Row Hammer bug allows kernel privileges

[Rorin]

Since I haven't found much info in hardware forums, I want to share with you the recent findings of Google Project Zero

http://googleprojectzero.blogspot.de/2015/03/exploiting-dram-rowhammer-bug-to-gain.html

You can test if your RAM is affected by using memtest 6.0

http://www.memtest86.com/download.htm

You have to use Test 13 of memtest86.

It is a hardware defect.

I found out that you can lower the chance of being affected just by decreasing

tREF

value in system BIOS.

E.G. my value was 7280.

I lowered the value to 3640 and the error rate decreased significantly.

 

[new boy]

And the potential implications of this are?

 

[Rorin]

And the potential implications of this are?

The referenced documents explain it much better than I would ever be able to.

Additionally, as far as I am concerned, I do not like the possibility of unintended bit errors in my memory, one way or another.

That is the whole point of testing memory for stability, isn't it?

 

[new boy]

The referenced documents explain it much better than I would ever be able to.

Additionally, as far as I am concerned, I do not like the possibility of unintended bit errors in my memory, one way or another.

That is the whole point of testing memory for stability, isn't it?

Maybe, but there are lots of words. I'll just continue on in ignorance I think.

I've been trying to stabilise my memory OC for a few days now. Keeps crashing out after 3 hours of prime blend with 6gb ram used. I've probably flipped way more bits on my windows install doing that I reckon.