kernel: [4020055.079557] [Hardware Error]: CPU:0 (15:2:0) MC4_STATUS[-|CE|MiscV|-|AddrV|-|Poison|CECC]

mfoley · 05-29-2019, 05:09 PM

I have a computer with MSI 970A-G46 Motherboard, AMI V2.6 BIOS, AMD FX-6300 3.5GHz processor and 4G memory. I'm running Linux Slackware64 14.2, kernel 4.4.172. I got the following messages:

Code:

Message from syslogd@hiram at Fri May 24 12:19:24 2019 ...
hiram kernel: [4020055.079380] [Hardware Error]: Corrected error, no action required.

Message from syslogd@hiram at Fri May 24 12:19:24 2019 ...
hiram kernel: [4020055.079557] [Hardware Error]: CPU:0 (15:2:0) MC4_STATUS[-|CE|MiscV|-|AddrV|-|Poison|CECC]: 0x9d614805001c011b

Message from syslogd@hiram at Fri May 24 12:19:24 2019 ...
hiram kernel: [4020055.079737] [Hardware Error]: MC4 Error Address: 0x00000000bd46aa48

Message from syslogd@hiram at Fri May 24 12:19:24 2019 ...

Message from syslogd@hiram at Fri May 24 12:19:24 2019 ...
hiramhiram  kernel: [4020055.079917] [Hardware Error]: cache level: L3/GEN, tx: GEN, mem-tx: RDkernel: [4020055.079826] [Hardware Error]: MC4 Error (node 0): L3 data cache ECC error.

Based on web searches, I tried to run 'mcelog --ascii' to try and figure out the problem, but that command just sat there doing nothing for nearly 24 hours before I killed it.

Does anyone have an idea on this? Is my CPU [going] bad? I'd like to be able to interpret the meaning of MC4, CE, MiscV, AddrV, Poison, CECC, etc.

RickDeckard · 05-30-2019, 12:58 AM

Machine check exceptions should indicate whether or not execution can reliably continue in some form. The fact that it's not, and is hanging for 24 hours, implies that something is going bad, but what that "something" is without being at your computer is up for debate IMHO. It could be your CPU, it could be your RAM.

Start with on-board system memory diagnostics if you have those, or even Memtest86+, then move to dmidecode and to ECC checks with edac-util. If you can't find a conclusive answer after that, my guess would be CPU.

I'd start looking on Amazon or Newegg for replacement parts.