me_cleaner status #3
Comments
|
Working on:
Working for more than a month now. Everything works perfectly, and the MEI device has disappeared from the PCI bus. |
I can confirm it works on the Lenovo Thinkpad X220, but coreboot then recognizes only one of my two 8GB RAM modules. I'm currently investigating. |
Hey there, I build and flashed an image of coreboot for my chromebook (C720p) running a Haswell 2955U. There is no MEI entry in the lspci list. Passed the 30 minutes mark, seems to work. Thanks |
|
Working on:
Everything works, the HECI (formerly MEI) device disapperars, a screen at boot notifies that the ME firmware is corrupted, but pressing F2 lets the boot continue. |
^^ Yes that's right, the experimental branch works on this board folks! No lzma modules! |
|
Working on
|
|
Working on
|
ASRock's bios packages are all in a proprietary format, but Windows based tools, specifically the UBU pack (http://www.win-raid.com/t154f16-Tool-Guide-News-quot-UEFI-BIOS-Updater-quot-UBU.html), allow them to be extracted and the firmware inside upgraded or downgraded. One may flash this modified file directly from the UEFI settings themselves, as it doesn't validate them. It is unclear whether or not ME is properly disabled, as the kernel module loads but is not really usable, and the tools to check ME status segfault. |
BIOS file name must be same as Instant Flash bios name, or else instant flash in bios does not detect it. In this case, Z97Ex62.70 /dev/mei0 does not exist, intelmetool reports it doesnt support my system (maybe it doesn't?), mei/mei_me modules still required by some ASRock Intel ME pci listing
But everything seems to be working properly so far. |
On my system that I've never flashed before (I haven't used me_cleaner yet) I get the same message when trying to read from /dev/mei0. It does not mean that the ME is disabled |
|
Working on
MEI is no more present in lspci output. dmesg says: The problem seems identical to the one reported by this user: This is the related bug on the ubuntu kernel bug tracker: |
|
Working on:
OEM BIOS: MEI device has disappeared from the PCI bus. **Bad news, you have a MEI not hidden on PCI, checking if visible ME Status : 0x304181 ME: FW Partition Table : OK ME: Extend Register not valid ME: has a broken implementation on your board with this BIOS |
|
Working on:
MEI device has disappeared from the PCI bus. |
|
Working on:
Sadly I failed to extract a valid OEM BIOS image this time. MEI device has disappeared from the PCI bus initially, but after programming back from the scheme below the MEI reappears and keeps present. ME is confirmed broken. |
|
Working on:
MEI device won't go away and confirmed that ME is broken, and integrated graphic card conpletely ceases to work, and goes away. |
|
Works on:
|
Everything works, |
|
Works on:
MEI device has disappeared from the PCI bus. However, the ethernet card needs a warm reboot to be functional. |
|
Working on
MEI is no more present in lspci output. However, the ethernet card needs a warm reboot to be functional. |
|
Working on:
|
|
Working on
MEI is no more present in lspci output. However, the ethernet card needs a warm reboot to be functional. |
|
Working on
|
|
Working on:
The MEI Controller device still appears in |
|
Working on:
After the KERNEL module of ME is removed, the integrated NIC works after a COLD reboot now. |
|
|
|
internally flashed with the instructions from 62
|
|
Working On: TPM no longer functions. |
successful, BIOS shows N/A as expected unlike my previous test, no 30 minute shutdown and network card is fully operational. but one small problem BCLK overclocking is not functional, whenever I set a cosmetic BCLK like 101.0 it treats it as if it is set to 100.0, clock monitoring tools from inside windows (HWINFO) confirm the BCLK is unchanged from 100.0 even when set to 101.0 |
|
Working on:
(The Output with |
|
(Not) working on:
One must disconnect all probes from the SPI flash before the machine can boot (stray capacitance is an issue).
|
|
Success on an ASRock Z270 Killer SLI/ac.
The Intel ME region was left unlocked, while the flash descriptor was not. So originally I dumped the ME region, and flashed it back with the removed contents. Then I went back, using the BIOS found online for the motherboard, and tried both of the flags and was able to flash both of them using the Instant Flash (internal) BIOS feature. Boot time is the same, so no noticable differences except for of course the Intel PTT not working. |
|
Works on Asus X502C notebook
The board is used as a standalone, extracted from the laptop chassis. keyboard and monitor are standard system peripherals (USB and VGA respectively), cannot test with the original LVDS screen and the integrated keyboard. interestingly, the OEM ASUS logo disappeared from the POST screen despite no relevant changes applied in the OEM BIOS. Linux loading remains normal, albeit the kernel sometimes stops for varying times (fraction of second to minutes) at "verifying modloop". Cannot determine whether me_cleaner related or not as the board was not extensively tested before the modification outside the scope of basic functionality. IME is no longer visible in lspci output, /dev/mei* no longer exists |
|
Works on ThinkStation E31 desktop
the board has at least 3 SPI flash chips. The one next to the proprietary power connector did not report type. This leaves the two chips next to the SATA connectors: the one closer to the CMOS battery (U5) is a 2M one with content I couldn't identify. The one relevant in the present case is the one further from the battery holder (U4), this is a 8M chip. the ME version and build number is N/A in the OEM BIOS setup utility after flashing. The particular device I had already had issues with some setup data corruption, this remained in the first boot but disappeared at the second, not sure if related, needs further investigation. The board has a ME_DISABLE jumper, it was left set for the entire time, did not investigate the possible effect (only realized after the flashing was done). It is worth mentioning originally the ME version was still visible even when this jumper was set. Boots normally, Ethernet controller available, lspci doesn't show IME and /dev/mei* not present after flashing (albeit I have no definite record of it being present before) |
|
Success on Asrock Rack E3C226D2I motherboard
IPMI works just fine, no abnormal boot messages. ubuntu 22.04.1 and FreeBSD 13.1 seem to run just fine. BIOS setup reports: Some oddities: If I try uppercase
|
|
Working for me using Internal flashing with unlocked OEM BIOS.
Bios shows ME version 0.0.0.0 Note: I haven't tried -s flag or using a stock oem bios. I had to Enable Me FW Image Reflash in bios as well. I am not 100% sure why or maybe i was invoking the FPTw.exe command incorrectly |
|
Works on my laptop, had to flash externally using CH341a programmer
Note: Output of |
|
Works on the following devices:
|
|
ECS H55H-CM Motherboard
running me_cleaner with "-s" does not seem to have any effect, at least
No side effects noticed so far (sorry for the bad quality)
|
|
Working on
After flashing modded image the ME version number disappears in the UEFI menu and CSME PCI device could no longer be found |
|
Hi, Thinkpad T440p w/ i7-4800MQ
thanks to: |
|
Hello,
30 min restriction gone. Works from 8 December. So far so good. Thanks!! :) |
I used the internal flash method following this guide: Everything seems to be working fine. |
Everything seems to be working. Stock firmware (pre flash):Post flash: |
|
Asus Z10PE-D8 WS (C612 chipset) |
|
MacbookAir7,2 (2017 A1466 Model) Broadwell Mobile Per issue #230, but not having seen knowledge on if this worked with a MattCard, I got one myself (it's the entry all the way at the bottom for my model). Using the method described (sans the installation disk, I used my Pomona 5250 and a spare Pi 4 since the chip on the MattCard is a common Winbond chip), I cleaned my ME region and installed the MattCard. Bootup takes substantially longer (roughly a minute in my testing), and I occasionally don't hear the chime, but I get past the 30 minute mark without issues. YMMV. |
|
Samsung Slate 7 (XE700T1A) Sandy Bridge Mobile Uses a common Winbond chip, makes it past the 30 minute mark. Kind of a lethargic tablet considering, but I had it handy, wanted to try it. |
|
Working:
|
|
|
|
I wanted to do this one for a while, had to wait and get much MUCH shorter cables as I had just enough voltage drop with my current setup. Passes the 30 minute mark. EDIT: ME version also reports N/A, so good enough indicator for me. |
Boots to Windows fine, result: |
|
and others
Please comment here if me_cleaner works on your device.
If this tool does not work on your PC (or it does not behave as expected), don't comment here but open an issue instead.
Specify:
-s/-Sflag in me_cleanerThanks
The text was updated successfully, but these errors were encountered: