[U-Boot] Kernel OOPS on memcpy - bad memory setup?

[Kristoffer Ericson]

On Thu, Oct 28, 2010 at 09:15:55PM +0200, Wolfgang Denk wrote:
> Dear Kristoffer Ericson,
> 
> In message <20101028190714.GD893 at boggieman.bredbandsbolaget.se> you wrote:
> >
> > > No. U-Boot is long dead and gone when Linux boots.
> > 
> > Is that true even if I were to setup the memory registers incorrect?
> 
> It is truye that U-Boot ceases to exist as soon as Linux starts
> executing.
> 
> > I thought that linux pretty much expected the bootloader
> > to setup everything regarding the memory settings?
> 
> Of course Linux relies on proper hardware initialization by the boot
> loader.
> 

alright.

> > Oh, and IF I configured the memory incorrect, I should be able
> > to see somekind of fault when doing a memtest in u-boot, correct?
> 
> No. Usually this is NOT the case.
> 
> > Or atleast that the write != read values?
> 
> No. Please read the FAQ:
> http://www.denx.de/wiki/DULG/UBootCrashAfterRelocation
> 

Very useful information. So in short (just to make it crystal clear):
Even if u-boot works fine and linux starts without
any issues and can run some binaries (busybox from what I see) okey. 
If I am seeing some kernel oops when running bigger binaries (e.g apt/ipkg) / doing heavier stuff 
(trying to ftp/lynx through pcmcia card), it could be cause by bad memory configuration and/or bad initialization?

> Best regards,
> 
> Wolfgang Denk
> 
> -- 
> DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
> HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
> Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
> "He was so narrow minded he could see through  a  keyhole  with  both
> eyes ..."