Hans, OK. Thanks. Those messages are gone. However, now it's exposing something new that wasn't a problem with 0.10.0rc1. I started losing a lot of data, and getting another LARGE set of log messages.
I restarted my app, started reading in VBI from 8 tuners (no problem there), and then I started reading /dev/video from 3 tuners. In about 48 seconds, I started getting many of the following: Feb 16 09:07:35 server2 kernel: ivtv1: All encoder MPEG stream buffers are full. Dropping data. Feb 16 09:07:35 server2 kernel: ivtv1: Cause: the application is not reading fast enough. Then 10 seconds later, VBI buffer messages joined the MPEG messages: Feb 16 09:07:35 server2 kernel: ivtv0: All encoder VBI stream buffers are full. Dropping data. Feb 16 09:07:35 server2 kernel: ivtv0: Cause: the application is not reading fast enough. I got about 200,000 lines of these messages between 09:00:48 (local) and 09:07:38, and then suddenly they stopped (the log messages stopped). I checked the video from one of the tuners, and I lost about 7 minutes of video. The captions appear to be synchronized with the video, so I think I lost about 7 minutes of captions as well. Doing a vmstat at this point, the load appears to be the same as it was with the previous version of the driver. So, I'm a little confused, but it looks like IVTV had a buffer problem, and then tried to recover, dropping data continuously until it finally did recover about 7 minutes later. And in the process, it created excessive amounts of log messages. It's worth noting that the program from 09:30 to 10:00 recorded the full 30 minutes of video just fine. Maybe this was a one-time glitch. I will continue to monitor it and let you know if it happens again. But it's coincidental that it happened after you fixed the other VBI problem. Thanks, Rich -- Rich Kadel Appeligo, Inc. (858) 433-1747 www.appeligo.com -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Hans Verkuil Sent: Friday, February 16, 2007 6:05 AM To: Discussion list for development of the IVTV driver Subject: Re: [ivtv-devel] System (practically) non-responsive after several hours...no clues Hi Rich, Please check out the latest trunk from subversion. This bug is fixed. It was a subtle error resulting in these spurious errors. Thanks for reporting this, Hans On Friday 16 February 2007 02:11, Rich Kadel wrote: > OK... Some more data for you, based on your questions: > > 1. It is sliced. I am reading it from /dev/vbi[x]. > > 2. I tried some tests. If I just cat /dev/video >/dev/null, no > messages appear. In fact, I ran "cat /dev/video[x] >/dev/null" > simultaneously with 5 tuners and saw no messages. > > But, then I added "cat /dev/vbi0 >/dev/null" and the following > messages showed up: > > Feb 15 17:03:10 server2 kernel: ivtv0 warning: Starting VBI after > starting an en > coding, seems to not work. > Feb 15 17:03:10 server2 kernel: ivtv0 warning: encoder VBI: Couldn't > find start > of buffer within the first 256 bytes > Feb 15 17:03:18 server2 last message repeated 18 times > > I stopped all but one tuner (/dev/video0), and started "cat > /dev/vbi0" again, and got the same message. > > Feb 15 17:03:35 server2 kernel: ivtv0 warning: Starting VBI after > starting an en > coding, seems to not work. > Feb 15 17:03:36 server2 kernel: ivtv0 warning: encoder VBI: Couldn't > find start > of buffer within the first 256 bytes > > Then I tried starting "cat /dev/vbi0" first, followed by "cat > /dev/video0", and I got only the message: > > Feb 15 17:03:36 server2 kernel: ivtv0 warning: encoder VBI: Couldn't > find start > of buffer within the first 256 bytes > > (Ignore the timestamps) > > So, the bottom line is, these messages appear when I read /dev/vbi > and /dev/video at the same time. > > Very reproducible, and fairly constant as I said. > > Thanks, > Rich > > -- > > Rich Kadel > Appeligo, Inc. > (858) 433-1747 > www.appeligo.com > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Hans Verkuil > Sent: Thursday, February 15, 2007 3:42 PM > To: Discussion list for development of the IVTV driver > Subject: Re: [ivtv-devel] System (practically) non-responsive after > several hours...no clues > > On Thursday 15 February 2007 17:42, Rich Kadel wrote: > > So far so good with 0.10.0rc1. At least, it hasn't crashed in a > > couple of days of continuous reading, 3 video streams, and 8 VBI. > > > > Now I have a new problem I'd like to resolve. My logs are getting > > filled up with the following message: > > > > Feb 12 16:00:00 server2 kernel: ivtv2 warning: encoder VBI: > > Couldn't find start > > of buffer within the first 256 bytes > > > > Occasionally, I'll notice some dropped characters in the VBI. Not > > often. But it could be a symptom. > > > > I'd like to fix the VBI problem if there's a way to do that. If > > not, is there a way to reduce the log messages? > > How are you capturing VBI? Is it sliced VBI or raw VBI? And if it is > sliced VBI, is it embedded in the MPEG stream or captured as a > separate stream? > > Do you always have these messages, or do they appear suddenly? It is > reproducable with a simple 'cat /dev/video0 >x.mpg' command? > > BTW, there is no need to increase the mpeg buffers, it's not related > to that. Basically this message indicates that something weird is > going on with the VBI transfers. It can occur once in a while, but it > shouldn't happen continuously. > > Regards, > > Hans > > _______________________________________________ > ivtv-devel mailing list > [email protected] > http://ivtvdriver.org/mailman/listinfo/ivtv-devel > > > _______________________________________________ > ivtv-devel mailing list > [email protected] > http://ivtvdriver.org/mailman/listinfo/ivtv-devel _______________________________________________ ivtv-devel mailing list [email protected] http://ivtvdriver.org/mailman/listinfo/ivtv-devel _______________________________________________ ivtv-devel mailing list [email protected] http://ivtvdriver.org/mailman/listinfo/ivtv-devel
