[mythtv-users] IVTV Problem

[Myster Man]

On Tue, 2003-10-28 at 19:20, Jarod C. Wilson wrote:
> On Sunday, Oct 26, 2003, at 02:34 US/Pacific, Torsten Schenkel wrote:
> 
> > On Sun, 2003-10-26 at 07:52, Jarod C. Wilson wrote:
> >
> >> So people successfully running an nForce2 board all seem to either 
> >> have
> >> a 2.4.20 kernel with APIC disabled, or a 2.4.22 kernel, in which APIC
> >> actually works. Or a Red Hat 2.4.20 kernel, with APIC in an unknown
> >> state. My presumption is that Red Hat backported the vital APIC stuff
> >> from 2.4.22, so their kernels are safe with nForce2 boards. Either
> >> that, or APIC is disabled. I'll have to look into that (as well as
> >> whether APIC is enabled or disabled in my BIOS)...
> >
> > no the 2.4.22 is NOT NForce2 save. You still have to disable local APIC
> > in the Kernel. I have two NForce2 machines running. One with 2.4.22 
> > WITH
> > APIC stable for two months uptime (new machine so only two months), MSI
> > K7... board (don't know the exact number, have to look it up on monday,
> > if you are interested, the boards no longer available). Because of the
> > good performance of this machine, I decided to go for another NForce2
> > machine, despite of the annoying binary drivers. The second machine 
> > (MSI
> > K7... Delta, the sucessor board to the other's) installed and came up
> > nicely, but heavy DMA traffic made it freeze, which, being in a number
> > crunching environment, led to a maximum uptime of 30 minutes. Changed
> > the board to the ASUS A7N8X Deluxe, no change. Disabled APIC in kernel
> > and BIOS and this second machine got stable.
> 
> Huh. Well, for kicks, I checked, and I do NOT have APIC disabled on my 
> system, and didn't do anything special with the kernel to disable it 
> either, and I'm rock-solid. I guess I'm just lucky. No problems w/ivtv 
> and recording multiple shows at once, while watching a recording. Go 
> figure...
> 
> > The problem with NForce2 seems to be that NVidia is not very thorough
> > with their revision numbering, with changes WITHIN one minor revision
> > which are not being documented. I heard stories of changes in the
> > endian-ness of the chipset that weren't documented and didn't lead to
> > other denominations, don't know if that's true, you'll have to ask 
> > Allan
> > Cox, I take it he's really pissed of by the NForce2.
> 
> Hrm... Changes of endian-ness, huh? Possibly some nForce2 Mac 
> prototypes involved in that...
> 
> > So, while everyone is going from VIA to NVidia, those chipsets aren't
> > stable either (the other way round, never had problems with VIA, except
> > ivtv).
> 
> And meanwhile, I never had any problems using a VIA chipset board 
> (KM266 chipset) w/ivtv...

If you actually did have the via working with no dma problems you might
want to let the ivtv people know the specifics bacause from everything
I've searched your the only one to ever get a via chipset less than 333
to work with ivtv.

> 
> > Hmm, the only really stable board in my park, that has no known
> > chipset issues is the ASUS Dual Athlon MP with AMD chipset (don't know
> > exactly which one, but NO Via southbridge).
> 
> AMD 760 MP, I believe. I've got the same chipset on my Tyan Thunder K7 
> board. Or possibly the 760 MPX (which I have on my Tyan Tiger MPX 
> board), but if it is the Asus board I'm thinking of, it is the 760 MP.
> 
> > I don't want to go back to
> > Intel, but if this situation won't change, there will be no choice.
> 
> Not good. I don't wanna go back either! But then I haven't had any 
> issues with either a VIA or nForce2 board. Perhaps there is more to 
> distribution selection than meets the eye...
> 
> --Jarod
> 
> ______________________________________________________________________
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users