[mythtv-users] Long delays with remote occasionally

[Josh Mastronarde]

On Tue, Mar 31, 2009 at 4:02 PM, Michael T. Dean
<mtdean at thirdcontact.com> wrote:
> On 03/31/2009 04:49 PM, Larry on the Dell Laptop wrote:
>>
>> Michael T. Dean wrote:
>>>
>>> On 03/31/2009 08:12 AM, Larry on the Dell Laptop wrote:
>>>>
>>>> True to Josh's prediction,  last night I noticed an occasional  "IR
>>>> Remote button press" delay even though I was running with the playback
>>>> profile set to "High Quality" on version 18722,
>>>
>>> Again, you do realize that there's an issue in Myth that affects -fixes
>>> users with versions prior to r19222, right?  So, it's likely that now that
>>> you've downgraded Myth to pre-r19222, the occasional delay was the
>>> screensaver one, not the other issue.
>>>
>>> Unless, of course, you did as I said and got rid of
>>> gnome-screensaver-command, that is.
>>
>> Thank you again Mike.   I 'll  kill the screen saver again and see if my
>> 18722 shows the remote button press delay.  Maybe the delay is prompted from
>> two separate causes, this one and the one Josh has identified.
>
> Not kill the screensaver, but prevent gnome-screensaver-command from ever
> executing--i.e. get it off your system or mv it or something.

I'm pretty sure there are at least 3 distinct causes of the issue of
remote and/or keypress inputs getting stalled, dependent on system
configuration, types of encoders, etc.

1)  Pre-r19222 Myth combined with gnome-screensaver -- I don't run
gnome-screensaver at all, so didn't have this issue.

2)  Nvidia driver without 'Option "UseEvents" "True"' in the "Device"
section -- adding this solved the delays for me back when I had just
PVR150 (mpeg ps) recordings, until they reappeared when I added the
HDHomerun.

3)  Mixed recording types (in my case mpeg ps from a PVR150, mpeg ts
from a HDHomerun, and mp4 h.264) causing file detection/probing issues
-- in my case with separate symptoms (one type failing to play, then
playing the 2nd time it's selected, then another type not playing
after that, etc).  When I root caused that and fixed it, the lirc
delays went away too (which I'm thankful for, because as a fleeting
issue that was going to be much harder to properly trace).

I suspect these are all tied together in the way they cause the player
loop to get stalled somehow, but it's not just raw CPU utilization --
I had the issues on both an old Pentium 4 and a 3.15GHz dual-core
E8500.

Josh