On Saturday 08 December 2001 15:47, Alan Cox wrote:
> There are a set of old machines that do this sort of stuff with all 2.4.x
> kernels. Right now I don't know why. The Digital celebris has the same
> bug.
Well it's nice to see I'm not alone in this. I gave up on my SuperMicro
Micro ATX 370 boards a few weeks ago. They were used primarily with WinNT so
they don't have a well establishied Linux track record. (Though one of them
did run a mid 2.2.?? for a short time, and I didn't notice any issues
with it.)
I'd like to use them as routers, but they die with heavy network activity.
>From my experience it looks like an IRQ problem, but no amount of BIOS
fiddling will get them stable.
My findings after playing around with them tonight
2.4.10 or 2.4.16 seems maybe a bit then 2.4.8
With a 'default' config it may take a few minutes of ping -f and
scp, it bring it down, but it will die suddenly rebooting.
The primary board I'm using is a D-Link Quad Tulip. By 'default' it
scans the eth's in reverse. pci=nosort fixes that. Something strange
is eth2 alwasy picks up irq 5, while the rest are 11 (it
always shifts to the logical eth2 depending on the sort)
Playing with various pci= options will get it so it dies instantly
from `ping -f`. pci=biosirq seems to be a sure bet.
> Turn both of these off
> And these
Didn't help. In fact it dies instantly with `ping -f`
-- The time is now 22:54 (Totalitarian) - http://www.ccops.org/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Dec 15 2001 - 21:00:19 EST