[mythtv-users] Updated ceton patch for 0.24

[David Engel]

On Sat, Jan 07, 2012 at 09:12:08PM -0500, Ronald Frazier wrote:
> A while ago I rewrote my ceton patch to work in trunk. In doing so, I
> was able to make a number of design changes to address bugs, increase
> performance/reliability, and add new features. I have now backported
> all of those changes to 0.24. I've been running this patch on my
> production system all week, and it's been working great. Here are some
> of the key changes that were made
> 
> 1) Fixed some bugs that could cause the backend to hang if you tuned
> an invalid channel (or if a channel was not broadcasting any data). If
> you ever had to restart the backend because it said the tuner was
> busy, even though it wasn't, this should fix that
> 2) You no longer need to run the channel mapper to configure the
> channels if you have a cablecard installed. Simply letting
> mythtv-setup or mythfilldatabase create the channels should be all
> that is necessary.
> 3) In addition to the old method of reading from the /dev/ceton/*
> device entries, you can also now use RTP/RTSP to stream the data. This
> method can also be useful when the Ceton card is installed in another
> system (as long as that system has the networks bridged so that the
> Ceton is accessible externally).
> 4) Multiple Ceton cards can now be used, either using the direct
> device entry method, or using RTP/RTSP.
> 5) Channel changing performance has been improved. Under normal
> circumstances, this may not be noticable, or only slightly noticable.
> However, under heavy loads some users previously experienced slow
> channel changing. This should fix that.
> 
> 
> The new patch is available here:
> http://www.ronfrazier.net/mythtv/0.24/index.html#CetonSupport

Hi Ron,

I gave your ver008 patch a whirl today.  There might be a problem.  

I scheduled 4 blocks of 4 30-minute recordings.  IOW, I intended to
keep all 4 tuners busy for 2 hours.  I had two separate failures where
the recorder couldn't find the desired program in the mpeg stream and
got stuck in the rsTuning state.

Here are the relevant snippets from the backend log.  I can provide
more complete logs if I trimmed them too aggresively.

2012-01-09 11:30:06.522 Started recording: "Majesty Outdoors":"Christian in Georgia Part 2": channel 6590 on cardid 7, sourceid 6
2012-01-09 11:30:06.586 TVRec(8): Changing from RecordingOnly to None
2012-01-09 11:30:06.607 Finished recording Recon: channel 6124
2012-01-09 11:30:06.883 Finished recording Recon: channel 6124
2012-01-09 11:30:07.577 TVRec(8): Changing from None to RecordingOnly
2012-01-09 11:30:07.619 TVRec(8): HW Tuner: 8->8
2012-01-09 11:30:07.797 Program #0 not found in PAT!
Program Association Table
 PSIP tableID(0x0) length(17) extension(0x12a2)
      version(23) current(1) section(0) last_section(0)
         tsid: 4770
 programCount: 2
  program number   682 has PID 0x1aa4   data  0x 2 0xaa 0xfa 0xa4
  program number   741 has PID 0x1cf2   data  0x 2 0xe5 0xfc 0xf2

2012-01-09 11:30:07.921 AutoExpire: CalcParams(): Max required Free Space: 3.0 GB w/freq: 3 min
2012-01-09 11:30:08.001 Started recording: "Around the Services": channel 6124 on cardid 8, sourceid 6
2012-01-09 11:30:08.061 TVRec(5): Changing from RecordingOnly to None
2012-01-09 11:30:08.094 Finished recording Yes, Dear "Dominic's Buddy": channel 6552
2012-01-09 11:30:08.240 Finished recording Yes, Dear "Dominic's Buddy": channel 6552
2012-01-09 11:30:08.415 ProcessPAT: Program not found in PAT. 
			Rescan your transports.
2012-01-09 11:30:08.450 Desired program #0 not found in PAT.
			Cannot create single program PAT.
2012-01-09 11:30:08.965 TVRec(8): rec->GetPathname(): '/tux3/shows/6124_20120109113000.mpg'

2012-01-09 12:30:08.103 TVRec(5): HW Tuner: 5->5
2012-01-09 12:30:08.289 AutoExpire: CalcParams(): Max required Free Space: 3.0 GB w/freq: 3 min
2012-01-09 12:30:08.310 Started recording: "My Name Is Earl":"Something to Live For": channel 6552 on cardid 5, sourceid 6
2012-01-09 12:30:08.385 TVRec(7): Changing from RecordingOnly to None
2012-01-09 12:30:08.411 Finished recording A Baby Story "Baby Breen": channel 6639
2012-01-09 12:30:08.523 Finished recording A Baby Story "Baby Breen": channel 6639
2012-01-09 12:30:09.161 Program #0 not found in PAT!
Program Association Table
 PSIP tableID(0x0) length(17) extension(0x128b)
      version(7) current(1) section(0) last_section(0)
         tsid: 4747
 programCount: 2
  program number   684 has PID 0x1ab8   data  0x 2 0xac 0xfa 0xb8
  program number   800 has PID 0x1f40   data  0x 3 0x20 0xff 0x40

2012-01-09 12:30:09.374 TVRec(7): Changing from None to RecordingOnly
2012-01-09 12:30:09.415 TVRec(7): HW Tuner: 7->7
2012-01-09 12:30:09.605 AutoExpire: CalcParams(): Max required Free Space: 3.0 GB w/freq: 3 min
2012-01-09 12:30:09.648 Started recording: "A Baby Story":"Baby Mcgrath": channel 6639 on cardid 7, sourceid 6
2012-01-09 12:30:09.679 ProcessPAT: Program not found in PAT. 
			Rescan your transports.
2012-01-09 12:30:09.721 Updating status for "Income Property" on cardid 8 (Recording => Recorded)
2012-01-09 12:30:09.721 Desired program #0 not found in PAT.
			Cannot create single program PAT.

The two problems occurred on different tuners and different channels.
I have not seen a failure since reverting back to your ver007 patch.

David
-- 
David Engel
david at istwok.net