I'm using the OMAP-L137. I've been working on a project for it for about 1.5 years. The boot process has not caused any problems until I got the new boards. I am trying to boot from SPI flash, and the Winbond parts work, but the Numonyx parts flash, but won't boot. These are the same Numonyx parts that worked on the previous rev of the board. I'm trying to figure out what is different. I'm pretty sure it is a HW change that is causing the problem. Here is what I have so far:
The DSP boots.
Armubl starts. I've captured the SPI traffic on a scope and I see the chip select bracketing a command to read followed by 7 reads. The first 3 seem to be the address width check, suggesting that the code things the address width is 24 bits. (This is correct for the part.) I do not see the chip select deassert/assert, but I see an addition 4 reads. I can't match that up to the code I see in armubl.
I would like to trace the activity from the width check to whereever the code goes next, because I think this may tell me something useful about the failure mode. It is unclear to me how to do this. (I'm willing to do the tracing in assembly, but I'm not sure how to connect to the ARM before it runs past the point.)
On the DSP side, I put a "while (pleaseWait)" at the beginning of the code and, after connecting, I change the value of pleaseWait from TRUE to FALSE. I would like to do something similar on the ARM side, but I have no symbol table info. Still, I'm guessing it can be done.
Remember, u-boot is not yet running, so the memory map is pretty much controlled by armubl and the gel file. Has anyone does this before?