Hi,
Jonathan McDowell wrote:
I've actually compiled up a kernel more recently than that to check it still built ok, but haven't done anything much with the E3 recently.
Using OpenEmbedded, linux-omap_git.bb with slight amsdelta related additions actually works for me. Power nanagement must be off, otherwise omapfb framebuffer does not work correctly. Missing drivers: camera, keyboard, modem (should be trivial to port from 2.6.19 patch), sound, hook switch (in progress), probably more. I am going to see what I can do to narrow this list.
- Did you try to create a driver for hook switch hanging off GPIO
input 4, as Matt Callow detected it once? If not, what hints could you give me for creating one?
I had this hooked up to the OMAP gpio switch framework, but I found it conflicted with the FIQ keyboard patch and caused crashes. Possibly the gpio-keys driver is the way to do this these days?
As I have already written in a second thread, both gpio-switch and gpio-keys based devices actually work for me if without FIQ patch.
For now, gpio-switch device gives me the switch status via sysfs, but does not provide any events, so not much different than exporting the pin over gpiolib sysfs interface, but I think that adding kobject CHANGE event to gpio-switch interrupt handler should be trivial.
OTOH, gpio-keys based device generates input events obviously, but does not allow for status check (or I don't know how to check for it), that I would rather find an essential feature.
Cheers, Janusz