hardware issue, nothing to do for us

Hi Mika,

I run Memtest86+ all night and it did not detect any memory failures.

If I rebuild the harddrive with the large_8BG_image.img in an other PC with a
different chipset than sha256sum from grml64 gives the same checksum for each
run. So I can not reproduce the bug on an other 64bit machine.

I suspect that there is a bug in the chipsetdriver of the 64bit kernel for AMD
890FX +
SB850.

regards,
ZombB

Hi,

this sounds like broken hardware. Please run memcheck on your box to verify the RAM is really 
working as expected.

Marking the bug as resolved for now, please reopen the bug if you can reproduce the bug on a 
second system as well or if the hardware checks claim your system isn't broken.

regards,
Mika

Steps to reproduce the problem:
My computer has a Gigabyte_GA-890FXA-UD7_(rev. 2.1) mainboard with AMD 890FX +
SB850 chipset. I run grml96-full_2012.05.iso or grml96-full_2013.01-rc1.iso and
choose grml64 from the boot menu. I boot into the console and mount a partition
on a sata harddrive (mount /media/sda2). On the partition I create a image of
another partition with dd which is 8GB. When I call sha256sum on that 8GB file I
get a different checksum for each run (sha256sum -b ./large_8GB_image.img). This
also happens with sha1sum and md5sum. Thus generating the hashsum of the SSD
which is the source of the image gives the same checksum fpr every run
(sha256sum -b /dev/sde).

This only affects grml64. Choosing grml32 from the bootmenu instead will show
the expected behavior: sha256sum -b ./large_8GB_image.img gives the same hashsum
for every run and it is identically to the hashsum of the source drive.