The bus error thing generally happens in my experience when RAM is not
initialized yet -- that is, the PXA doesn't think there's any RAM to
talk to, and you then try to load your SREC (or write a word) to some
address that's not mapped by the CPU, leading to the bus error
exception. Try editing the macro script to write a word to some RAM
address like 0xA2000000 or something instead of doing the SREC download
at the end of the script -- if the write word to a RAM address works,
then it might be some odd problem with the SREC (though your procedure
for generating the SREC looks OK). If it's not working, then
something's going wrong with the RAM initialization code when we poke
the controller earlier in the macro script.
On Aug 6, 2004, at 9:16 PM, Holly Gates wrote:
> Hello, I am trying to reload u-boot since I just hosed a board by
> trying to load the u-boot which got made from the latest buildroot
> (which apparently doesn't work, or I messed up loading somehow).
> I am using a genuine wiggler with the OCD Commander software on a
> windows PC, with the wiggler hooked up to the waysmall board.
> OCD Commander starts, and I can apparently manually set the PC, read
> the registers, etc. But when I run the macro, it says that the
> download of u-boot.srec failed, with the following on a pop-up
> "Write large, target bus error (40:11)"
> I am using a u-boot.srec which I made by getting the
> u-boot.bin.r48-400MHz file off of sourceforge and running:
> arm-linux-objcopy -I binary -O srec u-boot.bin.r48-400MHz
> Is this the right way to make the srec? Any other ideas on what could
> be wrong?
> This SF.Net email is sponsored by OSTG. Have you noticed the changes on
> Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
> one more big change to announce. We are now OSTG- Open Source
> Group. Come see the changes on the new OSTG site. http://www.ostg.com
> gumstix-users mailing list