[mythtv] Ticket #2077 Lossless Transcode Deadlock detected.

[John Pilkington]

Geoffrey Hausheer wrote:
> On 10/11/06, John Pilkington wrote:
>> I have just seen this for the first time, on two recordings both from
>> UKHistory.  Both show a pattern similar to that in the ticket.
>> ---------------
>> ...
>> 2006-10-11 15:00:21.843 Deadlock detected.  One buffer is full when
>>                  the other is empty!  Aborting
> Just because you are getting the same error message doesn't mean that
> it has the same cause.  Please open a new ticket.  As always, I need a
> log '-v all' to see what is going on.  Please either attach to the
> ticekt, or send me off list.

It wasn't just the error message but the three lines shortly before that 
were similar.  ie

2006-10-11 15:00:18.220 #0 PTS:10:28:17.390 Delta: 0.0ms queue: 9
2006-10-11 15:00:18.221 #1 PTS:10:28:17.372 Delta: 18.2222ms queue: 2
2006-10-11 15:00:18.221 #3 PTS:10:28:18.216 Delta: -826.167ms queue: 30

and ok, I'll send more info if I can understand how to get it!
> 
>> I'm also affected by Ticket #1970 : limit number of open pid filters.
>> This drops 'unimportant' pids in a round robin. Might there be a connection?
> My understanding is this only happens when first tuning a channel, is
> this not correct?  I'm not familiar with the issue, but they will only
> be related if myth is changing which pids it captures after the
> recording has begun.

No, my undersanding from danielk is that one pid is dropped and another 
picked up in sequence throughout the recording process.  In my case the 
switch happens every 2 seconds.  It's a partially broken hardware issue.
Daniel's explanation is at the foot of the ticket.

John P