Bug 95141

Summary:	strange errors with Amilo Xi 2550 after a possible evil maid attack
Product:	Platform Specific/Hardware	Reporter:	Elmar Stellnberger (estellnb)
Component:	x86-64	Assignee:	platform_x86_64 (platform_x86_64)
Status:	RESOLVED OBSOLETE
Severity:	high
Priority:	P1
Hardware:	x86-64
OS:	Linux
Kernel Version:	3.2.0-4-amd64	Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	ScreenShot_1 ScreenShot_2 ScreenShot_3 ScreenShot_4 /var/log/messages after reboot

Description Elmar Stellnberger 2015-03-20 10:16:53 UTC

bug description:
After a holiday I came back to my Amilo Xi 2550 machine, resumed it from s2disk as I always do and tried to continue my work. At first I noted sudden system freezes where the mouse pointer did not move. A few minutes later the whole system freezed. It was neither possible to move the mouse pointer nor to alight or unlight the keyboard leds by pressing num lock. The only way out was a Alt-Print-S/U/B. Unfortunately I do not have a backtrace of the freeze (should have pressed the right SysRQ keys).
After reboot the screen was swamped with a couple of error messages like 'port 2 disabled by hub (EMI?)'; see for the attachements. Up to now the machine has succeeded to reboot. I have never experienced similar errors on that machine though the Debian installation is a couple of month old.

Dear developers of kernel.org. The reason why I suspect that this error may even be due to manipulated hardware is because of two reasons:
* before the last summer the machine was running stable but yielded computational errors which were not reproducible on other machines with the exactly same setup (same version of Debian, same version of MySql: http://bugs.mysql.com/bug.php?id=73231).
* After exchanging the mainboard all errors were gone (I had to reinstall the bootloader because the machine did not boot after the exchange by an exactly similar mainboard; only chip numbers and usb connectors were different in their width). Additionally the guy with whom I have exchanged the mainboard pointed me out that the mainboard was already scratched and had possibly already been mounted and unmounted several times though the notebook was bought new and the mainboard never exchanged before.

Comment 1 Elmar Stellnberger 2015-03-20 10:17:46 UTC

Created attachment 171411 [details]
ScreenShot_1

Comment 2 Elmar Stellnberger 2015-03-20 10:18:23 UTC

Created attachment 171421 [details]
ScreenShot_2

Comment 3 Elmar Stellnberger 2015-03-20 10:19:12 UTC

Created attachment 171431 [details]
ScreenShot_3

Comment 4 Elmar Stellnberger 2015-03-20 10:19:55 UTC

Created attachment 171441 [details]
ScreenShot_4

Comment 5 Elmar Stellnberger 2015-03-20 10:23:35 UTC

Created attachment 171451 [details]
/var/log/messages after reboot

Comment 6 Elmar Stellnberger 2015-03-20 10:26:53 UTC

I am not sure on what you can do though any hint on how to analyse the described kernel issues are very welcome! This is my primary working machine (always kept offline) and my whole work depends on it.

Comment 7 Elmar Stellnberger 2015-03-20 15:47:14 UTC

Unfortunately no other SysRq keys than Alt-PrnScr-s/u/b work here. I have just rebooted and it freezed again not even responding upon the SysRq keys Alt-PrnScr-l/c/m. In its current state the machine can no more be used for work. It was working perfectly before I left it alone while nobody should have touched it (again this points to an 'evil maid' attack). Please help me as good as you can. I have been testing for Linux since years now and have written some well used little scripts for Linux like xchroot, confinedrv, debcheckroot, etc.

Comment 8 Elmar Stellnberger 2015-03-31 09:54:10 UTC

  In order to exclude the possibility of manipulated software (the machine is always kept offline) I have installed Linux new from scratch. The mouse pointer and keyboard input does not freeze if I avoid those USB ports which are directly on-board; there is a third USB2 port powered by an extension card; this is the only working one.
  It would be very nice if anyone could have a look at it!

Comment 9 Elmar Stellnberger 2015-04-13 05:33:42 UTC

In addition to the random freezes of mouse and keyboard (the suddenly don`t accept any input for a while when plugged into the two onboard USB connectors) the error from bug 12118 has aggravated: In spite of the processor.max_cstate=1 and nohz=off kernel parameters the quality of music playback is suddenly very bad.

Comment 10 Elmar Stellnberger 2015-04-13 05:41:42 UTC

500€ for anyone who analyses the problem and who will be able to tell me convincingly what is wrong with this machine! Note that also other users may be affected as is the case by bug 12118.

Comment 11 Elmar Stellnberger 2016-01-21 18:47:03 UTC

Will close this as worksform now.