John Drescher wrote: > On 5/2/07, Brian Walter <[EMAIL PROTECTED]> wrote: > >> Greetings, >> >> I'm working on getting the last few issues cleaned up with a MythTV >> setup. One of the issues I have, is that whenever I access the first >> tuner on my PVR-500, I get a series of DMA warnings/errors: >> >> May 1 00:06:08 mythtvserver1 kernel: ivtv1 warning: ENC DMA ERROR b >> May 1 00:06:08 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 0 -> 128 >> May 1 00:06:08 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 128 -> 104 >> May 1 00:06:10 mythtvserver1 kernel: ivtv1 warning: ENC DMA ERROR b >> May 1 00:06:10 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 104 -> 128 >> May 1 00:06:10 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 128 -> 36 >> May 1 00:06:15 mythtvserver1 kernel: ivtv1 warning: ENC DMA ERROR b >> May 1 00:06:15 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 36 -> 128 >> May 1 00:06:15 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 128 -> 0 >> May 1 00:06:22 mythtvserver1 kernel: ivtv1 warning: ENC DMA ERROR b >> May 1 00:06:22 mythtvserver1 kernel: ivtv1 warning: encoder MPEG: >> offset 0 -> 128 >> May 1 00:06:22 mythtvserver1 kernel: ivtv1: DMA TIMEOUT 00000001 0 >> >> Now, the above seems to only to occur on ivtv1 (the first tuner of the >> pvr-500). I also have a pvr-150 installed, which consumes ivtv0 (video0). >> >> With video0 in use (recording), if I cat from video1 (first tuner), I >> almost immediately get the above. If I then turn around and cat from >> video2 (second tuner), it works fine with no errors. >> >> I notice that all three ivtv devices are currently on the same >> interrupt. Is this a cause for concern? >> >> Would moving the cards impact this (the problem, not the interrupts)? >> >> Changing the interrupts? >> >> Indications of a bad board? >> >> All of the above? >> >> None of the above? >> >> > Did you solve this? I had to move an old pvr250 from old master > backend (where it ran for almost a year without a reboot or any > problems) to a newer box that had one of those pvr500s with the bad > samsung tuner but now I experience these dma errors and it ends up > terminating the stream until I reboot which basically means I can not > use the card. > John > > I haven't 'solved' it....I'm waiting for the latest and greatest driver - 0.10.2 I'm running a trunk version of it now, with a few less errors - but I ended up writing a script that checks every 30 minutes at 10 minutes after to see if myth is busy (recording or comflaging) and if not, reboots the server.
It's a brute force method but until I know it's not the driver (and I suspect that it is, because prior to this, I had *horrendous* clock drift), this is the solution I'm living with. If after the driver is fixed, and the issue still persists, I might either try a replacement card (I have a newer card, which has the Samsung tuner - Hauppauge *only* ships Samsung now..), or pursue it here - to see if it's a different failure of the driver... Brian _______________________________________________ ivtv-users mailing list [email protected] http://ivtvdriver.org/mailman/listinfo/ivtv-users
