Wednesday, February 24, 2010


I have been working on u-boot since last week. I need to port u-boot to our bcmring platform.

We have some example from another team of Broadcom to support u-boot on the 2153 platform, which has a similar ARM core with our bcmring platform.

I got the u-boot to build last week. Since this week, I learned to use JTAG to load the u-boot elf file into the memory of the platform after ddr initialization. I ported the UART driver and then the timer driver. Yesterday, I was able to boot to u-boot shell with JTAG. That's a good milestone.

Today, I studied our 1st boot loader (boot1) and the relation between the boot1 and boot2, the second stage boot loader. As I was trying to replace boot2 with u-boot, after I figured out the memory map of boot2, with the help of Darwin, I managed to load u-boot from boot1 and boot to u-boot shell without using JTAG. That's a big success. I can then work on the nand driver to get u-boot to boot our kernel.

The ethernet driver, sd/mmc support will be followed after the nand driver.