[mythtv-users] Using devinput but devinput lircd.conf file does not contain all my remote buttons on Harmony One

[Jarod Wilson]

On Jun 8, 2011, at 5:15 PM, Gabe Rubin wrote:
...
>>> I will also try changing the UNKNOWN to RC-5.
>>> When exactly do I need to execute the command?  With the lirc service
>>> off?  With the correct module loaded?  At any time?  And do I need to
>>> do it on each reboot?
>> 
>> In theory, the udev rule should trigger automatically after you load
>> up the ir-kbd-i2c module, but maybe its not enabled in F14... In which
>> case, you'd want to modprobe ir-kbd-i2c, then run ir-keytable, from
>> somewhere such as /etc/sysconfig/modules/lirc.modules. (There's no
>> auto-loading of ir-kbd-i2c, so it has to be loaded by hand before there
>> is anything for ir-keytable and/or lirc to talk to).
>> 
>> 
>>> Still not understanding why some keys that are mapped in that haupp
>>> file show up in irw when I have the right module loaded but others
>>> don't.
>> 
>> That file doesn't necessarily match what's actually in the kernel.
>> After loading ir-kbd-i2c, ir-keytable -r will dump the keytable that
>> is actually present in the kernel.
>> 
> 
> Thanks.  I will try this.  Either tonight or next week when I have
> free time.  Is there a way to automate this on boot up once I confirm
> everything works?

The previously suggested lirc.modules route. :) Just needs to be an
executable shell script, you can call modprobe and ir-keytable from in
it, and it'll get run by rc.sysinit on boot.

(The ir-keytable call should be redundant if there's a v4l-utils'
udev rule installed and active though, just not sure if that is the
case on fedora 14 or not).


> As an aside, why did support for lirc_i2c get dropped?  I do
> appreciate all the work that has gone into lirc, but it seems like a
> huge hassle when things were working before.  As a result, I am still
> using an older kernel until I get things working (but I am sure I will
> get things working soon and don't want this to sound like a
> complaint).

There's a policy against having multiple drivers for the same hardware,
and ir-kbd-i2c has been in the kernel for some number of years now. Its
just recently gotten a major facelift as part of the rc-core effort,
and *should* give 100% feature-parity with lirc_i2c in the latest
upstream kernels. Basically, lirc_i2c was redundant dead weight, with
no chance of ever going upstream.


-- 
Jarod Wilson
jarod at wilsonet.com