Hi! I am new in here, so first a kind 'hello' to everyone in here.

I wanted to linux-ify a given board that is very similar to a AT91SAM9RL-EK. Similar but not equal as it misses the LEDs and the dataflash. But I'd like to add network later.

So here is a short summary of the story how I proceeded so you know what way I took and whre you might jump in and help me a bit.

I got openocd more or less working with the board. I.e. I can control the CPU and up-/download something to SRAM. So I was able to download a small application.

When this app worked, I adapted the original Atmel bootloader to my board, as I have no dataflash and a so far unknown NAND Flash and I modified the SDRAM setup. Loading the modded bootloader and executing it works fine. A self written little memory test shows a working SDRAM.

As the NAND is still empty I let the memory test active in the bootloader to get a long boot delay. In that delay I could HALT the CPU and upload the u-boot version 1.3.4 from the linux4sam buildroot demo code. With a resume 0x23f00000 the u-boot started. Yea! I thought. With that u-boot I copied the bootloader and the u-boot itself into NAND and setup up some enviroment variables. Saved and did a reset...

Outch: After the bootloader starts u-boot the u-boot fails at some NAND things: U-Boot 1.3.4 (Sep 24 2009 - 16:56:53)

DRAM: 64 MB NAND: nand_scan(): Cannot allocate oob_buf 0 MiB Error Timeout

Now, If I reset the system, let the bootloader start, HALT the CPU and upload u-boot via JTAG, and again resume it, it again works fine!

So I crawled the u-boot code and saw that there are some NAND chips listed but my one is not in. So I flt I have to compile a version myself.

I had used buildroot with at91rm9200 two years ago with success, so I gave it a try. But the version from linux4sam is horribly outdated. I think I could spend hours by replacing links and recompile. It failed at the beginning of the toolchain already...

So I started with a fresh checked out git version of u-boot and tried mixed linux/windows work with yagarto. Unfortunately the initial configuration doesn't work under windows, so I switched to linux completely and started with a make at91sam9rlek_config, make.

I didn't get very far until the first compiler errors appear. Looks like at91sam9rl is very old for u-boot too.

It started with a message that I had to define CONFIG_AT91FAMILY for that chip. I added that one to the config.mk as it was with other at91sam9xxx cpus.

Later I got 'lost in the Matrix' :) as the legacy driver or lowlevel_init.S likes to initialize the matrix controller of the CPU abut the defines for SAM9RL are missing. I could find only two of the needed lines, but there is one with an offset that I didn't understand.

Then I found out that in some older code there was no line telling about legacy, so I disabled this define: //#define CONFIG_AT91_LEGACY

Now it got very far but then compiling aborted with lots of missing hardware defines like AT91_PIN_PBx, so the GPIOs where broken without the legacy info. And the USARTs/DBGU where too. It claims about dataflash related things, so I disabled the dataflash, I thougt... But that doesn't help, I cannot find any #define or #undef that declares dataflash to not exist on that board.

Now I'm a bit lost. Does someone have a working version of u-boot for at91sam9rl, working in the meaning of compiling through? Or can someone give me some hints of how to reactivate the at91sam9rl again.

Many thanks in advance and best regards Ulrich

----- Best regards Ulrich