[mythtv-users] Can't capture using PVR250
brtb at brtb.org
brtb at brtb.org
Tue Mar 30 19:06:49 EST 2004
Any problems with KT333 or KT400? I've been having a few problems with the
waited-for-data-for-15-sec deal which results in backend lockup and
truncated recordings, all it takes is a "killall mythbackend &&
mythbackend -d -l /path/to/log" but it's completely unpredictable.
Happens in LiveTV and recording modes. I really want to point my finger at
my GeForce2MX-200 I just put in to swap with a Radeon 7500 (long
story), but it's an encoder fault, not decode.
[obligatory system details: PVR250 rev 3, ivtv 0.1.9 unpatched, and a
KT400 at the moment, Asus A7V8X-X to be specific. 512MB PC3200 ddr,
athlon 1700+, tv storage is a 2-drive striped-raid IDE, 160gb total,
XFS. front/backend on same machine.]
On Tue, 30 Mar 2004, Coax wrote:
> Interesting!
>
> Check lspci.
>
> Given your failure mode, i'm not entirely sure that either of these 2
> things i'm about to point out - is really your problem. But these are at
> least 2 things to check.. (I'm not sure if you've posted your hardware
> configuration, so if you have, I apologize in advance.)
>
> Make *SURE* you don't have a VIA KT-133 (or KT-266!) chipset on your
> motherboard. (if you do, you'll see stuff about it in the lspci output.)
> If you do, trash it (or use it for some other application) and get a new
> motherboard. The Hauppauge PVR-250/PVR-350 will NOT work well with it.
> There are ways to try to minimize the lockups and crashes, but in the end,
> there is no definitive workaround. (The VIA chipset in question has a bug
> that will be triggered when large amounts of dma data is transferred
> through the chipset.. Makes sense, too, because with this type of
> application, about 99.9% of it is all done via dma.. video (both hardware
> encoder and video card itself), disk, etc.)
>
> Also, check if you have an NFORCE2 motherboard. (lspci will report a bunch
> of nVidia devices if you have one) If you do, you'll need to recompile
> your kernel to A) disable SMP support (Isnt needed anyway unless you have
> a multiprocessor capable board) and B) disable Local APIC support. (If
> you do indeed have a NFORCE2 board, I have a 2.4.25 kernel built that will
> probably work for you so let me know.)
>
> I've not seen the I/O error problem on either of these boards. But
> rather, lockups within X - and sometimes, even the whole system. (I use a
> PVR-350 w/ the TV-Out port on that heading to my TV.) But the idea here
> is that if you do indeed have one of these motherboards, the "failure
> condition" almost always appears the same: Random problems that end up
> not-so-random when you really try to use your MythTV setup. (But hey, at
> least the NFORCE2 problem can easily be worked around.)
>
> Hope that helps!
>
> Chad
>
>
> > ivtv_debug=1 is turned on. The only error that is 'spewed' into dmesg is...
> > ivtv: Timeout waiting for data!
> >
> > ----- Original Message -----
> > From: "Coax" <coax at cornernet.com>
> > To: "Discussion about mythtv" <mythtv-users at mythtv.org>
> > Sent: Tuesday, March 30, 2004 4:00 PM
> > Subject: Re: [mythtv-users] Can't capture using PVR250
> >
> >
> > I'd definitely load your ivtv module with ivtv_debug=1 - so you can
> > 'dmesg' and capture any errors resulting from your 'cat /dev/video0 >
> > /tmp/test.mpg' command.
> >
> > Any errors spewed (if there are any) might point out the problem.
> >
>
>
More information about the mythtv-users
mailing list