Re: ethernet issue with lombard

On Tue, Dec 19, 2000 at 11:33:26PM -0600, jeramy b smith wrote:

Me too. There are 3 new lines in the bmac driver that are in the
dbdma_reset .  Ben added them because what they do is also done in
darwin. I haven't had a chance to downgrade my driver and try
though. Running in 100tx mode I have seen it happen whil my
bandwidth utilization was very low. I was actually actively
communication each time it got 'gummed up.'

Actually, with 2.2.18, I've had random network lockups as well when
running in 100tx mode.

For a while I thought it only happened in 100 tx, but it just happened
in 10 randomly, so I tried to trigger it with ping -f.  Locks up the
network in a matter of 20 seconds usually.

I've downloaded the darwin driver source, and I've been comparing --
but...why do they have to use C++!?!?  oh well :)

ttyl,

-- 
Josh Huber
1024D/6B21489A 61F0 6138 BE7B FEBF A223  E9D1 BFE1 2065 6B21 489A

Hello all,

I got kernel 2.2.18 from kernel.org yesterday, and built it as an SMP kernel on my Dual G4 450 machine.  This is the first Linux kernel I've run in SMP 'mode'.  Both
processors are detected, and it seems to be using the both of them.  I've noticed a downside though; programs seem to coredump out of the blue.  Mozilla 0.6 which
hadn't crashed since I installed it a week ago suddenly core dumps every hour or so.  The GNOME Panel dies periodically now, and the X Server once core dumped.

My question is, are these weird core dumps due to flakey SMP support for the Dual G4 in the kernel, or are these bugs in these programs that don't show up on a
Uni-Processor machine/kernel?

I'm not sure whether I should be reporting these crashes or not ;)

Thanks!
Chris Backas


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/