>>>>> "HV" == Hans Verkuil <[EMAIL PROTECTED]> writes:
HV> In that case try with -D72 instead. It's a lot less debugging, but
HV> hopefully it will show me something.
Well... It's been a few weeks, but here's my update on trying with
-D72: I still couldn't reproduce the error with even -D72 debugging
turned on.
A week ago I really needed to update the system in question so I crossed
my fingers that it wouldn't get worse (the .22 kernel on my fc6 box
seemed to be worse than the .20 with 0.10.6 so I'd been running .20 for
quite a while).
Now the system is running is running F8 with 2.6.23.8 and what I believe
to be the stock ivtv 1.0.0 according to the startup info (ivtvctl says
1.0.3 though). It still occasionally has issues, but it actually seems
to be mildly better. Some new symptoms though:
- The playback video actually starts being jumpy after ivtv has seen
issues. That sort of leads me to believe something odd is definitely
happening internally with respect to buffering and something is
locking the system.
- With 2 tuners, only one of them actually starts failing with DMA
timeouts. Those files get messed up, but the ones recorded on the
other tuner turn out ok (which is new; I was sure with the older
kernel they'd both always go at once).
- I see less error messages. It used to flood the messages file with
tons of them but now it's only occasionally sending the TIMEOUT
errors. Note though: the system is still "broken" and has to be
rebooted, but there are less error messages. Example timings:
Dec 16 08:29:48 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 08:29:58 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 08:30:07 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 08:59:17 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:00:12 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:01:26 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:01:35 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:02:39 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:03:25 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:04:39 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:04:48 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:06:20 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:06:48 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:08:02 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:09:15 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:09:25 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:10:11 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:11:15 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:12:20 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:12:29 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:12:38 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
Dec 16 09:14:19 machine kernel: ivtv0: DMA TIMEOUT 00000001 0
As you can see, they're every few minutes (not seconds like before).
- New error messages appear that I hadn't seen previously:
Dec 16 09:40:12 machine kernel: ivtv0 dma: encoder MPEG: Last DMA of size
0x00000000
Dec 16 10:30:46 machine kernel: ivtv0 dma: encoder MPEG: Last DMA of size
0x00000000
Dec 17 07:00:03 machine kernel: ivtv0 dma: encoder MPEG: Last DMA of size
0x00000000
Dec 17 07:27:22 machine kernel: ivtv1 dma: encoder MPEG: Last DMA of size
0x0000000
- The debugging doesn't turn anything on. -D72 produces no output. I
suspect this is because the stock driver doesn't include the
debugging output turned on.
I'd be happy to upgrade to ivtv 1.0.3 (1.0.4?) if you can tell me where
to get the source code for; the standard download page says the ivtv
base is now in the kernel and only the FB code is in the tar-ball.
I'm open to suggestions?
--
"In the bathtub of history the truth is harder to hold than the soap,
and much more difficult to find." -- Terry Pratchett
_______________________________________________
ivtv-users mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-users
