Re: [openwince-list] PXA255 with Am29LV800B?

[Andrew D. <am...@gm...>]

On 7/21/05, KaruppuSwamy.T <swa...@ka...> wrote:
> Yes, Mr.Marcel. OPERATION SUCCESS but PATIENT DEAD :) :)
> There is a LED in my target, which should glow when there is a flash
> access. While I am writing the image on flash, it is glowing. After
> writing, when i restart my target, it is not at all glowing. That means
> there is no flash access started?? I am in trouble, can anybody suggest
> me a way to proceed/debug further into it?

some random thoughts:

assuming the LED doesn't have any kind of one-shot circuit on it to stretch=
 the
CS/ pulses out, if it only toggles a few times, you won't see it with your =
eyes.

Put a o'scope on the CS/ line to the flash and make sure it goes at least o=
nce
or twice after a reset.  If it only toggles a few times, chances are the CP=
U is
getting back bad instruction data for some reason and crashing.  Hook a log=
ic
analyzer to the flash and see the first few addresses and data.

Make sure the image you loaded was in the right place in flash - I don't
know about the pxa255, but sometimes the address you give to the
flashmem command isn't the same as the virtual address the CPU is
using.  It depends on how the 'bus driver' part of jtag for the pxa255
changes jtag addresses into hardware addressing.

Are you sure the image you burned into the flash isn't crashing from some
s/w bug?  Write a short assy program to endlessly loop, load that, see what
happens with an o'scope. Try filling the entire flash with NOP and a single
jump to the start as the last instruction.

make sure your jtag cable or something else isn't holding the board in rese=
t.=20

get a jtag probe for the pxa255 with software to single step and see what t=
he=20
core thinks it's getting.

--=20
Hardware, n.:
        The parts of a computer system that can be kicked.