Honored Contributor
Honored ContributorHi Max,
> But when I restart am I supposed to have a sort of u-boot config prompt ? Yes. Normally, u-boot is stored in flash at the reset address. At startup, u-boot copies itself to SDRAM (in this case 018e_0000), then jumps into SDRAM. From there, everything is the same. A few things to check if u-boot doesn't start on reset/power-on: 1. Make sure u-boot is stored at the actual reset address. 2. When saving u-boot to flash ... make sure you write the entire image. If you use loadb/loads to download the binary/srec, the actual size of the binary will be reported -- use the reported size. 3. To check the binary image you downloaded to SRAM, you can just use the 'go' command. Based on your last post, you would execute: ==> go 800000 Or to confirm that the image you stored in flash behaves properly: ==> go 0 If your new image doesn't run ... you have other problems ;-) > Another thing : "bootm X" command returns > ... The magic number is invalid in the image header at address 0x0.: The bootm command is used to boot a u-boot image file generated with the mkimage utility (it's not the same as the 'go' command). The bootm command does the following: 0. Checks the header magic number. 1. Performs a 32-bit CRC check of the image header. 2. Performs a 32-bit CRC check of the image data. 3. Executes other consistency checks (e.g. image type, architecture, etc). 4. Optionally decompresses the data. 5. Relocates the data to the load address. 6. Disables interrupts. 7. Finally, jumps to entry point address. So, it's normally used to boot a gzipped linux kernel image. > ... > docs about U-BOOT implementation > ... This has information for a PowerPC-based board, but it's loaded with target-independent info as well: http://www.denx.de/twiki/bin/view/dulg/uboot (http://www.denx.de/twiki/bin/view/dulg/uboot) Regards, --Scott
