[Arm-netbook] Tutorial: kernel + android + livesuite image creation for A10 tablets

[Alejandro Martínez]

On Sun, Feb 12, 2012 at 3:43 PM, "Sztupák Sz. Zsolt" <mail at sztupy.hu> wrote:

> ADBD can work on it's own, it's only dependency is having a
> /system/bin/sh avialable. I sometimes start adbd even before the real
> init (simply create an init.sh with a shebang, run the commands inside
> it (mounting /dev, /sys and /proc, starting adbd and running the old
> init at the end)) just to get very early access to the device. And It's
> probably simpler to debug the device this way as you don't have to
> reconfigure the pins in script.bin, which can actually be another source
> of problem. Besides adb pull and push is probably faster than using
> serial console to get files on/off the device, and you can run multiple
> amounts of it (having one cat /proc/kmsg is very useful, especially if
> insmod restarts the device)

Good to know! Thanks for the tip, I'll try it, I'm now using ssh & scp
through WiFi to copy things around and access the terminal.

script.bin doesn't seem as the source for the framebuffer problem as
everything works fine with the same kernel and .bin on Android..


> /dev/block/nandi is currently the sdcard part (at least on the GB
> partitioniong scheme), but it doesn't get flashed with my tools, as it's
> . For that you would need to create a different partitioning scheme,
> where you have to add another "partition mirror". Never tried this, but
> as I would also try to get ubuntu running I will sometime :) For testing
> though I think it would be easier to put /usr on the sdcard and try to
> mount it. If you use adbd instead of a serial console, you will have
> access to the sdcard port, and it's much easier to copy files to the
> sdcard from your PC.

I didn't touch the sdcard partition because it was FAT32, although I
didn't think on formatting it from Android or the PC via USB to Ext4..
I might try that later



Alejandro