As you may have already seen in a separate response to Bryan Mayland on
the list, I hadn't tried recording anything for several months and when I
actually did based on his observation that it was working for him with
essentially the same setup as what I have, I gave it another try and it
was now working. I was rather surprised by this as it had stopped working
previously very reliably after 20-30 minutes. I had not tried your patch
and don't know what the difference was the allowed it to now work. I had
upgrade Kubuntu 5.04 to 5.10, changing the kernel from 2.6.10-5-386 to
2.6.12-9-K7 as well as going from ivtv-0.3.6someletter to ivtv-0.4.0. I'm
guessing it wasn't the change in ivtv since I never heard of anything on
the list and suppose it was probably the kernel update that must have made
the difference. Can you tell me what kernel version you are using?
Yes I saw that after I wrote the reply. If it works for you, that's
great -- I wouldn't worry about looking at it any further. I'm not sure
that the ivtv version really matters in fact -- the i2c is all handled
by the in-kernel bit banging stuff, so it may be that something was
fixed between the two kernel versions you tried.
Personally I'm currently using the 2.6.14 kernel with v4l-linux/dvb CVS
+ ivtv 0.5.0. I'll try using the kernel i2c implementation and see if
it works.
For all I know it may have only made the remote loss far less frequent for
me and it will die tomorrow, but for now I am happy.
I found that it would not happen at all if the card was not recording at
the time. There are also a whole bunch of people who don't report any
issues, so I guess it must work for some people and not others. I
thought it might be card specific, but maybe it is in fact kernel
specific. Thanks for pointing that out, it's an interesting data point.
I appreciate you updating your patch and making it available and I'm sure
that several others will be able to benefit from it if it is added to the
main ivtv driver.
It's maybe unnecessary if a kernel upgrade takes care of it. The newer
algorithm is slightly faster (as in consumes less CPU) but does need
much wider testing on a range of cards. I've tried it on 2xPVR-150 and
1xPVR-350, and it's been run on about another 5 PVR-150s without issue,
so not much coverage so far.
Tnaks,
Mark
_______________________________________________
ivtv-devel mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-devel
