I have Shuttle Omninas kd20, and I managed to screw its u-boot image in NAND. So it does not boot into u-boot from NAND anymore. The board does have serial header (which I am using), but no JTAG header. And obviously it has 2 SATA interfaces (it's a NAS box).
I am trying to use this guide to boot from SATA disk, and it works. Almost. First, using different stage1 images result in reported frequecies that do not match those in the names of the images. I need to use stage1/stage1.wrapped800 to get reported 750 MHz that is "right" for the board.
After booting into the u-boot, I can do almost anything using tftp: chain-load another u-boot, or boot a kernel. So far so good. But NAND operation is unstable. Erase works, reading usually works, writing always hangs. So I cannot recreate the u-boot partition in NAND, and presumably ubifs won't work if I try.
NAND access (obviously) worked fine when I was booting from NAND, including when I chain-loaded an external u-boot over tftp.
All of the above makes me think that the problem is in stage1, that does not suit my board, and incorrectly initializes something in the hardware.
I (obviously) have u-boot built from the source that works when chain-loaded over tftp.
Now the question: how can I constuct the stage1 and u-boot images that are suitable for writing onto the SATA disk, and do I need to modify the contents of the first sector (the one constructed by perl "print"s in disk_create)? I did not find any resources explaining this.
Thanks!
I am trying to use this guide to boot from SATA disk, and it works. Almost. First, using different stage1 images result in reported frequecies that do not match those in the names of the images. I need to use stage1/stage1.wrapped800 to get reported 750 MHz that is "right" for the board.
After booting into the u-boot, I can do almost anything using tftp: chain-load another u-boot, or boot a kernel. So far so good. But NAND operation is unstable. Erase works, reading usually works, writing always hangs. So I cannot recreate the u-boot partition in NAND, and presumably ubifs won't work if I try.
NAND access (obviously) worked fine when I was booting from NAND, including when I chain-loaded an external u-boot over tftp.
All of the above makes me think that the problem is in stage1, that does not suit my board, and incorrectly initializes something in the hardware.
I (obviously) have u-boot built from the source that works when chain-loaded over tftp.
Now the question: how can I constuct the stage1 and u-boot images that are suitable for writing onto the SATA disk, and do I need to modify the contents of the first sector (the one constructed by perl "print"s in disk_create)? I did not find any resources explaining this.
Thanks!