From: Wolfgang Denk <wd@de...> - 2002-11-09 15:45:29
In message <B9F29444.6F5A%tchen@...> you wrote:
> but when i burn the software in the FLASH and then execute from there,
> it stop working.... (in fact, it is stuck while trying to read the
> chip's manufacturer ID/..... maybe this is the reason the original
> INTEL FLASH code did not bother to read manufacturer ID???)
No. You CANNOT execute code from the same flash device you are trying
to identify / erase / program.
> so the big question is whether it is possible to do EXECUTE-IN-PLACE
> (XIP) straight from flash.... and still perform operation to FLASH itself?
Not with most flash types.
> because the INSTRUCTION fetch may get in the way of these consecutive
> READ/WRITE operation
This has nothing to do with instruction fetch or read/write, but only
with the internal state of the flash; it's state meechine will
provide normal data (like necessary to read instructions) only in
read array mode; as soon as you send any command you cannot read any
data (or instructions) any more.
> if not.. i dont know how do people implement JFFS2? how do people run
JFFS2 has nothing to do with it. Code is just read into RAM before
you execute it (you cannot execute compressed code anyway).
> boot in FLASH and burn code in the SAME FLASH (in different section...)
> is this not possible physically?
Not with most flashes. At least not without running the programming
code from RAM.
Software Engineering: Embedded and Realtime Systems, Embedded Linux
Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd@...
See us @ electronica 2002 in Munich, Nov 12-15, Hall A3, Booth A3.325