Re: [PATCH 5/9] IR: extend interfaces to support more device settings

[lirc@bartelmus.de (Christoph Bartelmus)]

Hi Maxim,

on 29 Jul 10 at 18:27, Maxim Levitsky wrote:
> On Thu, 2010-07-29 at 09:25 +0200, Christoph Bartelmus wrote:
>> Hi!
>>
>> Maxim Levitsky "maximlevitsky@gmail.com" wrote:
>>
>>> Also reuse LIRC_SET_MEASURE_CARRIER_MODE as LIRC_SET_LEARN_MODE
>>> (LIRC_SET_LEARN_MODE will start carrier reports if possible, and
>>> tune receiver to wide band mode)
>>
>> I don't like the rename of the ioctl. The ioctl should enable carrier
>> reports. Anything else is hardware specific. Learn mode gives a somewhat
>> wrong association to me. irrecord always has been using "learn mode"
>> without ever using this ioctl.

> Why?

If an ioctl enables/disables measuring of the carrier, then call it  
LIRC_SET_MEASURE_CARRIER_MODE and not LIRC_SET_LEARN_MODE.

Whether we need a LIRC_ENABLE_WIDE_BAND_RECEIVER ioctl is another  
question.

> Carrier measure (if supported by hardware I think should always be
> enabled, because it can help in-kernel decoders).

That does not work in the real-world scenario. All receivers with a high  
range demodulate the signal and you won't get the carrier.

[...]
> Another thing is reporting these results to lirc.
> By default lirc shouldn't get carrier reports, but as soon as irrecord
> starts, it can place device in special mode that allows it to capture
> input better, and optionally do carrier reports.

And that's what LIRC_SET_MEASURE_CARRIER_MODE is made for.

> Do you think carrier reports are needed by lircd?

No.

Christoph