[mythtv-users] mythbuntu 11.04 lirc mceusb module - how?

[Jarod Wilson]

On Jun 14, 2011, at 3:59 AM, Tim Draper wrote:

> I've had a reply from TBS support:
> 
> hi Tim,
> 
> thank you for all the details. so, what you're experiencing is something related to part of the information here (about 'ir_core' and 'rc_core'):
> 
> http://www.buydvb.net/forum/viewtopic.php?f=22&t=110&start=90#p2251
> 
> that i've just commented in another e-mail:
> 
> "what Jarod Wilson told you is correct in general, but there is one big "but" - we deliberately keep 'ir_core' instead of the new 'rc_core' (as far as i know Jarod Wilson  is one of the developers of 'rc_core'), because on our tests with 'rc_core' even on the latest kernels there are constant crashes of the whole system when the new 'rc_core' is in use and that's why we kept the old 'ir_core' - for better user-experience. so, when 'rc_core' reached the necessary stability then we can re-consider to use the new 'rc_core' instead the old 'ir_core'. also, the message you see about 'ir_core' is just a warning and doesn't indicate any error as it might sound from the Jarod's comment."
> 
> so, now the whole dilemma is that if we use the new remote control infrastructure, which will result in working 'mceusb' for you then using the 'cx23885' remote control results in unstablility of the whole system and that's what most of our Linux customers use. it seems the only way is 'rc_core' stability with other chips like 'cx23885' to be improved first. i understand that from different points of view it looks different, but from our point of view 'rc_core' broke the stability of 'cx23885' remote control support and that applies not only for TBS hardware, but for all hardware out there that relies on 'cx23885' driver code and thus it's general regression.

Fantastic. So they've apparently encountered a bug, but not bothered to
actually report it, since I've never seen what they're talking about.
How exactly do they expect it to get fixed if they don't report it? I'm
not even sure exactly what the nature of the problem they claim exists
*is*, so I can't even begin to try to fix it... Maybe they've contacted
the cx2388x driver via back channels? Dunno.


> doesnt alot of the hauppauge cards use the cx23885 module?

There are some, but I'm not sure how many folks are actually using IR
on them. I have a pair of cx2388x Hauppauge cards myself, but neither
actually has onboard IR.


> i only have my TBS card, so cant test with something else that uses cx23885. i guess that would prove where the problem lays; TBS drivers specifically, or with cx23885 as a whole...
> i've dropped back to mythbuntu 10.10 which is working fine, and im on the lastest 0.24.1, so i'm currently quite happy were I am. I guess something will need to be sorted by someone (who?) before myth 0.25 is released, but that seems to be a while off yet :)

They could get the ball rolling by actually sending a report upstream
(to linux-media at vger.kernel.org) that describes the problem they're
seeing, and maybe once the right people actually know there's a problem,
then might have a chance of fixing it.

This is one of many reasons why keeping stuff out-of-tree blows.

-- 
Jarod Wilson
jarod at wilsonet.com