[mythtv-users] lirc device load order changed - tidiest way to recover?

[Alec Leamas]

On 21/05/18 05:00, Stephen Worthington wrote:
> On Sun, 20 May 2018 16:14:37 -0400, you wrote:
> 
>> On Sun, May 20, 2018 at 3:39 PM, Andrew Codrington <
>> andrew.codrington at fidenti.ca> wrote:
>>
>>>
>>> On Sun, May 20, 2018 at 2:43 AM, Stephen Worthington <
>>> stephen_agent at jsw.gen.nz> wrote:
>>>
>>>> On Sat, 19 May 2018 13:03:06 -0400, you wrote:
>>>>
>>>> <snipped>
>>>>
>>>
>>>> So I could use
>>>>
>>>> REMOTE_DEVICE="/dev/input/by-id/usb-SMK_eHome_Infrared_Trans
>>>> ceiver_SM005h4M-event-if00"
>>>>
>>>> and avoid having a udev rule.  The by-id option is only available in
>>>> recent kernels, but has been available for at least a couple of years.
>>>>
>>>
>>> Thank you very much for a detailed response!
>>>
>>> My machine has by-id so I started with that.
>>>
>>> First pass doesn't look successful though. lirc didn't look happy about
>>> restarting.
>>> I'll keep googling....
>>>
>>
>> Since your response made it clear the only place I needed to edit was in
>> the hardware.conf file I changed the REMOTE_DEVICE to point at /dev/lirc1
>> and will just hope the load order stays stable...
>>
>> I'm back in business. Thank you!
>>
>> Cheers,
>> Andrew
> 
> I would not rely on the order of the IR receivers remaining stable,
> unless you can figure out why it changed in the first place. 

Indeed, neither  would I. The kernel explicitly does not guarantee the
load order - it's more or less random and can change for all sorts of
reasons.

> I am also surprised that the by-id link did not work. 

That cannot work as I understand it. The setup seems to use the raw
timing data in the /dev/lirc? devides, probably consumed by the lirc
(silly named) default driver.

OTOH, the /dev/input/* links outputs decoded events provided by the
kernel. Such a setup in a different beast using the lirc devinput driver.

> So you should try a udev rule.

Agreed. There are some hints at [1], focused on this case (as opposed
to the /dev/input devices)

Cheers!

--alec

[1] http://lirc.org/html/configuration-guide.html#appendix-9