[mythtv-users] Question about ivtv messages and backend crashes

[Brian Walter]

Greetings,

I've started noticing the following when Myth is recording from two 
different pvr sources (1xpvr-500 and 1xpvr15):

Apr 24 11:20:20 mythtvserver1 kernel: ivtv1 warning: ENC DMA ERROR b
Apr 24 11:20:20 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: 
offset 0 -> 128
Apr 24 11:20:20 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: 
offset 128 -> 32

And when myth crashes, the following is logged:
2007-04-24 00:04:57.739 MPEGRec(/dev/video2) Error: select timeout - ivtv
driver has stopped responding
2007-04-24 00:05:02.047 MainServer::HandleAnnounce Monitor
2007-04-24 00:05:02.158 adding: mythtvserver1 as a client (events: 0)
2007-04-24 00:05:04.043 TVRec(3): Changing from None to RecordingOnly
2007-04-24 00:05:04.453 TVRec(3): HW Tuner: 3->3
2007-04-24 00:05:04.926 MPEGRec(/dev/video1) Warning: Audio sample rate 
32000
Hz
                        is not supported by ivtv driver, using 48000 Hz
instead.
2007-04-24 00:05:04.928 Started recording: Law & Order: Criminal Intent
"Slither": channel 2035 on cardid 3, sourceid 2
2007-04-24 00:05:05.444 scheduler: Started recording: Law & Order: Criminal
Intent "Slither": channel 2035 on cardid 3, sourceid 2
2007-04-24 00:05:07.535 MPEGRec(/dev/video2) Error: select timeout - ivtv
driver has stopped responding
2007-04-24 00:05:11.943 Event socket closed. No connection to the backend.
2007-04-24 00:13:08.043 RingBuf(/video/3307_20070423233000.mpg): Waited 1.0
seconds for data to become available...
[mpeg2video @ 0x2aaaac0b03d0]invalid cbp at 26 7
[mpeg2video @ 0x2aaaac0b03d0]Warning MVs not available
2007-04-24 00:13:08.455 Connecting to backend server: 192.168.0.10:6543 
(try 1
of 5)
2007-04-24 00:13:08.465 Connection timed out.


The pvr-500 is running off of analog cable, and the pvr-150 is running 
off of s-video from a directv stb.

This is the alleged latest and greatest from ivtv (0.10.1).

On the whole, this is *usually* benign.   However,  every now and then 
it seems to cause myth to die, and, on at least one occasion, made the 
server inaccessible  from the network,  requiring a power cycle to recover.

Anyone else seeing this, or, know what the meaning is?

Thanks!
Brian