> Ahmon Dancy wrote:
> >
> > I suspect that going to UDMA2 just makes the problem harder
> > to reproduce due to the reduced load.
>
> Or it could be something intrinsic about PCI busmastering
> timings. I would try different PCI latencies in the BIOS.
> Higher numbers mean a device can hog the bus for longer,
> and dump buffers quicker but other devices might have to
> wait. Lower numbers chop up each device's access so nobody
> waits long, but total transfers are slower.
Do you have BP6 board - are you sure you could select these values
somewhere in BP6 BIOS ?
I've found only two things which are similir to this in Chipset Feature Menu
Passive Release & Delayed Release - I've tried all 4 combination
and the machine has always locked - however according to my experiments
the longest surviving has been with both this option enabled.
But I'm not that sure as I've not made that many test for this.
Anyway is there a way that linux could set these numbers during the boot.
(As maybe longer "Delayed Release" could possible make the machine more
stable) ?
--
There are three types of people in the world:
those who can count, and those who can't.
Zdenek Kabelac http://i.am/kabi/ [EMAIL PROTECTED] {debian.org; fi.muni.cz}
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
