Lan Barnes wrote:
[snip]
So if I have this box with a VIA chipset (and I'll take your word for it
that VIA is a POS), and it runs FC 4 just fine _but_ Myth on FC4 on this
box has found a number of creative ways to choke and lock up, why am I not
justified in saying that Myth has a way to go before it's production
ready? Because taken to its logical extreme, you're saying that if Myth
works fine on your machine, then my machine must be bad, which is a
variation of the old developer's mantra, "it works on _my_ machine."
Sorry. Myth has a way to go.
You don't seem to understand that there is a limit to the extent that
software can make up for bad hardware. The problem is not Myth, it's not
the kernel, and it's not Linux. When you turn over control of the
hardware to some other piece of hardware without software intervention
(which is what DMA does), there is no way to fix it.
The PVR-150 video capture card that you use requires DMA to work, even
though the actual data rates don't need it, and there is no way to turn
it off. DMA is mediated by hardware in the chipset. If the DMA transfers
are faulty because of bad hardware, then bad things are going to happen.
You have symptoms of other bad things happening that you are also
ignoring that are also related to the DMA issue, namely the filesystem
corruption. The hard drive is operating in DMA mode. Faults with the DMA
controller will cause data corruption because data goes directly to/from
the drive into/out of memory via the DMA controller, bypassing software
controls.
And I think you meant MythDora 4 which is based on Fedora Core 6, not
Fedora Core 4 which is now quite old.
Gus
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
