[coreboot] KGPE-D16 and Samsung DDR3L memory sticks

[Daniel Kulesz]

Hi Paul,

> Am Samstag, den 11.02.2017, 16:14 +0100 schrieb Daniel Kulesz:
> > To answer my question myself: It works partially.
> > 
> > > 1.) Samsung M393B1K70DH0-YK0
> > > 
> > > Type: DDR3 DIMM 240-Pin, reg ECC ? Ranks/Banks: dual rank, x4 ? Modules: 1x 8GB ? JEDEC: PC3L-12800R ? Voltage: 1.35V
> > 
> > I got 8 of these working with one CPU package on the KGPE-D16 with
> > one of the latest Coreboot master versions:
> 
> (Please note, that coreboot is officially spelled all lowercase.)
> 
> How much did you pay for these modules?
> 

One of them costs around 12-25 Euros (used) or ~65 Euros (new).

> Just to be sure. You got 64 GB of RAM working. If you plug in more
> modules, it fails, right?
> 

No. Using 64 GB worked more or less fine, but using *less* than 64 GB caused issues - depending in which slots I had put them.

> > Version: 4.5-963-gf57a768
> 
> Please upload your logs to the board status repository. (This commit
> has not been uploaded by REACTS, so it wouldn?t overwrite anything.)
> 

I uploaded one of the tested configurations here:

https://review.coreboot.org/cgit/board-status.git/commit/asus/kgpe-d16/4.5-964-gd96669e9db/2017-02-11T13_58_51Z/cbfs.txt?id=8e7fe74b18ecd903623447ac150efe08e1ec03b2

I was not able to upload more configurations since the build_status output folder is named after the build, and I didn't want to rename too much manually.

> Without the logs, it?s hard to debug anything. Verbose logs are one of
> the biggest advantages of coreboot. So please upload them, or attach
> them.

I attached the serial logs of one of the failing configurations (all orange slots populated).

However, I was unable to reproduce the exact failure because meanwhile I populated the second CPU socket and didn't find a way to deactivate the CPU (after pulling the second power connector the system did not boot up at all).

I also noticed the following messages in dmesg:

[ 1561.833618] [Hardware Error]: Corrected error, no action required.
[ 1561.840026] [Hardware Error]: CPU:16 (15:1:2) MC4_STATUS[Over|CE|MiscV|-|AddrV|-|-|CECC]: 0xdc644000ea080a13
[ 1561.851687] [Hardware Error]: Error Addr: 0x0000000ffe8a4c70
[ 1561.860313] [Hardware Error]: MC4 Error (node 2): DRAM ECC error detected on the NB.
[ 1561.870995] [Hardware Error]: cache level: L3/GEN, mem/io: MEM, mem-tx: RD, part-proc: RES (no timeout)
[ 1873.121721] mce: [Hardware Error]: Machine check events logged
[ 1873.121777] [Hardware Error]: Corrected error, no action required.
[ 1873.128184] [Hardware Error]: CPU:16 (15:1:2) MC4_STATUS[-|CE|MiscV|-|AddrV|-|-|CECC]: 0x9c644000ea080a13
[ 1873.142264] [Hardware Error]: Error Addr: 0x0000000ff95c9230
[ 1873.150857] [Hardware Error]: MC4 Error (node 2): DRAM ECC error detected on the NB.
[ 1873.161573] [Hardware Error]: cache level: L3/GEN, mem/io: MEM, mem-tx: RD, part-proc: RES (no timeout)

I didn't notice any of these before the installation of the second CPU module. Don't know if this is related or not. But I'll run a memtest next to check if maybe one of the modules could be faulty.

Cheers, Daniel
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: boot2.txt
URL: <http://www.coreboot.org/pipermail/coreboot/attachments/20170212/5527d9cc/attachment-0001.txt>