u-boot git doesn’t have the s3c2440 support somehow. I don’t know why but these patches were never upgraded. of course u-boot evolved considerably since, and a plain patch won’t work.
As for the kernel, there is support for s3c2440, but not this particular board. I’m asking here because since it’s cheap ($125 with a 3.5" screen) and seemingly plentiful someone must have ran into it…
Thats what I thought too. They have a 7" screen version as well, and a board that has a VGA output (1024x768).
All in all worth the effort of getting it to work, that ARM9 is no slouch !
I’ll try to get a more modern kernel working this week-end, hopefully it’s less problematic to integrate independant drivers to the existing CPU support. For a distro, I already have openwrt built…
I managed to get a recent u-boot working on the board (1.3.4 base); I took the openmoko branch that already has s2c2440 base support, and added the necessary bits to it.
So now I have a modern bootloader with nand, mmc, tftp and all that. I decided to ignore the NOR an boot u0boot from nand. I’ll keep the crap vivi bootloader on the nor for emergency
I also managed to upload code via openocd, but the nand support is no longer accessible.
Kernels are more problematic. I can build kernels but they won’t run… they hang after the decompression. I tried 3 toolchains (opensourcery, openwrt and sandbox2) and they all behave the same.
The binary 2.6.13 from the cd starts up fine… most annoying…
I now got a git-kernel (2.6.29-rc3) working fine on the board. Got pretty much everything working, but I need to find a nagging problem when I run it at “full” speed (400/100), the DM9000 ethernet fails somehow.
It’s possible it can’t run that fast anyway, and I might have to use more conservative bus speed.
I also got a openwrt board support working, so it’s a nice little distro. No graphical apps so far, I’ll debunk the rest of the drivers before bothering