angstrom+omap3503 EVM+beagleboard

I originally posted this to the OE list without much luck. I would like
to get the EVM up and running using the OE/angstrom/beagleboard. I then
could help out the project in whatever area is needed patches, porting,
testing. My company will be using the omap3 for several products for a
client and is willing to help out the community (as we are immensely
benefited by the community) in order to get the omap3 well supported for
embedded development.

Below is original post.

Ben Anderson

However when I try to run the kernel (via tftp, or from flash) it does
the following.

======================================================================
OMAP3EVM # run bootcmd
Reseting CHIP... Done
LAN9x18 (0x01150002) detected.
start Auto negotiation... (take ~2sec)
Auto negotiation complete, 100BaseTX, full duplex
Hit any key to stop autoboot: 0
OMAP3EVM # setenv get_kernel run tftp_kernel
OMAP3EVM # run bootcmd
Reseting CHIP... Done
LAN9x18 (0x01150002) detected.
start Auto negotiation... (take ~2sec)
Auto negotiation complete, 100BaseTX, full duplex
TFTP from server 192.168.20.176; our IP address is 192.168.20.201
Filename 'uImage'.
Load address: 0x80000000
Loading:
#################################################################
...
#################################################################
done
Bytes transferred = 2155668 (20e494 hex)
## Booting image at 80000000 ...
   Image Name: Angstrom/2.6.25+2.6.26-rc4+git24
   Image Type: ARM Linux Kernel Image (uncompressed)
   Data Size: 2155604 Bytes = 2.1 MB
   Load Address: 80008000
   Entry Point: 80008000
   Verifying Checksum ... OK
OK

Starting kernel ...

Then it appears to lock up.

If it locks up or at least appears to lock up at this phase then the two most likely culprits are:

1). Your "console" argument in the bootargs is missing or incorrect.

2). There is a mis-match between the machine number that u-boot has and the kernel has. U-boot passes a machine number to the kernel on start-up and the machine number the kernel has must match what is passed by u-boot.

If you enable the "kernel low level debugging functions" under "kernel hacking" when building the kernel this would tell you if there was a mis-match in the machine number when you attempt to boot the kernel. This is probably most likely as the beagle board will have a different machine number than the omap3evm.

Cheers
Jon

Beagle and the EVM are different enough that you really don't want to
try to use a Beagle kernel on the EVM.
http://www.sakoman.net/omap3/evm

I really need to feed my machine file & kernel recipe mods to koen so
he can put them in oe.dev. Then you can build them for yourself!

Steve

That would be great if there was a machine recipe for the evm!

I am also hoping to get a beagleboard as well. They both will work
equally as well for what I need them for.

Is there any way I can help with getting the EVM working with OE?

Testing/Documentation/etc?

I tested the kernel uImage from http://www.sakoman.net/omap3/evm and it
did the same thing the beagleboard image did (see previous post). Is it
possible that I have a machine ID mismatch between my u-boot and the
uImage you provided?

Ben

You will typically get an error message if the machine id is wrong.

I'm not sure what is wrong with your system. I'm running with the
stock xload and u-boot that ship with the evm. I've overwritten the
linux and rootfs partitions with the images on my site and haven't had
any issues.

The EVM actually works quite well with OE images. There's some
tweaking that needs to be done with X11 stuff (screen rotation for
example), but overall it isn't too bad.

Steve

I must be doing something wrong because I still can't get the
uImage/rotofs for evm to work.

Here is my flash script (modified the scripts from
http://www.sakoman.net/omap3/flash%20procedure.txt )