"Machine Check Exception: 4 Bank 0: 9000000020000003" when boot RHEL 7, with Intel CPU E3845, and BIOS option "EIST" enable.

Chaoyong chaoyong at huawei.com
Thu Mar 5 06:28:54 UTC 2015


My hardware is:
CPU: Intel(R) Atom(TM) CPU  E3845  @ 1.91GHz
and it supports EIST(Enhanced Intel SpeedStep Technology).
OS: RHEL 7(kernel-3.10)

When I set the BIOS options "CPU configuration >> EIST" to "enable", it appears a kernel panic when the system is starting up. Here is what I am seeing:
......
Fast TSC calibration failed.
[Hardware Error] CPU 3: Machine Check Exception: 4 Bank 0: 9000000020000003
......
Machine check: Invalid.
Panic - not syncing. Fatal machine check on current CPU.
Rebooting in 30 seconds.

By reading the Intel-programming-manual, the MCE means:
Machine Check Exception: 4 -- MCG_STATUS_MCIP
Bank 0: 9000000020000003   --
         0x9000 -- 1001 0000 0000 0000 -- IA32_MCi_Status MSR register
         0x0003 -- 0000 0000 0000 0011 -- External Error, The BINIT# from another processor caused this processor to enter machine check.
and seems it is an external bus error.

If I DISABLED the BIOS option "EIST", there will be no error or panic.
If I change to CentOS 6.5/6.6(kernel 2.6), there will be no error or panic even when "EIST" is ON.

So, it should be a bug in kernel source of RHLR 7/CentOS 7(kernel-3.10).
Ps: kernel source v3.10 from kernel.org is OK

Any help would be greatly appreciated.

Thanks.



More information about the redhat-list mailing list