Deekay: you are correct, you can write a user-space program if you wanted, but keep in mind that it's not possible to have something that runs from initial boot all the way through to user-land in the kernel. You'd want to enable the watchdog early on (x-loader...), and prevent it being re-configured in u-boot and the kernel (the state gets nuked during initialisation of the kernel...).

One reason you probably just want to rebuild the kernel is that it keeps things tidy: the only way you can write to the twl manually with i2c is to use the I2C_SLAVE_FORCE macro - which basically lets you use a slave device even if a driver has reserved it (which it will have, because the kernel sets things up and communicates with the tps at boot...).  I've not had any nasty collisions when doing this in the past (when controlling the LED drivers on the TPS), but I would still approach it with caution.  

On Thu, Jan 31, 2013 at 6:15 PM, deekay <> wrote:
Akram: Many thanks for you reply. I was worried that I need to re-build the
thing. But I guess that the "driver" reads and writes to TWL via I2C. I
spotted the source code for the TWL4030_watchdog.c, so my guess was that it
would be possible to build a stand alone program for accessing the device.
Without modifying the kernel that is.

Well, well... I might need to do some backward engineering after all.

I will however, give you the uname -a information shortly.


View this message in context:
Sent from the Gumstix mailing list archive at

Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
gumstix-users mailing list