Cliff Lawson wrote:
It took us about a YEAR(!) to get the bloody keyboard working reliably so you're on the early stage of your journey.
Great! :-) It seems wise to follow the FIQ route then, since you've probably already explored all other options. I will try to get the FIQ handler working with the 2.6 kernel. Hopefully having a working 2.4 version will hopefully bring the development time down to less that a year!
www.ourcottage.plus.com/ams_keyb_mod.c (from kernel/drivers/char)
I'm not sure what dependent .h you might need so holler when it won't build!
Thanks for that. Probably not going to try to compile it, as the matrix keyboard is already handled in the omap code, and I'm planing to write a driver that will fit in with the serio framework in the 2.6 kernel. But it's still useful to have. Whilst looking at the code, I noticed that the qwe interrupt handler writes to latch2 AMS_DELTA_HDW_OP_SCARD_2_MASK (I'm assuming AMS_DELTA_HDW_OP_SCARD_2_MASK is 0x400), setting it at the start and clearing it at the end. Is this for debug purposes only? I noticed a similar thing in the FIQ handler using GPIO6.
I also notice that the driver has some code for writing to the keyboard, but data never seems to be sent. Is my understanding correct? Does the keyboard accept data?
Matt