On Saturday 03 December 2005 17:05, Matt Klepcyk wrote:
> on 11/20/2005 06:55 Hans Verkuil said the following:
> >On Sunday 20 November 2005 05:02, Matt Klepcyk wrote:
> >>on 11/19/2005 12:08 Hans Verkuil said the following:
> >>>On Friday 18 November 2005 20:23, Matt Klepcyk wrote:
> >>>>Hans Verkuil wrote:
> >>>>>On Friday 18 November 2005 16:41, Matt Klepcyk wrote:
> >>>>
> >>>><snip>
> >>>>
> >>>>>My initial guess is that there is some conflict with another PCI
> >>>>>device. Another option is that there is very little memory
> >>>>>available.
> >>>>>
> >>>>> Hans
> >>>>
> >>>>That's what I was thinking, but tuner 1 (according to lspci -vv:
> >>>>http://stuff.mjk.org/myth/lspci-vv.txt) is on IRQ 7, which
> >>>> nothing else is using. Even when ivtv is loaded up, IRQ 3 (for
> >>>> tuner 2) shows up in /proc/interrupts, but IRQ 7 does not
> >>>> (indicating it's not in use?).
> >>>>
> >>>>Other than IRQs, what other conflicts should I look for?
> >>>
> >>>What is the output of 'cat /proc/iomem'? I think there is a
> >>> conflict with the I/O memory regions. (I couldn't remember where
> >>> to find this info, it turns out it can be read from /proc/iomem).
> >>
> >>Here is the output. As I said in my previous post, I've since
> >> added an old PVR250, which has allowed both tuners to be
> >> recognized on the PVR500. As this /proc/iomem is different from
> >> when the card wasn't loading, please let me know if you'd like me
> >> to try and re-create those conditions and ship you a new
> >> /proc/iomem. Of course, I'm ok with not doing it too. :)
> >
> >Please try again without the PVR250. That's really the interesting
> > case to look at, although it is nice to have this as a comparison.
> >
> >Thanks,
>
> Hans,
>
> Sorry for the delay, it's been a busy last few weeks. Follows is
> /proc/iomem WITHOUT the PVR250. I also snapshotted new logs for
> pretty much everything else and posted them here (w/ 12.03.2005
> dates): http://stuff.mjk.org/myth/
>
> Hope this helps, thanks!
>
> -Matt
Yes, at least it is clear what is wrong.
Compare the iomem output of the FAIL case:
> f4400000-f44fffff : PCI Bus #05
> f8000000-fa00ffff : ivtv0
> fed13000-fed19fff : reserved
> fed1c000-fed9ffff : reserved
with the OK case:
cc000000-d3ffffff : PCI Bus #07
cc000000-cfffffff : 0000:07:08.0
cc000000-ce00ffff : ivtv1
d0000000-d3ffffff : 0000:07:09.0
d0000000-d200ffff : ivtv2
d4000000-d7ffffff : 0000:06:02.0
d4000000-d600ffff : ivtv0
It's clear that a PVR500 unit is not detected in the FAIL case. The
PVR500 have a kind of internal PCI bridge in order to support two PCI
devices on one PCI card. Apparently some systems have problems with
that. Where the cause of the problem is in this case I do not know. See
if there is a BIOS update for your motherboard and try that. There is
also the possibility that it is a bug in the linux PCI subsystem. This
may be a question for the kernel mailinglist.
Anyway, I've seen enough to know that it is not an ivtv bug.
Hans
_______________________________________________
ivtv-users mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-users
