[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: NSLU2 Linux

On Fri, Aug 18, 2006 at 01:18:59AM +0200, Martin Michlmayr wrote:
> * Szafran <szafran_cien@poczta.onet.pl> [2006-08-18 00:01]:
> > I'm using Debian on NSLU2. Great job by the way. I was wondering if
> > i replace the memory chips and place in larger replacements will the
> > debian recognize and use the extended memory ??
> Not by default, no.  I assume that you can change the cmd line passed
> to the kernel by RedBoot if you have a serial console but I'm not
> entirely sure.  I'm CCing Rod Whitby who probably knows.
> Once we use move to using APEX as the 2nd stage boot loader, I assume
> it will automatically recognize your memory if you simply added another
> bank and pass the correct parameter to the kernel.  Marc, is this
> correct?

There is another way that I neglected to mention.

APEX will always initialize the SDRAM controller when it is executed
from flash.  So, if a version APEX were written into flash that was
capable of initializing 64MiB of SDRAM, one could arrange for RedBoot
to execute this version of APEX in flash.  Some trickery would be
necessary, though not overly complex.

There is a possible gotcha and because I have no fat-slug, I cannot
verify whether or not this is a problem.  I believe it is OK to setup
the SDRAM controller for 32MiB and use the memory even though there is
really 64MiB installed.  Users have reported that this is the case.
What I don't know is if this is true in all cases where the memory has
been upgraded.  I don't know if these fatslugs were four-chip 64MiB
slugs, or two-chip 64MiB slugs.  Similarly, a 128MiB slug may or may
not work in this scenario.


Reply to: