Bad memory

Adam Jackson ajax at
Fri Jan 18 19:10:41 UTC 2013

On Fri, 2013-01-18 at 20:49 +0200, cornel panceac wrote:

> 2013/1/18 Justin M. Forbes <jmforbes at>

>         There are a lof of reasons that this isn't feasible on a
>         running system.
>         You just don't have access to all of the memory for a full
>         check.
>         Justin
> why not? is there something that can not be relocated in RAM?

The kernel image itself.  DMA buffers.  Memory made inaccessible by PCI
BARs being in the way.  Probably some other stuff I'm forgetting.  Even
if you fixed all that you'd still be Heisenberg-uncertain because you'd
no longer be the only thing exerting cache pressure.

> on the other hand, wasn't there supposed to be a way to exclude a
> damaged RAM area by passing some arguments to the kernel?

See linux/Documentation/bad_memory.txt

- ajax

More information about the test mailing list