reducing its ability to interact with the system.

Intel ME

Intel ME is a co-processor integrated in all post-2006 Intel boards, which is the base hardware for many Intel features like Intel AMT, Intel Boot Guard, Intel PAVP and many others. To provide such features, it requires full access to the system, including memory (through DMA) and network access (transparent to the user).

Unlike many other firmware components, the Intel ME firmware can't be neither disabled nor reimplemented, as it is tightly integrated in the boot process and it is signed.

This poses an issue both to the free firmware implementations like coreboot, which are forced to rely on a proprietary, obscure and always-on blob, and to the privacy-aware users, who are reasonably worried about such firmware, running on the lowest privilege ring on x86.

What can be done

Before Nehalem (ME version 6, 2008/2009) the ME firmware could be removed completely from the flash chip by setting a couple of bits inside the flash descriptor, effectively disabling it.

Starting from Nehalem the Intel ME firmware can't be removed anymore: without a valid firmware the PC shuts off forcefully after 30 minutes, probably as an attempt to enforce the Intel Anti-Theft policies.

However, while Intel ME can't be turned off completely, it is still possible to modify its firmware up to a point where Intel ME is active only during the boot process, effectively disabling it during the normal operation, which is what me_cleaner tries to accomplish.

Platform support

me_cleaner currently works on most of the Intel platforms; while this doesn't mean it works on all the boards (due to the different firmware implementations), it has been proven quite reliable on a great number of them.

Usage

me_cleaner should handle all the steps necessary to the modification of an Intel ME firmware with the command:

  $ python me_cleaner.py -S -O modified_image.bin original_dump.bin

However, obtaining the original firmware and flashing back the modified one is usually not trivial, as the Intel ME firmware region is often non-writable from the OS (and it's not a safe option anyways), requiring the use of an external SPI programmer.

Results

For generation 1 (before Nehalem, ME version <= 5) this tool removes the whole ME firmware and disables it completely.

For generation 2 (Nehalem-Broadwell, ME version between 6 and 10) this tool removes almost everything, leaving only the two fundamental modules needed for the correct boot, ROMP and BUP. The firmware size is reduced from 1.5 MB (non-AMT firmware) or 5 MB (AMT firmware) to ~90 kB.

For generation 3 (from Skylake onwards, ME version >= 11) the ME subsystem and the firmware structure have changed, requiring substantial changes in me_cleaner. The fundamental modules required for the correct boot are now four (rbe, kernel, syslib and bup) and the minimum firmware size is ~300 kB (from the 2 MB of the non-AMT firmware and the 7 MB of the AMT one).

On some boards the OEM firmware fails to boot without a valid Intel ME firmware; in the other cases the system should work with minor inconveniences (like longer boot times or warning messages) or without issues at all.

Obviously, the features provided by Intel ME won't be functional anymore after the modifications.

Documentation

The detailed documentation about the working of me_cleaner can be found on the page "How does it work?" page.

Various guides and tutorials are available on the Internet, however a good starting point is the "How to apply me_cleaner" guide.

About

Tool for partial deblobbing of Intel ME/TXE firmware images

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 100.0%