Comment 9 for bug 454371

I've been screwing with this since lirc 0.8.4a became obseleted with > 2.2.28 kernel changes. I use a PVR-1600 which has been providing me with a fine new recreational activity that regularly replaces watching tv with fixing tv.

Don't get me wrong; tons and tons of fine work in all the ivtv,cx18 code, etc etc - It's been fun reading code/posts from Hans, Andy etc as my card went from entirely supported to functional in mainline kernel drivers.

I have fiddled with 0.8.6 release and cvs pulled the mercurial repo from v4l, remade any variety of modules and was generally stonewalled. I thought I was onto something with some very odd looking KERNEL_VERSION_IFDEFMADNESS in i2c_dev but it turns out I was just tired.

Helplessly googling instead of watching House I did come across a interesting read starting out with:

ir-kbd-i2c: Switch to the new-style device binding model 2009-04-04
 (http://patchwork.kernel.org/patch/16323/)
-fascinating 3 way match between kernel maintainers, various v4l people and in an uninvolved/offhand way lirc.
-the only summary I can come up with is that there is great confusion and discussion on what style of standard can be applied to something as diverse as an i2c device, the failings and merits of various pre-existing schemas.
-abolishing i2c-dev, or ir-kbd-i2c or perhaps all of them seem to be opinioned as options.

Re: [RFC] Anticipating lirc breakage 2009-04-07
 (http://osdir.com/ml/linux-media/2009-04/msg00363.html)
-a bit of the fallout discussed in this ominous looking post. I never really did find discussion around this on the sourceforge hosted lirc mailing list, but I have not the patience to wait for them..

Somewhere in the next 6 months a variety of other things happen; new kernels arrive and disappear with and without regressions between the deadly triumvirate of kernel/v4l/lirc. Turns out that i2c drivers are still all over the road, and again I stop watching TV.

Re: cx18 i2c changes for 2.6.31.x? 2009-10-20
 (http://<email address hidden>/msg10856.html)
-makes clear the necessity of a recent pull of v4l to get the cx18 cards working with ir-kbd-i2c

On a whim, I loaded ir-kbd-i2c today.

Nov 3 21:10:45 indulgent kernel: [ ] input: i2c IR (CX23418 Z8F0811 Hauppau as /devices/virtual/input/input7
Nov 3 21:10:45 indulgent kernel: [ ] ir-kbd-i2c: i2c IR (CX23418 Z8F0811 Hauppau detected at i2c-0/0-0071/ir0 [cx18 i2c driver #0-0]

Note that in one of the lengthy discussions about not breaking anything they decided to leave ir-kbd-i2c arranged such that nothing loads it.. but it loads.

Very interesting!

Quick like a fox, I recompiled lircd with the "Linux input layer (/dev/input/eventX)" selected.

sudo lircd -n --device=/dev/input/event7 --output=/var/run/lirc/lircd

lircd-0.8.6[]: lircd(devinput) ready, using /var/run/lirc/lircd
lircd-0.8.6[]: accepted new client on /var/run/lirc/lircd
lircd-0.8.6[]: initializing '/dev/input/event6'

Promising! With great zeal I dove into the corners of my computer room, searching for the long buried remote.

indi@indulgent:~$irw
0000000080010161 00 Go HVR-1100

Go HVR-.. well, I do need to set this up (again).

I am tempted to assume that at some point /dev/lirc0 will get created again by one of the drivers, but I don't care. The remote actually works - I can lounge and watch TV instead of staring at all i2c related diffs.

Ubuntu 9.04
2.6.31-02063105 x86_64 (2.6.31.5 stable from kernel.ubuntu.com/~kernel-ppa )
LIRC 0.8.6 (release)
clone from linuxtv.org/hg/v4l-dvb (screwed around with firedtv to get a compile)