[U-Boot] Need Help in bulding U-boot-2010-06 for OpenRD client

[Albert ARIBAUD]

Le 06/07/2010 17:47, kalyan karnati a écrit :
> Hi Albert,
> Thanks for your reply.
>
> I compiled U-boot with the patch that enables support for
> OpenRD_client and OpenRD_ultimate given by Tanmay.
> OpenRD Client is connected to PC/Host with FTDI programmed for Channel
> A as CPU FIFO.
>
> I'm using the binary for OpenOCD that I got from Marvell to tranfer
> uboot.bin into board's NAND.
> I copied the 'u-boot-2010.06.bin' in to
> ~\openocd\openocd.binaries.libftdi/bin as 'uboot.bin'.
> Then executed the command ' ./openocd –f target/board/openrd.cfg –c
> init -c openrd_reflash_uboot'
>
> Logs show:
> wrote file uboot.bin to NAND flash 0 up to offset 0x0004e800 in 50.190395s
> Info : JTAG tap: feroceon.cpu tap/device found: 0x20a023d3
> (Manufacturer: 0x1e9, Part: 0x0a02, Version: 0x2)
> Info : JTAG Tap/device matched
>
> But then the board doesn't boot and the console is all blank :(
> Please let me know if I'm doing something wrong..
>
> Thanks,
> Kalyan.

Hi Kalyan,

At the moment the openrd image will only run if loaded into RAM through 
openocd.

I'd discussed with Tanmay about running the openrd binary: I'd tried to 
load it into RAM from the resident u-boot via TFTP, and that fails too.

In your case it will also certainly fail because the board config file 
contains #define CONFIG_SKIP_LOWLEVEL_INIT, which as the name implies 
skips very early and critical initialization code -- and without this 
code u-boot won't start from Flash (this config option is used precisely 
when running u-boot from RAM).

Can you try removing the #define and let us know if that fixes things up?

I'm not sure it will, though, because it fails when run in RAM from 
u-boot with the #define, which means it probably does not do some 
critical init --one that openocd does when it runs u-voot from RAM, 
since in that case it works.

Amicalement,
-- 
Albert.