Re: [Kernel-discuss] h4350 keyboard patch

On Thu, Mar 23, 2006 at 08:03:09PM +0200, Paul Sokolovsky wrote:
> Hello Shawn,
>
> Thanks for your patch! It's not committed as of yet, so I'd like
> to discuss it a bit. First of all, h4100_buttons.c now is copy of
> h1900_buttons.c modulo symbol prefixes, so the are good candidates for
> merging into common module. Patching h4100_buttons.c with 4350 code
> would mean it just will need to be extracted then. What do you think
> about converting this code into separate module, like h4300_kbd.c ? If
> you like I can take the chore, though of course I'll need you help
> with testing and debugging (I'm on 4150).
>

yeh, that would probably be better with a seperate module, so then the
h4100 and h1900 button code could be shared.. I would just be worried if
a h4100 user loaded the h4300_kbd module, It would make their pda
unusable till they rebooted.. (see my comments below)

I have a version of this patch that makes the h4300 keyboard a seperate
module, I'll clean it up a little, and email it here it for review.

> And another question which for sure needs to be answered if go
> separate module route, what can go wrong if we just ignore 4100 vs
> 4300 issue, i.e. just have both buttons and keyboard handlers running
> unconditionally. As far as I see, IRQ lines used are different and are
> not reused in two models, so it should be ok. Do you have idea why it
> might not?
>

I got a hold of a h4150 to test the h4300 keyboard code on it,
unfortunatly the h4150 doesnt like it at all. When I setup the keyboard
interrupt on the h4150, the asic3 peripheral interrupt status register
gets stuck, which causes the asic3_irq_demux function to keep looping
while spitting out "interrupt processing overrun"...

> Thanks,
> Paul

cya,
Shawn Anderson
Received on Thu Mar 23 2006 - 21:47:39 EST