On Nov 3, 2010, at 6:19 PM, Martin Bunyard wrote:

On Wed, 2010-11-03 at 01:20 -0400, Jarod Wilson wrote:
On Nov 2, 2010, at 7:51 PM, Martin Bunyard wrote:

On Thu, 2010-10-28 at 21:49 -0400, Jarod Wilson wrote:
On Thu, Oct 28, 2010 at 9:24 PM, Martin Bunyard
<martin.bunyard@googlemail.com> wrote:
...
As can be seen there is no /dev/lirc0

There's no evidence there that the nuvoton-cir driver found anything
to bind to, which would certainly explain why you have no /dev/lirc0.
Check through the output of lspnp and/or lshal for pnp ids of either
WEC0530 or NTN0530. If neither are present, that would explain why the
driver isn't binding. If that's indeed the case, you've either got a
different pnp id, that might be added to the driver and it'd magically
Just Work, or perhaps you do have the same pnp id, but its not
exposed, because the IR part is disabled in your bios.


Well I have found that the pnp id is different for this motherboard.

lshal shows:

udi = '/org/freedesktop/Hal/devices/pnp_PNP0510'
info.parent = '/org/freedesktop/Hal/devices/computer'  (string)
info.product = 'Generic IRDA-compatible device'  (string)
info.subsystem = 'pnp'  (string)
info.udi = '/org/freedesktop/Hal/devices/pnp_PNP0510'  (string)
linux.hotplug_type = 2  (0x2)  (int)
linux.subsystem = 'pnp'  (string)
linux.sysfs_path = '/sys/devices/pnp0/00:09'  (string)
pnp.description = 'Generic IRDA-compatible device'  (string)
pnp.id = 'PNP0510'  (string)

I added PNP0510 to pnp_device_id (in nuvoton-cir.c) in the same style as
the other pnp ids but am now confronted with another error.

That's not a Consumer IR device, its an IrDA device, which is something different.


dmesg shows:

[    6.921771] nuvoton-cir 00:09: IR PNP Port not valid!
[    6.922984] nuvoton-cir 00:09: disabled

I took this to mean that the relevant port address had already been
allocated to another device.  The BIOS is configured to allocate
IRQ3/02f8 for the IR device so I checked in dmesg which shows:

[    0.994658] serial8250: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
[    0.994829] serial8250: ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
[    0.995410] 00:08: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A

Which seems to demonstrate a Before and After setserial configuration
(as per my /etc/serial.conf).  
In other words ttyS1 was allocated to the relevant irq and port but was
then de-allocated as I intended.  I have fiddled around
with /etc/serial.conf even to the point of assigning dummy port
addresses in the hope that this would release 0x02f8 but to no avail.

The BIOS does not help by forcing the IR device to take on the same
irq/port as a typical COM port (COM1 to COM4).

Perhaps I am 'barking up the wrong tree', can you help?

Unfortunately, it sounds like you're out of luck and you don't actually have the CIR bits available on your board. Or does your bios perhaps give you the ability to specify CIR instead of IrDA?

Also, can you confirm exactly what model of Nuvoton chip you have on your board? Might be able to dig up some further information in the documentation...


So I have had a look at the board and there quite clearly is:
   Nuvoton
   NCT6775F
   937G9B2936B013-ZXFB

I'd be very interested to hear what you find out.

Definitely looks to be the same chip as in the asrock ion330 kit, and I see actual evidence its supposed to have functional cir support. Unless there's an option somewhere in the bios to select between irda and cir though, I'm out of ideas where to look.

-- 
Jarod Wilson
jarod@wilsonet.com