I currenlty need to load things into SDRAM from a paged flash that is
larger than the 1Meg window available. To survive this, I currently
put a copy of linuxbios at the top of every page.
This is annoying because it wastes flash space and I have to break up
large blocks of data into 960K chunks, putting one chunk on each page
with the copy of linuxbios.
It seems it would be better to be executing code from SDRAM when this
accessing the other pages of the flash.
I see two options:
- Copy the functions the are needed into SDRAM
- Copy all of Linuxbios into SDRAM and run all of it from there.
Questions:
- What will bite me if I try the second option? The only absolute jump
that I know about is when switching to protected mode.
- How relocatable is the code?
- What about references to static contants?
- Do I need to create a special linker segment for this?
- Should I compile linuxbios to be located someplace else from the start
and copy it there before doing much of anything?
I'm sure I can figure this out, but I suspect that someone on this list
already knows all the answers. ;-) Doesn't the DoC stuff already deal
with this sort of problem?
Pointers and suggestions appreciated!
Thanks!
Ty
--
Tyson D Sawyer iRobot Corporation
Senior Systems Engineer Real World Interface Div.
[EMAIL PROTECTED] Robots for the Real World
603-532-6900 ext 206 http://www.irobot.com
