Pergunta

I started studying emebedded device on a freescale mx5x board, and I am now exploring the boot-mods of the board. depending on the boot-mod choosed, the proper uboot image is needed. for the mmc boot-mod, the image is uboot-no-padding.bin, which is generated from uboot.bin. so , my questions : why is no-padding uboot needed,what is the deffence between the two images ?

Please help, thanks

Foi útil?

Solução

I am just guessing but normally uboot is put on "naked" flash, where every write happen in erase block size quanta. Thus it makes sense to pad the uboot.bin image to erase block size. MMC is flash based, but the flash is managed by firmware, so you don't know (or care) about erase block sizes, hence the padding just makes the image bigger for no reason.

Outras dicas

According to my knowledge, Flash storage devices like SD, NAND and NOR-Flash memories into which images are loaded do not accept images that are smaller than a specific size (128k as minimum size from what I know). If the generated u-boot image is smaller than that threshold, padding must be added in order to be able to load it.

I am using dd command to remove 1k of padding and I got at the end a u-boot.bin with size 128k which is appropriate for loading. I am using a microSD of 4GB. This is brought from my personal experience. Please, take a look at this link. It talks about building and flashing u-boot for colibri target. In the page 6, it talks about u-boot padding.

http://antmicro.com/files/release/u-boot-colibri/U-Boot_Colibri_Manual.pdf

I hope helpful.

Licenciado em: CC-BY-SA com atribuição
Não afiliado a StackOverflow
scroll top